Mercurial > libavcodec.hg
annotate snow.c @ 7904:47717d5239fa libavcodec
AltiVec detection support for OpenBSD, patch by Brad, brad comstyle com.
author | diego |
---|---|
date | Mon, 22 Sep 2008 05:38:50 +0000 |
parents | 5a3907818652 |
children | c5276ad92ff8 |
rev | line source |
---|---|
2138 | 1 /* |
2 * Copyright (C) 2004 Michael Niedermayer <michaelni@gmx.at> | |
3 * | |
3947
c8c591fe26f8
Change license headers to say 'FFmpeg' instead of 'this program/this library'
diego
parents:
3920
diff
changeset
|
4 * This file is part of FFmpeg. |
c8c591fe26f8
Change license headers to say 'FFmpeg' instead of 'this program/this library'
diego
parents:
3920
diff
changeset
|
5 * |
c8c591fe26f8
Change license headers to say 'FFmpeg' instead of 'this program/this library'
diego
parents:
3920
diff
changeset
|
6 * FFmpeg is free software; you can redistribute it and/or |
2138 | 7 * modify it under the terms of the GNU Lesser General Public |
8 * License as published by the Free Software Foundation; either | |
3947
c8c591fe26f8
Change license headers to say 'FFmpeg' instead of 'this program/this library'
diego
parents:
3920
diff
changeset
|
9 * version 2.1 of the License, or (at your option) any later version. |
2138 | 10 * |
3947
c8c591fe26f8
Change license headers to say 'FFmpeg' instead of 'this program/this library'
diego
parents:
3920
diff
changeset
|
11 * FFmpeg is distributed in the hope that it will be useful, |
2138 | 12 * but WITHOUT ANY WARRANTY; without even the implied warranty of |
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
14 * Lesser General Public License for more details. | |
15 * | |
16 * You should have received a copy of the GNU Lesser General Public | |
3947
c8c591fe26f8
Change license headers to say 'FFmpeg' instead of 'this program/this library'
diego
parents:
3920
diff
changeset
|
17 * License along with FFmpeg; if not, write to the Free Software |
3036
0b546eab515d
Update licensing information: The FSF changed postal address.
diego
parents:
3035
diff
changeset
|
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
2138 | 19 */ |
20 | |
21 #include "avcodec.h" | |
22 #include "dsputil.h" | |
3198
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
23 #include "snow.h" |
2335 | 24 |
25 #include "rangecoder.h" | |
2138 | 26 |
27 #include "mpegvideo.h" | |
28 | |
29 #undef NDEBUG | |
30 #include <assert.h> | |
31 | |
32 static const int8_t quant3[256]={ | |
33 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
34 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
35 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
36 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
37 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
38 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
39 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
40 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
41 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
42 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
43 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
44 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
45 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
46 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
47 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
48 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, 0, | |
49 }; | |
50 static const int8_t quant3b[256]={ | |
51 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
52 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
53 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
54 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
55 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
56 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
57 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
58 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, | |
59 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
60 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
61 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
62 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
63 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
64 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
65 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
66 -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, | |
67 }; | |
2596 | 68 static const int8_t quant3bA[256]={ |
69 0, 0, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
70 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
71 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
72 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
73 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
74 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
75 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
76 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
77 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
78 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
79 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
80 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
81 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
82 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
83 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
84 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, 1,-1, | |
85 }; | |
2138 | 86 static const int8_t quant5[256]={ |
87 0, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, | |
88 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, | |
89 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, | |
90 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, | |
91 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, | |
92 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, | |
93 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, | |
94 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, | |
95 -2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2, | |
96 -2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2, | |
97 -2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2, | |
98 -2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2, | |
99 -2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2, | |
100 -2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2, | |
101 -2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2, | |
102 -2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-1,-1,-1, | |
103 }; | |
104 static const int8_t quant7[256]={ | |
105 0, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, | |
106 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, | |
107 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, | |
108 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, | |
109 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, | |
110 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, | |
111 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, | |
112 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, | |
113 -3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3, | |
114 -3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3, | |
115 -3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3, | |
116 -3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3, | |
117 -3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-3, | |
118 -3,-3,-3,-3,-3,-3,-3,-3,-3,-2,-2,-2,-2,-2,-2,-2, | |
119 -2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2, | |
120 -2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-2,-1,-1, | |
121 }; | |
122 static const int8_t quant9[256]={ | |
123 0, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, | |
124 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, | |
125 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, | |
126 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, | |
127 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, | |
128 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, | |
129 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, | |
130 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, | |
131 -4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4, | |
132 -4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4, | |
133 -4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4, | |
134 -4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4, | |
135 -4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4, | |
136 -4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4, | |
137 -4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-3,-3,-3,-3, | |
138 -3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-2,-2,-2,-2,-1,-1, | |
139 }; | |
140 static const int8_t quant11[256]={ | |
141 0, 1, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, | |
142 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, | |
143 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, | |
144 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, | |
145 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, | |
146 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, | |
147 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, | |
148 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, | |
149 -5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5, | |
150 -5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5, | |
151 -5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5, | |
152 -5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5, | |
153 -5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5, | |
154 -5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-4,-4, | |
155 -4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4,-4, | |
156 -4,-4,-4,-4,-4,-3,-3,-3,-3,-3,-3,-3,-2,-2,-2,-1, | |
157 }; | |
158 static const int8_t quant13[256]={ | |
159 0, 1, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, | |
160 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, | |
161 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, | |
162 5, 5, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, | |
163 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, | |
164 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, | |
165 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, | |
166 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, | |
167 -6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6, | |
168 -6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6, | |
169 -6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6, | |
170 -6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6, | |
171 -6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-6,-5, | |
172 -5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5, | |
173 -5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5,-5, | |
174 -4,-4,-4,-4,-4,-4,-4,-4,-4,-3,-3,-3,-3,-2,-2,-1, | |
175 }; | |
176 | |
177 #if 0 //64*cubic | |
178 static const uint8_t obmc32[1024]={ | |
5659
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
179 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
180 0, 0, 0, 0, 0, 4, 4, 4, 4, 4, 4, 4, 4, 8, 8, 8, 8, 8, 8, 4, 4, 4, 4, 4, 4, 4, 4, 0, 0, 0, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
181 0, 0, 0, 4, 4, 4, 4, 8, 8, 12, 12, 12, 16, 16, 16, 16, 16, 16, 16, 16, 12, 12, 12, 8, 8, 4, 4, 4, 4, 0, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
182 0, 0, 4, 4, 8, 8, 12, 16, 16, 20, 24, 24, 28, 28, 32, 32, 32, 32, 28, 28, 24, 24, 20, 16, 16, 12, 8, 8, 4, 4, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
183 0, 0, 4, 8, 8, 12, 16, 24, 28, 32, 36, 40, 44, 48, 48, 48, 48, 48, 48, 44, 40, 36, 32, 28, 24, 16, 12, 8, 8, 4, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
184 0, 4, 4, 8, 12, 20, 24, 32, 40, 44, 52, 56, 60, 64, 68, 72, 72, 68, 64, 60, 56, 52, 44, 40, 32, 24, 20, 12, 8, 4, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
185 0, 4, 4, 12, 16, 24, 32, 40, 52, 60, 68, 76, 80, 88, 88, 92, 92, 88, 88, 80, 76, 68, 60, 52, 40, 32, 24, 16, 12, 4, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
186 0, 4, 8, 16, 24, 32, 40, 52, 64, 76, 84, 92,100,108,112,116,116,112,108,100, 92, 84, 76, 64, 52, 40, 32, 24, 16, 8, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
187 0, 4, 8, 16, 28, 40, 52, 64, 76, 88,100,112,124,132,136,140,140,136,132,124,112,100, 88, 76, 64, 52, 40, 28, 16, 8, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
188 0, 4, 12, 20, 32, 44, 60, 76, 88,104,120,132,144,152,160,164,164,160,152,144,132,120,104, 88, 76, 60, 44, 32, 20, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
189 0, 4, 12, 24, 36, 48, 68, 84,100,120,136,152,164,176,180,184,184,180,176,164,152,136,120,100, 84, 68, 48, 36, 24, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
190 0, 4, 12, 24, 40, 56, 76, 92,112,132,152,168,180,192,204,208,208,204,192,180,168,152,132,112, 92, 76, 56, 40, 24, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
191 0, 4, 16, 28, 44, 60, 80,100,124,144,164,180,196,208,220,224,224,220,208,196,180,164,144,124,100, 80, 60, 44, 28, 16, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
192 0, 8, 16, 28, 48, 64, 88,108,132,152,176,192,208,224,232,240,240,232,224,208,192,176,152,132,108, 88, 64, 48, 28, 16, 8, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
193 0, 4, 16, 32, 48, 68, 88,112,136,160,180,204,220,232,244,248,248,244,232,220,204,180,160,136,112, 88, 68, 48, 32, 16, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
194 1, 8, 16, 32, 48, 72, 92,116,140,164,184,208,224,240,248,255,255,248,240,224,208,184,164,140,116, 92, 72, 48, 32, 16, 8, 1, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
195 1, 8, 16, 32, 48, 72, 92,116,140,164,184,208,224,240,248,255,255,248,240,224,208,184,164,140,116, 92, 72, 48, 32, 16, 8, 1, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
196 0, 4, 16, 32, 48, 68, 88,112,136,160,180,204,220,232,244,248,248,244,232,220,204,180,160,136,112, 88, 68, 48, 32, 16, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
197 0, 8, 16, 28, 48, 64, 88,108,132,152,176,192,208,224,232,240,240,232,224,208,192,176,152,132,108, 88, 64, 48, 28, 16, 8, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
198 0, 4, 16, 28, 44, 60, 80,100,124,144,164,180,196,208,220,224,224,220,208,196,180,164,144,124,100, 80, 60, 44, 28, 16, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
199 0, 4, 12, 24, 40, 56, 76, 92,112,132,152,168,180,192,204,208,208,204,192,180,168,152,132,112, 92, 76, 56, 40, 24, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
200 0, 4, 12, 24, 36, 48, 68, 84,100,120,136,152,164,176,180,184,184,180,176,164,152,136,120,100, 84, 68, 48, 36, 24, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
201 0, 4, 12, 20, 32, 44, 60, 76, 88,104,120,132,144,152,160,164,164,160,152,144,132,120,104, 88, 76, 60, 44, 32, 20, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
202 0, 4, 8, 16, 28, 40, 52, 64, 76, 88,100,112,124,132,136,140,140,136,132,124,112,100, 88, 76, 64, 52, 40, 28, 16, 8, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
203 0, 4, 8, 16, 24, 32, 40, 52, 64, 76, 84, 92,100,108,112,116,116,112,108,100, 92, 84, 76, 64, 52, 40, 32, 24, 16, 8, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
204 0, 4, 4, 12, 16, 24, 32, 40, 52, 60, 68, 76, 80, 88, 88, 92, 92, 88, 88, 80, 76, 68, 60, 52, 40, 32, 24, 16, 12, 4, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
205 0, 4, 4, 8, 12, 20, 24, 32, 40, 44, 52, 56, 60, 64, 68, 72, 72, 68, 64, 60, 56, 52, 44, 40, 32, 24, 20, 12, 8, 4, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
206 0, 0, 4, 8, 8, 12, 16, 24, 28, 32, 36, 40, 44, 48, 48, 48, 48, 48, 48, 44, 40, 36, 32, 28, 24, 16, 12, 8, 8, 4, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
207 0, 0, 4, 4, 8, 8, 12, 16, 16, 20, 24, 24, 28, 28, 32, 32, 32, 32, 28, 28, 24, 24, 20, 16, 16, 12, 8, 8, 4, 4, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
208 0, 0, 0, 4, 4, 4, 4, 8, 8, 12, 12, 12, 16, 16, 16, 16, 16, 16, 16, 16, 12, 12, 12, 8, 8, 4, 4, 4, 4, 0, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
209 0, 0, 0, 0, 0, 4, 4, 4, 4, 4, 4, 4, 4, 8, 8, 8, 8, 8, 8, 4, 4, 4, 4, 4, 4, 4, 4, 0, 0, 0, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
210 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
2138 | 211 //error:0.000022 |
212 }; | |
213 static const uint8_t obmc16[256]={ | |
5659
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
214 0, 0, 0, 0, 0, 0, 4, 4, 4, 4, 0, 0, 0, 0, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
215 0, 4, 4, 8, 16, 20, 20, 24, 24, 20, 20, 16, 8, 4, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
216 0, 4, 16, 24, 36, 44, 52, 60, 60, 52, 44, 36, 24, 16, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
217 0, 8, 24, 44, 60, 80, 96,104,104, 96, 80, 60, 44, 24, 8, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
218 0, 16, 36, 60, 92,116,136,152,152,136,116, 92, 60, 36, 16, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
219 0, 20, 44, 80,116,152,180,196,196,180,152,116, 80, 44, 20, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
220 4, 20, 52, 96,136,180,212,228,228,212,180,136, 96, 52, 20, 4, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
221 4, 24, 60,104,152,196,228,248,248,228,196,152,104, 60, 24, 4, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
222 4, 24, 60,104,152,196,228,248,248,228,196,152,104, 60, 24, 4, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
223 4, 20, 52, 96,136,180,212,228,228,212,180,136, 96, 52, 20, 4, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
224 0, 20, 44, 80,116,152,180,196,196,180,152,116, 80, 44, 20, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
225 0, 16, 36, 60, 92,116,136,152,152,136,116, 92, 60, 36, 16, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
226 0, 8, 24, 44, 60, 80, 96,104,104, 96, 80, 60, 44, 24, 8, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
227 0, 4, 16, 24, 36, 44, 52, 60, 60, 52, 44, 36, 24, 16, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
228 0, 4, 4, 8, 16, 20, 20, 24, 24, 20, 20, 16, 8, 4, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
229 0, 0, 0, 0, 0, 0, 4, 4, 4, 4, 0, 0, 0, 0, 0, 0, |
2138 | 230 //error:0.000033 |
231 }; | |
232 #elif 1 // 64*linear | |
233 static const uint8_t obmc32[1024]={ | |
3206 | 234 0, 0, 0, 0, 4, 4, 4, 4, 4, 4, 4, 4, 8, 8, 8, 8, 8, 8, 8, 8, 4, 4, 4, 4, 4, 4, 4, 4, 0, 0, 0, 0, |
235 0, 4, 4, 4, 8, 8, 8, 12, 12, 16, 16, 16, 20, 20, 20, 24, 24, 20, 20, 20, 16, 16, 16, 12, 12, 8, 8, 8, 4, 4, 4, 0, | |
236 0, 4, 8, 8, 12, 12, 16, 20, 20, 24, 28, 28, 32, 32, 36, 40, 40, 36, 32, 32, 28, 28, 24, 20, 20, 16, 12, 12, 8, 8, 4, 0, | |
237 0, 4, 8, 12, 16, 20, 24, 28, 28, 32, 36, 40, 44, 48, 52, 56, 56, 52, 48, 44, 40, 36, 32, 28, 28, 24, 20, 16, 12, 8, 4, 0, | |
238 4, 8, 12, 16, 20, 24, 28, 32, 40, 44, 48, 52, 56, 60, 64, 68, 68, 64, 60, 56, 52, 48, 44, 40, 32, 28, 24, 20, 16, 12, 8, 4, | |
239 4, 8, 12, 20, 24, 32, 36, 40, 48, 52, 56, 64, 68, 76, 80, 84, 84, 80, 76, 68, 64, 56, 52, 48, 40, 36, 32, 24, 20, 12, 8, 4, | |
240 4, 8, 16, 24, 28, 36, 44, 48, 56, 60, 68, 76, 80, 88, 96,100,100, 96, 88, 80, 76, 68, 60, 56, 48, 44, 36, 28, 24, 16, 8, 4, | |
241 4, 12, 20, 28, 32, 40, 48, 56, 64, 72, 80, 88, 92,100,108,116,116,108,100, 92, 88, 80, 72, 64, 56, 48, 40, 32, 28, 20, 12, 4, | |
242 4, 12, 20, 28, 40, 48, 56, 64, 72, 80, 88, 96,108,116,124,132,132,124,116,108, 96, 88, 80, 72, 64, 56, 48, 40, 28, 20, 12, 4, | |
243 4, 16, 24, 32, 44, 52, 60, 72, 80, 92,100,108,120,128,136,148,148,136,128,120,108,100, 92, 80, 72, 60, 52, 44, 32, 24, 16, 4, | |
244 4, 16, 28, 36, 48, 56, 68, 80, 88,100,112,120,132,140,152,164,164,152,140,132,120,112,100, 88, 80, 68, 56, 48, 36, 28, 16, 4, | |
245 4, 16, 28, 40, 52, 64, 76, 88, 96,108,120,132,144,156,168,180,180,168,156,144,132,120,108, 96, 88, 76, 64, 52, 40, 28, 16, 4, | |
246 8, 20, 32, 44, 56, 68, 80, 92,108,120,132,144,156,168,180,192,192,180,168,156,144,132,120,108, 92, 80, 68, 56, 44, 32, 20, 8, | |
247 8, 20, 32, 48, 60, 76, 88,100,116,128,140,156,168,184,196,208,208,196,184,168,156,140,128,116,100, 88, 76, 60, 48, 32, 20, 8, | |
248 8, 20, 36, 52, 64, 80, 96,108,124,136,152,168,180,196,212,224,224,212,196,180,168,152,136,124,108, 96, 80, 64, 52, 36, 20, 8, | |
249 8, 24, 40, 56, 68, 84,100,116,132,148,164,180,192,208,224,240,240,224,208,192,180,164,148,132,116,100, 84, 68, 56, 40, 24, 8, | |
250 8, 24, 40, 56, 68, 84,100,116,132,148,164,180,192,208,224,240,240,224,208,192,180,164,148,132,116,100, 84, 68, 56, 40, 24, 8, | |
251 8, 20, 36, 52, 64, 80, 96,108,124,136,152,168,180,196,212,224,224,212,196,180,168,152,136,124,108, 96, 80, 64, 52, 36, 20, 8, | |
252 8, 20, 32, 48, 60, 76, 88,100,116,128,140,156,168,184,196,208,208,196,184,168,156,140,128,116,100, 88, 76, 60, 48, 32, 20, 8, | |
253 8, 20, 32, 44, 56, 68, 80, 92,108,120,132,144,156,168,180,192,192,180,168,156,144,132,120,108, 92, 80, 68, 56, 44, 32, 20, 8, | |
254 4, 16, 28, 40, 52, 64, 76, 88, 96,108,120,132,144,156,168,180,180,168,156,144,132,120,108, 96, 88, 76, 64, 52, 40, 28, 16, 4, | |
255 4, 16, 28, 36, 48, 56, 68, 80, 88,100,112,120,132,140,152,164,164,152,140,132,120,112,100, 88, 80, 68, 56, 48, 36, 28, 16, 4, | |
256 4, 16, 24, 32, 44, 52, 60, 72, 80, 92,100,108,120,128,136,148,148,136,128,120,108,100, 92, 80, 72, 60, 52, 44, 32, 24, 16, 4, | |
257 4, 12, 20, 28, 40, 48, 56, 64, 72, 80, 88, 96,108,116,124,132,132,124,116,108, 96, 88, 80, 72, 64, 56, 48, 40, 28, 20, 12, 4, | |
258 4, 12, 20, 28, 32, 40, 48, 56, 64, 72, 80, 88, 92,100,108,116,116,108,100, 92, 88, 80, 72, 64, 56, 48, 40, 32, 28, 20, 12, 4, | |
259 4, 8, 16, 24, 28, 36, 44, 48, 56, 60, 68, 76, 80, 88, 96,100,100, 96, 88, 80, 76, 68, 60, 56, 48, 44, 36, 28, 24, 16, 8, 4, | |
260 4, 8, 12, 20, 24, 32, 36, 40, 48, 52, 56, 64, 68, 76, 80, 84, 84, 80, 76, 68, 64, 56, 52, 48, 40, 36, 32, 24, 20, 12, 8, 4, | |
261 4, 8, 12, 16, 20, 24, 28, 32, 40, 44, 48, 52, 56, 60, 64, 68, 68, 64, 60, 56, 52, 48, 44, 40, 32, 28, 24, 20, 16, 12, 8, 4, | |
262 0, 4, 8, 12, 16, 20, 24, 28, 28, 32, 36, 40, 44, 48, 52, 56, 56, 52, 48, 44, 40, 36, 32, 28, 28, 24, 20, 16, 12, 8, 4, 0, | |
263 0, 4, 8, 8, 12, 12, 16, 20, 20, 24, 28, 28, 32, 32, 36, 40, 40, 36, 32, 32, 28, 28, 24, 20, 20, 16, 12, 12, 8, 8, 4, 0, | |
264 0, 4, 4, 4, 8, 8, 8, 12, 12, 16, 16, 16, 20, 20, 20, 24, 24, 20, 20, 20, 16, 16, 16, 12, 12, 8, 8, 8, 4, 4, 4, 0, | |
265 0, 0, 0, 0, 4, 4, 4, 4, 4, 4, 4, 4, 8, 8, 8, 8, 8, 8, 8, 8, 4, 4, 4, 4, 4, 4, 4, 4, 0, 0, 0, 0, | |
2138 | 266 //error:0.000020 |
267 }; | |
268 static const uint8_t obmc16[256]={ | |
3206 | 269 0, 4, 4, 8, 8, 12, 12, 16, 16, 12, 12, 8, 8, 4, 4, 0, |
270 4, 8, 16, 20, 28, 32, 40, 44, 44, 40, 32, 28, 20, 16, 8, 4, | |
271 4, 16, 24, 36, 44, 56, 64, 76, 76, 64, 56, 44, 36, 24, 16, 4, | |
272 8, 20, 36, 48, 64, 76, 92,104,104, 92, 76, 64, 48, 36, 20, 8, | |
273 8, 28, 44, 64, 80,100,116,136,136,116,100, 80, 64, 44, 28, 8, | |
274 12, 32, 56, 76,100,120,144,164,164,144,120,100, 76, 56, 32, 12, | |
275 12, 40, 64, 92,116,144,168,196,196,168,144,116, 92, 64, 40, 12, | |
276 16, 44, 76,104,136,164,196,224,224,196,164,136,104, 76, 44, 16, | |
277 16, 44, 76,104,136,164,196,224,224,196,164,136,104, 76, 44, 16, | |
278 12, 40, 64, 92,116,144,168,196,196,168,144,116, 92, 64, 40, 12, | |
279 12, 32, 56, 76,100,120,144,164,164,144,120,100, 76, 56, 32, 12, | |
280 8, 28, 44, 64, 80,100,116,136,136,116,100, 80, 64, 44, 28, 8, | |
281 8, 20, 36, 48, 64, 76, 92,104,104, 92, 76, 64, 48, 36, 20, 8, | |
282 4, 16, 24, 36, 44, 56, 64, 76, 76, 64, 56, 44, 36, 24, 16, 4, | |
283 4, 8, 16, 20, 28, 32, 40, 44, 44, 40, 32, 28, 20, 16, 8, 4, | |
284 0, 4, 4, 8, 8, 12, 12, 16, 16, 12, 12, 8, 8, 4, 4, 0, | |
2138 | 285 //error:0.000015 |
286 }; | |
287 #else //64*cos | |
288 static const uint8_t obmc32[1024]={ | |
5659
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
289 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
290 0, 0, 0, 0, 0, 0, 4, 4, 4, 4, 4, 4, 4, 4, 8, 4, 4, 8, 4, 4, 4, 4, 4, 4, 4, 4, 0, 0, 0, 0, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
291 0, 0, 0, 4, 4, 4, 4, 8, 8, 12, 12, 12, 12, 16, 16, 16, 16, 16, 16, 12, 12, 12, 12, 8, 8, 4, 4, 4, 4, 0, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
292 0, 0, 4, 4, 4, 8, 8, 12, 16, 20, 20, 24, 28, 28, 28, 28, 28, 28, 28, 28, 24, 20, 20, 16, 12, 8, 8, 4, 4, 4, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
293 0, 0, 4, 4, 8, 12, 16, 20, 24, 28, 36, 40, 44, 44, 48, 48, 48, 48, 44, 44, 40, 36, 28, 24, 20, 16, 12, 8, 4, 4, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
294 0, 0, 4, 8, 12, 20, 24, 32, 36, 44, 48, 56, 60, 64, 68, 68, 68, 68, 64, 60, 56, 48, 44, 36, 32, 24, 20, 12, 8, 4, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
295 0, 4, 4, 8, 16, 24, 32, 40, 48, 60, 68, 76, 80, 84, 88, 92, 92, 88, 84, 80, 76, 68, 60, 48, 40, 32, 24, 16, 8, 4, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
296 0, 4, 8, 12, 20, 32, 40, 52, 64, 76, 84, 96,104,108,112,116,116,112,108,104, 96, 84, 76, 64, 52, 40, 32, 20, 12, 8, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
297 0, 4, 8, 16, 24, 36, 48, 64, 76, 92,104,116,124,132,136,140,140,136,132,124,116,104, 92, 76, 64, 48, 36, 24, 16, 8, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
298 0, 4, 12, 20, 28, 44, 60, 76, 92,104,120,136,148,156,160,164,164,160,156,148,136,120,104, 92, 76, 60, 44, 28, 20, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
299 0, 4, 12, 20, 36, 48, 68, 84,104,120,140,152,168,176,184,188,188,184,176,168,152,140,120,104, 84, 68, 48, 36, 20, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
300 0, 4, 12, 24, 36, 56, 76, 96,116,136,152,172,184,196,204,208,208,204,196,184,172,152,136,116, 96, 76, 56, 36, 24, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
301 0, 4, 12, 24, 44, 60, 80,104,124,148,168,184,200,212,224,228,228,224,212,200,184,168,148,124,104, 80, 60, 44, 24, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
302 0, 4, 12, 28, 44, 64, 84,108,132,156,176,196,212,228,236,240,240,236,228,212,196,176,156,132,108, 84, 64, 44, 28, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
303 0, 4, 16, 28, 48, 68, 88,112,136,160,184,204,224,236,244,252,252,244,236,224,204,184,160,136,112, 88, 68, 48, 28, 16, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
304 1, 4, 16, 28, 48, 68, 92,116,140,164,188,208,228,240,252,255,255,252,240,228,208,188,164,140,116, 92, 68, 48, 28, 16, 4, 1, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
305 1, 4, 16, 28, 48, 68, 92,116,140,164,188,208,228,240,252,255,255,252,240,228,208,188,164,140,116, 92, 68, 48, 28, 16, 4, 1, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
306 0, 4, 16, 28, 48, 68, 88,112,136,160,184,204,224,236,244,252,252,244,236,224,204,184,160,136,112, 88, 68, 48, 28, 16, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
307 0, 4, 12, 28, 44, 64, 84,108,132,156,176,196,212,228,236,240,240,236,228,212,196,176,156,132,108, 84, 64, 44, 28, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
308 0, 4, 12, 24, 44, 60, 80,104,124,148,168,184,200,212,224,228,228,224,212,200,184,168,148,124,104, 80, 60, 44, 24, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
309 0, 4, 12, 24, 36, 56, 76, 96,116,136,152,172,184,196,204,208,208,204,196,184,172,152,136,116, 96, 76, 56, 36, 24, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
310 0, 4, 12, 20, 36, 48, 68, 84,104,120,140,152,168,176,184,188,188,184,176,168,152,140,120,104, 84, 68, 48, 36, 20, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
311 0, 4, 12, 20, 28, 44, 60, 76, 92,104,120,136,148,156,160,164,164,160,156,148,136,120,104, 92, 76, 60, 44, 28, 20, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
312 0, 4, 8, 16, 24, 36, 48, 64, 76, 92,104,116,124,132,136,140,140,136,132,124,116,104, 92, 76, 64, 48, 36, 24, 16, 8, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
313 0, 4, 8, 12, 20, 32, 40, 52, 64, 76, 84, 96,104,108,112,116,116,112,108,104, 96, 84, 76, 64, 52, 40, 32, 20, 12, 8, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
314 0, 4, 4, 8, 16, 24, 32, 40, 48, 60, 68, 76, 80, 84, 88, 92, 92, 88, 84, 80, 76, 68, 60, 48, 40, 32, 24, 16, 8, 4, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
315 0, 0, 4, 8, 12, 20, 24, 32, 36, 44, 48, 56, 60, 64, 68, 68, 68, 68, 64, 60, 56, 48, 44, 36, 32, 24, 20, 12, 8, 4, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
316 0, 0, 4, 4, 8, 12, 16, 20, 24, 28, 36, 40, 44, 44, 48, 48, 48, 48, 44, 44, 40, 36, 28, 24, 20, 16, 12, 8, 4, 4, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
317 0, 0, 4, 4, 4, 8, 8, 12, 16, 20, 20, 24, 28, 28, 28, 28, 28, 28, 28, 28, 24, 20, 20, 16, 12, 8, 8, 4, 4, 4, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
318 0, 0, 0, 4, 4, 4, 4, 8, 8, 12, 12, 12, 12, 16, 16, 16, 16, 16, 16, 12, 12, 12, 12, 8, 8, 4, 4, 4, 4, 0, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
319 0, 0, 0, 0, 0, 0, 4, 4, 4, 4, 4, 4, 4, 4, 8, 4, 4, 8, 4, 4, 4, 4, 4, 4, 4, 4, 0, 0, 0, 0, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
320 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
2138 | 321 //error:0.000022 |
322 }; | |
323 static const uint8_t obmc16[256]={ | |
5659
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
324 0, 0, 0, 0, 0, 4, 4, 4, 4, 4, 4, 0, 0, 0, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
325 0, 0, 4, 8, 12, 16, 20, 20, 20, 20, 16, 12, 8, 4, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
326 0, 4, 12, 24, 32, 44, 52, 56, 56, 52, 44, 32, 24, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
327 0, 8, 24, 40, 60, 80, 96,104,104, 96, 80, 60, 40, 24, 8, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
328 0, 12, 32, 64, 92,120,140,152,152,140,120, 92, 64, 32, 12, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
329 4, 16, 44, 80,120,156,184,196,196,184,156,120, 80, 44, 16, 4, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
330 4, 20, 52, 96,140,184,216,232,232,216,184,140, 96, 52, 20, 4, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
331 0, 20, 56,104,152,196,232,252,252,232,196,152,104, 56, 20, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
332 0, 20, 56,104,152,196,232,252,252,232,196,152,104, 56, 20, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
333 4, 20, 52, 96,140,184,216,232,232,216,184,140, 96, 52, 20, 4, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
334 4, 16, 44, 80,120,156,184,196,196,184,156,120, 80, 44, 16, 4, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
335 0, 12, 32, 64, 92,120,140,152,152,140,120, 92, 64, 32, 12, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
336 0, 8, 24, 40, 60, 80, 96,104,104, 96, 80, 60, 40, 24, 8, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
337 0, 4, 12, 24, 32, 44, 52, 56, 56, 52, 44, 32, 24, 12, 4, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
338 0, 0, 4, 8, 12, 16, 20, 20, 20, 20, 16, 12, 8, 4, 0, 0, |
ff44e77914ca
scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
michael
parents:
5652
diff
changeset
|
339 0, 0, 0, 0, 0, 4, 4, 4, 4, 4, 4, 0, 0, 0, 0, 0, |
2138 | 340 //error:0.000022 |
341 }; | |
5910 | 342 #endif /* 0 */ |
2138 | 343 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
344 //linear *64 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
345 static const uint8_t obmc8[64]={ |
3206 | 346 4, 12, 20, 28, 28, 20, 12, 4, |
347 12, 36, 60, 84, 84, 60, 36, 12, | |
348 20, 60,100,140,140,100, 60, 20, | |
349 28, 84,140,196,196,140, 84, 28, | |
350 28, 84,140,196,196,140, 84, 28, | |
351 20, 60,100,140,140,100, 60, 20, | |
352 12, 36, 60, 84, 84, 60, 36, 12, | |
353 4, 12, 20, 28, 28, 20, 12, 4, | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
354 //error:0.000000 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
355 }; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
356 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
357 //linear *64 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
358 static const uint8_t obmc4[16]={ |
3206 | 359 16, 48, 48, 16, |
360 48,144,144, 48, | |
361 48,144,144, 48, | |
362 16, 48, 48, 16, | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
363 //error:0.000000 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
364 }; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
365 |
7129 | 366 static const uint8_t * const obmc_tab[4]={ |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
367 obmc32, obmc16, obmc8, obmc4 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
368 }; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
369 |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
370 static int scale_mv_ref[MAX_REF_FRAMES][MAX_REF_FRAMES]; |
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
371 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
372 typedef struct BlockNode{ |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
373 int16_t mx; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
374 int16_t my; |
3314 | 375 uint8_t ref; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
376 uint8_t color[3]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
377 uint8_t type; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
378 //#define TYPE_SPLIT 1 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
379 #define BLOCK_INTRA 1 |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
380 #define BLOCK_OPT 2 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
381 //#define TYPE_NOCOLOR 4 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
382 uint8_t level; //FIXME merge into type? |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
383 }BlockNode; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
384 |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
385 static const BlockNode null_block= { //FIXME add border maybe |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
386 .color= {128,128,128}, |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
387 .mx= 0, |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
388 .my= 0, |
3314 | 389 .ref= 0, |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
390 .type= 0, |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
391 .level= 0, |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
392 }; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
393 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
394 #define LOG2_MB_SIZE 4 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
395 #define MB_SIZE (1<<LOG2_MB_SIZE) |
5575 | 396 #define ENCODER_EXTRA_BITS 4 |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
397 #define HTAPS_MAX 8 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
398 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
399 typedef struct x_and_coeff{ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
400 int16_t x; |
2596 | 401 uint16_t coeff; |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
402 } x_and_coeff; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
403 |
2138 | 404 typedef struct SubBand{ |
405 int level; | |
406 int stride; | |
407 int width; | |
408 int height; | |
6412 | 409 int qlog; ///< log(qscale)/log[2^(1/6)] |
2138 | 410 DWTELEM *buf; |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
411 IDWTELEM *ibuf; |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
412 int buf_x_offset; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
413 int buf_y_offset; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
414 int stride_line; ///< Stride measured in lines, not pixels. |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
415 x_and_coeff * x_coeff; |
2138 | 416 struct SubBand *parent; |
417 uint8_t state[/*7*2*/ 7 + 512][32]; | |
418 }SubBand; | |
419 | |
420 typedef struct Plane{ | |
421 int width; | |
422 int height; | |
423 SubBand band[MAX_DECOMPOSITIONS][4]; | |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
424 |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
425 int htaps; |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
426 int8_t hcoeff[HTAPS_MAX/2]; |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
427 int diag_mc; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
428 int fast_mc; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
429 |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
430 int last_htaps; |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
431 int8_t last_hcoeff[HTAPS_MAX/2]; |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
432 int last_diag_mc; |
2138 | 433 }Plane; |
434 | |
435 typedef struct SnowContext{ | |
6412 | 436 // MpegEncContext m; // needed for motion estimation, should not be used for anything else, the idea is to eventually make the motion estimation independent of MpegEncContext, so this will be removed then (FIXME/XXX) |
2138 | 437 |
438 AVCodecContext *avctx; | |
2335 | 439 RangeCoder c; |
2138 | 440 DSPContext dsp; |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
441 AVFrame new_picture; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
442 AVFrame input_picture; ///< new_picture with the internal linesizes |
2138 | 443 AVFrame current_picture; |
3314 | 444 AVFrame last_picture[MAX_REF_FRAMES]; |
5633
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
445 uint8_t *halfpel_plane[MAX_REF_FRAMES][4][4]; |
2138 | 446 AVFrame mconly_picture; |
447 // uint8_t q_context[16]; | |
448 uint8_t header_state[32]; | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
449 uint8_t block_state[128 + 32*128]; |
2138 | 450 int keyframe; |
2199 | 451 int always_reset; |
2138 | 452 int version; |
453 int spatial_decomposition_type; | |
4331
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
454 int last_spatial_decomposition_type; |
2138 | 455 int temporal_decomposition_type; |
456 int spatial_decomposition_count; | |
5670
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
457 int last_spatial_decomposition_count; |
2138 | 458 int temporal_decomposition_count; |
3314 | 459 int max_ref_frames; |
460 int ref_frames; | |
461 int16_t (*ref_mvs[MAX_REF_FRAMES])[2]; | |
462 uint32_t *ref_scores[MAX_REF_FRAMES]; | |
2138 | 463 DWTELEM *spatial_dwt_buffer; |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
464 IDWTELEM *spatial_idwt_buffer; |
2138 | 465 int colorspace_type; |
466 int chroma_h_shift; | |
467 int chroma_v_shift; | |
468 int spatial_scalability; | |
469 int qlog; | |
4331
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
470 int last_qlog; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
471 int lambda; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
472 int lambda2; |
3313 | 473 int pass1_rc; |
2138 | 474 int mv_scale; |
4331
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
475 int last_mv_scale; |
2138 | 476 int qbias; |
4331
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
477 int last_qbias; |
2138 | 478 #define QBIAS_SHIFT 3 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
479 int b_width; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
480 int b_height; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
481 int block_max_depth; |
4331
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
482 int last_block_max_depth; |
2138 | 483 Plane plane[MAX_PLANES]; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
484 BlockNode *block; |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
485 #define ME_CACHE_SIZE 1024 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
486 int me_cache[ME_CACHE_SIZE]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
487 int me_cache_generation; |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
488 slice_buffer sb; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
489 |
6412 | 490 MpegEncContext m; // needed for motion estimation, should not be used for anything else, the idea is to eventually make the motion estimation independent of MpegEncContext, so this will be removed then (FIXME/XXX) |
2138 | 491 }SnowContext; |
492 | |
2562 | 493 typedef struct { |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
494 IDWTELEM *b0; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
495 IDWTELEM *b1; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
496 IDWTELEM *b2; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
497 IDWTELEM *b3; |
2562 | 498 int y; |
499 } dwt_compose_t; | |
500 | |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
501 #define slice_buffer_get_line(slice_buf, line_num) ((slice_buf)->line[line_num] ? (slice_buf)->line[line_num] : slice_buffer_load_line((slice_buf), (line_num))) |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
502 //#define slice_buffer_get_line(slice_buf, line_num) (slice_buffer_load_line((slice_buf), (line_num))) |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
503 |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
504 static void iterative_me(SnowContext *s); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
505 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
506 static void slice_buffer_init(slice_buffer * buf, int line_count, int max_allocated_lines, int line_width, IDWTELEM * base_buffer) |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
507 { |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
508 int i; |
2967 | 509 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
510 buf->base_buffer = base_buffer; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
511 buf->line_count = line_count; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
512 buf->line_width = line_width; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
513 buf->data_count = max_allocated_lines; |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
514 buf->line = av_mallocz (sizeof(IDWTELEM *) * line_count); |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
515 buf->data_stack = av_malloc (sizeof(IDWTELEM *) * max_allocated_lines); |
2967 | 516 |
6474
e6995f3fbf7f
cosmetics: Normalize {} placement after for, while, if.
diego
parents:
6437
diff
changeset
|
517 for(i = 0; i < max_allocated_lines; i++){ |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
518 buf->data_stack[i] = av_malloc (sizeof(IDWTELEM) * line_width); |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
519 } |
2967 | 520 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
521 buf->data_stack_top = max_allocated_lines - 1; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
522 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
523 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
524 static IDWTELEM * slice_buffer_load_line(slice_buffer * buf, int line) |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
525 { |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
526 int offset; |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
527 IDWTELEM * buffer; |
2967 | 528 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
529 assert(buf->data_stack_top >= 0); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
530 // assert(!buf->line[line]); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
531 if (buf->line[line]) |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
532 return buf->line[line]; |
2967 | 533 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
534 offset = buf->line_width * line; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
535 buffer = buf->data_stack[buf->data_stack_top]; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
536 buf->data_stack_top--; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
537 buf->line[line] = buffer; |
2967 | 538 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
539 return buffer; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
540 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
541 |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
542 static void slice_buffer_release(slice_buffer * buf, int line) |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
543 { |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
544 int offset; |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
545 IDWTELEM * buffer; |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
546 |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
547 assert(line >= 0 && line < buf->line_count); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
548 assert(buf->line[line]); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
549 |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
550 offset = buf->line_width * line; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
551 buffer = buf->line[line]; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
552 buf->data_stack_top++; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
553 buf->data_stack[buf->data_stack_top] = buffer; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
554 buf->line[line] = NULL; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
555 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
556 |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
557 static void slice_buffer_flush(slice_buffer * buf) |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
558 { |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
559 int i; |
6474
e6995f3fbf7f
cosmetics: Normalize {} placement after for, while, if.
diego
parents:
6437
diff
changeset
|
560 for(i = 0; i < buf->line_count; i++){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
561 if (buf->line[i]) |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
562 slice_buffer_release(buf, i); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
563 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
564 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
565 |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
566 static void slice_buffer_destroy(slice_buffer * buf) |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
567 { |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
568 int i; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
569 slice_buffer_flush(buf); |
2967 | 570 |
6474
e6995f3fbf7f
cosmetics: Normalize {} placement after for, while, if.
diego
parents:
6437
diff
changeset
|
571 for(i = buf->data_count - 1; i >= 0; i--){ |
3190 | 572 av_freep(&buf->data_stack[i]); |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
573 } |
3190 | 574 av_freep(&buf->data_stack); |
575 av_freep(&buf->line); | |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
576 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
577 |
2979 | 578 #ifdef __sgi |
2368
a7ac68734a91
fix for build on IRIX by (Michel Bardiaux {mbardiaux peaktime be})
michael
parents:
2335
diff
changeset
|
579 // Avoid a name clash on SGI IRIX |
2979 | 580 #undef qexp |
2368
a7ac68734a91
fix for build on IRIX by (Michel Bardiaux {mbardiaux peaktime be})
michael
parents:
2335
diff
changeset
|
581 #endif |
2246 | 582 #define QEXPSHIFT (7-FRAC_BITS+8) //FIXME try to change this to 0 |
2600 | 583 static uint8_t qexp[QROOT]; |
2138 | 584 |
585 static inline int mirror(int v, int m){ | |
2998 | 586 while((unsigned)v > (unsigned)m){ |
587 v=-v; | |
588 if(v<0) v+= 2*m; | |
589 } | |
590 return v; | |
2138 | 591 } |
592 | |
2335 | 593 static inline void put_symbol(RangeCoder *c, uint8_t *state, int v, int is_signed){ |
2138 | 594 int i; |
595 | |
596 if(v){ | |
4001 | 597 const int a= FFABS(v); |
2138 | 598 const int e= av_log2(a); |
599 #if 1 | |
2967 | 600 const int el= FFMIN(e, 10); |
2335 | 601 put_rac(c, state+0, 0); |
2138 | 602 |
603 for(i=0; i<el; i++){ | |
2335 | 604 put_rac(c, state+1+i, 1); //1..10 |
2138 | 605 } |
606 for(; i<e; i++){ | |
2335 | 607 put_rac(c, state+1+9, 1); //1..10 |
2138 | 608 } |
2335 | 609 put_rac(c, state+1+FFMIN(i,9), 0); |
2138 | 610 |
611 for(i=e-1; i>=el; i--){ | |
2335 | 612 put_rac(c, state+22+9, (a>>i)&1); //22..31 |
2138 | 613 } |
614 for(; i>=0; i--){ | |
2335 | 615 put_rac(c, state+22+i, (a>>i)&1); //22..31 |
2138 | 616 } |
617 | |
618 if(is_signed) | |
2335 | 619 put_rac(c, state+11 + el, v < 0); //11..21 |
2138 | 620 #else |
2967 | 621 |
2335 | 622 put_rac(c, state+0, 0); |
2138 | 623 if(e<=9){ |
624 for(i=0; i<e; i++){ | |
2335 | 625 put_rac(c, state+1+i, 1); //1..10 |
2138 | 626 } |
2335 | 627 put_rac(c, state+1+i, 0); |
2138 | 628 |
629 for(i=e-1; i>=0; i--){ | |
2335 | 630 put_rac(c, state+22+i, (a>>i)&1); //22..31 |
2138 | 631 } |
632 | |
633 if(is_signed) | |
2335 | 634 put_rac(c, state+11 + e, v < 0); //11..21 |
2138 | 635 }else{ |
636 for(i=0; i<e; i++){ | |
2335 | 637 put_rac(c, state+1+FFMIN(i,9), 1); //1..10 |
2138 | 638 } |
2335 | 639 put_rac(c, state+1+FFMIN(i,9), 0); |
2138 | 640 |
641 for(i=e-1; i>=0; i--){ | |
2335 | 642 put_rac(c, state+22+FFMIN(i,9), (a>>i)&1); //22..31 |
2138 | 643 } |
644 | |
645 if(is_signed) | |
2335 | 646 put_rac(c, state+11 + FFMIN(e,10), v < 0); //11..21 |
2138 | 647 } |
5910 | 648 #endif /* 1 */ |
2138 | 649 }else{ |
2335 | 650 put_rac(c, state+0, 1); |
2138 | 651 } |
652 } | |
653 | |
2335 | 654 static inline int get_symbol(RangeCoder *c, uint8_t *state, int is_signed){ |
655 if(get_rac(c, state+0)) | |
2138 | 656 return 0; |
657 else{ | |
2240
c46fed9b7575
simplify getsymbol patch by (Loren Merritt <lorenm at u dot washington dot edu>)
michael
parents:
2232
diff
changeset
|
658 int i, e, a; |
c46fed9b7575
simplify getsymbol patch by (Loren Merritt <lorenm at u dot washington dot edu>)
michael
parents:
2232
diff
changeset
|
659 e= 0; |
2335 | 660 while(get_rac(c, state+1 + FFMIN(e,9))){ //1..10 |
2240
c46fed9b7575
simplify getsymbol patch by (Loren Merritt <lorenm at u dot washington dot edu>)
michael
parents:
2232
diff
changeset
|
661 e++; |
2138 | 662 } |
663 | |
2240
c46fed9b7575
simplify getsymbol patch by (Loren Merritt <lorenm at u dot washington dot edu>)
michael
parents:
2232
diff
changeset
|
664 a= 1; |
c46fed9b7575
simplify getsymbol patch by (Loren Merritt <lorenm at u dot washington dot edu>)
michael
parents:
2232
diff
changeset
|
665 for(i=e-1; i>=0; i--){ |
2335 | 666 a += a + get_rac(c, state+22 + FFMIN(i,9)); //22..31 |
2240
c46fed9b7575
simplify getsymbol patch by (Loren Merritt <lorenm at u dot washington dot edu>)
michael
parents:
2232
diff
changeset
|
667 } |
c46fed9b7575
simplify getsymbol patch by (Loren Merritt <lorenm at u dot washington dot edu>)
michael
parents:
2232
diff
changeset
|
668 |
2335 | 669 if(is_signed && get_rac(c, state+11 + FFMIN(e,10))) //11..21 |
2138 | 670 return -a; |
671 else | |
672 return a; | |
673 } | |
674 } | |
675 | |
2335 | 676 static inline void put_symbol2(RangeCoder *c, uint8_t *state, int v, int log2){ |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
677 int i; |
2159
7f42295c1517
improved magnitude coding, 0.2% lower bitrate (foreman@352x288 qscale 1 and 8)
michael
parents:
2156
diff
changeset
|
678 int r= log2>=0 ? 1<<log2 : 1; |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
679 |
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
680 assert(v>=0); |
2159
7f42295c1517
improved magnitude coding, 0.2% lower bitrate (foreman@352x288 qscale 1 and 8)
michael
parents:
2156
diff
changeset
|
681 assert(log2>=-4); |
7f42295c1517
improved magnitude coding, 0.2% lower bitrate (foreman@352x288 qscale 1 and 8)
michael
parents:
2156
diff
changeset
|
682 |
7f42295c1517
improved magnitude coding, 0.2% lower bitrate (foreman@352x288 qscale 1 and 8)
michael
parents:
2156
diff
changeset
|
683 while(v >= r){ |
2335 | 684 put_rac(c, state+4+log2, 1); |
2159
7f42295c1517
improved magnitude coding, 0.2% lower bitrate (foreman@352x288 qscale 1 and 8)
michael
parents:
2156
diff
changeset
|
685 v -= r; |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
686 log2++; |
2159
7f42295c1517
improved magnitude coding, 0.2% lower bitrate (foreman@352x288 qscale 1 and 8)
michael
parents:
2156
diff
changeset
|
687 if(log2>0) r+=r; |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
688 } |
2335 | 689 put_rac(c, state+4+log2, 0); |
2967 | 690 |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
691 for(i=log2-1; i>=0; i--){ |
2335 | 692 put_rac(c, state+31-i, (v>>i)&1); |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
693 } |
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
694 } |
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
695 |
2335 | 696 static inline int get_symbol2(RangeCoder *c, uint8_t *state, int log2){ |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
697 int i; |
2159
7f42295c1517
improved magnitude coding, 0.2% lower bitrate (foreman@352x288 qscale 1 and 8)
michael
parents:
2156
diff
changeset
|
698 int r= log2>=0 ? 1<<log2 : 1; |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
699 int v=0; |
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
700 |
2159
7f42295c1517
improved magnitude coding, 0.2% lower bitrate (foreman@352x288 qscale 1 and 8)
michael
parents:
2156
diff
changeset
|
701 assert(log2>=-4); |
7f42295c1517
improved magnitude coding, 0.2% lower bitrate (foreman@352x288 qscale 1 and 8)
michael
parents:
2156
diff
changeset
|
702 |
2335 | 703 while(get_rac(c, state+4+log2)){ |
2159
7f42295c1517
improved magnitude coding, 0.2% lower bitrate (foreman@352x288 qscale 1 and 8)
michael
parents:
2156
diff
changeset
|
704 v+= r; |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
705 log2++; |
2159
7f42295c1517
improved magnitude coding, 0.2% lower bitrate (foreman@352x288 qscale 1 and 8)
michael
parents:
2156
diff
changeset
|
706 if(log2>0) r+=r; |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
707 } |
2967 | 708 |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
709 for(i=log2-1; i>=0; i--){ |
2335 | 710 v+= get_rac(c, state+31-i)<<i; |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
711 } |
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
712 |
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
713 return v; |
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
714 } |
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
715 |
5627 | 716 static av_always_inline void |
717 lift(DWTELEM *dst, DWTELEM *src, DWTELEM *ref, | |
718 int dst_step, int src_step, int ref_step, | |
719 int width, int mul, int add, int shift, | |
720 int highpass, int inverse){ | |
2138 | 721 const int mirror_left= !highpass; |
722 const int mirror_right= (width&1) ^ highpass; | |
723 const int w= (width>>1) - 1 + (highpass & width); | |
724 int i; | |
725 | |
726 #define LIFT(src, ref, inv) ((src) + ((inv) ? - (ref) : + (ref))) | |
727 if(mirror_left){ | |
728 dst[0] = LIFT(src[0], ((mul*2*ref[0]+add)>>shift), inverse); | |
729 dst += dst_step; | |
730 src += src_step; | |
731 } | |
2967 | 732 |
2138 | 733 for(i=0; i<w; i++){ |
5627 | 734 dst[i*dst_step] = |
735 LIFT(src[i*src_step], | |
736 ((mul*(ref[i*ref_step] + ref[(i+1)*ref_step])+add)>>shift), | |
737 inverse); | |
2138 | 738 } |
2967 | 739 |
2138 | 740 if(mirror_right){ |
5627 | 741 dst[w*dst_step] = |
742 LIFT(src[w*src_step], | |
743 ((mul*2*ref[w*ref_step]+add)>>shift), | |
744 inverse); | |
2138 | 745 } |
746 } | |
747 | |
5627 | 748 static av_always_inline void |
749 inv_lift(IDWTELEM *dst, IDWTELEM *src, IDWTELEM *ref, | |
750 int dst_step, int src_step, int ref_step, | |
751 int width, int mul, int add, int shift, | |
752 int highpass, int inverse){ | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
753 const int mirror_left= !highpass; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
754 const int mirror_right= (width&1) ^ highpass; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
755 const int w= (width>>1) - 1 + (highpass & width); |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
756 int i; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
757 |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
758 #define LIFT(src, ref, inv) ((src) + ((inv) ? - (ref) : + (ref))) |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
759 if(mirror_left){ |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
760 dst[0] = LIFT(src[0], ((mul*2*ref[0]+add)>>shift), inverse); |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
761 dst += dst_step; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
762 src += src_step; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
763 } |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
764 |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
765 for(i=0; i<w; i++){ |
5627 | 766 dst[i*dst_step] = |
767 LIFT(src[i*src_step], | |
768 ((mul*(ref[i*ref_step] + ref[(i+1)*ref_step])+add)>>shift), | |
769 inverse); | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
770 } |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
771 |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
772 if(mirror_right){ |
5627 | 773 dst[w*dst_step] = |
774 LIFT(src[w*src_step], | |
775 ((mul*2*ref[w*ref_step]+add)>>shift), | |
776 inverse); | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
777 } |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
778 } |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
779 |
3198
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
780 #ifndef liftS |
5627 | 781 static av_always_inline void |
782 liftS(DWTELEM *dst, DWTELEM *src, DWTELEM *ref, | |
783 int dst_step, int src_step, int ref_step, | |
784 int width, int mul, int add, int shift, | |
785 int highpass, int inverse){ | |
2602
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
786 const int mirror_left= !highpass; |
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
787 const int mirror_right= (width&1) ^ highpass; |
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
788 const int w= (width>>1) - 1 + (highpass & width); |
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
789 int i; |
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
790 |
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
791 assert(shift == 4); |
5627 | 792 #define LIFTS(src, ref, inv) \ |
793 ((inv) ? \ | |
794 (src) + (((ref) + 4*(src))>>shift): \ | |
795 -((-16*(src) + (ref) + add/4 + 1 + (5<<25))/(5*4) - (1<<23))) | |
2602
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
796 if(mirror_left){ |
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
797 dst[0] = LIFTS(src[0], mul*2*ref[0]+add, inverse); |
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
798 dst += dst_step; |
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
799 src += src_step; |
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
800 } |
2967 | 801 |
2602
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
802 for(i=0; i<w; i++){ |
5627 | 803 dst[i*dst_step] = |
804 LIFTS(src[i*src_step], | |
805 mul*(ref[i*ref_step] + ref[(i+1)*ref_step])+add, | |
806 inverse); | |
2602
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
807 } |
2967 | 808 |
2602
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
809 if(mirror_right){ |
5627 | 810 dst[w*dst_step] = |
811 LIFTS(src[w*src_step], mul*2*ref[w*ref_step]+add, inverse); | |
2602
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
812 } |
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
813 } |
5627 | 814 static av_always_inline void |
815 inv_liftS(IDWTELEM *dst, IDWTELEM *src, IDWTELEM *ref, | |
816 int dst_step, int src_step, int ref_step, | |
817 int width, int mul, int add, int shift, | |
818 int highpass, int inverse){ | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
819 const int mirror_left= !highpass; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
820 const int mirror_right= (width&1) ^ highpass; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
821 const int w= (width>>1) - 1 + (highpass & width); |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
822 int i; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
823 |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
824 assert(shift == 4); |
5627 | 825 #define LIFTS(src, ref, inv) \ |
826 ((inv) ? \ | |
827 (src) + (((ref) + 4*(src))>>shift): \ | |
828 -((-16*(src) + (ref) + add/4 + 1 + (5<<25))/(5*4) - (1<<23))) | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
829 if(mirror_left){ |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
830 dst[0] = LIFTS(src[0], mul*2*ref[0]+add, inverse); |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
831 dst += dst_step; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
832 src += src_step; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
833 } |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
834 |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
835 for(i=0; i<w; i++){ |
5627 | 836 dst[i*dst_step] = |
837 LIFTS(src[i*src_step], | |
838 mul*(ref[i*ref_step] + ref[(i+1)*ref_step])+add, | |
839 inverse); | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
840 } |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
841 |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
842 if(mirror_right){ |
5627 | 843 dst[w*dst_step] = |
844 LIFTS(src[w*src_step], mul*2*ref[w*ref_step]+add, inverse); | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
845 } |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
846 } |
6413 | 847 #endif /* ! liftS */ |
2602
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
848 |
2241
c26038875ebc
consistent use of types patch by (D Richard Felker III <dalias at aerifal dot cx>)
michael
parents:
2240
diff
changeset
|
849 static void horizontal_decompose53i(DWTELEM *b, int width){ |
c26038875ebc
consistent use of types patch by (D Richard Felker III <dalias at aerifal dot cx>)
michael
parents:
2240
diff
changeset
|
850 DWTELEM temp[width]; |
2138 | 851 const int width2= width>>1; |
2893 | 852 int x; |
2138 | 853 const int w2= (width+1)>>1; |
854 | |
855 for(x=0; x<width2; x++){ | |
856 temp[x ]= b[2*x ]; | |
857 temp[x+w2]= b[2*x + 1]; | |
858 } | |
859 if(width&1) | |
860 temp[x ]= b[2*x ]; | |
861 #if 0 | |
2893 | 862 { |
863 int A1,A2,A3,A4; | |
2138 | 864 A2= temp[1 ]; |
865 A4= temp[0 ]; | |
866 A1= temp[0+width2]; | |
867 A1 -= (A2 + A4)>>1; | |
868 A4 += (A1 + 1)>>1; | |
869 b[0+width2] = A1; | |
870 b[0 ] = A4; | |
871 for(x=1; x+1<width2; x+=2){ | |
872 A3= temp[x+width2]; | |
873 A4= temp[x+1 ]; | |
874 A3 -= (A2 + A4)>>1; | |
875 A2 += (A1 + A3 + 2)>>2; | |
876 b[x+width2] = A3; | |
877 b[x ] = A2; | |
878 | |
879 A1= temp[x+1+width2]; | |
880 A2= temp[x+2 ]; | |
881 A1 -= (A2 + A4)>>1; | |
882 A4 += (A1 + A3 + 2)>>2; | |
883 b[x+1+width2] = A1; | |
884 b[x+1 ] = A4; | |
885 } | |
886 A3= temp[width-1]; | |
887 A3 -= A2; | |
888 A2 += (A1 + A3 + 2)>>2; | |
889 b[width -1] = A3; | |
890 b[width2-1] = A2; | |
2893 | 891 } |
2967 | 892 #else |
2138 | 893 lift(b+w2, temp+w2, temp, 1, 1, 1, width, -1, 0, 1, 1, 0); |
894 lift(b , temp , b+w2, 1, 1, 1, width, 1, 2, 2, 0, 0); | |
6413 | 895 #endif /* 0 */ |
2138 | 896 } |
897 | |
2241
c26038875ebc
consistent use of types patch by (D Richard Felker III <dalias at aerifal dot cx>)
michael
parents:
2240
diff
changeset
|
898 static void vertical_decompose53iH0(DWTELEM *b0, DWTELEM *b1, DWTELEM *b2, int width){ |
2138 | 899 int i; |
2967 | 900 |
2138 | 901 for(i=0; i<width; i++){ |
902 b1[i] -= (b0[i] + b2[i])>>1; | |
903 } | |
904 } | |
905 | |
2241
c26038875ebc
consistent use of types patch by (D Richard Felker III <dalias at aerifal dot cx>)
michael
parents:
2240
diff
changeset
|
906 static void vertical_decompose53iL0(DWTELEM *b0, DWTELEM *b1, DWTELEM *b2, int width){ |
2138 | 907 int i; |
2967 | 908 |
2138 | 909 for(i=0; i<width; i++){ |
910 b1[i] += (b0[i] + b2[i] + 2)>>2; | |
911 } | |
912 } | |
913 | |
2241
c26038875ebc
consistent use of types patch by (D Richard Felker III <dalias at aerifal dot cx>)
michael
parents:
2240
diff
changeset
|
914 static void spatial_decompose53i(DWTELEM *buffer, int width, int height, int stride){ |
2198 | 915 int y; |
2138 | 916 DWTELEM *b0= buffer + mirror(-2-1, height-1)*stride; |
917 DWTELEM *b1= buffer + mirror(-2 , height-1)*stride; | |
2967 | 918 |
2138 | 919 for(y=-2; y<height; y+=2){ |
920 DWTELEM *b2= buffer + mirror(y+1, height-1)*stride; | |
921 DWTELEM *b3= buffer + mirror(y+2, height-1)*stride; | |
922 | |
2998 | 923 if(y+1<(unsigned)height) horizontal_decompose53i(b2, width); |
924 if(y+2<(unsigned)height) horizontal_decompose53i(b3, width); | |
6405 | 925 |
2998 | 926 if(y+1<(unsigned)height) vertical_decompose53iH0(b1, b2, b3, width); |
927 if(y+0<(unsigned)height) vertical_decompose53iL0(b0, b1, b2, width); | |
2967 | 928 |
2138 | 929 b0=b2; |
930 b1=b3; | |
931 } | |
932 } | |
933 | |
2241
c26038875ebc
consistent use of types patch by (D Richard Felker III <dalias at aerifal dot cx>)
michael
parents:
2240
diff
changeset
|
934 static void horizontal_decompose97i(DWTELEM *b, int width){ |
c26038875ebc
consistent use of types patch by (D Richard Felker III <dalias at aerifal dot cx>)
michael
parents:
2240
diff
changeset
|
935 DWTELEM temp[width]; |
2138 | 936 const int w2= (width+1)>>1; |
937 | |
5565
93082c591c8b
Change rounding of the horizontal DWT to match the vertical one.
michael
parents:
5551
diff
changeset
|
938 lift (temp+w2, b +1, b , 1, 2, 2, width, W_AM, W_AO, W_AS, 1, 1); |
93082c591c8b
Change rounding of the horizontal DWT to match the vertical one.
michael
parents:
5551
diff
changeset
|
939 liftS(temp , b , temp+w2, 1, 2, 1, width, W_BM, W_BO, W_BS, 0, 0); |
5589
946c2db0a093
cleanup (remove some old experimentation related code)
michael
parents:
5588
diff
changeset
|
940 lift (b +w2, temp+w2, temp , 1, 1, 1, width, W_CM, W_CO, W_CS, 1, 0); |
2138 | 941 lift (b , temp , b +w2, 1, 1, 1, width, W_DM, W_DO, W_DS, 0, 0); |
942 } | |
943 | |
944 | |
2241
c26038875ebc
consistent use of types patch by (D Richard Felker III <dalias at aerifal dot cx>)
michael
parents:
2240
diff
changeset
|
945 static void vertical_decompose97iH0(DWTELEM *b0, DWTELEM *b1, DWTELEM *b2, int width){ |
2138 | 946 int i; |
2967 | 947 |
2138 | 948 for(i=0; i<width; i++){ |
949 b1[i] -= (W_AM*(b0[i] + b2[i])+W_AO)>>W_AS; | |
950 } | |
951 } | |
952 | |
2241
c26038875ebc
consistent use of types patch by (D Richard Felker III <dalias at aerifal dot cx>)
michael
parents:
2240
diff
changeset
|
953 static void vertical_decompose97iH1(DWTELEM *b0, DWTELEM *b1, DWTELEM *b2, int width){ |
2138 | 954 int i; |
2967 | 955 |
2138 | 956 for(i=0; i<width; i++){ |
957 b1[i] += (W_CM*(b0[i] + b2[i])+W_CO)>>W_CS; | |
958 } | |
959 } | |
960 | |
2241
c26038875ebc
consistent use of types patch by (D Richard Felker III <dalias at aerifal dot cx>)
michael
parents:
2240
diff
changeset
|
961 static void vertical_decompose97iL0(DWTELEM *b0, DWTELEM *b1, DWTELEM *b2, int width){ |
2138 | 962 int i; |
2967 | 963 |
2138 | 964 for(i=0; i<width; i++){ |
2602
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
965 #ifdef liftS |
2138 | 966 b1[i] -= (W_BM*(b0[i] + b2[i])+W_BO)>>W_BS; |
2602
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
967 #else |
5565
93082c591c8b
Change rounding of the horizontal DWT to match the vertical one.
michael
parents:
5551
diff
changeset
|
968 b1[i] = (16*4*b1[i] - 4*(b0[i] + b2[i]) + W_BO*5 + (5<<27)) / (5*16) - (1<<23); |
2602
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
969 #endif |
2138 | 970 } |
971 } | |
972 | |
2241
c26038875ebc
consistent use of types patch by (D Richard Felker III <dalias at aerifal dot cx>)
michael
parents:
2240
diff
changeset
|
973 static void vertical_decompose97iL1(DWTELEM *b0, DWTELEM *b1, DWTELEM *b2, int width){ |
2138 | 974 int i; |
2967 | 975 |
2138 | 976 for(i=0; i<width; i++){ |
977 b1[i] += (W_DM*(b0[i] + b2[i])+W_DO)>>W_DS; | |
978 } | |
979 } | |
980 | |
2241
c26038875ebc
consistent use of types patch by (D Richard Felker III <dalias at aerifal dot cx>)
michael
parents:
2240
diff
changeset
|
981 static void spatial_decompose97i(DWTELEM *buffer, int width, int height, int stride){ |
2198 | 982 int y; |
2138 | 983 DWTELEM *b0= buffer + mirror(-4-1, height-1)*stride; |
984 DWTELEM *b1= buffer + mirror(-4 , height-1)*stride; | |
985 DWTELEM *b2= buffer + mirror(-4+1, height-1)*stride; | |
986 DWTELEM *b3= buffer + mirror(-4+2, height-1)*stride; | |
2967 | 987 |
2138 | 988 for(y=-4; y<height; y+=2){ |
989 DWTELEM *b4= buffer + mirror(y+3, height-1)*stride; | |
990 DWTELEM *b5= buffer + mirror(y+4, height-1)*stride; | |
991 | |
2998 | 992 if(y+3<(unsigned)height) horizontal_decompose97i(b4, width); |
993 if(y+4<(unsigned)height) horizontal_decompose97i(b5, width); | |
6405 | 994 |
2998 | 995 if(y+3<(unsigned)height) vertical_decompose97iH0(b3, b4, b5, width); |
996 if(y+2<(unsigned)height) vertical_decompose97iL0(b2, b3, b4, width); | |
997 if(y+1<(unsigned)height) vertical_decompose97iH1(b1, b2, b3, width); | |
998 if(y+0<(unsigned)height) vertical_decompose97iL1(b0, b1, b2, width); | |
2138 | 999 |
1000 b0=b2; | |
1001 b1=b3; | |
1002 b2=b4; | |
1003 b3=b5; | |
1004 } | |
1005 } | |
1006 | |
2241
c26038875ebc
consistent use of types patch by (D Richard Felker III <dalias at aerifal dot cx>)
michael
parents:
2240
diff
changeset
|
1007 void ff_spatial_dwt(DWTELEM *buffer, int width, int height, int stride, int type, int decomposition_count){ |
2138 | 1008 int level; |
2967 | 1009 |
2164 | 1010 for(level=0; level<decomposition_count; level++){ |
1011 switch(type){ | |
3326 | 1012 case DWT_97: spatial_decompose97i(buffer, width>>level, height>>level, stride<<level); break; |
1013 case DWT_53: spatial_decompose53i(buffer, width>>level, height>>level, stride<<level); break; | |
2138 | 1014 } |
1015 } | |
1016 } | |
1017 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1018 static void horizontal_compose53i(IDWTELEM *b, int width){ |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1019 IDWTELEM temp[width]; |
2138 | 1020 const int width2= width>>1; |
1021 const int w2= (width+1)>>1; | |
2893 | 1022 int x; |
2138 | 1023 |
1024 #if 0 | |
2893 | 1025 int A1,A2,A3,A4; |
2138 | 1026 A2= temp[1 ]; |
1027 A4= temp[0 ]; | |
1028 A1= temp[0+width2]; | |
1029 A1 -= (A2 + A4)>>1; | |
1030 A4 += (A1 + 1)>>1; | |
1031 b[0+width2] = A1; | |
1032 b[0 ] = A4; | |
1033 for(x=1; x+1<width2; x+=2){ | |
1034 A3= temp[x+width2]; | |
1035 A4= temp[x+1 ]; | |
1036 A3 -= (A2 + A4)>>1; | |
1037 A2 += (A1 + A3 + 2)>>2; | |
1038 b[x+width2] = A3; | |
1039 b[x ] = A2; | |
1040 | |
1041 A1= temp[x+1+width2]; | |
1042 A2= temp[x+2 ]; | |
1043 A1 -= (A2 + A4)>>1; | |
1044 A4 += (A1 + A3 + 2)>>2; | |
1045 b[x+1+width2] = A1; | |
1046 b[x+1 ] = A4; | |
1047 } | |
1048 A3= temp[width-1]; | |
1049 A3 -= A2; | |
1050 A2 += (A1 + A3 + 2)>>2; | |
1051 b[width -1] = A3; | |
1052 b[width2-1] = A2; | |
2967 | 1053 #else |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1054 inv_lift(temp , b , b+w2, 1, 1, 1, width, 1, 2, 2, 0, 1); |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1055 inv_lift(temp+w2, b+w2, temp, 1, 1, 1, width, -1, 0, 1, 1, 1); |
5910 | 1056 #endif /* 0 */ |
2138 | 1057 for(x=0; x<width2; x++){ |
1058 b[2*x ]= temp[x ]; | |
1059 b[2*x + 1]= temp[x+w2]; | |
1060 } | |
1061 if(width&1) | |
1062 b[2*x ]= temp[x ]; | |
1063 } | |
1064 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1065 static void vertical_compose53iH0(IDWTELEM *b0, IDWTELEM *b1, IDWTELEM *b2, int width){ |
2138 | 1066 int i; |
2967 | 1067 |
2138 | 1068 for(i=0; i<width; i++){ |
1069 b1[i] += (b0[i] + b2[i])>>1; | |
1070 } | |
1071 } | |
1072 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1073 static void vertical_compose53iL0(IDWTELEM *b0, IDWTELEM *b1, IDWTELEM *b2, int width){ |
2138 | 1074 int i; |
2967 | 1075 |
2138 | 1076 for(i=0; i<width; i++){ |
1077 b1[i] -= (b0[i] + b2[i] + 2)>>2; | |
1078 } | |
1079 } | |
1080 | |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1081 static void spatial_compose53i_buffered_init(dwt_compose_t *cs, slice_buffer * sb, int height, int stride_line){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1082 cs->b0 = slice_buffer_get_line(sb, mirror(-1-1, height-1) * stride_line); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1083 cs->b1 = slice_buffer_get_line(sb, mirror(-1 , height-1) * stride_line); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1084 cs->y = -1; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1085 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1086 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1087 static void spatial_compose53i_init(dwt_compose_t *cs, IDWTELEM *buffer, int height, int stride){ |
2562 | 1088 cs->b0 = buffer + mirror(-1-1, height-1)*stride; |
1089 cs->b1 = buffer + mirror(-1 , height-1)*stride; | |
1090 cs->y = -1; | |
1091 } | |
1092 | |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1093 static void spatial_compose53i_dy_buffered(dwt_compose_t *cs, slice_buffer * sb, int width, int height, int stride_line){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1094 int y= cs->y; |
2967 | 1095 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1096 IDWTELEM *b0= cs->b0; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1097 IDWTELEM *b1= cs->b1; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1098 IDWTELEM *b2= slice_buffer_get_line(sb, mirror(y+1, height-1) * stride_line); |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1099 IDWTELEM *b3= slice_buffer_get_line(sb, mirror(y+2, height-1) * stride_line); |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1100 |
2998 | 1101 if(y+1<(unsigned)height) vertical_compose53iL0(b1, b2, b3, width); |
1102 if(y+0<(unsigned)height) vertical_compose53iH0(b0, b1, b2, width); | |
6405 | 1103 |
2998 | 1104 if(y-1<(unsigned)height) horizontal_compose53i(b0, width); |
1105 if(y+0<(unsigned)height) horizontal_compose53i(b1, width); | |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1106 |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1107 cs->b0 = b2; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1108 cs->b1 = b3; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1109 cs->y += 2; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1110 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1111 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1112 static void spatial_compose53i_dy(dwt_compose_t *cs, IDWTELEM *buffer, int width, int height, int stride){ |
2562 | 1113 int y= cs->y; |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1114 IDWTELEM *b0= cs->b0; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1115 IDWTELEM *b1= cs->b1; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1116 IDWTELEM *b2= buffer + mirror(y+1, height-1)*stride; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1117 IDWTELEM *b3= buffer + mirror(y+2, height-1)*stride; |
2138 | 1118 |
2998 | 1119 if(y+1<(unsigned)height) vertical_compose53iL0(b1, b2, b3, width); |
1120 if(y+0<(unsigned)height) vertical_compose53iH0(b0, b1, b2, width); | |
6405 | 1121 |
2998 | 1122 if(y-1<(unsigned)height) horizontal_compose53i(b0, width); |
1123 if(y+0<(unsigned)height) horizontal_compose53i(b1, width); | |
2138 | 1124 |
2562 | 1125 cs->b0 = b2; |
1126 cs->b1 = b3; | |
1127 cs->y += 2; | |
1128 } | |
1129 | |
5989
382aa8b243c5
prevent warnings about functions being possibly unused
michael
parents:
5988
diff
changeset
|
1130 static void av_unused spatial_compose53i(IDWTELEM *buffer, int width, int height, int stride){ |
5988
919193c63fb8
thank you for your efforts to keep my code clean, but i would like to
michael
parents:
5934
diff
changeset
|
1131 dwt_compose_t cs; |
919193c63fb8
thank you for your efforts to keep my code clean, but i would like to
michael
parents:
5934
diff
changeset
|
1132 spatial_compose53i_init(&cs, buffer, height, stride); |
919193c63fb8
thank you for your efforts to keep my code clean, but i would like to
michael
parents:
5934
diff
changeset
|
1133 while(cs.y <= height) |
919193c63fb8
thank you for your efforts to keep my code clean, but i would like to
michael
parents:
5934
diff
changeset
|
1134 spatial_compose53i_dy(&cs, buffer, width, height, stride); |
919193c63fb8
thank you for your efforts to keep my code clean, but i would like to
michael
parents:
5934
diff
changeset
|
1135 } |
919193c63fb8
thank you for your efforts to keep my code clean, but i would like to
michael
parents:
5934
diff
changeset
|
1136 |
2967 | 1137 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1138 void ff_snow_horizontal_compose97i(IDWTELEM *b, int width){ |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1139 IDWTELEM temp[width]; |
2138 | 1140 const int w2= (width+1)>>1; |
1141 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1142 inv_lift (temp , b , b +w2, 1, 1, 1, width, W_DM, W_DO, W_DS, 0, 1); |
5589
946c2db0a093
cleanup (remove some old experimentation related code)
michael
parents:
5588
diff
changeset
|
1143 inv_lift (temp+w2, b +w2, temp , 1, 1, 1, width, W_CM, W_CO, W_CS, 1, 1); |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1144 inv_liftS(b , temp , temp+w2, 2, 1, 1, width, W_BM, W_BO, W_BS, 0, 1); |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1145 inv_lift (b+1 , temp+w2, b , 2, 1, 2, width, W_AM, W_AO, W_AS, 1, 0); |
2138 | 1146 } |
1147 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1148 static void vertical_compose97iH0(IDWTELEM *b0, IDWTELEM *b1, IDWTELEM *b2, int width){ |
2138 | 1149 int i; |
2967 | 1150 |
2138 | 1151 for(i=0; i<width; i++){ |
1152 b1[i] += (W_AM*(b0[i] + b2[i])+W_AO)>>W_AS; | |
1153 } | |
1154 } | |
1155 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1156 static void vertical_compose97iH1(IDWTELEM *b0, IDWTELEM *b1, IDWTELEM *b2, int width){ |
2138 | 1157 int i; |
2967 | 1158 |
2138 | 1159 for(i=0; i<width; i++){ |
1160 b1[i] -= (W_CM*(b0[i] + b2[i])+W_CO)>>W_CS; | |
1161 } | |
1162 } | |
1163 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1164 static void vertical_compose97iL0(IDWTELEM *b0, IDWTELEM *b1, IDWTELEM *b2, int width){ |
2138 | 1165 int i; |
2967 | 1166 |
2138 | 1167 for(i=0; i<width; i++){ |
2602
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
1168 #ifdef liftS |
2138 | 1169 b1[i] += (W_BM*(b0[i] + b2[i])+W_BO)>>W_BS; |
2602
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
1170 #else |
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
1171 b1[i] += (W_BM*(b0[i] + b2[i])+4*b1[i]+W_BO)>>W_BS; |
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
1172 #endif |
2138 | 1173 } |
1174 } | |
1175 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1176 static void vertical_compose97iL1(IDWTELEM *b0, IDWTELEM *b1, IDWTELEM *b2, int width){ |
2138 | 1177 int i; |
2967 | 1178 |
2138 | 1179 for(i=0; i<width; i++){ |
1180 b1[i] -= (W_DM*(b0[i] + b2[i])+W_DO)>>W_DS; | |
1181 } | |
1182 } | |
1183 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1184 void ff_snow_vertical_compose97i(IDWTELEM *b0, IDWTELEM *b1, IDWTELEM *b2, IDWTELEM *b3, IDWTELEM *b4, IDWTELEM *b5, int width){ |
2592 | 1185 int i; |
2967 | 1186 |
2592 | 1187 for(i=0; i<width; i++){ |
1188 b4[i] -= (W_DM*(b3[i] + b5[i])+W_DO)>>W_DS; | |
1189 b3[i] -= (W_CM*(b2[i] + b4[i])+W_CO)>>W_CS; | |
2602
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
1190 #ifdef liftS |
2592 | 1191 b2[i] += (W_BM*(b1[i] + b3[i])+W_BO)>>W_BS; |
2602
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
1192 #else |
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
1193 b2[i] += (W_BM*(b1[i] + b3[i])+4*b2[i]+W_BO)>>W_BS; |
5ec55feb6fdd
rescale coefficients during IDWT, that way the lifting steps are much simpler and faster
michael
parents:
2601
diff
changeset
|
1194 #endif |
2592 | 1195 b1[i] += (W_AM*(b0[i] + b2[i])+W_AO)>>W_AS; |
1196 } | |
1197 } | |
1198 | |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1199 static void spatial_compose97i_buffered_init(dwt_compose_t *cs, slice_buffer * sb, int height, int stride_line){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1200 cs->b0 = slice_buffer_get_line(sb, mirror(-3-1, height-1) * stride_line); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1201 cs->b1 = slice_buffer_get_line(sb, mirror(-3 , height-1) * stride_line); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1202 cs->b2 = slice_buffer_get_line(sb, mirror(-3+1, height-1) * stride_line); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1203 cs->b3 = slice_buffer_get_line(sb, mirror(-3+2, height-1) * stride_line); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1204 cs->y = -3; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1205 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1206 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1207 static void spatial_compose97i_init(dwt_compose_t *cs, IDWTELEM *buffer, int height, int stride){ |
2562 | 1208 cs->b0 = buffer + mirror(-3-1, height-1)*stride; |
1209 cs->b1 = buffer + mirror(-3 , height-1)*stride; | |
1210 cs->b2 = buffer + mirror(-3+1, height-1)*stride; | |
1211 cs->b3 = buffer + mirror(-3+2, height-1)*stride; | |
1212 cs->y = -3; | |
1213 } | |
2138 | 1214 |
3198
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
1215 static void spatial_compose97i_dy_buffered(DSPContext *dsp, dwt_compose_t *cs, slice_buffer * sb, int width, int height, int stride_line){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1216 int y = cs->y; |
2967 | 1217 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1218 IDWTELEM *b0= cs->b0; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1219 IDWTELEM *b1= cs->b1; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1220 IDWTELEM *b2= cs->b2; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1221 IDWTELEM *b3= cs->b3; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1222 IDWTELEM *b4= slice_buffer_get_line(sb, mirror(y + 3, height - 1) * stride_line); |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1223 IDWTELEM *b5= slice_buffer_get_line(sb, mirror(y + 4, height - 1) * stride_line); |
2967 | 1224 |
2592 | 1225 if(y>0 && y+4<height){ |
3198
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
1226 dsp->vertical_compose97i(b0, b1, b2, b3, b4, b5, width); |
2592 | 1227 }else{ |
2998 | 1228 if(y+3<(unsigned)height) vertical_compose97iL1(b3, b4, b5, width); |
1229 if(y+2<(unsigned)height) vertical_compose97iH1(b2, b3, b4, width); | |
1230 if(y+1<(unsigned)height) vertical_compose97iL0(b1, b2, b3, width); | |
1231 if(y+0<(unsigned)height) vertical_compose97iH0(b0, b1, b2, width); | |
2592 | 1232 } |
6405 | 1233 |
3198
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
1234 if(y-1<(unsigned)height) dsp->horizontal_compose97i(b0, width); |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
1235 if(y+0<(unsigned)height) dsp->horizontal_compose97i(b1, width); |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1236 |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1237 cs->b0=b2; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1238 cs->b1=b3; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1239 cs->b2=b4; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1240 cs->b3=b5; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1241 cs->y += 2; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1242 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1243 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1244 static void spatial_compose97i_dy(dwt_compose_t *cs, IDWTELEM *buffer, int width, int height, int stride){ |
2562 | 1245 int y = cs->y; |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1246 IDWTELEM *b0= cs->b0; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1247 IDWTELEM *b1= cs->b1; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1248 IDWTELEM *b2= cs->b2; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1249 IDWTELEM *b3= cs->b3; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1250 IDWTELEM *b4= buffer + mirror(y+3, height-1)*stride; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1251 IDWTELEM *b5= buffer + mirror(y+4, height-1)*stride; |
2138 | 1252 |
2998 | 1253 if(y+3<(unsigned)height) vertical_compose97iL1(b3, b4, b5, width); |
1254 if(y+2<(unsigned)height) vertical_compose97iH1(b2, b3, b4, width); | |
1255 if(y+1<(unsigned)height) vertical_compose97iL0(b1, b2, b3, width); | |
1256 if(y+0<(unsigned)height) vertical_compose97iH0(b0, b1, b2, width); | |
6405 | 1257 |
3198
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
1258 if(y-1<(unsigned)height) ff_snow_horizontal_compose97i(b0, width); |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
1259 if(y+0<(unsigned)height) ff_snow_horizontal_compose97i(b1, width); |
2562 | 1260 |
1261 cs->b0=b2; | |
1262 cs->b1=b3; | |
1263 cs->b2=b4; | |
1264 cs->b3=b5; | |
1265 cs->y += 2; | |
1266 } | |
1267 | |
5989
382aa8b243c5
prevent warnings about functions being possibly unused
michael
parents:
5988
diff
changeset
|
1268 static void av_unused spatial_compose97i(IDWTELEM *buffer, int width, int height, int stride){ |
5988
919193c63fb8
thank you for your efforts to keep my code clean, but i would like to
michael
parents:
5934
diff
changeset
|
1269 dwt_compose_t cs; |
919193c63fb8
thank you for your efforts to keep my code clean, but i would like to
michael
parents:
5934
diff
changeset
|
1270 spatial_compose97i_init(&cs, buffer, height, stride); |
919193c63fb8
thank you for your efforts to keep my code clean, but i would like to
michael
parents:
5934
diff
changeset
|
1271 while(cs.y <= height) |
919193c63fb8
thank you for your efforts to keep my code clean, but i would like to
michael
parents:
5934
diff
changeset
|
1272 spatial_compose97i_dy(&cs, buffer, width, height, stride); |
919193c63fb8
thank you for your efforts to keep my code clean, but i would like to
michael
parents:
5934
diff
changeset
|
1273 } |
919193c63fb8
thank you for your efforts to keep my code clean, but i would like to
michael
parents:
5934
diff
changeset
|
1274 |
3075 | 1275 static void ff_spatial_idwt_buffered_init(dwt_compose_t *cs, slice_buffer * sb, int width, int height, int stride_line, int type, int decomposition_count){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1276 int level; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1277 for(level=decomposition_count-1; level>=0; level--){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1278 switch(type){ |
3326 | 1279 case DWT_97: spatial_compose97i_buffered_init(cs+level, sb, height>>level, stride_line<<level); break; |
1280 case DWT_53: spatial_compose53i_buffered_init(cs+level, sb, height>>level, stride_line<<level); break; | |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1281 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1282 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1283 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1284 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1285 static void ff_spatial_idwt_init(dwt_compose_t *cs, IDWTELEM *buffer, int width, int height, int stride, int type, int decomposition_count){ |
2562 | 1286 int level; |
1287 for(level=decomposition_count-1; level>=0; level--){ | |
1288 switch(type){ | |
3326 | 1289 case DWT_97: spatial_compose97i_init(cs+level, buffer, height>>level, stride<<level); break; |
1290 case DWT_53: spatial_compose53i_init(cs+level, buffer, height>>level, stride<<level); break; | |
2562 | 1291 } |
1292 } | |
1293 } | |
1294 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1295 static void ff_spatial_idwt_slice(dwt_compose_t *cs, IDWTELEM *buffer, int width, int height, int stride, int type, int decomposition_count, int y){ |
2562 | 1296 const int support = type==1 ? 3 : 5; |
1297 int level; | |
1298 if(type==2) return; | |
1299 | |
1300 for(level=decomposition_count-1; level>=0; level--){ | |
1301 while(cs[level].y <= FFMIN((y>>level)+support, height>>level)){ | |
1302 switch(type){ | |
3326 | 1303 case DWT_97: spatial_compose97i_dy(cs+level, buffer, width>>level, height>>level, stride<<level); |
6414 | 1304 break; |
3326 | 1305 case DWT_53: spatial_compose53i_dy(cs+level, buffer, width>>level, height>>level, stride<<level); |
6414 | 1306 break; |
2562 | 1307 } |
1308 } | |
2138 | 1309 } |
1310 } | |
1311 | |
3198
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
1312 static void ff_spatial_idwt_buffered_slice(DSPContext *dsp, dwt_compose_t *cs, slice_buffer * slice_buf, int width, int height, int stride_line, int type, int decomposition_count, int y){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1313 const int support = type==1 ? 3 : 5; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1314 int level; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1315 if(type==2) return; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1316 |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1317 for(level=decomposition_count-1; level>=0; level--){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1318 while(cs[level].y <= FFMIN((y>>level)+support, height>>level)){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1319 switch(type){ |
3326 | 1320 case DWT_97: spatial_compose97i_dy_buffered(dsp, cs+level, slice_buf, width>>level, height>>level, stride_line<<level); |
6414 | 1321 break; |
3326 | 1322 case DWT_53: spatial_compose53i_dy_buffered(cs+level, slice_buf, width>>level, height>>level, stride_line<<level); |
6414 | 1323 break; |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1324 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1325 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1326 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1327 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1328 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1329 static void ff_spatial_idwt(IDWTELEM *buffer, int width, int height, int stride, int type, int decomposition_count){ |
2562 | 1330 dwt_compose_t cs[MAX_DECOMPOSITIONS]; |
1331 int y; | |
1332 ff_spatial_idwt_init(cs, buffer, width, height, stride, type, decomposition_count); | |
1333 for(y=0; y<height; y+=4) | |
1334 ff_spatial_idwt_slice(cs, buffer, width, height, stride, type, decomposition_count, y); | |
2138 | 1335 } |
1336 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1337 static int encode_subband_c0run(SnowContext *s, SubBand *b, IDWTELEM *src, IDWTELEM *parent, int stride, int orientation){ |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
1338 const int w= b->width; |
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
1339 const int h= b->height; |
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
1340 int x, y; |
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
1341 |
2138 | 1342 if(1){ |
1343 int run=0; | |
2149 | 1344 int runs[w*h]; |
2138 | 1345 int run_index=0; |
2609
0f74a379a890
store the number of runs to avoid storing the last run value
michael
parents:
2608
diff
changeset
|
1346 int max_index; |
2967 | 1347 |
2138 | 1348 for(y=0; y<h; y++){ |
1349 for(x=0; x<w; x++){ | |
2148
678be5a8f282
use parent sample to predict significance & magnitude
michael
parents:
2146
diff
changeset
|
1350 int v, p=0; |
2144 | 1351 int /*ll=0, */l=0, lt=0, t=0, rt=0; |
2149 | 1352 v= src[x + y*stride]; |
2138 | 1353 |
1354 if(y){ | |
2149 | 1355 t= src[x + (y-1)*stride]; |
2138 | 1356 if(x){ |
2149 | 1357 lt= src[x - 1 + (y-1)*stride]; |
2138 | 1358 } |
1359 if(x + 1 < w){ | |
2149 | 1360 rt= src[x + 1 + (y-1)*stride]; |
2138 | 1361 } |
1362 } | |
1363 if(x){ | |
2149 | 1364 l= src[x - 1 + y*stride]; |
2144 | 1365 /*if(x > 1){ |
1366 if(orientation==1) ll= src[y + (x-2)*stride]; | |
1367 else ll= src[x - 2 + y*stride]; | |
2138 | 1368 }*/ |
1369 } | |
2148
678be5a8f282
use parent sample to predict significance & magnitude
michael
parents:
2146
diff
changeset
|
1370 if(parent){ |
2149 | 1371 int px= x>>1; |
1372 int py= y>>1; | |
2967 | 1373 if(px<b->parent->width && py<b->parent->height) |
2148
678be5a8f282
use parent sample to predict significance & magnitude
michael
parents:
2146
diff
changeset
|
1374 p= parent[px + py*2*stride]; |
678be5a8f282
use parent sample to predict significance & magnitude
michael
parents:
2146
diff
changeset
|
1375 } |
678be5a8f282
use parent sample to predict significance & magnitude
michael
parents:
2146
diff
changeset
|
1376 if(!(/*ll|*/l|lt|t|rt|p)){ |
2138 | 1377 if(v){ |
1378 runs[run_index++]= run; | |
1379 run=0; | |
1380 }else{ | |
1381 run++; | |
1382 } | |
1383 } | |
1384 } | |
1385 } | |
2609
0f74a379a890
store the number of runs to avoid storing the last run value
michael
parents:
2608
diff
changeset
|
1386 max_index= run_index; |
2138 | 1387 runs[run_index++]= run; |
1388 run_index=0; | |
1389 run= runs[run_index++]; | |
1390 | |
2609
0f74a379a890
store the number of runs to avoid storing the last run value
michael
parents:
2608
diff
changeset
|
1391 put_symbol2(&s->c, b->state[30], max_index, 0); |
0f74a379a890
store the number of runs to avoid storing the last run value
michael
parents:
2608
diff
changeset
|
1392 if(run_index <= max_index) |
0f74a379a890
store the number of runs to avoid storing the last run value
michael
parents:
2608
diff
changeset
|
1393 put_symbol2(&s->c, b->state[1], run, 3); |
2967 | 1394 |
2138 | 1395 for(y=0; y<h; y++){ |
2435
c89ac0e70c66
10l patch by (matthieu castet <castet.matthieu free fr>)
michael
parents:
2422
diff
changeset
|
1396 if(s->c.bytestream_end - s->c.bytestream < w*40){ |
2422 | 1397 av_log(s->avctx, AV_LOG_ERROR, "encoded frame too large\n"); |
1398 return -1; | |
1399 } | |
2138 | 1400 for(x=0; x<w; x++){ |
2148
678be5a8f282
use parent sample to predict significance & magnitude
michael
parents:
2146
diff
changeset
|
1401 int v, p=0; |
2144 | 1402 int /*ll=0, */l=0, lt=0, t=0, rt=0; |
2149 | 1403 v= src[x + y*stride]; |
2138 | 1404 |
1405 if(y){ | |
2149 | 1406 t= src[x + (y-1)*stride]; |
2138 | 1407 if(x){ |
2149 | 1408 lt= src[x - 1 + (y-1)*stride]; |
2138 | 1409 } |
1410 if(x + 1 < w){ | |
2149 | 1411 rt= src[x + 1 + (y-1)*stride]; |
2138 | 1412 } |
1413 } | |
1414 if(x){ | |
2149 | 1415 l= src[x - 1 + y*stride]; |
2144 | 1416 /*if(x > 1){ |
1417 if(orientation==1) ll= src[y + (x-2)*stride]; | |
1418 else ll= src[x - 2 + y*stride]; | |
2138 | 1419 }*/ |
1420 } | |
2148
678be5a8f282
use parent sample to predict significance & magnitude
michael
parents:
2146
diff
changeset
|
1421 if(parent){ |
2149 | 1422 int px= x>>1; |
1423 int py= y>>1; | |
2967 | 1424 if(px<b->parent->width && py<b->parent->height) |
2148
678be5a8f282
use parent sample to predict significance & magnitude
michael
parents:
2146
diff
changeset
|
1425 p= parent[px + py*2*stride]; |
678be5a8f282
use parent sample to predict significance & magnitude
michael
parents:
2146
diff
changeset
|
1426 } |
678be5a8f282
use parent sample to predict significance & magnitude
michael
parents:
2146
diff
changeset
|
1427 if(/*ll|*/l|lt|t|rt|p){ |
4001 | 1428 int context= av_log2(/*FFABS(ll) + */3*FFABS(l) + FFABS(lt) + 2*FFABS(t) + FFABS(rt) + FFABS(p)); |
2144 | 1429 |
2335 | 1430 put_rac(&s->c, &b->state[0][context], !!v); |
2138 | 1431 }else{ |
1432 if(!run){ | |
1433 run= runs[run_index++]; | |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
1434 |
2609
0f74a379a890
store the number of runs to avoid storing the last run value
michael
parents:
2608
diff
changeset
|
1435 if(run_index <= max_index) |
0f74a379a890
store the number of runs to avoid storing the last run value
michael
parents:
2608
diff
changeset
|
1436 put_symbol2(&s->c, b->state[1], run, 3); |
2138 | 1437 assert(v); |
1438 }else{ | |
1439 run--; | |
1440 assert(!v); | |
1441 } | |
1442 } | |
1443 if(v){ | |
4001 | 1444 int context= av_log2(/*FFABS(ll) + */3*FFABS(l) + FFABS(lt) + 2*FFABS(t) + FFABS(rt) + FFABS(p)); |
1445 int l2= 2*FFABS(l) + (l<0); | |
1446 int t2= 2*FFABS(t) + (t<0); | |
1447 | |
1448 put_symbol2(&s->c, b->state[context + 2], FFABS(v)-1, context-4); | |
2596 | 1449 put_rac(&s->c, &b->state[0][16 + 1 + 3 + quant3bA[l2&0xFF] + 3*quant3bA[t2&0xFF]], v<0); |
2138 | 1450 } |
1451 } | |
1452 } | |
1453 } | |
2422 | 1454 return 0; |
2138 | 1455 } |
1456 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1457 static int encode_subband(SnowContext *s, SubBand *b, IDWTELEM *src, IDWTELEM *parent, int stride, int orientation){ |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
1458 // encode_subband_qtree(s, b, src, parent, stride, orientation); |
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
1459 // encode_subband_z0run(s, b, src, parent, stride, orientation); |
2422 | 1460 return encode_subband_c0run(s, b, src, parent, stride, orientation); |
2155
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
1461 // encode_subband_dzr(s, b, src, parent, stride, orientation); |
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
1462 } |
274a01d80f4a
various subband encoders (all either worse or complicated so they are commented out)
michael
parents:
2152
diff
changeset
|
1463 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1464 static inline void unpack_coeffs(SnowContext *s, SubBand *b, SubBand * parent, int orientation){ |
2138 | 1465 const int w= b->width; |
1466 const int h= b->height; | |
1467 int x,y; | |
2967 | 1468 |
2138 | 1469 if(1){ |
2609
0f74a379a890
store the number of runs to avoid storing the last run value
michael
parents:
2608
diff
changeset
|
1470 int run, runs; |
2607
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1471 x_and_coeff *xc= b->x_coeff; |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1472 x_and_coeff *prev_xc= NULL; |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1473 x_and_coeff *prev2_xc= xc; |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1474 x_and_coeff *parent_xc= parent ? parent->x_coeff : NULL; |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1475 x_and_coeff *prev_parent_xc= parent_xc; |
2138 | 1476 |
2609
0f74a379a890
store the number of runs to avoid storing the last run value
michael
parents:
2608
diff
changeset
|
1477 runs= get_symbol2(&s->c, b->state[30], 0); |
0f74a379a890
store the number of runs to avoid storing the last run value
michael
parents:
2608
diff
changeset
|
1478 if(runs-- > 0) run= get_symbol2(&s->c, b->state[1], 3); |
0f74a379a890
store the number of runs to avoid storing the last run value
michael
parents:
2608
diff
changeset
|
1479 else run= INT_MAX; |
0f74a379a890
store the number of runs to avoid storing the last run value
michael
parents:
2608
diff
changeset
|
1480 |
2138 | 1481 for(y=0; y<h; y++){ |
2193 | 1482 int v=0; |
1483 int lt=0, t=0, rt=0; | |
1484 | |
2607
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1485 if(y && prev_xc->x == 0){ |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1486 rt= prev_xc->coeff; |
2193 | 1487 } |
2138 | 1488 for(x=0; x<w; x++){ |
2193 | 1489 int p=0; |
1490 const int l= v; | |
2967 | 1491 |
2193 | 1492 lt= t; t= rt; |
1493 | |
2194 | 1494 if(y){ |
2607
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1495 if(prev_xc->x <= x) |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1496 prev_xc++; |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1497 if(prev_xc->x == x + 1) |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1498 rt= prev_xc->coeff; |
2194 | 1499 else |
1500 rt=0; | |
1501 } | |
2607
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1502 if(parent_xc){ |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1503 if(x>>1 > parent_xc->x){ |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1504 parent_xc++; |
2192 | 1505 } |
2607
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1506 if(x>>1 == parent_xc->x){ |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1507 p= parent_xc->coeff; |
2194 | 1508 } |
2148
678be5a8f282
use parent sample to predict significance & magnitude
michael
parents:
2146
diff
changeset
|
1509 } |
678be5a8f282
use parent sample to predict significance & magnitude
michael
parents:
2146
diff
changeset
|
1510 if(/*ll|*/l|lt|t|rt|p){ |
4001 | 1511 int context= av_log2(/*FFABS(ll) + */3*(l>>1) + (lt>>1) + (t&~1) + (rt>>1) + (p>>1)); |
2144 | 1512 |
2335 | 1513 v=get_rac(&s->c, &b->state[0][context]); |
2605 | 1514 if(v){ |
1515 v= 2*(get_symbol2(&s->c, b->state[context + 2], context-4) + 1); | |
1516 v+=get_rac(&s->c, &b->state[0][16 + 1 + 3 + quant3bA[l&0xFF] + 3*quant3bA[t&0xFF]]); | |
2967 | 1517 |
2607
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1518 xc->x=x; |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1519 (xc++)->coeff= v; |
2605 | 1520 } |
2138 | 1521 }else{ |
1522 if(!run){ | |
2609
0f74a379a890
store the number of runs to avoid storing the last run value
michael
parents:
2608
diff
changeset
|
1523 if(runs-- > 0) run= get_symbol2(&s->c, b->state[1], 3); |
0f74a379a890
store the number of runs to avoid storing the last run value
michael
parents:
2608
diff
changeset
|
1524 else run= INT_MAX; |
2605 | 1525 v= 2*(get_symbol2(&s->c, b->state[0 + 2], 0-4) + 1); |
1526 v+=get_rac(&s->c, &b->state[0][16 + 1 + 3]); | |
2967 | 1527 |
2607
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1528 xc->x=x; |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1529 (xc++)->coeff= v; |
2138 | 1530 }else{ |
2606 | 1531 int max_run; |
2138 | 1532 run--; |
1533 v=0; | |
2191 | 1534 |
2607
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1535 if(y) max_run= FFMIN(run, prev_xc->x - x - 2); |
2606 | 1536 else max_run= FFMIN(run, w-x-1); |
2607
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1537 if(parent_xc) |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1538 max_run= FFMIN(max_run, 2*parent_xc->x - x - 1); |
2606 | 1539 x+= max_run; |
1540 run-= max_run; | |
2138 | 1541 } |
1542 } | |
2192 | 1543 } |
2607
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1544 (xc++)->x= w+1; //end marker |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1545 prev_xc= prev2_xc; |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1546 prev2_xc= xc; |
2967 | 1547 |
2607
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1548 if(parent_xc){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1549 if(y&1){ |
2607
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1550 while(parent_xc->x != parent->width+1) |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1551 parent_xc++; |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1552 parent_xc++; |
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1553 prev_parent_xc= parent_xc; |
2192 | 1554 }else{ |
2607
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1555 parent_xc= prev_parent_xc; |
2138 | 1556 } |
1557 } | |
1558 } | |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1559 |
2607
fde7b6fe2aaf
replace complicated pointer dereference + index stuff by pointers in unpack_coeffs()
michael
parents:
2606
diff
changeset
|
1560 (xc++)->x= w+1; //end marker |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1561 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1562 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1563 |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1564 static inline void decode_subband_slice_buffered(SnowContext *s, SubBand *b, slice_buffer * sb, int start_y, int h, int save_state[1]){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1565 const int w= b->width; |
2893 | 1566 int y; |
4594 | 1567 const int qlog= av_clip(s->qlog + b->qlog, 0, QROOT*16); |
2600 | 1568 int qmul= qexp[qlog&(QROOT-1)]<<(qlog>>QSHIFT); |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1569 int qadd= (s->qbias*qmul)>>QBIAS_SHIFT; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1570 int new_index = 0; |
2967 | 1571 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1572 if(b->ibuf == s->spatial_idwt_buffer || s->qlog == LOSSLESS_QLOG){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1573 qadd= 0; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1574 qmul= 1<<QEXPSHIFT; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1575 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1576 |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1577 /* If we are on the second or later slice, restore our index. */ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1578 if (start_y != 0) |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1579 new_index = save_state[0]; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1580 |
2967 | 1581 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1582 for(y=start_y; y<h; y++){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1583 int x = 0; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1584 int v; |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1585 IDWTELEM * line = slice_buffer_get_line(sb, y * b->stride_line + b->buf_y_offset) + b->buf_x_offset; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
1586 memset(line, 0, b->width*sizeof(IDWTELEM)); |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1587 v = b->x_coeff[new_index].coeff; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1588 x = b->x_coeff[new_index++].x; |
6474
e6995f3fbf7f
cosmetics: Normalize {} placement after for, while, if.
diego
parents:
6437
diff
changeset
|
1589 while(x < w){ |
2596 | 1590 register int t= ( (v>>1)*qmul + qadd)>>QEXPSHIFT; |
1591 register int u= -(v&1); | |
1592 line[x] = (t^u) - u; | |
1593 | |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1594 v = b->x_coeff[new_index].coeff; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1595 x = b->x_coeff[new_index++].x; |
2138 | 1596 } |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1597 } |
2967 | 1598 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1599 /* Save our variables for the next slice. */ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1600 save_state[0] = new_index; |
2967 | 1601 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
1602 return; |
2138 | 1603 } |
1604 | |
4331
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
1605 static void reset_contexts(SnowContext *s){ //FIXME better initial contexts |
2138 | 1606 int plane_index, level, orientation; |
1607 | |
2199 | 1608 for(plane_index=0; plane_index<3; plane_index++){ |
5665
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
1609 for(level=0; level<MAX_DECOMPOSITIONS; level++){ |
2138 | 1610 for(orientation=level ? 1:0; orientation<4; orientation++){ |
2335 | 1611 memset(s->plane[plane_index].band[level][orientation].state, MID_STATE, sizeof(s->plane[plane_index].band[level][orientation].state)); |
2138 | 1612 } |
1613 } | |
1614 } | |
2335 | 1615 memset(s->header_state, MID_STATE, sizeof(s->header_state)); |
1616 memset(s->block_state, MID_STATE, sizeof(s->block_state)); | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1617 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1618 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1619 static int alloc_blocks(SnowContext *s){ |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1620 int w= -((-s->avctx->width )>>LOG2_MB_SIZE); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1621 int h= -((-s->avctx->height)>>LOG2_MB_SIZE); |
2967 | 1622 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1623 s->b_width = w; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1624 s->b_height= h; |
2967 | 1625 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1626 s->block= av_mallocz(w * h * sizeof(BlockNode) << (s->block_max_depth*2)); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1627 return 0; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1628 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1629 |
2335 | 1630 static inline void copy_rac_state(RangeCoder *d, RangeCoder *s){ |
1631 uint8_t *bytestream= d->bytestream; | |
1632 uint8_t *bytestream_start= d->bytestream_start; | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1633 *d= *s; |
2335 | 1634 d->bytestream= bytestream; |
1635 d->bytestream_start= bytestream_start; | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1636 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1637 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1638 //near copy & paste from dsputil, FIXME |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1639 static int pix_sum(uint8_t * pix, int line_size, int w) |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1640 { |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1641 int s, i, j; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1642 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1643 s = 0; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1644 for (i = 0; i < w; i++) { |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1645 for (j = 0; j < w; j++) { |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1646 s += pix[0]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1647 pix ++; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1648 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1649 pix += line_size - w; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1650 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1651 return s; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1652 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1653 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1654 //near copy & paste from dsputil, FIXME |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1655 static int pix_norm1(uint8_t * pix, int line_size, int w) |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1656 { |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1657 int s, i, j; |
4179 | 1658 uint32_t *sq = ff_squareTbl + 256; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1659 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1660 s = 0; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1661 for (i = 0; i < w; i++) { |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1662 for (j = 0; j < w; j ++) { |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1663 s += sq[pix[0]]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1664 pix ++; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1665 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1666 pix += line_size - w; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1667 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1668 return s; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1669 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1670 |
3314 | 1671 static inline void set_blocks(SnowContext *s, int level, int x, int y, int l, int cb, int cr, int mx, int my, int ref, int type){ |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1672 const int w= s->b_width << s->block_max_depth; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1673 const int rem_depth= s->block_max_depth - level; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1674 const int index= (x + y*w) << rem_depth; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1675 const int block_w= 1<<rem_depth; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1676 BlockNode block; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1677 int i,j; |
2967 | 1678 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1679 block.color[0]= l; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1680 block.color[1]= cb; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1681 block.color[2]= cr; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1682 block.mx= mx; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1683 block.my= my; |
3314 | 1684 block.ref= ref; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1685 block.type= type; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1686 block.level= level; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1687 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1688 for(j=0; j<block_w; j++){ |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1689 for(i=0; i<block_w; i++){ |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1690 s->block[index + i + j*w]= block; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1691 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1692 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1693 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1694 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1695 static inline void init_ref(MotionEstContext *c, uint8_t *src[3], uint8_t *ref[3], uint8_t *ref2[3], int x, int y, int ref_index){ |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1696 const int offset[3]= { |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1697 y*c-> stride + x, |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1698 ((y*c->uvstride + x)>>1), |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1699 ((y*c->uvstride + x)>>1), |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1700 }; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1701 int i; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1702 for(i=0; i<3; i++){ |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1703 c->src[0][i]= src [i]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1704 c->ref[0][i]= ref [i] + offset[i]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1705 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1706 assert(!ref_index); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1707 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1708 |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1709 static inline void pred_mv(SnowContext *s, int *mx, int *my, int ref, |
4408
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
1710 const BlockNode *left, const BlockNode *top, const BlockNode *tr){ |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1711 if(s->ref_frames == 1){ |
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1712 *mx = mid_pred(left->mx, top->mx, tr->mx); |
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1713 *my = mid_pred(left->my, top->my, tr->my); |
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1714 }else{ |
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1715 const int *scale = scale_mv_ref[ref]; |
4407
15688bdfe1b5
Brings down the number of snow.c warnings from 33 to 27 by putting parentheses
takis
parents:
4360
diff
changeset
|
1716 *mx = mid_pred((left->mx * scale[left->ref] + 128) >>8, |
15688bdfe1b5
Brings down the number of snow.c warnings from 33 to 27 by putting parentheses
takis
parents:
4360
diff
changeset
|
1717 (top ->mx * scale[top ->ref] + 128) >>8, |
15688bdfe1b5
Brings down the number of snow.c warnings from 33 to 27 by putting parentheses
takis
parents:
4360
diff
changeset
|
1718 (tr ->mx * scale[tr ->ref] + 128) >>8); |
15688bdfe1b5
Brings down the number of snow.c warnings from 33 to 27 by putting parentheses
takis
parents:
4360
diff
changeset
|
1719 *my = mid_pred((left->my * scale[left->ref] + 128) >>8, |
15688bdfe1b5
Brings down the number of snow.c warnings from 33 to 27 by putting parentheses
takis
parents:
4360
diff
changeset
|
1720 (top ->my * scale[top ->ref] + 128) >>8, |
15688bdfe1b5
Brings down the number of snow.c warnings from 33 to 27 by putting parentheses
takis
parents:
4360
diff
changeset
|
1721 (tr ->my * scale[tr ->ref] + 128) >>8); |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1722 } |
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1723 } |
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1724 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1725 //FIXME copy&paste |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1726 #define P_LEFT P[1] |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1727 #define P_TOP P[2] |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1728 #define P_TOPRIGHT P[3] |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1729 #define P_MEDIAN P[4] |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1730 #define P_MV1 P[9] |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1731 #define FLAG_QPEL 1 //must be 1 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1732 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1733 static int encode_q_branch(SnowContext *s, int level, int x, int y){ |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1734 uint8_t p_buffer[1024]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1735 uint8_t i_buffer[1024]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1736 uint8_t p_state[sizeof(s->block_state)]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1737 uint8_t i_state[sizeof(s->block_state)]; |
2335 | 1738 RangeCoder pc, ic; |
1739 uint8_t *pbbak= s->c.bytestream; | |
1740 uint8_t *pbbak_start= s->c.bytestream_start; | |
5082 | 1741 int score, score2, iscore, i_len, p_len, block_s, sum, base_bits; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1742 const int w= s->b_width << s->block_max_depth; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1743 const int h= s->b_height << s->block_max_depth; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1744 const int rem_depth= s->block_max_depth - level; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1745 const int index= (x + y*w) << rem_depth; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1746 const int block_w= 1<<(LOG2_MB_SIZE - level); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1747 int trx= (x+1)<<rem_depth; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1748 int try= (y+1)<<rem_depth; |
4408
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
1749 const BlockNode *left = x ? &s->block[index-1] : &null_block; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
1750 const BlockNode *top = y ? &s->block[index-w] : &null_block; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
1751 const BlockNode *right = trx<w ? &s->block[index+1] : &null_block; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
1752 const BlockNode *bottom= try<h ? &s->block[index+w] : &null_block; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
1753 const BlockNode *tl = y && x ? &s->block[index-w-1] : left; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
1754 const BlockNode *tr = y && trx<w && ((x&1)==0 || level==0) ? &s->block[index-w+(1<<rem_depth)] : tl; //FIXME use lt |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1755 int pl = left->color[0]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1756 int pcb= left->color[1]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1757 int pcr= left->color[2]; |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1758 int pmx, pmy; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1759 int mx=0, my=0; |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1760 int l,cr,cb; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1761 const int stride= s->current_picture.linesize[0]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1762 const int uvstride= s->current_picture.linesize[1]; |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1763 uint8_t *current_data[3]= { s->input_picture.data[0] + (x + y* stride)*block_w, |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1764 s->input_picture.data[1] + (x + y*uvstride)*block_w/2, |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1765 s->input_picture.data[2] + (x + y*uvstride)*block_w/2}; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1766 int P[10][2]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1767 int16_t last_mv[3][2]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1768 int qpel= !!(s->avctx->flags & CODEC_FLAG_QPEL); //unused |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1769 const int shift= 1+qpel; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1770 MotionEstContext *c= &s->m.me; |
3314 | 1771 int ref_context= av_log2(2*left->ref) + av_log2(2*top->ref); |
4001 | 1772 int mx_context= av_log2(2*FFABS(left->mx - top->mx)); |
1773 int my_context= av_log2(2*FFABS(left->my - top->my)); | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1774 int s_context= 2*left->level + 2*top->level + tl->level + tr->level; |
3314 | 1775 int ref, best_ref, ref_score, ref_mx, ref_my; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1776 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1777 assert(sizeof(s->block_state) >= 256); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1778 if(s->keyframe){ |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1779 set_blocks(s, level, x, y, pl, pcb, pcr, 0, 0, 0, BLOCK_INTRA); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1780 return 0; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1781 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1782 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1783 // clip predictors / edge ? |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1784 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1785 P_LEFT[0]= left->mx; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1786 P_LEFT[1]= left->my; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1787 P_TOP [0]= top->mx; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1788 P_TOP [1]= top->my; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1789 P_TOPRIGHT[0]= tr->mx; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1790 P_TOPRIGHT[1]= tr->my; |
2967 | 1791 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1792 last_mv[0][0]= s->block[index].mx; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1793 last_mv[0][1]= s->block[index].my; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1794 last_mv[1][0]= right->mx; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1795 last_mv[1][1]= right->my; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1796 last_mv[2][0]= bottom->mx; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1797 last_mv[2][1]= bottom->my; |
2967 | 1798 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1799 s->m.mb_stride=2; |
2967 | 1800 s->m.mb_x= |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1801 s->m.mb_y= 0; |
4360
5f887a3c7281
Change the Snow encoder to always use the available MotionEstContext pointer.
takis
parents:
4332
diff
changeset
|
1802 c->skip= 0; |
5f887a3c7281
Change the Snow encoder to always use the available MotionEstContext pointer.
takis
parents:
4332
diff
changeset
|
1803 |
5f887a3c7281
Change the Snow encoder to always use the available MotionEstContext pointer.
takis
parents:
4332
diff
changeset
|
1804 assert(c-> stride == stride); |
5f887a3c7281
Change the Snow encoder to always use the available MotionEstContext pointer.
takis
parents:
4332
diff
changeset
|
1805 assert(c->uvstride == uvstride); |
2967 | 1806 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1807 c->penalty_factor = get_penalty_factor(s->lambda, s->lambda2, c->avctx->me_cmp); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1808 c->sub_penalty_factor= get_penalty_factor(s->lambda, s->lambda2, c->avctx->me_sub_cmp); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1809 c->mb_penalty_factor = get_penalty_factor(s->lambda, s->lambda2, c->avctx->mb_cmp); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1810 c->current_mv_penalty= c->mv_penalty[s->m.f_code=1] + MAX_MV; |
2967 | 1811 |
2206 | 1812 c->xmin = - x*block_w - 16+2; |
1813 c->ymin = - y*block_w - 16+2; | |
1814 c->xmax = - (x+1)*block_w + (w<<(LOG2_MB_SIZE - s->block_max_depth)) + 16-2; | |
1815 c->ymax = - (y+1)*block_w + (h<<(LOG2_MB_SIZE - s->block_max_depth)) + 16-2; | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1816 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1817 if(P_LEFT[0] > (c->xmax<<shift)) P_LEFT[0] = (c->xmax<<shift); |
2967 | 1818 if(P_LEFT[1] > (c->ymax<<shift)) P_LEFT[1] = (c->ymax<<shift); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1819 if(P_TOP[0] > (c->xmax<<shift)) P_TOP[0] = (c->xmax<<shift); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1820 if(P_TOP[1] > (c->ymax<<shift)) P_TOP[1] = (c->ymax<<shift); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1821 if(P_TOPRIGHT[0] < (c->xmin<<shift)) P_TOPRIGHT[0]= (c->xmin<<shift); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1822 if(P_TOPRIGHT[0] > (c->xmax<<shift)) P_TOPRIGHT[0]= (c->xmax<<shift); //due to pmx no clip |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1823 if(P_TOPRIGHT[1] > (c->ymax<<shift)) P_TOPRIGHT[1]= (c->ymax<<shift); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1824 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1825 P_MEDIAN[0]= mid_pred(P_LEFT[0], P_TOP[0], P_TOPRIGHT[0]); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1826 P_MEDIAN[1]= mid_pred(P_LEFT[1], P_TOP[1], P_TOPRIGHT[1]); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1827 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1828 if (!y) { |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1829 c->pred_x= P_LEFT[0]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1830 c->pred_y= P_LEFT[1]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1831 } else { |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1832 c->pred_x = P_MEDIAN[0]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1833 c->pred_y = P_MEDIAN[1]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1834 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1835 |
3314 | 1836 score= INT_MAX; |
1837 best_ref= 0; | |
1838 for(ref=0; ref<s->ref_frames; ref++){ | |
1839 init_ref(c, current_data, s->last_picture[ref].data, NULL, block_w*x, block_w*y, 0); | |
1840 | |
1841 ref_score= ff_epzs_motion_search(&s->m, &ref_mx, &ref_my, P, 0, /*ref_index*/ 0, last_mv, | |
1842 (1<<16)>>shift, level-LOG2_MB_SIZE+4, block_w); | |
1843 | |
1844 assert(ref_mx >= c->xmin); | |
1845 assert(ref_mx <= c->xmax); | |
1846 assert(ref_my >= c->ymin); | |
1847 assert(ref_my <= c->ymax); | |
1848 | |
4360
5f887a3c7281
Change the Snow encoder to always use the available MotionEstContext pointer.
takis
parents:
4332
diff
changeset
|
1849 ref_score= c->sub_motion_search(&s->m, &ref_mx, &ref_my, ref_score, 0, 0, level-LOG2_MB_SIZE+4, block_w); |
3314 | 1850 ref_score= ff_get_mb_score(&s->m, ref_mx, ref_my, 0, 0, level-LOG2_MB_SIZE+4, block_w, 0); |
1851 ref_score+= 2*av_log2(2*ref)*c->penalty_factor; | |
1852 if(s->ref_mvs[ref]){ | |
1853 s->ref_mvs[ref][index][0]= ref_mx; | |
1854 s->ref_mvs[ref][index][1]= ref_my; | |
1855 s->ref_scores[ref][index]= ref_score; | |
1856 } | |
1857 if(score > ref_score){ | |
1858 score= ref_score; | |
1859 best_ref= ref; | |
1860 mx= ref_mx; | |
1861 my= ref_my; | |
1862 } | |
1863 } | |
5127 | 1864 //FIXME if mb_cmp != SSE then intra cannot be compared currently and mb_penalty vs. lambda2 |
2967 | 1865 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1866 // subpel search |
5085 | 1867 base_bits= get_rac_count(&s->c) - 8*(s->c.bytestream - s->c.bytestream_start); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1868 pc= s->c; |
2335 | 1869 pc.bytestream_start= |
1870 pc.bytestream= p_buffer; //FIXME end/start? and at the other stoo | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1871 memcpy(p_state, s->block_state, sizeof(s->block_state)); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1872 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1873 if(level!=s->block_max_depth) |
2335 | 1874 put_rac(&pc, &p_state[4 + s_context], 1); |
1875 put_rac(&pc, &p_state[1 + left->type + top->type], 0); | |
3314 | 1876 if(s->ref_frames > 1) |
1877 put_symbol(&pc, &p_state[128 + 1024 + 32*ref_context], best_ref, 0); | |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1878 pred_mv(s, &pmx, &pmy, best_ref, left, top, tr); |
3314 | 1879 put_symbol(&pc, &p_state[128 + 32*(mx_context + 16*!!best_ref)], mx - pmx, 1); |
1880 put_symbol(&pc, &p_state[128 + 32*(my_context + 16*!!best_ref)], my - pmy, 1); | |
2335 | 1881 p_len= pc.bytestream - pc.bytestream_start; |
5082 | 1882 score += (s->lambda2*(get_rac_count(&pc)-base_bits))>>FF_LAMBDA_SHIFT; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1883 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1884 block_s= block_w*block_w; |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1885 sum = pix_sum(current_data[0], stride, block_w); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1886 l= (sum + block_s/2)/block_s; |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1887 iscore = pix_norm1(current_data[0], stride, block_w) - 2*l*sum + l*l*block_s; |
2967 | 1888 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1889 block_s= block_w*block_w>>2; |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1890 sum = pix_sum(current_data[1], uvstride, block_w>>1); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1891 cb= (sum + block_s/2)/block_s; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1892 // iscore += pix_norm1(¤t_mb[1][0], uvstride, block_w>>1) - 2*cb*sum + cb*cb*block_s; |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1893 sum = pix_sum(current_data[2], uvstride, block_w>>1); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1894 cr= (sum + block_s/2)/block_s; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1895 // iscore += pix_norm1(¤t_mb[2][0], uvstride, block_w>>1) - 2*cr*sum + cr*cr*block_s; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1896 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1897 ic= s->c; |
2335 | 1898 ic.bytestream_start= |
1899 ic.bytestream= i_buffer; //FIXME end/start? and at the other stoo | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1900 memcpy(i_state, s->block_state, sizeof(s->block_state)); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1901 if(level!=s->block_max_depth) |
2335 | 1902 put_rac(&ic, &i_state[4 + s_context], 1); |
1903 put_rac(&ic, &i_state[1 + left->type + top->type], 1); | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1904 put_symbol(&ic, &i_state[32], l-pl , 1); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1905 put_symbol(&ic, &i_state[64], cb-pcb, 1); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1906 put_symbol(&ic, &i_state[96], cr-pcr, 1); |
2335 | 1907 i_len= ic.bytestream - ic.bytestream_start; |
5082 | 1908 iscore += (s->lambda2*(get_rac_count(&ic)-base_bits))>>FF_LAMBDA_SHIFT; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1909 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1910 // assert(score==256*256*256*64-1); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1911 assert(iscore < 255*255*256 + s->lambda2*10); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1912 assert(iscore >= 0); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1913 assert(l>=0 && l<=255); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1914 assert(pl>=0 && pl<=255); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1915 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1916 if(level==0){ |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1917 int varc= iscore >> 8; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1918 int vard= score >> 8; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1919 if (vard <= 64 || vard < varc) |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1920 c->scene_change_score+= ff_sqrt(vard) - ff_sqrt(varc); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1921 else |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1922 c->scene_change_score+= s->m.qscale; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1923 } |
2967 | 1924 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1925 if(level!=s->block_max_depth){ |
2335 | 1926 put_rac(&s->c, &s->block_state[4 + s_context], 0); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1927 score2 = encode_q_branch(s, level+1, 2*x+0, 2*y+0); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1928 score2+= encode_q_branch(s, level+1, 2*x+1, 2*y+0); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1929 score2+= encode_q_branch(s, level+1, 2*x+0, 2*y+1); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1930 score2+= encode_q_branch(s, level+1, 2*x+1, 2*y+1); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1931 score2+= s->lambda2>>FF_LAMBDA_SHIFT; //FIXME exact split overhead |
2967 | 1932 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1933 if(score2 < score && score2 < iscore) |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1934 return score2; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1935 } |
2967 | 1936 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1937 if(iscore < score){ |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1938 pred_mv(s, &pmx, &pmy, 0, left, top, tr); |
2335 | 1939 memcpy(pbbak, i_buffer, i_len); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1940 s->c= ic; |
2335 | 1941 s->c.bytestream_start= pbbak_start; |
1942 s->c.bytestream= pbbak + i_len; | |
3314 | 1943 set_blocks(s, level, x, y, l, cb, cr, pmx, pmy, 0, BLOCK_INTRA); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1944 memcpy(s->block_state, i_state, sizeof(s->block_state)); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1945 return iscore; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1946 }else{ |
2335 | 1947 memcpy(pbbak, p_buffer, p_len); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1948 s->c= pc; |
2335 | 1949 s->c.bytestream_start= pbbak_start; |
1950 s->c.bytestream= pbbak + p_len; | |
3314 | 1951 set_blocks(s, level, x, y, pl, pcb, pcr, mx, my, best_ref, 0); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1952 memcpy(s->block_state, p_state, sizeof(s->block_state)); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1953 return score; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1954 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1955 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
1956 |
4283
d6f83e2f8804
rename always_inline to av_always_inline and move to common.h
mru
parents:
4197
diff
changeset
|
1957 static av_always_inline int same_block(BlockNode *a, BlockNode *b){ |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1958 if((a->type&BLOCK_INTRA) && (b->type&BLOCK_INTRA)){ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1959 return !((a->color[0] - b->color[0]) | (a->color[1] - b->color[1]) | (a->color[2] - b->color[2])); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1960 }else{ |
3314 | 1961 return !((a->mx - b->mx) | (a->my - b->my) | (a->ref - b->ref) | ((a->type ^ b->type)&BLOCK_INTRA)); |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1962 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1963 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1964 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1965 static void encode_q_branch2(SnowContext *s, int level, int x, int y){ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1966 const int w= s->b_width << s->block_max_depth; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1967 const int rem_depth= s->block_max_depth - level; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1968 const int index= (x + y*w) << rem_depth; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1969 int trx= (x+1)<<rem_depth; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1970 BlockNode *b= &s->block[index]; |
4408
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
1971 const BlockNode *left = x ? &s->block[index-1] : &null_block; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
1972 const BlockNode *top = y ? &s->block[index-w] : &null_block; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
1973 const BlockNode *tl = y && x ? &s->block[index-w-1] : left; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
1974 const BlockNode *tr = y && trx<w && ((x&1)==0 || level==0) ? &s->block[index-w+(1<<rem_depth)] : tl; //FIXME use lt |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1975 int pl = left->color[0]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1976 int pcb= left->color[1]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1977 int pcr= left->color[2]; |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1978 int pmx, pmy; |
3314 | 1979 int ref_context= av_log2(2*left->ref) + av_log2(2*top->ref); |
4001 | 1980 int mx_context= av_log2(2*FFABS(left->mx - top->mx)) + 16*!!b->ref; |
1981 int my_context= av_log2(2*FFABS(left->my - top->my)) + 16*!!b->ref; | |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1982 int s_context= 2*left->level + 2*top->level + tl->level + tr->level; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1983 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1984 if(s->keyframe){ |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
1985 set_blocks(s, level, x, y, pl, pcb, pcr, 0, 0, 0, BLOCK_INTRA); |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1986 return; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1987 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1988 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1989 if(level!=s->block_max_depth){ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1990 if(same_block(b,b+1) && same_block(b,b+w) && same_block(b,b+w+1)){ |
2995
dfc271b90fe6
4mv + iter ME 10l fix (still not bugfree but better)
michael
parents:
2994
diff
changeset
|
1991 put_rac(&s->c, &s->block_state[4 + s_context], 1); |
dfc271b90fe6
4mv + iter ME 10l fix (still not bugfree but better)
michael
parents:
2994
diff
changeset
|
1992 }else{ |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1993 put_rac(&s->c, &s->block_state[4 + s_context], 0); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1994 encode_q_branch2(s, level+1, 2*x+0, 2*y+0); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1995 encode_q_branch2(s, level+1, 2*x+1, 2*y+0); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1996 encode_q_branch2(s, level+1, 2*x+0, 2*y+1); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1997 encode_q_branch2(s, level+1, 2*x+1, 2*y+1); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1998 return; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
1999 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2000 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2001 if(b->type & BLOCK_INTRA){ |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
2002 pred_mv(s, &pmx, &pmy, 0, left, top, tr); |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2003 put_rac(&s->c, &s->block_state[1 + (left->type&1) + (top->type&1)], 1); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2004 put_symbol(&s->c, &s->block_state[32], b->color[0]-pl , 1); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2005 put_symbol(&s->c, &s->block_state[64], b->color[1]-pcb, 1); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2006 put_symbol(&s->c, &s->block_state[96], b->color[2]-pcr, 1); |
3314 | 2007 set_blocks(s, level, x, y, b->color[0], b->color[1], b->color[2], pmx, pmy, 0, BLOCK_INTRA); |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2008 }else{ |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
2009 pred_mv(s, &pmx, &pmy, b->ref, left, top, tr); |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2010 put_rac(&s->c, &s->block_state[1 + (left->type&1) + (top->type&1)], 0); |
3314 | 2011 if(s->ref_frames > 1) |
2012 put_symbol(&s->c, &s->block_state[128 + 1024 + 32*ref_context], b->ref, 0); | |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2013 put_symbol(&s->c, &s->block_state[128 + 32*mx_context], b->mx - pmx, 1); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2014 put_symbol(&s->c, &s->block_state[128 + 32*my_context], b->my - pmy, 1); |
3314 | 2015 set_blocks(s, level, x, y, pl, pcb, pcr, b->mx, b->my, b->ref, 0); |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2016 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2017 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2018 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2019 static void decode_q_branch(SnowContext *s, int level, int x, int y){ |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2020 const int w= s->b_width << s->block_max_depth; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2021 const int rem_depth= s->block_max_depth - level; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2022 const int index= (x + y*w) << rem_depth; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2023 int trx= (x+1)<<rem_depth; |
4408
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
2024 const BlockNode *left = x ? &s->block[index-1] : &null_block; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
2025 const BlockNode *top = y ? &s->block[index-w] : &null_block; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
2026 const BlockNode *tl = y && x ? &s->block[index-w-1] : left; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
2027 const BlockNode *tr = y && trx<w && ((x&1)==0 || level==0) ? &s->block[index-w+(1<<rem_depth)] : tl; //FIXME use lt |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2028 int s_context= 2*left->level + 2*top->level + tl->level + tr->level; |
2967 | 2029 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2030 if(s->keyframe){ |
3314 | 2031 set_blocks(s, level, x, y, null_block.color[0], null_block.color[1], null_block.color[2], null_block.mx, null_block.my, null_block.ref, BLOCK_INTRA); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2032 return; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2033 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2034 |
2335 | 2035 if(level==s->block_max_depth || get_rac(&s->c, &s->block_state[4 + s_context])){ |
4332 | 2036 int type, mx, my; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2037 int l = left->color[0]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2038 int cb= left->color[1]; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2039 int cr= left->color[2]; |
3314 | 2040 int ref = 0; |
2041 int ref_context= av_log2(2*left->ref) + av_log2(2*top->ref); | |
4001 | 2042 int mx_context= av_log2(2*FFABS(left->mx - top->mx)) + 0*av_log2(2*FFABS(tr->mx - top->mx)); |
2043 int my_context= av_log2(2*FFABS(left->my - top->my)) + 0*av_log2(2*FFABS(tr->my - top->my)); | |
2967 | 2044 |
2335 | 2045 type= get_rac(&s->c, &s->block_state[1 + left->type + top->type]) ? BLOCK_INTRA : 0; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2046 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2047 if(type){ |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
2048 pred_mv(s, &mx, &my, 0, left, top, tr); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2049 l += get_symbol(&s->c, &s->block_state[32], 1); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2050 cb+= get_symbol(&s->c, &s->block_state[64], 1); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2051 cr+= get_symbol(&s->c, &s->block_state[96], 1); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2052 }else{ |
3314 | 2053 if(s->ref_frames > 1) |
2054 ref= get_symbol(&s->c, &s->block_state[128 + 1024 + 32*ref_context], 0); | |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
2055 pred_mv(s, &mx, &my, ref, left, top, tr); |
3314 | 2056 mx+= get_symbol(&s->c, &s->block_state[128 + 32*(mx_context + 16*!!ref)], 1); |
2057 my+= get_symbol(&s->c, &s->block_state[128 + 32*(my_context + 16*!!ref)], 1); | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2058 } |
3314 | 2059 set_blocks(s, level, x, y, l, cb, cr, mx, my, ref, type); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2060 }else{ |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2061 decode_q_branch(s, level+1, 2*x+0, 2*y+0); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2062 decode_q_branch(s, level+1, 2*x+1, 2*y+0); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2063 decode_q_branch(s, level+1, 2*x+0, 2*y+1); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2064 decode_q_branch(s, level+1, 2*x+1, 2*y+1); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2065 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2066 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2067 |
3661
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
2068 static void encode_blocks(SnowContext *s, int search){ |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2069 int x, y; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2070 int w= s->b_width; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2071 int h= s->b_height; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2072 |
3661
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
2073 if(s->avctx->me_method == ME_ITER && !s->keyframe && search) |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2074 iterative_me(s); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2075 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2076 for(y=0; y<h; y++){ |
2435
c89ac0e70c66
10l patch by (matthieu castet <castet.matthieu free fr>)
michael
parents:
2422
diff
changeset
|
2077 if(s->c.bytestream_end - s->c.bytestream < w*MB_SIZE*MB_SIZE*3){ //FIXME nicer limit |
2422 | 2078 av_log(s->avctx, AV_LOG_ERROR, "encoded frame too large\n"); |
2079 return; | |
2080 } | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2081 for(x=0; x<w; x++){ |
3661
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
2082 if(s->avctx->me_method == ME_ITER || !search) |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2083 encode_q_branch2(s, 0, x, y); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2084 else |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2085 encode_q_branch (s, 0, x, y); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2086 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2087 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2088 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2089 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2090 static void decode_blocks(SnowContext *s){ |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2091 int x, y; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2092 int w= s->b_width; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2093 int h= s->b_height; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2094 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2095 for(y=0; y<h; y++){ |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2096 for(x=0; x<w; x++){ |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2097 decode_q_branch(s, 0, x, y); |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2098 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2099 } |
2138 | 2100 } |
2101 | |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2102 static void mc_block(Plane *p, uint8_t *dst, const uint8_t *src, uint8_t *tmp, int stride, int b_w, int b_h, int dx, int dy){ |
7622 | 2103 static const uint8_t weight[64]={ |
5648 | 2104 8,7,6,5,4,3,2,1, |
2105 7,7,0,0,0,0,0,1, | |
2106 6,0,6,0,0,0,2,0, | |
2107 5,0,0,5,0,3,0,0, | |
2108 4,0,0,0,4,0,0,0, | |
2109 3,0,0,5,0,3,0,0, | |
2110 2,0,6,0,0,0,2,0, | |
2111 1,7,0,0,0,0,0,1, | |
2112 }; | |
2113 | |
7622 | 2114 static const uint8_t brane[256]={ |
5648 | 2115 0x00,0x01,0x01,0x01,0x01,0x01,0x01,0x01,0x11,0x12,0x12,0x12,0x12,0x12,0x12,0x12, |
2116 0x04,0x05,0xcc,0xcc,0xcc,0xcc,0xcc,0x41,0x15,0x16,0xcc,0xcc,0xcc,0xcc,0xcc,0x52, | |
2117 0x04,0xcc,0x05,0xcc,0xcc,0xcc,0x41,0xcc,0x15,0xcc,0x16,0xcc,0xcc,0xcc,0x52,0xcc, | |
2118 0x04,0xcc,0xcc,0x05,0xcc,0x41,0xcc,0xcc,0x15,0xcc,0xcc,0x16,0xcc,0x52,0xcc,0xcc, | |
2119 0x04,0xcc,0xcc,0xcc,0x41,0xcc,0xcc,0xcc,0x15,0xcc,0xcc,0xcc,0x16,0xcc,0xcc,0xcc, | |
2120 0x04,0xcc,0xcc,0x41,0xcc,0x05,0xcc,0xcc,0x15,0xcc,0xcc,0x52,0xcc,0x16,0xcc,0xcc, | |
2121 0x04,0xcc,0x41,0xcc,0xcc,0xcc,0x05,0xcc,0x15,0xcc,0x52,0xcc,0xcc,0xcc,0x16,0xcc, | |
2122 0x04,0x41,0xcc,0xcc,0xcc,0xcc,0xcc,0x05,0x15,0x52,0xcc,0xcc,0xcc,0xcc,0xcc,0x16, | |
2123 0x44,0x45,0x45,0x45,0x45,0x45,0x45,0x45,0x55,0x56,0x56,0x56,0x56,0x56,0x56,0x56, | |
2124 0x48,0x49,0xcc,0xcc,0xcc,0xcc,0xcc,0x85,0x59,0x5A,0xcc,0xcc,0xcc,0xcc,0xcc,0x96, | |
2125 0x48,0xcc,0x49,0xcc,0xcc,0xcc,0x85,0xcc,0x59,0xcc,0x5A,0xcc,0xcc,0xcc,0x96,0xcc, | |
2126 0x48,0xcc,0xcc,0x49,0xcc,0x85,0xcc,0xcc,0x59,0xcc,0xcc,0x5A,0xcc,0x96,0xcc,0xcc, | |
2127 0x48,0xcc,0xcc,0xcc,0x49,0xcc,0xcc,0xcc,0x59,0xcc,0xcc,0xcc,0x96,0xcc,0xcc,0xcc, | |
2128 0x48,0xcc,0xcc,0x85,0xcc,0x49,0xcc,0xcc,0x59,0xcc,0xcc,0x96,0xcc,0x5A,0xcc,0xcc, | |
2129 0x48,0xcc,0x85,0xcc,0xcc,0xcc,0x49,0xcc,0x59,0xcc,0x96,0xcc,0xcc,0xcc,0x5A,0xcc, | |
2130 0x48,0x85,0xcc,0xcc,0xcc,0xcc,0xcc,0x49,0x59,0x96,0xcc,0xcc,0xcc,0xcc,0xcc,0x5A, | |
2131 }; | |
2132 | |
7622 | 2133 static const uint8_t needs[16]={ |
5648 | 2134 0,1,0,0, |
2135 2,4,2,0, | |
2136 0,1,0,0, | |
2137 15 | |
2138 }; | |
2139 | |
2140 int x, y, b, r, l; | |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2141 int16_t tmpIt [64*(32+HTAPS_MAX)]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2142 uint8_t tmp2t[3][stride*(32+HTAPS_MAX)]; |
5648 | 2143 int16_t *tmpI= tmpIt; |
2144 uint8_t *tmp2= tmp2t[0]; | |
6205 | 2145 const uint8_t *hpel[11]; |
5648 | 2146 assert(dx<16 && dy<16); |
2147 r= brane[dx + 16*dy]&15; | |
2148 l= brane[dx + 16*dy]>>4; | |
2149 | |
2150 b= needs[l] | needs[r]; | |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2151 if(p && !p->diag_mc) |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2152 b= 15; |
5648 | 2153 |
2154 if(b&5){ | |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2155 for(y=0; y < b_h+HTAPS_MAX-1; y++){ |
5649 | 2156 for(x=0; x < b_w; x++){ |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2157 int a_1=src[x + HTAPS_MAX/2-4]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2158 int a0= src[x + HTAPS_MAX/2-3]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2159 int a1= src[x + HTAPS_MAX/2-2]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2160 int a2= src[x + HTAPS_MAX/2-1]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2161 int a3= src[x + HTAPS_MAX/2+0]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2162 int a4= src[x + HTAPS_MAX/2+1]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2163 int a5= src[x + HTAPS_MAX/2+2]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2164 int a6= src[x + HTAPS_MAX/2+3]; |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2165 int am=0; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2166 if(!p || p->fast_mc){ |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2167 am= 20*(a2+a3) - 5*(a1+a4) + (a0+a5); |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2168 tmpI[x]= am; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2169 am= (am+16)>>5; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2170 }else{ |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2171 am= p->hcoeff[0]*(a2+a3) + p->hcoeff[1]*(a1+a4) + p->hcoeff[2]*(a0+a5) + p->hcoeff[3]*(a_1+a6); |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2172 tmpI[x]= am; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2173 am= (am+32)>>6; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2174 } |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2175 |
5649 | 2176 if(am&(~255)) am= ~(am>>31); |
2177 tmp2[x]= am; | |
2178 } | |
2179 tmpI+= 64; | |
2180 tmp2+= stride; | |
2181 src += stride; | |
2138 | 2182 } |
5649 | 2183 src -= stride*y; |
5648 | 2184 } |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2185 src += HTAPS_MAX/2 - 1; |
5648 | 2186 tmp2= tmp2t[1]; |
2187 | |
2188 if(b&2){ | |
5649 | 2189 for(y=0; y < b_h; y++){ |
2190 for(x=0; x < b_w+1; x++){ | |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2191 int a_1=src[x + (HTAPS_MAX/2-4)*stride]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2192 int a0= src[x + (HTAPS_MAX/2-3)*stride]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2193 int a1= src[x + (HTAPS_MAX/2-2)*stride]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2194 int a2= src[x + (HTAPS_MAX/2-1)*stride]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2195 int a3= src[x + (HTAPS_MAX/2+0)*stride]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2196 int a4= src[x + (HTAPS_MAX/2+1)*stride]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2197 int a5= src[x + (HTAPS_MAX/2+2)*stride]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2198 int a6= src[x + (HTAPS_MAX/2+3)*stride]; |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2199 int am=0; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2200 if(!p || p->fast_mc) |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2201 am= (20*(a2+a3) - 5*(a1+a4) + (a0+a5) + 16)>>5; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2202 else |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2203 am= (p->hcoeff[0]*(a2+a3) + p->hcoeff[1]*(a1+a4) + p->hcoeff[2]*(a0+a5) + p->hcoeff[3]*(a_1+a6) + 32)>>6; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2204 |
5649 | 2205 if(am&(~255)) am= ~(am>>31); |
2206 tmp2[x]= am; | |
2207 } | |
2208 src += stride; | |
2209 tmp2+= stride; | |
5648 | 2210 } |
5649 | 2211 src -= stride*y; |
5648 | 2212 } |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2213 src += stride*(HTAPS_MAX/2 - 1); |
5648 | 2214 tmp2= tmp2t[2]; |
2215 tmpI= tmpIt; | |
2216 if(b&4){ | |
2217 for(y=0; y < b_h; y++){ | |
2218 for(x=0; x < b_w; x++){ | |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2219 int a_1=tmpI[x + (HTAPS_MAX/2-4)*64]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2220 int a0= tmpI[x + (HTAPS_MAX/2-3)*64]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2221 int a1= tmpI[x + (HTAPS_MAX/2-2)*64]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2222 int a2= tmpI[x + (HTAPS_MAX/2-1)*64]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2223 int a3= tmpI[x + (HTAPS_MAX/2+0)*64]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2224 int a4= tmpI[x + (HTAPS_MAX/2+1)*64]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2225 int a5= tmpI[x + (HTAPS_MAX/2+2)*64]; |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2226 int a6= tmpI[x + (HTAPS_MAX/2+3)*64]; |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2227 int am=0; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2228 if(!p || p->fast_mc) |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2229 am= (20*(a2+a3) - 5*(a1+a4) + (a0+a5) + 512)>>10; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2230 else |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2231 am= (p->hcoeff[0]*(a2+a3) + p->hcoeff[1]*(a1+a4) + p->hcoeff[2]*(a0+a5) + p->hcoeff[3]*(a_1+a6) + 2048)>>12; |
5648 | 2232 if(am&(~255)) am= ~(am>>31); |
2233 tmp2[x]= am; | |
2234 } | |
2235 tmpI+= 64; | |
2236 tmp2+= stride; | |
2138 | 2237 } |
5648 | 2238 } |
2239 | |
2240 hpel[ 0]= src; | |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2241 hpel[ 1]= tmp2t[0] + stride*(HTAPS_MAX/2-1); |
5648 | 2242 hpel[ 2]= src + 1; |
2243 | |
2244 hpel[ 4]= tmp2t[1]; | |
2245 hpel[ 5]= tmp2t[2]; | |
2246 hpel[ 6]= tmp2t[1] + 1; | |
2247 | |
2248 hpel[ 8]= src + stride; | |
2249 hpel[ 9]= hpel[1] + stride; | |
2250 hpel[10]= hpel[8] + 1; | |
2251 | |
2252 if(b==15){ | |
6205 | 2253 const uint8_t *src1= hpel[dx/8 + dy/8*4 ]; |
2254 const uint8_t *src2= hpel[dx/8 + dy/8*4+1]; | |
2255 const uint8_t *src3= hpel[dx/8 + dy/8*4+4]; | |
2256 const uint8_t *src4= hpel[dx/8 + dy/8*4+5]; | |
5648 | 2257 dx&=7; |
2258 dy&=7; | |
2259 for(y=0; y < b_h; y++){ | |
2260 for(x=0; x < b_w; x++){ | |
2261 dst[x]= ((8-dx)*(8-dy)*src1[x] + dx*(8-dy)*src2[x]+ | |
2262 (8-dx)* dy *src3[x] + dx* dy *src4[x]+32)>>6; | |
2263 } | |
2264 src1+=stride; | |
2265 src2+=stride; | |
2266 src3+=stride; | |
2267 src4+=stride; | |
2268 dst +=stride; | |
2269 } | |
2270 }else{ | |
6205 | 2271 const uint8_t *src1= hpel[l]; |
2272 const uint8_t *src2= hpel[r]; | |
5648 | 2273 int a= weight[((dx&7) + (8*(dy&7)))]; |
2274 int b= 8-a; | |
2275 for(y=0; y < b_h; y++){ | |
2276 for(x=0; x < b_w; x++){ | |
2277 dst[x]= (a*src1[x] + b*src2[x] + 4)>>3; | |
2278 } | |
2279 src1+=stride; | |
2280 src2+=stride; | |
2281 dst +=stride; | |
2282 } | |
2138 | 2283 } |
2284 } | |
2285 | |
2286 #define mca(dx,dy,b_w)\ | |
5254 | 2287 static void mc_block_hpel ## dx ## dy ## b_w(uint8_t *dst, const uint8_t *src, int stride, int h){\ |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2288 uint8_t tmp[stride*(b_w+HTAPS_MAX-1)];\ |
2138 | 2289 assert(h==b_w);\ |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2290 mc_block(NULL, dst, src-(HTAPS_MAX/2-1)-(HTAPS_MAX/2-1)*stride, tmp, stride, b_w, b_w, dx, dy);\ |
2138 | 2291 } |
2292 | |
2293 mca( 0, 0,16) | |
2294 mca( 8, 0,16) | |
2295 mca( 0, 8,16) | |
2296 mca( 8, 8,16) | |
2224
11d54cb7ac4e
100l (forgot to set the 8x8 mc functions, so ME did use the mpeg ones ...)
michael
parents:
2223
diff
changeset
|
2297 mca( 0, 0,8) |
11d54cb7ac4e
100l (forgot to set the 8x8 mc functions, so ME did use the mpeg ones ...)
michael
parents:
2223
diff
changeset
|
2298 mca( 8, 0,8) |
11d54cb7ac4e
100l (forgot to set the 8x8 mc functions, so ME did use the mpeg ones ...)
michael
parents:
2223
diff
changeset
|
2299 mca( 0, 8,8) |
11d54cb7ac4e
100l (forgot to set the 8x8 mc functions, so ME did use the mpeg ones ...)
michael
parents:
2223
diff
changeset
|
2300 mca( 8, 8,8) |
2138 | 2301 |
3314 | 2302 static void pred_block(SnowContext *s, uint8_t *dst, uint8_t *tmp, int stride, int sx, int sy, int b_w, int b_h, BlockNode *block, int plane_index, int w, int h){ |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2303 if(block->type & BLOCK_INTRA){ |
2206 | 2304 int x, y; |
3018 | 2305 const int color = block->color[plane_index]; |
2306 const int color4= color*0x01010101; | |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2307 if(b_w==32){ |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2308 for(y=0; y < b_h; y++){ |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2309 *(uint32_t*)&dst[0 + y*stride]= color4; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2310 *(uint32_t*)&dst[4 + y*stride]= color4; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2311 *(uint32_t*)&dst[8 + y*stride]= color4; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2312 *(uint32_t*)&dst[12+ y*stride]= color4; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2313 *(uint32_t*)&dst[16+ y*stride]= color4; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2314 *(uint32_t*)&dst[20+ y*stride]= color4; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2315 *(uint32_t*)&dst[24+ y*stride]= color4; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2316 *(uint32_t*)&dst[28+ y*stride]= color4; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2317 } |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2318 }else if(b_w==16){ |
3018 | 2319 for(y=0; y < b_h; y++){ |
2320 *(uint32_t*)&dst[0 + y*stride]= color4; | |
2321 *(uint32_t*)&dst[4 + y*stride]= color4; | |
2322 *(uint32_t*)&dst[8 + y*stride]= color4; | |
2323 *(uint32_t*)&dst[12+ y*stride]= color4; | |
2324 } | |
2325 }else if(b_w==8){ | |
2326 for(y=0; y < b_h; y++){ | |
2327 *(uint32_t*)&dst[0 + y*stride]= color4; | |
2328 *(uint32_t*)&dst[4 + y*stride]= color4; | |
2329 } | |
2330 }else if(b_w==4){ | |
2331 for(y=0; y < b_h; y++){ | |
2332 *(uint32_t*)&dst[0 + y*stride]= color4; | |
2333 } | |
2334 }else{ | |
2335 for(y=0; y < b_h; y++){ | |
2336 for(x=0; x < b_w; x++){ | |
2337 dst[x + y*stride]= color; | |
2338 } | |
2138 | 2339 } |
2340 } | |
2341 }else{ | |
3314 | 2342 uint8_t *src= s->last_picture[block->ref].data[plane_index]; |
2206 | 2343 const int scale= plane_index ? s->mv_scale : 2*s->mv_scale; |
2344 int mx= block->mx*scale; | |
2345 int my= block->my*scale; | |
2223 | 2346 const int dx= mx&15; |
2347 const int dy= my&15; | |
3020
c75fb0747e74
use h264 MC functions for 2xX Xx2 blocks in snow too
michael
parents:
3018
diff
changeset
|
2348 const int tab_index= 3 - (b_w>>2) + (b_w>>4); |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2349 sx += (mx>>4) - (HTAPS_MAX/2-1); |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2350 sy += (my>>4) - (HTAPS_MAX/2-1); |
2206 | 2351 src += sx + sy*stride; |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2352 if( (unsigned)sx >= w - b_w - (HTAPS_MAX-2) |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2353 || (unsigned)sy >= h - b_h - (HTAPS_MAX-2)){ |
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2354 ff_emulated_edge_mc(tmp + MB_SIZE, src, stride, b_w+HTAPS_MAX-1, b_h+HTAPS_MAX-1, sx, sy, w, h); |
2206 | 2355 src= tmp + MB_SIZE; |
2138 | 2356 } |
3189 | 2357 // assert(b_w == b_h || 2*b_w == b_h || b_w == 2*b_h); |
2358 // assert(!(b_w&(b_w-1))); | |
3018 | 2359 assert(b_w>1 && b_h>1); |
6206
ea591bc7522a
add parenthesis, fix warning: snow.c:2413: warning: suggest parentheses around && within ||
bcoudurier
parents:
6205
diff
changeset
|
2360 assert((tab_index>=0 && tab_index<4) || b_w==32); |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2361 if((dx&3) || (dy&3) || !(b_w == b_h || 2*b_w == b_h || b_w == 2*b_h) || (b_w&(b_w-1)) || !s->plane[plane_index].fast_mc ) |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2362 mc_block(&s->plane[plane_index], dst, src, tmp, stride, b_w, b_h, dx, dy); |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2363 else if(b_w==32){ |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2364 int y; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2365 for(y=0; y<b_h; y+=16){ |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2366 s->dsp.put_h264_qpel_pixels_tab[0][dy+(dx>>2)](dst + y*stride, src + 3 + (y+3)*stride,stride); |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2367 s->dsp.put_h264_qpel_pixels_tab[0][dy+(dx>>2)](dst + 16 + y*stride, src + 19 + (y+3)*stride,stride); |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2368 } |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2369 }else if(b_w==b_h) |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2370 s->dsp.put_h264_qpel_pixels_tab[tab_index ][dy+(dx>>2)](dst,src + 3 + 3*stride,stride); |
3018 | 2371 else if(b_w==2*b_h){ |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2372 s->dsp.put_h264_qpel_pixels_tab[tab_index+1][dy+(dx>>2)](dst ,src + 3 + 3*stride,stride); |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2373 s->dsp.put_h264_qpel_pixels_tab[tab_index+1][dy+(dx>>2)](dst+b_h,src + 3 + b_h + 3*stride,stride); |
3018 | 2374 }else{ |
2375 assert(2*b_w==b_h); | |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2376 s->dsp.put_h264_qpel_pixels_tab[tab_index ][dy+(dx>>2)](dst ,src + 3 + 3*stride ,stride); |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
2377 s->dsp.put_h264_qpel_pixels_tab[tab_index ][dy+(dx>>2)](dst+b_w*stride,src + 3 + 3*stride+b_w*stride,stride); |
3018 | 2378 } |
2138 | 2379 } |
2380 } | |
2381 | |
4436
d3e389536b0a
Add the const specifier as needed to reduce the number of warnings.
takis
parents:
4409
diff
changeset
|
2382 void ff_snow_inner_add_yblock(const uint8_t *obmc, const int obmc_stride, uint8_t * * block, int b_w, int b_h, |
3198
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2383 int src_x, int src_y, int src_stride, slice_buffer * sb, int add, uint8_t * dst8){ |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2384 int y, x; |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2385 IDWTELEM * dst; |
3198
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2386 for(y=0; y<b_h; y++){ |
5409 | 2387 //FIXME ugly misuse of obmc_stride |
4436
d3e389536b0a
Add the const specifier as needed to reduce the number of warnings.
takis
parents:
4409
diff
changeset
|
2388 const uint8_t *obmc1= obmc + y*obmc_stride; |
d3e389536b0a
Add the const specifier as needed to reduce the number of warnings.
takis
parents:
4409
diff
changeset
|
2389 const uint8_t *obmc2= obmc1+ (obmc_stride>>1); |
d3e389536b0a
Add the const specifier as needed to reduce the number of warnings.
takis
parents:
4409
diff
changeset
|
2390 const uint8_t *obmc3= obmc1+ obmc_stride*(obmc_stride>>1); |
d3e389536b0a
Add the const specifier as needed to reduce the number of warnings.
takis
parents:
4409
diff
changeset
|
2391 const uint8_t *obmc4= obmc3+ (obmc_stride>>1); |
3198
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2392 dst = slice_buffer_get_line(sb, src_y + y); |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2393 for(x=0; x<b_w; x++){ |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2394 int v= obmc1[x] * block[3][x + y*src_stride] |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2395 +obmc2[x] * block[2][x + y*src_stride] |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2396 +obmc3[x] * block[1][x + y*src_stride] |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2397 +obmc4[x] * block[0][x + y*src_stride]; |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2398 |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2399 v <<= 8 - LOG2_OBMC_MAX; |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2400 if(FRAC_BITS != 8){ |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2401 v >>= 8 - FRAC_BITS; |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2402 } |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2403 if(add){ |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2404 v += dst[x + src_x]; |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2405 v = (v + (1<<(FRAC_BITS-1))) >> FRAC_BITS; |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2406 if(v&(~255)) v= ~(v>>31); |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2407 dst8[x + y*src_stride] = v; |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2408 }else{ |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2409 dst[x + src_x] -= v; |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2410 } |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2411 } |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2412 } |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2413 } |
6b9f0c4fbdbe
First part of a series of speed-enchancing patches.
gpoirier
parents:
3197
diff
changeset
|
2414 |
6412 | 2415 //FIXME name cleanup (b_w, block_w, b_width stuff) |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2416 static av_always_inline void add_yblock(SnowContext *s, int sliced, slice_buffer *sb, IDWTELEM *dst, uint8_t *dst8, const uint8_t *obmc, int src_x, int src_y, int b_w, int b_h, int w, int h, int dst_stride, int src_stride, int obmc_stride, int b_x, int b_y, int add, int offset_dst, int plane_index){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2417 const int b_width = s->b_width << s->block_max_depth; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2418 const int b_height= s->b_height << s->block_max_depth; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2419 const int b_stride= b_width; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2420 BlockNode *lt= &s->block[b_x + b_y*b_stride]; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2421 BlockNode *rt= lt+1; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2422 BlockNode *lb= lt+b_stride; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2423 BlockNode *rb= lb+1; |
2967 | 2424 uint8_t *block[4]; |
2842 | 2425 int tmp_step= src_stride >= 7*MB_SIZE ? MB_SIZE : MB_SIZE*src_stride; |
2426 uint8_t tmp[src_stride*7*MB_SIZE]; //FIXME align | |
2427 uint8_t *ptmp; | |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2428 int x,y; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2429 |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2430 if(b_x<0){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2431 lt= rt; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2432 lb= rb; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2433 }else if(b_x + 1 >= b_width){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2434 rt= lt; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2435 rb= lb; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2436 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2437 if(b_y<0){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2438 lt= lb; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2439 rt= rb; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2440 }else if(b_y + 1 >= b_height){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2441 lb= lt; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2442 rb= rt; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2443 } |
2967 | 2444 |
6412 | 2445 if(src_x<0){ //FIXME merge with prev & always round internal width up to *16 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2446 obmc -= src_x; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2447 b_w += src_x; |
3662
fc714e9a5419
snow cosmetics: merge the sliced and non-sliced versions of add_yblock
lorenm
parents:
3661
diff
changeset
|
2448 if(!sliced && !offset_dst) |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2449 dst -= src_x; |
2206 | 2450 src_x=0; |
2451 }else if(src_x + b_w > w){ | |
2452 b_w = w - src_x; | |
2453 } | |
2454 if(src_y<0){ | |
2455 obmc -= src_y*obmc_stride; | |
2456 b_h += src_y; | |
3662
fc714e9a5419
snow cosmetics: merge the sliced and non-sliced versions of add_yblock
lorenm
parents:
3661
diff
changeset
|
2457 if(!sliced && !offset_dst) |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2458 dst -= src_y*dst_stride; |
2206 | 2459 src_y=0; |
2460 }else if(src_y + b_h> h){ | |
2461 b_h = h - src_y; | |
2462 } | |
2967 | 2463 |
2206 | 2464 if(b_w<=0 || b_h<=0) return; |
2465 | |
6414 | 2466 assert(src_stride > 2*MB_SIZE + 5); |
2467 | |
3662
fc714e9a5419
snow cosmetics: merge the sliced and non-sliced versions of add_yblock
lorenm
parents:
3661
diff
changeset
|
2468 if(!sliced && offset_dst) |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2469 dst += src_x + src_y*dst_stride; |
2249
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2470 dst8+= src_x + src_y*src_stride; |
2206 | 2471 // src += src_x + src_y*src_stride; |
2472 | |
2842 | 2473 ptmp= tmp + 3*tmp_step; |
2474 block[0]= ptmp; | |
2475 ptmp+=tmp_step; | |
3314 | 2476 pred_block(s, block[0], tmp, src_stride, src_x, src_y, b_w, b_h, lt, plane_index, w, h); |
2206 | 2477 |
2478 if(same_block(lt, rt)){ | |
2479 block[1]= block[0]; | |
2480 }else{ | |
2842 | 2481 block[1]= ptmp; |
2482 ptmp+=tmp_step; | |
3314 | 2483 pred_block(s, block[1], tmp, src_stride, src_x, src_y, b_w, b_h, rt, plane_index, w, h); |
2206 | 2484 } |
2967 | 2485 |
2206 | 2486 if(same_block(lt, lb)){ |
2487 block[2]= block[0]; | |
2488 }else if(same_block(rt, lb)){ | |
2489 block[2]= block[1]; | |
2490 }else{ | |
2842 | 2491 block[2]= ptmp; |
2492 ptmp+=tmp_step; | |
3314 | 2493 pred_block(s, block[2], tmp, src_stride, src_x, src_y, b_w, b_h, lb, plane_index, w, h); |
2206 | 2494 } |
2495 | |
2496 if(same_block(lt, rb) ){ | |
2497 block[3]= block[0]; | |
2498 }else if(same_block(rt, rb)){ | |
2499 block[3]= block[1]; | |
2500 }else if(same_block(lb, rb)){ | |
2501 block[3]= block[2]; | |
2502 }else{ | |
2842 | 2503 block[3]= ptmp; |
3314 | 2504 pred_block(s, block[3], tmp, src_stride, src_x, src_y, b_w, b_h, rb, plane_index, w, h); |
2206 | 2505 } |
2506 #if 0 | |
2507 for(y=0; y<b_h; y++){ | |
2508 for(x=0; x<b_w; x++){ | |
2509 int v= obmc [x + y*obmc_stride] * block[3][x + y*src_stride] * (256/OBMC_MAX); | |
2510 if(add) dst[x + y*dst_stride] += v; | |
2511 else dst[x + y*dst_stride] -= v; | |
2512 } | |
2513 } | |
2514 for(y=0; y<b_h; y++){ | |
2515 uint8_t *obmc2= obmc + (obmc_stride>>1); | |
2516 for(x=0; x<b_w; x++){ | |
2517 int v= obmc2[x + y*obmc_stride] * block[2][x + y*src_stride] * (256/OBMC_MAX); | |
2518 if(add) dst[x + y*dst_stride] += v; | |
2519 else dst[x + y*dst_stride] -= v; | |
2520 } | |
2521 } | |
2522 for(y=0; y<b_h; y++){ | |
2523 uint8_t *obmc3= obmc + obmc_stride*(obmc_stride>>1); | |
2524 for(x=0; x<b_w; x++){ | |
2525 int v= obmc3[x + y*obmc_stride] * block[1][x + y*src_stride] * (256/OBMC_MAX); | |
2526 if(add) dst[x + y*dst_stride] += v; | |
2527 else dst[x + y*dst_stride] -= v; | |
2528 } | |
2529 } | |
2530 for(y=0; y<b_h; y++){ | |
2531 uint8_t *obmc3= obmc + obmc_stride*(obmc_stride>>1); | |
2532 uint8_t *obmc4= obmc3+ (obmc_stride>>1); | |
2533 for(x=0; x<b_w; x++){ | |
2534 int v= obmc4[x + y*obmc_stride] * block[0][x + y*src_stride] * (256/OBMC_MAX); | |
2535 if(add) dst[x + y*dst_stride] += v; | |
2536 else dst[x + y*dst_stride] -= v; | |
2537 } | |
2538 } | |
2539 #else | |
3662
fc714e9a5419
snow cosmetics: merge the sliced and non-sliced versions of add_yblock
lorenm
parents:
3661
diff
changeset
|
2540 if(sliced){ |
fc714e9a5419
snow cosmetics: merge the sliced and non-sliced versions of add_yblock
lorenm
parents:
3661
diff
changeset
|
2541 s->dsp.inner_add_yblock(obmc, obmc_stride, block, b_w, b_h, src_x,src_y, src_stride, sb, add, dst8); |
6475 | 2542 }else{ |
2543 for(y=0; y<b_h; y++){ | |
2544 //FIXME ugly misuse of obmc_stride | |
2545 const uint8_t *obmc1= obmc + y*obmc_stride; | |
2546 const uint8_t *obmc2= obmc1+ (obmc_stride>>1); | |
2547 const uint8_t *obmc3= obmc1+ obmc_stride*(obmc_stride>>1); | |
2548 const uint8_t *obmc4= obmc3+ (obmc_stride>>1); | |
2549 for(x=0; x<b_w; x++){ | |
2550 int v= obmc1[x] * block[3][x + y*src_stride] | |
2551 +obmc2[x] * block[2][x + y*src_stride] | |
2552 +obmc3[x] * block[1][x + y*src_stride] | |
2553 +obmc4[x] * block[0][x + y*src_stride]; | |
2554 | |
2555 v <<= 8 - LOG2_OBMC_MAX; | |
2556 if(FRAC_BITS != 8){ | |
2557 v >>= 8 - FRAC_BITS; | |
2558 } | |
2559 if(add){ | |
2560 v += dst[x + y*dst_stride]; | |
2561 v = (v + (1<<(FRAC_BITS-1))) >> FRAC_BITS; | |
2562 if(v&(~255)) v= ~(v>>31); | |
2563 dst8[x + y*src_stride] = v; | |
2564 }else{ | |
2565 dst[x + y*dst_stride] -= v; | |
2566 } | |
2249
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2567 } |
2206 | 2568 } |
2569 } | |
5910 | 2570 #endif /* 0 */ |
2206 | 2571 } |
2572 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2573 static av_always_inline void predict_slice_buffered(SnowContext *s, slice_buffer * sb, IDWTELEM * old_buffer, int plane_index, int add, int mb_y){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2574 Plane *p= &s->plane[plane_index]; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2575 const int mb_w= s->b_width << s->block_max_depth; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2576 const int mb_h= s->b_height << s->block_max_depth; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2577 int x, y, mb_x; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2578 int block_size = MB_SIZE >> s->block_max_depth; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2579 int block_w = plane_index ? block_size/2 : block_size; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2580 const uint8_t *obmc = plane_index ? obmc_tab[s->block_max_depth+1] : obmc_tab[s->block_max_depth]; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2581 int obmc_stride= plane_index ? block_size : 2*block_size; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2582 int ref_stride= s->current_picture.linesize[plane_index]; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2583 uint8_t *dst8= s->current_picture.data[plane_index]; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2584 int w= p->width; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2585 int h= p->height; |
2967 | 2586 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2587 if(s->keyframe || (s->avctx->debug&512)){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2588 if(mb_y==mb_h) |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2589 return; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2590 |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2591 if(add){ |
6474
e6995f3fbf7f
cosmetics: Normalize {} placement after for, while, if.
diego
parents:
6437
diff
changeset
|
2592 for(y=block_w*mb_y; y<FFMIN(h,block_w*(mb_y+1)); y++){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2593 // DWTELEM * line = slice_buffer_get_line(sb, y); |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2594 IDWTELEM * line = sb->line[y]; |
6474
e6995f3fbf7f
cosmetics: Normalize {} placement after for, while, if.
diego
parents:
6437
diff
changeset
|
2595 for(x=0; x<w; x++){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2596 // int v= buf[x + y*w] + (128<<FRAC_BITS) + (1<<(FRAC_BITS-1)); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2597 int v= line[x] + (128<<FRAC_BITS) + (1<<(FRAC_BITS-1)); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2598 v >>= FRAC_BITS; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2599 if(v&(~255)) v= ~(v>>31); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2600 dst8[x + y*ref_stride]= v; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2601 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2602 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2603 }else{ |
6474
e6995f3fbf7f
cosmetics: Normalize {} placement after for, while, if.
diego
parents:
6437
diff
changeset
|
2604 for(y=block_w*mb_y; y<FFMIN(h,block_w*(mb_y+1)); y++){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2605 // DWTELEM * line = slice_buffer_get_line(sb, y); |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2606 IDWTELEM * line = sb->line[y]; |
6474
e6995f3fbf7f
cosmetics: Normalize {} placement after for, while, if.
diego
parents:
6437
diff
changeset
|
2607 for(x=0; x<w; x++){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2608 line[x] -= 128 << FRAC_BITS; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2609 // buf[x + y*w]-= 128<<FRAC_BITS; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2610 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2611 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2612 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2613 |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2614 return; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2615 } |
2967 | 2616 |
6475 | 2617 for(mb_x=0; mb_x<=mb_w; mb_x++){ |
2618 add_yblock(s, 1, sb, old_buffer, dst8, obmc, | |
2619 block_w*mb_x - block_w/2, | |
2620 block_w*mb_y - block_w/2, | |
2621 block_w, block_w, | |
2622 w, h, | |
2623 w, ref_stride, obmc_stride, | |
2624 mb_x - 1, mb_y - 1, | |
2625 add, 0, plane_index); | |
2626 } | |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2627 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
2628 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2629 static av_always_inline void predict_slice(SnowContext *s, IDWTELEM *buf, int plane_index, int add, int mb_y){ |
2138 | 2630 Plane *p= &s->plane[plane_index]; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2631 const int mb_w= s->b_width << s->block_max_depth; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2632 const int mb_h= s->b_height << s->block_max_depth; |
2562 | 2633 int x, y, mb_x; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2634 int block_size = MB_SIZE >> s->block_max_depth; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
2635 int block_w = plane_index ? block_size/2 : block_size; |
2206 | 2636 const uint8_t *obmc = plane_index ? obmc_tab[s->block_max_depth+1] : obmc_tab[s->block_max_depth]; |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2637 const int obmc_stride= plane_index ? block_size : 2*block_size; |
2249
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2638 int ref_stride= s->current_picture.linesize[plane_index]; |
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2639 uint8_t *dst8= s->current_picture.data[plane_index]; |
2138 | 2640 int w= p->width; |
2641 int h= p->height; | |
2967 | 2642 |
2206 | 2643 if(s->keyframe || (s->avctx->debug&512)){ |
2562 | 2644 if(mb_y==mb_h) |
2645 return; | |
2646 | |
2249
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2647 if(add){ |
2604
b7e6c3d31c65
Snow segfault bug in revision 1.43 patch by (Yartrebo /yartrebo earthlink net/)
michael
parents:
2602
diff
changeset
|
2648 for(y=block_w*mb_y; y<FFMIN(h,block_w*(mb_y+1)); y++){ |
2249
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2649 for(x=0; x<w; x++){ |
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2650 int v= buf[x + y*w] + (128<<FRAC_BITS) + (1<<(FRAC_BITS-1)); |
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2651 v >>= FRAC_BITS; |
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2652 if(v&(~255)) v= ~(v>>31); |
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2653 dst8[x + y*ref_stride]= v; |
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2654 } |
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2655 } |
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2656 }else{ |
2604
b7e6c3d31c65
Snow segfault bug in revision 1.43 patch by (Yartrebo /yartrebo earthlink net/)
michael
parents:
2602
diff
changeset
|
2657 for(y=block_w*mb_y; y<FFMIN(h,block_w*(mb_y+1)); y++){ |
2249
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2658 for(x=0; x<w; x++){ |
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2659 buf[x + y*w]-= 128<<FRAC_BITS; |
2b1a5e1fd449
merge predict_plane() with DWTELEM->8bit conversation (21% faster)
michael
parents:
2246
diff
changeset
|
2660 } |
2206 | 2661 } |
2138 | 2662 } |
2206 | 2663 |
2664 return; | |
2138 | 2665 } |
2967 | 2666 |
6414 | 2667 for(mb_x=0; mb_x<=mb_w; mb_x++){ |
2668 add_yblock(s, 0, NULL, buf, dst8, obmc, | |
2669 block_w*mb_x - block_w/2, | |
2670 block_w*mb_y - block_w/2, | |
2671 block_w, block_w, | |
2672 w, h, | |
2673 w, ref_stride, obmc_stride, | |
2674 mb_x - 1, mb_y - 1, | |
2675 add, 1, plane_index); | |
2676 } | |
2562 | 2677 } |
2678 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2679 static av_always_inline void predict_plane(SnowContext *s, IDWTELEM *buf, int plane_index, int add){ |
2562 | 2680 const int mb_h= s->b_height << s->block_max_depth; |
2681 int mb_y; | |
2682 for(mb_y=0; mb_y<=mb_h; mb_y++) | |
2683 predict_slice(s, buf, plane_index, add, mb_y); | |
2138 | 2684 } |
2685 | |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2686 static int get_dc(SnowContext *s, int mb_x, int mb_y, int plane_index){ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2687 int i, x2, y2; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2688 Plane *p= &s->plane[plane_index]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2689 const int block_size = MB_SIZE >> s->block_max_depth; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2690 const int block_w = plane_index ? block_size/2 : block_size; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2691 const uint8_t *obmc = plane_index ? obmc_tab[s->block_max_depth+1] : obmc_tab[s->block_max_depth]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2692 const int obmc_stride= plane_index ? block_size : 2*block_size; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2693 const int ref_stride= s->current_picture.linesize[plane_index]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2694 uint8_t *src= s-> input_picture.data[plane_index]; |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2695 IDWTELEM *dst= (IDWTELEM*)s->m.obmc_scratchpad + plane_index*block_size*block_size*4; //FIXME change to unsigned |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2696 const int b_stride = s->b_width << s->block_max_depth; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2697 const int w= p->width; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2698 const int h= p->height; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2699 int index= mb_x + mb_y*b_stride; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2700 BlockNode *b= &s->block[index]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2701 BlockNode backup= *b; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2702 int ab=0; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2703 int aa=0; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2704 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2705 b->type|= BLOCK_INTRA; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2706 b->color[plane_index]= 0; |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2707 memset(dst, 0, obmc_stride*obmc_stride*sizeof(IDWTELEM)); |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2708 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2709 for(i=0; i<4; i++){ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2710 int mb_x2= mb_x + (i &1) - 1; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2711 int mb_y2= mb_y + (i>>1) - 1; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2712 int x= block_w*mb_x2 + block_w/2; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2713 int y= block_w*mb_y2 + block_w/2; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2714 |
3662
fc714e9a5419
snow cosmetics: merge the sliced and non-sliced versions of add_yblock
lorenm
parents:
3661
diff
changeset
|
2715 add_yblock(s, 0, NULL, dst + ((i&1)+(i>>1)*obmc_stride)*block_w, NULL, obmc, |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2716 x, y, block_w, block_w, w, h, obmc_stride, ref_stride, obmc_stride, mb_x2, mb_y2, 0, 0, plane_index); |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2717 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2718 for(y2= FFMAX(y, 0); y2<FFMIN(h, y+block_w); y2++){ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2719 for(x2= FFMAX(x, 0); x2<FFMIN(w, x+block_w); x2++){ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2720 int index= x2-(block_w*mb_x - block_w/2) + (y2-(block_w*mb_y - block_w/2))*obmc_stride; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2721 int obmc_v= obmc[index]; |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2722 int d; |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2723 if(y<0) obmc_v += obmc[index + block_w*obmc_stride]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2724 if(x<0) obmc_v += obmc[index + block_w]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2725 if(y+block_w>h) obmc_v += obmc[index - block_w*obmc_stride]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2726 if(x+block_w>w) obmc_v += obmc[index - block_w]; |
6412 | 2727 //FIXME precalculate this or simplify it somehow else |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2728 |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2729 d = -dst[index] + (1<<(FRAC_BITS-1)); |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2730 dst[index] = d; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2731 ab += (src[x2 + y2*ref_stride] - (d>>FRAC_BITS)) * obmc_v; |
6412 | 2732 aa += obmc_v * obmc_v; //FIXME precalculate this |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2733 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2734 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2735 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2736 *b= backup; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2737 |
5127 | 2738 return av_clip(((ab<<LOG2_OBMC_MAX) + aa/2)/aa, 0, 255); //FIXME we should not need clipping |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2739 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2740 |
3051 | 2741 static inline int get_block_bits(SnowContext *s, int x, int y, int w){ |
2742 const int b_stride = s->b_width << s->block_max_depth; | |
2743 const int b_height = s->b_height<< s->block_max_depth; | |
2744 int index= x + y*b_stride; | |
4408
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
2745 const BlockNode *b = &s->block[index]; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
2746 const BlockNode *left = x ? &s->block[index-1] : &null_block; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
2747 const BlockNode *top = y ? &s->block[index-b_stride] : &null_block; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
2748 const BlockNode *tl = y && x ? &s->block[index-b_stride-1] : left; |
c3be13eac2d6
Bring down the number of snow.c warnings from 27 to 17 by using the const
takis
parents:
4407
diff
changeset
|
2749 const BlockNode *tr = y && x+w<b_stride ? &s->block[index-b_stride+w] : tl; |
3051 | 2750 int dmx, dmy; |
4001 | 2751 // int mx_context= av_log2(2*FFABS(left->mx - top->mx)); |
2752 // int my_context= av_log2(2*FFABS(left->my - top->my)); | |
3051 | 2753 |
2754 if(x<0 || x>=b_stride || y>=b_height) | |
2755 return 0; | |
2756 /* | |
2757 1 0 0 | |
2758 01X 1-2 1 | |
2759 001XX 3-6 2-3 | |
2760 0001XXX 7-14 4-7 | |
2761 00001XXXX 15-30 8-15 | |
2762 */ | |
2763 //FIXME try accurate rate | |
6412 | 2764 //FIXME intra and inter predictors if surrounding blocks are not the same type |
3051 | 2765 if(b->type & BLOCK_INTRA){ |
4001 | 2766 return 3+2*( av_log2(2*FFABS(left->color[0] - b->color[0])) |
2767 + av_log2(2*FFABS(left->color[1] - b->color[1])) | |
2768 + av_log2(2*FFABS(left->color[2] - b->color[2]))); | |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
2769 }else{ |
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
2770 pred_mv(s, &dmx, &dmy, b->ref, left, top, tr); |
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
2771 dmx-= b->mx; |
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
2772 dmy-= b->my; |
4001 | 2773 return 2*(1 + av_log2(2*FFABS(dmx)) //FIXME kill the 2* can be merged in lambda |
2774 + av_log2(2*FFABS(dmy)) | |
3314 | 2775 + av_log2(2*b->ref)); |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
2776 } |
3051 | 2777 } |
2778 | |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2779 static int get_block_rd(SnowContext *s, int mb_x, int mb_y, int plane_index, const uint8_t *obmc_edged){ |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2780 Plane *p= &s->plane[plane_index]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2781 const int block_size = MB_SIZE >> s->block_max_depth; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2782 const int block_w = plane_index ? block_size/2 : block_size; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2783 const int obmc_stride= plane_index ? block_size : 2*block_size; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2784 const int ref_stride= s->current_picture.linesize[plane_index]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2785 uint8_t *dst= s->current_picture.data[plane_index]; |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2786 uint8_t *src= s-> input_picture.data[plane_index]; |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2787 IDWTELEM *pred= (IDWTELEM*)s->m.obmc_scratchpad + plane_index*block_size*block_size*4; |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2788 uint8_t cur[ref_stride*2*MB_SIZE]; //FIXME alignment |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
2789 uint8_t tmp[ref_stride*(2*MB_SIZE+HTAPS_MAX-1)]; |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2790 const int b_stride = s->b_width << s->block_max_depth; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2791 const int b_height = s->b_height<< s->block_max_depth; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2792 const int w= p->width; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2793 const int h= p->height; |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2794 int distortion; |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2795 int rate= 0; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2796 const int penalty_factor= get_penalty_factor(s->lambda, s->lambda2, s->avctx->me_cmp); |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2797 int sx= block_w*mb_x - block_w/2; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2798 int sy= block_w*mb_y - block_w/2; |
3206 | 2799 int x0= FFMAX(0,-sx); |
2800 int y0= FFMAX(0,-sy); | |
2801 int x1= FFMIN(block_w*2, w-sx); | |
2802 int y1= FFMIN(block_w*2, h-sy); | |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2803 int i,x,y; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2804 |
3314 | 2805 pred_block(s, cur, tmp, ref_stride, sx, sy, block_w*2, block_w*2, &s->block[mb_x + mb_y*b_stride], plane_index, w, h); |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2806 |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2807 for(y=y0; y<y1; y++){ |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2808 const uint8_t *obmc1= obmc_edged + y*obmc_stride; |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2809 const IDWTELEM *pred1 = pred + y*obmc_stride; |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2810 uint8_t *cur1 = cur + y*ref_stride; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2811 uint8_t *dst1 = dst + sx + (sy+y)*ref_stride; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2812 for(x=x0; x<x1; x++){ |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2813 #if FRAC_BITS >= LOG2_OBMC_MAX |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2814 int v = (cur1[x] * obmc1[x]) << (FRAC_BITS - LOG2_OBMC_MAX); |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2815 #else |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2816 int v = (cur1[x] * obmc1[x] + (1<<(LOG2_OBMC_MAX - FRAC_BITS-1))) >> (LOG2_OBMC_MAX - FRAC_BITS); |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
2817 #endif |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2818 v = (v + pred1[x]) >> FRAC_BITS; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2819 if(v&(~255)) v= ~(v>>31); |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2820 dst1[x] = v; |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2821 } |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2822 } |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2823 |
3206 | 2824 /* copy the regions where obmc[] = (uint8_t)256 */ |
2825 if(LOG2_OBMC_MAX == 8 | |
2826 && (mb_x == 0 || mb_x == b_stride-1) | |
2827 && (mb_y == 0 || mb_y == b_height-1)){ | |
2828 if(mb_x == 0) | |
2829 x1 = block_w; | |
2830 else | |
2831 x0 = block_w; | |
2832 if(mb_y == 0) | |
2833 y1 = block_w; | |
2834 else | |
2835 y0 = block_w; | |
2836 for(y=y0; y<y1; y++) | |
2837 memcpy(dst + sx+x0 + (sy+y)*ref_stride, cur + x0 + y*ref_stride, x1-x0); | |
2838 } | |
2839 | |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2840 if(block_w==16){ |
3323
87c54a3f8d19
Snow: fix subband weighting in wavelet cmp functions. use 32x32 cmp in iterative motion estimation.
lorenm
parents:
3322
diff
changeset
|
2841 /* FIXME rearrange dsputil to fit 32x32 cmp functions */ |
87c54a3f8d19
Snow: fix subband weighting in wavelet cmp functions. use 32x32 cmp in iterative motion estimation.
lorenm
parents:
3322
diff
changeset
|
2842 /* FIXME check alignment of the cmp wavelet vs the encoding wavelet */ |
6412 | 2843 /* FIXME cmps overlap but do not cover the wavelet's whole support. |
2844 * So improving the score of one block is not strictly guaranteed | |
2845 * to improve the score of the whole frame, thus iterative motion | |
2846 * estimation does not always converge. */ | |
3323
87c54a3f8d19
Snow: fix subband weighting in wavelet cmp functions. use 32x32 cmp in iterative motion estimation.
lorenm
parents:
3322
diff
changeset
|
2847 if(s->avctx->me_cmp == FF_CMP_W97) |
4197 | 2848 distortion = w97_32_c(&s->m, src + sx + sy*ref_stride, dst + sx + sy*ref_stride, ref_stride, 32); |
3323
87c54a3f8d19
Snow: fix subband weighting in wavelet cmp functions. use 32x32 cmp in iterative motion estimation.
lorenm
parents:
3322
diff
changeset
|
2849 else if(s->avctx->me_cmp == FF_CMP_W53) |
4197 | 2850 distortion = w53_32_c(&s->m, src + sx + sy*ref_stride, dst + sx + sy*ref_stride, ref_stride, 32); |
3323
87c54a3f8d19
Snow: fix subband weighting in wavelet cmp functions. use 32x32 cmp in iterative motion estimation.
lorenm
parents:
3322
diff
changeset
|
2851 else{ |
87c54a3f8d19
Snow: fix subband weighting in wavelet cmp functions. use 32x32 cmp in iterative motion estimation.
lorenm
parents:
3322
diff
changeset
|
2852 distortion = 0; |
87c54a3f8d19
Snow: fix subband weighting in wavelet cmp functions. use 32x32 cmp in iterative motion estimation.
lorenm
parents:
3322
diff
changeset
|
2853 for(i=0; i<4; i++){ |
87c54a3f8d19
Snow: fix subband weighting in wavelet cmp functions. use 32x32 cmp in iterative motion estimation.
lorenm
parents:
3322
diff
changeset
|
2854 int off = sx+16*(i&1) + (sy+16*(i>>1))*ref_stride; |
87c54a3f8d19
Snow: fix subband weighting in wavelet cmp functions. use 32x32 cmp in iterative motion estimation.
lorenm
parents:
3322
diff
changeset
|
2855 distortion += s->dsp.me_cmp[0](&s->m, src + off, dst + off, ref_stride, 16); |
87c54a3f8d19
Snow: fix subband weighting in wavelet cmp functions. use 32x32 cmp in iterative motion estimation.
lorenm
parents:
3322
diff
changeset
|
2856 } |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2857 } |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2858 }else{ |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2859 assert(block_w==8); |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2860 distortion = s->dsp.me_cmp[0](&s->m, src + sx + sy*ref_stride, dst + sx + sy*ref_stride, ref_stride, block_w*2); |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2861 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2862 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2863 if(plane_index==0){ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2864 for(i=0; i<4; i++){ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2865 /* ..RRr |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2866 * .RXx. |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2867 * rxx.. |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2868 */ |
3051 | 2869 rate += get_block_bits(s, mb_x + (i&1) - (i>>1), mb_y + (i>>1), 1); |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2870 } |
3057 | 2871 if(mb_x == b_stride-2) |
2872 rate += get_block_bits(s, mb_x + 1, mb_y + 1, 1); | |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2873 } |
3051 | 2874 return distortion + rate*penalty_factor; |
2875 } | |
2876 | |
2877 static int get_4block_rd(SnowContext *s, int mb_x, int mb_y, int plane_index){ | |
2878 int i, y2; | |
2879 Plane *p= &s->plane[plane_index]; | |
2880 const int block_size = MB_SIZE >> s->block_max_depth; | |
2881 const int block_w = plane_index ? block_size/2 : block_size; | |
2882 const uint8_t *obmc = plane_index ? obmc_tab[s->block_max_depth+1] : obmc_tab[s->block_max_depth]; | |
2883 const int obmc_stride= plane_index ? block_size : 2*block_size; | |
2884 const int ref_stride= s->current_picture.linesize[plane_index]; | |
2885 uint8_t *dst= s->current_picture.data[plane_index]; | |
2886 uint8_t *src= s-> input_picture.data[plane_index]; | |
6426 | 2887 //FIXME zero_dst is const but add_yblock changes dst if add is 0 (this is never the case for dst=zero_dst |
2888 // const has only been removed from zero_dst to suppress a warning | |
2889 static IDWTELEM zero_dst[4096]; //FIXME | |
3051 | 2890 const int b_stride = s->b_width << s->block_max_depth; |
2891 const int w= p->width; | |
2892 const int h= p->height; | |
2893 int distortion= 0; | |
2894 int rate= 0; | |
2895 const int penalty_factor= get_penalty_factor(s->lambda, s->lambda2, s->avctx->me_cmp); | |
2896 | |
2897 for(i=0; i<9; i++){ | |
2898 int mb_x2= mb_x + (i%3) - 1; | |
2899 int mb_y2= mb_y + (i/3) - 1; | |
2900 int x= block_w*mb_x2 + block_w/2; | |
2901 int y= block_w*mb_y2 + block_w/2; | |
2902 | |
3662
fc714e9a5419
snow cosmetics: merge the sliced and non-sliced versions of add_yblock
lorenm
parents:
3661
diff
changeset
|
2903 add_yblock(s, 0, NULL, zero_dst, dst, obmc, |
3051 | 2904 x, y, block_w, block_w, w, h, /*dst_stride*/0, ref_stride, obmc_stride, mb_x2, mb_y2, 1, 1, plane_index); |
2905 | |
2906 //FIXME find a cleaner/simpler way to skip the outside stuff | |
2907 for(y2= y; y2<0; y2++) | |
2908 memcpy(dst + x + y2*ref_stride, src + x + y2*ref_stride, block_w); | |
2909 for(y2= h; y2<y+block_w; y2++) | |
2910 memcpy(dst + x + y2*ref_stride, src + x + y2*ref_stride, block_w); | |
2911 if(x<0){ | |
2912 for(y2= y; y2<y+block_w; y2++) | |
2913 memcpy(dst + x + y2*ref_stride, src + x + y2*ref_stride, -x); | |
2914 } | |
2915 if(x+block_w > w){ | |
2916 for(y2= y; y2<y+block_w; y2++) | |
2917 memcpy(dst + w + y2*ref_stride, src + w + y2*ref_stride, x+block_w - w); | |
2918 } | |
2919 | |
2920 assert(block_w== 8 || block_w==16); | |
2921 distortion += s->dsp.me_cmp[block_w==8](&s->m, src + x + y*ref_stride, dst + x + y*ref_stride, ref_stride, block_w); | |
2922 } | |
2923 | |
2924 if(plane_index==0){ | |
2925 BlockNode *b= &s->block[mb_x+mb_y*b_stride]; | |
2926 int merged= same_block(b,b+1) && same_block(b,b+b_stride) && same_block(b,b+b_stride+1); | |
2927 | |
2928 /* ..RRRr | |
2929 * .RXXx. | |
2930 * .RXXx. | |
2931 * rxxx. | |
2932 */ | |
2933 if(merged) | |
2934 rate = get_block_bits(s, mb_x, mb_y, 2); | |
2935 for(i=merged?4:0; i<9; i++){ | |
2936 static const int dxy[9][2] = {{0,0},{1,0},{0,1},{1,1},{2,0},{2,1},{-1,2},{0,2},{1,2}}; | |
2937 rate += get_block_bits(s, mb_x + dxy[i][0], mb_y + dxy[i][1], 1); | |
2938 } | |
2939 } | |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2940 return distortion + rate*penalty_factor; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2941 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2942 |
4283
d6f83e2f8804
rename always_inline to av_always_inline and move to common.h
mru
parents:
4197
diff
changeset
|
2943 static av_always_inline int check_block(SnowContext *s, int mb_x, int mb_y, int p[3], int intra, const uint8_t *obmc_edged, int *best_rd){ |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2944 const int b_stride= s->b_width << s->block_max_depth; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2945 BlockNode *block= &s->block[mb_x + mb_y * b_stride]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2946 BlockNode backup= *block; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2947 int rd, index, value; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2948 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2949 assert(mb_x>=0 && mb_y>=0); |
2994 | 2950 assert(mb_x<b_stride); |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2951 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2952 if(intra){ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2953 block->color[0] = p[0]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2954 block->color[1] = p[1]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2955 block->color[2] = p[2]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2956 block->type |= BLOCK_INTRA; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2957 }else{ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2958 index= (p[0] + 31*p[1]) & (ME_CACHE_SIZE-1); |
3314 | 2959 value= s->me_cache_generation + (p[0]>>10) + (p[1]<<6) + (block->ref<<12); |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2960 if(s->me_cache[index] == value) |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2961 return 0; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2962 s->me_cache[index]= value; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2963 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2964 block->mx= p[0]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2965 block->my= p[1]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2966 block->type &= ~BLOCK_INTRA; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2967 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2968 |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
2969 rd= get_block_rd(s, mb_x, mb_y, 0, obmc_edged); |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2970 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2971 //FIXME chroma |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2972 if(rd < *best_rd){ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2973 *best_rd= rd; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2974 return 1; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2975 }else{ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2976 *block= backup; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2977 return 0; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2978 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2979 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
2980 |
6412 | 2981 /* special case for int[2] args we discard afterwards, |
2982 * fixes compilation problem with gcc 2.95 */ | |
4283
d6f83e2f8804
rename always_inline to av_always_inline and move to common.h
mru
parents:
4197
diff
changeset
|
2983 static av_always_inline int check_block_inter(SnowContext *s, int mb_x, int mb_y, int p0, int p1, const uint8_t *obmc_edged, int *best_rd){ |
2999
bc83e7a080a5
this fixes compilation with gcc 2.95 which seems to not like array casting... special case inline of check_block which makes the array and passes it
mmu_man
parents:
2998
diff
changeset
|
2984 int p[2] = {p0, p1}; |
3197 | 2985 return check_block(s, mb_x, mb_y, p, 0, obmc_edged, best_rd); |
2999
bc83e7a080a5
this fixes compilation with gcc 2.95 which seems to not like array casting... special case inline of check_block which makes the array and passes it
mmu_man
parents:
2998
diff
changeset
|
2986 } |
bc83e7a080a5
this fixes compilation with gcc 2.95 which seems to not like array casting... special case inline of check_block which makes the array and passes it
mmu_man
parents:
2998
diff
changeset
|
2987 |
4283
d6f83e2f8804
rename always_inline to av_always_inline and move to common.h
mru
parents:
4197
diff
changeset
|
2988 static av_always_inline int check_4block_inter(SnowContext *s, int mb_x, int mb_y, int p0, int p1, int ref, int *best_rd){ |
3051 | 2989 const int b_stride= s->b_width << s->block_max_depth; |
2990 BlockNode *block= &s->block[mb_x + mb_y * b_stride]; | |
2991 BlockNode backup[4]= {block[0], block[1], block[b_stride], block[b_stride+1]}; | |
2992 int rd, index, value; | |
2993 | |
2994 assert(mb_x>=0 && mb_y>=0); | |
2995 assert(mb_x<b_stride); | |
2996 assert(((mb_x|mb_y)&1) == 0); | |
2997 | |
2998 index= (p0 + 31*p1) & (ME_CACHE_SIZE-1); | |
3314 | 2999 value= s->me_cache_generation + (p0>>10) + (p1<<6) + (block->ref<<12); |
3051 | 3000 if(s->me_cache[index] == value) |
3001 return 0; | |
3002 s->me_cache[index]= value; | |
3003 | |
3004 block->mx= p0; | |
3005 block->my= p1; | |
3314 | 3006 block->ref= ref; |
3051 | 3007 block->type &= ~BLOCK_INTRA; |
3008 block[1]= block[b_stride]= block[b_stride+1]= *block; | |
3009 | |
3010 rd= get_4block_rd(s, mb_x, mb_y, 0); | |
3011 | |
3012 //FIXME chroma | |
3013 if(rd < *best_rd){ | |
3014 *best_rd= rd; | |
3015 return 1; | |
3016 }else{ | |
3017 block[0]= backup[0]; | |
3018 block[1]= backup[1]; | |
3019 block[b_stride]= backup[2]; | |
3020 block[b_stride+1]= backup[3]; | |
3021 return 0; | |
3022 } | |
3023 } | |
3024 | |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3025 static void iterative_me(SnowContext *s){ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3026 int pass, mb_x, mb_y; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3027 const int b_width = s->b_width << s->block_max_depth; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3028 const int b_height= s->b_height << s->block_max_depth; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3029 const int b_stride= b_width; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3030 int color[3]; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3031 |
3194
c30e9bcbb716
seed iterative_me with mvs from conventional search.
lorenm
parents:
3193
diff
changeset
|
3032 { |
c30e9bcbb716
seed iterative_me with mvs from conventional search.
lorenm
parents:
3193
diff
changeset
|
3033 RangeCoder r = s->c; |
c30e9bcbb716
seed iterative_me with mvs from conventional search.
lorenm
parents:
3193
diff
changeset
|
3034 uint8_t state[sizeof(s->block_state)]; |
c30e9bcbb716
seed iterative_me with mvs from conventional search.
lorenm
parents:
3193
diff
changeset
|
3035 memcpy(state, s->block_state, sizeof(s->block_state)); |
c30e9bcbb716
seed iterative_me with mvs from conventional search.
lorenm
parents:
3193
diff
changeset
|
3036 for(mb_y= 0; mb_y<s->b_height; mb_y++) |
c30e9bcbb716
seed iterative_me with mvs from conventional search.
lorenm
parents:
3193
diff
changeset
|
3037 for(mb_x= 0; mb_x<s->b_width; mb_x++) |
c30e9bcbb716
seed iterative_me with mvs from conventional search.
lorenm
parents:
3193
diff
changeset
|
3038 encode_q_branch(s, 0, mb_x, mb_y); |
c30e9bcbb716
seed iterative_me with mvs from conventional search.
lorenm
parents:
3193
diff
changeset
|
3039 s->c = r; |
c30e9bcbb716
seed iterative_me with mvs from conventional search.
lorenm
parents:
3193
diff
changeset
|
3040 memcpy(s->block_state, state, sizeof(s->block_state)); |
c30e9bcbb716
seed iterative_me with mvs from conventional search.
lorenm
parents:
3193
diff
changeset
|
3041 } |
c30e9bcbb716
seed iterative_me with mvs from conventional search.
lorenm
parents:
3193
diff
changeset
|
3042 |
3323
87c54a3f8d19
Snow: fix subband weighting in wavelet cmp functions. use 32x32 cmp in iterative motion estimation.
lorenm
parents:
3322
diff
changeset
|
3043 for(pass=0; pass<25; pass++){ |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3044 int change= 0; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3045 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3046 for(mb_y= 0; mb_y<b_height; mb_y++){ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3047 for(mb_x= 0; mb_x<b_width; mb_x++){ |
3314 | 3048 int dia_change, i, j, ref; |
3049 int best_rd= INT_MAX, ref_rd; | |
3050 BlockNode backup, ref_b; | |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3051 const int index= mb_x + mb_y * b_stride; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3052 BlockNode *block= &s->block[index]; |
3324 | 3053 BlockNode *tb = mb_y ? &s->block[index-b_stride ] : NULL; |
3054 BlockNode *lb = mb_x ? &s->block[index -1] : NULL; | |
3055 BlockNode *rb = mb_x+1<b_width ? &s->block[index +1] : NULL; | |
3056 BlockNode *bb = mb_y+1<b_height ? &s->block[index+b_stride ] : NULL; | |
3057 BlockNode *tlb= mb_x && mb_y ? &s->block[index-b_stride-1] : NULL; | |
3058 BlockNode *trb= mb_x+1<b_width && mb_y ? &s->block[index-b_stride+1] : NULL; | |
3059 BlockNode *blb= mb_x && mb_y+1<b_height ? &s->block[index+b_stride-1] : NULL; | |
3060 BlockNode *brb= mb_x+1<b_width && mb_y+1<b_height ? &s->block[index+b_stride+1] : NULL; | |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3061 const int b_w= (MB_SIZE >> s->block_max_depth); |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3062 uint8_t obmc_edged[b_w*2][b_w*2]; |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3063 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3064 if(pass && (block->type & BLOCK_OPT)) |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3065 continue; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3066 block->type |= BLOCK_OPT; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3067 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3068 backup= *block; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3069 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3070 if(!s->me_cache_generation) |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3071 memset(s->me_cache, 0, sizeof(s->me_cache)); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3072 s->me_cache_generation += 1<<22; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3073 |
6412 | 3074 //FIXME precalculate |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3075 { |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3076 int x, y; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3077 memcpy(obmc_edged, obmc_tab[s->block_max_depth], b_w*b_w*4); |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3078 if(mb_x==0) |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3079 for(y=0; y<b_w*2; y++) |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3080 memset(obmc_edged[y], obmc_edged[y][0] + obmc_edged[y][b_w-1], b_w); |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3081 if(mb_x==b_stride-1) |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3082 for(y=0; y<b_w*2; y++) |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3083 memset(obmc_edged[y]+b_w, obmc_edged[y][b_w] + obmc_edged[y][b_w*2-1], b_w); |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3084 if(mb_y==0){ |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3085 for(x=0; x<b_w*2; x++) |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3086 obmc_edged[0][x] += obmc_edged[b_w-1][x]; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3087 for(y=1; y<b_w; y++) |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3088 memcpy(obmc_edged[y], obmc_edged[0], b_w*2); |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3089 } |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3090 if(mb_y==b_height-1){ |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3091 for(x=0; x<b_w*2; x++) |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3092 obmc_edged[b_w*2-1][x] += obmc_edged[b_w][x]; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3093 for(y=b_w; y<b_w*2-1; y++) |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3094 memcpy(obmc_edged[y], obmc_edged[b_w*2-1], b_w*2); |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3095 } |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3096 } |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3097 |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3098 //skip stuff outside the picture |
6474
e6995f3fbf7f
cosmetics: Normalize {} placement after for, while, if.
diego
parents:
6437
diff
changeset
|
3099 if(mb_x==0 || mb_y==0 || mb_x==b_width-1 || mb_y==b_height-1){ |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3100 uint8_t *src= s-> input_picture.data[0]; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3101 uint8_t *dst= s->current_picture.data[0]; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3102 const int stride= s->current_picture.linesize[0]; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3103 const int block_w= MB_SIZE >> s->block_max_depth; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3104 const int sx= block_w*mb_x - block_w/2; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3105 const int sy= block_w*mb_y - block_w/2; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3106 const int w= s->plane[0].width; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3107 const int h= s->plane[0].height; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3108 int y; |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3109 |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3110 for(y=sy; y<0; y++) |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3111 memcpy(dst + sx + y*stride, src + sx + y*stride, block_w*2); |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3112 for(y=h; y<sy+block_w*2; y++) |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3113 memcpy(dst + sx + y*stride, src + sx + y*stride, block_w*2); |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3114 if(sx<0){ |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3115 for(y=sy; y<sy+block_w*2; y++) |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3116 memcpy(dst + sx + y*stride, src + sx + y*stride, -sx); |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3117 } |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3118 if(sx+block_w*2 > w){ |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3119 for(y=sy; y<sy+block_w*2; y++) |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3120 memcpy(dst + w + y*stride, src + w + y*stride, sx+block_w*2 - w); |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3121 } |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3122 } |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3123 |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3124 // intra(black) = neighbors' contribution to the current block |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3125 for(i=0; i<3; i++) |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3126 color[i]= get_dc(s, mb_x, mb_y, i); |
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3127 |
5127 | 3128 // get previous score (cannot be cached due to OBMC) |
3057 | 3129 if(pass > 0 && (block->type&BLOCK_INTRA)){ |
3130 int color0[3]= {block->color[0], block->color[1], block->color[2]}; | |
3131 check_block(s, mb_x, mb_y, color0, 1, *obmc_edged, &best_rd); | |
3132 }else | |
3197 | 3133 check_block_inter(s, mb_x, mb_y, block->mx, block->my, *obmc_edged, &best_rd); |
3134 | |
3314 | 3135 ref_b= *block; |
3136 ref_rd= best_rd; | |
3137 for(ref=0; ref < s->ref_frames; ref++){ | |
3138 int16_t (*mvr)[2]= &s->ref_mvs[ref][index]; | |
3139 if(s->ref_scores[ref][index] > s->ref_scores[ref_b.ref][index]*3/2) //FIXME tune threshold | |
3140 continue; | |
3141 block->ref= ref; | |
3142 best_rd= INT_MAX; | |
3143 | |
3144 check_block_inter(s, mb_x, mb_y, mvr[0][0], mvr[0][1], *obmc_edged, &best_rd); | |
3145 check_block_inter(s, mb_x, mb_y, 0, 0, *obmc_edged, &best_rd); | |
3324 | 3146 if(tb) |
3314 | 3147 check_block_inter(s, mb_x, mb_y, mvr[-b_stride][0], mvr[-b_stride][1], *obmc_edged, &best_rd); |
3324 | 3148 if(lb) |
3314 | 3149 check_block_inter(s, mb_x, mb_y, mvr[-1][0], mvr[-1][1], *obmc_edged, &best_rd); |
3324 | 3150 if(rb) |
3314 | 3151 check_block_inter(s, mb_x, mb_y, mvr[1][0], mvr[1][1], *obmc_edged, &best_rd); |
3324 | 3152 if(bb) |
3314 | 3153 check_block_inter(s, mb_x, mb_y, mvr[b_stride][0], mvr[b_stride][1], *obmc_edged, &best_rd); |
3154 | |
3155 /* fullpel ME */ | |
6412 | 3156 //FIXME avoid subpel interpolation / round to nearest integer |
3314 | 3157 do{ |
3158 dia_change=0; | |
3159 for(i=0; i<FFMAX(s->avctx->dia_size, 1); i++){ | |
3160 for(j=0; j<i; j++){ | |
3161 dia_change |= check_block_inter(s, mb_x, mb_y, block->mx+4*(i-j), block->my+(4*j), *obmc_edged, &best_rd); | |
3162 dia_change |= check_block_inter(s, mb_x, mb_y, block->mx-4*(i-j), block->my-(4*j), *obmc_edged, &best_rd); | |
3163 dia_change |= check_block_inter(s, mb_x, mb_y, block->mx+4*(i-j), block->my-(4*j), *obmc_edged, &best_rd); | |
3164 dia_change |= check_block_inter(s, mb_x, mb_y, block->mx-4*(i-j), block->my+(4*j), *obmc_edged, &best_rd); | |
3165 } | |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3166 } |
3314 | 3167 }while(dia_change); |
3168 /* subpel ME */ | |
3169 do{ | |
3170 static const int square[8][2]= {{+1, 0},{-1, 0},{ 0,+1},{ 0,-1},{+1,+1},{-1,-1},{+1,-1},{-1,+1},}; | |
3171 dia_change=0; | |
3172 for(i=0; i<8; i++) | |
3173 dia_change |= check_block_inter(s, mb_x, mb_y, block->mx+square[i][0], block->my+square[i][1], *obmc_edged, &best_rd); | |
3174 }while(dia_change); | |
3175 //FIXME or try the standard 2 pass qpel or similar | |
3176 | |
3177 mvr[0][0]= block->mx; | |
3178 mvr[0][1]= block->my; | |
3179 if(ref_rd > best_rd){ | |
3180 ref_rd= best_rd; | |
3181 ref_b= *block; | |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3182 } |
3314 | 3183 } |
3184 best_rd= ref_rd; | |
3185 *block= ref_b; | |
2998 | 3186 #if 1 |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
3187 check_block(s, mb_x, mb_y, color, 1, *obmc_edged, &best_rd); |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3188 //FIXME RD style color selection |
2998 | 3189 #endif |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3190 if(!same_block(block, &backup)){ |
3324 | 3191 if(tb ) tb ->type &= ~BLOCK_OPT; |
3192 if(lb ) lb ->type &= ~BLOCK_OPT; | |
3193 if(rb ) rb ->type &= ~BLOCK_OPT; | |
3194 if(bb ) bb ->type &= ~BLOCK_OPT; | |
3195 if(tlb) tlb->type &= ~BLOCK_OPT; | |
3196 if(trb) trb->type &= ~BLOCK_OPT; | |
3197 if(blb) blb->type &= ~BLOCK_OPT; | |
3198 if(brb) brb->type &= ~BLOCK_OPT; | |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3199 change ++; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3200 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3201 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3202 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3203 av_log(NULL, AV_LOG_ERROR, "pass:%d changed:%d\n", pass, change); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3204 if(!change) |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3205 break; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3206 } |
3051 | 3207 |
3208 if(s->block_max_depth == 1){ | |
3209 int change= 0; | |
3210 for(mb_y= 0; mb_y<b_height; mb_y+=2){ | |
3211 for(mb_x= 0; mb_x<b_width; mb_x+=2){ | |
3324 | 3212 int i; |
3051 | 3213 int best_rd, init_rd; |
3214 const int index= mb_x + mb_y * b_stride; | |
3215 BlockNode *b[4]; | |
3216 | |
3217 b[0]= &s->block[index]; | |
3218 b[1]= b[0]+1; | |
3219 b[2]= b[0]+b_stride; | |
3220 b[3]= b[2]+1; | |
3221 if(same_block(b[0], b[1]) && | |
3222 same_block(b[0], b[2]) && | |
3223 same_block(b[0], b[3])) | |
3224 continue; | |
3225 | |
3226 if(!s->me_cache_generation) | |
3227 memset(s->me_cache, 0, sizeof(s->me_cache)); | |
3228 s->me_cache_generation += 1<<22; | |
3229 | |
3230 init_rd= best_rd= get_4block_rd(s, mb_x, mb_y, 0); | |
3231 | |
3314 | 3232 //FIXME more multiref search? |
3051 | 3233 check_4block_inter(s, mb_x, mb_y, |
3234 (b[0]->mx + b[1]->mx + b[2]->mx + b[3]->mx + 2) >> 2, | |
3314 | 3235 (b[0]->my + b[1]->my + b[2]->my + b[3]->my + 2) >> 2, 0, &best_rd); |
3051 | 3236 |
3237 for(i=0; i<4; i++) | |
3238 if(!(b[i]->type&BLOCK_INTRA)) | |
3314 | 3239 check_4block_inter(s, mb_x, mb_y, b[i]->mx, b[i]->my, b[i]->ref, &best_rd); |
3051 | 3240 |
3241 if(init_rd != best_rd) | |
3242 change++; | |
3243 } | |
3244 } | |
3245 av_log(NULL, AV_LOG_ERROR, "pass:4mv changed:%d\n", change*4); | |
3246 } | |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3247 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
3248 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3249 static void quantize(SnowContext *s, SubBand *b, IDWTELEM *dst, DWTELEM *src, int stride, int bias){ |
2138 | 3250 const int w= b->width; |
3251 const int h= b->height; | |
4594 | 3252 const int qlog= av_clip(s->qlog + b->qlog, 0, QROOT*16); |
5575 | 3253 const int qmul= qexp[qlog&(QROOT-1)]<<((qlog>>QSHIFT) + ENCODER_EXTRA_BITS); |
2150 | 3254 int x,y, thres1, thres2; |
2138 | 3255 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3256 if(s->qlog == LOSSLESS_QLOG){ |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3257 for(y=0; y<h; y++) |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3258 for(x=0; x<w; x++) |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3259 dst[x + y*stride]= src[x + y*stride]; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3260 return; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3261 } |
2967 | 3262 |
2138 | 3263 bias= bias ? 0 : (3*qmul)>>3; |
2150 | 3264 thres1= ((qmul - bias)>>QEXPSHIFT) - 1; |
3265 thres2= 2*thres1; | |
2967 | 3266 |
2138 | 3267 if(!bias){ |
3268 for(y=0; y<h; y++){ | |
3269 for(x=0; x<w; x++){ | |
2150 | 3270 int i= src[x + y*stride]; |
2967 | 3271 |
2150 | 3272 if((unsigned)(i+thres1) > thres2){ |
3273 if(i>=0){ | |
3274 i<<= QEXPSHIFT; | |
3275 i/= qmul; //FIXME optimize | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3276 dst[x + y*stride]= i; |
2150 | 3277 }else{ |
3278 i= -i; | |
3279 i<<= QEXPSHIFT; | |
3280 i/= qmul; //FIXME optimize | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3281 dst[x + y*stride]= -i; |
2150 | 3282 } |
3283 }else | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3284 dst[x + y*stride]= 0; |
2138 | 3285 } |
3286 } | |
3287 }else{ | |
3288 for(y=0; y<h; y++){ | |
3289 for(x=0; x<w; x++){ | |
2967 | 3290 int i= src[x + y*stride]; |
3291 | |
2150 | 3292 if((unsigned)(i+thres1) > thres2){ |
3293 if(i>=0){ | |
3294 i<<= QEXPSHIFT; | |
3295 i= (i + bias) / qmul; //FIXME optimize | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3296 dst[x + y*stride]= i; |
2150 | 3297 }else{ |
3298 i= -i; | |
3299 i<<= QEXPSHIFT; | |
3300 i= (i + bias) / qmul; //FIXME optimize | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3301 dst[x + y*stride]= -i; |
2150 | 3302 } |
3303 }else | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3304 dst[x + y*stride]= 0; |
2138 | 3305 } |
3306 } | |
3307 } | |
3308 } | |
3309 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3310 static void dequantize_slice_buffered(SnowContext *s, slice_buffer * sb, SubBand *b, IDWTELEM *src, int stride, int start_y, int end_y){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3311 const int w= b->width; |
4594 | 3312 const int qlog= av_clip(s->qlog + b->qlog, 0, QROOT*16); |
2600 | 3313 const int qmul= qexp[qlog&(QROOT-1)]<<(qlog>>QSHIFT); |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3314 const int qadd= (s->qbias*qmul)>>QBIAS_SHIFT; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3315 int x,y; |
2967 | 3316 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3317 if(s->qlog == LOSSLESS_QLOG) return; |
2967 | 3318 |
2634
337217ecbb3e
tighter snow slicing patch by (Yartrebo || yartrebo earthlink net)
michael
parents:
2628
diff
changeset
|
3319 for(y=start_y; y<end_y; y++){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3320 // DWTELEM * line = slice_buffer_get_line_from_address(sb, src + (y * stride)); |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3321 IDWTELEM * line = slice_buffer_get_line(sb, (y * b->stride_line) + b->buf_y_offset) + b->buf_x_offset; |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3322 for(x=0; x<w; x++){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3323 int i= line[x]; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3324 if(i<0){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3325 line[x]= -((-i*qmul + qadd)>>(QEXPSHIFT)); //FIXME try different bias |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3326 }else if(i>0){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3327 line[x]= (( i*qmul + qadd)>>(QEXPSHIFT)); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3328 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3329 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3330 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3331 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3332 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3333 static void dequantize(SnowContext *s, SubBand *b, IDWTELEM *src, int stride){ |
2138 | 3334 const int w= b->width; |
3335 const int h= b->height; | |
4594 | 3336 const int qlog= av_clip(s->qlog + b->qlog, 0, QROOT*16); |
2600 | 3337 const int qmul= qexp[qlog&(QROOT-1)]<<(qlog>>QSHIFT); |
2138 | 3338 const int qadd= (s->qbias*qmul)>>QBIAS_SHIFT; |
3339 int x,y; | |
2967 | 3340 |
2161 | 3341 if(s->qlog == LOSSLESS_QLOG) return; |
2967 | 3342 |
2138 | 3343 for(y=0; y<h; y++){ |
3344 for(x=0; x<w; x++){ | |
3345 int i= src[x + y*stride]; | |
3346 if(i<0){ | |
3347 src[x + y*stride]= -((-i*qmul + qadd)>>(QEXPSHIFT)); //FIXME try different bias | |
3348 }else if(i>0){ | |
3349 src[x + y*stride]= (( i*qmul + qadd)>>(QEXPSHIFT)); | |
3350 } | |
3351 } | |
3352 } | |
3353 } | |
3354 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3355 static void decorrelate(SnowContext *s, SubBand *b, IDWTELEM *src, int stride, int inverse, int use_median){ |
2138 | 3356 const int w= b->width; |
3357 const int h= b->height; | |
3358 int x,y; | |
2967 | 3359 |
2138 | 3360 for(y=h-1; y>=0; y--){ |
3361 for(x=w-1; x>=0; x--){ | |
3362 int i= x + y*stride; | |
2967 | 3363 |
2138 | 3364 if(x){ |
3365 if(use_median){ | |
3366 if(y && x+1<w) src[i] -= mid_pred(src[i - 1], src[i - stride], src[i - stride + 1]); | |
3367 else src[i] -= src[i - 1]; | |
3368 }else{ | |
3369 if(y) src[i] -= mid_pred(src[i - 1], src[i - stride], src[i - 1] + src[i - stride] - src[i - 1 - stride]); | |
3370 else src[i] -= src[i - 1]; | |
3371 } | |
3372 }else{ | |
3373 if(y) src[i] -= src[i - stride]; | |
3374 } | |
3375 } | |
3376 } | |
3377 } | |
3378 | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3379 static void correlate_slice_buffered(SnowContext *s, slice_buffer * sb, SubBand *b, IDWTELEM *src, int stride, int inverse, int use_median, int start_y, int end_y){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3380 const int w= b->width; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3381 int x,y; |
2967 | 3382 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3383 IDWTELEM * line=0; // silence silly "could be used without having been initialized" warning |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3384 IDWTELEM * prev; |
2967 | 3385 |
2634
337217ecbb3e
tighter snow slicing patch by (Yartrebo || yartrebo earthlink net)
michael
parents:
2628
diff
changeset
|
3386 if (start_y != 0) |
337217ecbb3e
tighter snow slicing patch by (Yartrebo || yartrebo earthlink net)
michael
parents:
2628
diff
changeset
|
3387 line = slice_buffer_get_line(sb, ((start_y - 1) * b->stride_line) + b->buf_y_offset) + b->buf_x_offset; |
2967 | 3388 |
2634
337217ecbb3e
tighter snow slicing patch by (Yartrebo || yartrebo earthlink net)
michael
parents:
2628
diff
changeset
|
3389 for(y=start_y; y<end_y; y++){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3390 prev = line; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3391 // line = slice_buffer_get_line_from_address(sb, src + (y * stride)); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3392 line = slice_buffer_get_line(sb, (y * b->stride_line) + b->buf_y_offset) + b->buf_x_offset; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3393 for(x=0; x<w; x++){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3394 if(x){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3395 if(use_median){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3396 if(y && x+1<w) line[x] += mid_pred(line[x - 1], prev[x], prev[x + 1]); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3397 else line[x] += line[x - 1]; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3398 }else{ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3399 if(y) line[x] += mid_pred(line[x - 1], prev[x], line[x - 1] + prev[x] - prev[x - 1]); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3400 else line[x] += line[x - 1]; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3401 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3402 }else{ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3403 if(y) line[x] += prev[x]; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3404 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3405 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3406 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3407 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3408 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3409 static void correlate(SnowContext *s, SubBand *b, IDWTELEM *src, int stride, int inverse, int use_median){ |
2138 | 3410 const int w= b->width; |
3411 const int h= b->height; | |
3412 int x,y; | |
2967 | 3413 |
2138 | 3414 for(y=0; y<h; y++){ |
3415 for(x=0; x<w; x++){ | |
3416 int i= x + y*stride; | |
2967 | 3417 |
2138 | 3418 if(x){ |
3419 if(use_median){ | |
3420 if(y && x+1<w) src[i] += mid_pred(src[i - 1], src[i - stride], src[i - stride + 1]); | |
3421 else src[i] += src[i - 1]; | |
3422 }else{ | |
3423 if(y) src[i] += mid_pred(src[i - 1], src[i - stride], src[i - 1] + src[i - stride] - src[i - 1 - stride]); | |
3424 else src[i] += src[i - 1]; | |
3425 } | |
3426 }else{ | |
3427 if(y) src[i] += src[i - stride]; | |
3428 } | |
3429 } | |
3430 } | |
3431 } | |
3432 | |
5666
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3433 static void encode_qlogs(SnowContext *s){ |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3434 int plane_index, level, orientation; |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3435 |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3436 for(plane_index=0; plane_index<2; plane_index++){ |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3437 for(level=0; level<s->spatial_decomposition_count; level++){ |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3438 for(orientation=level ? 1:0; orientation<4; orientation++){ |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3439 if(orientation==2) continue; |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3440 put_symbol(&s->c, s->header_state, s->plane[plane_index].band[level][orientation].qlog, 1); |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3441 } |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3442 } |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3443 } |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3444 } |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3445 |
2138 | 3446 static void encode_header(SnowContext *s){ |
5666
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3447 int plane_index, i; |
2967 | 3448 uint8_t kstate[32]; |
3449 | |
3450 memset(kstate, MID_STATE, sizeof(kstate)); | |
2138 | 3451 |
2335 | 3452 put_rac(&s->c, kstate, s->keyframe); |
4331
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3453 if(s->keyframe || s->always_reset){ |
2199 | 3454 reset_contexts(s); |
4331
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3455 s->last_spatial_decomposition_type= |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3456 s->last_qlog= |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3457 s->last_qbias= |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3458 s->last_mv_scale= |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3459 s->last_block_max_depth= 0; |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3460 for(plane_index=0; plane_index<2; plane_index++){ |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3461 Plane *p= &s->plane[plane_index]; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3462 p->last_htaps=0; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3463 p->last_diag_mc=0; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3464 memset(p->last_hcoeff, 0, sizeof(p->last_hcoeff)); |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3465 } |
4331
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3466 } |
2138 | 3467 if(s->keyframe){ |
3468 put_symbol(&s->c, s->header_state, s->version, 0); | |
2335 | 3469 put_rac(&s->c, s->header_state, s->always_reset); |
2138 | 3470 put_symbol(&s->c, s->header_state, s->temporal_decomposition_type, 0); |
3471 put_symbol(&s->c, s->header_state, s->temporal_decomposition_count, 0); | |
3472 put_symbol(&s->c, s->header_state, s->spatial_decomposition_count, 0); | |
3473 put_symbol(&s->c, s->header_state, s->colorspace_type, 0); | |
3474 put_symbol(&s->c, s->header_state, s->chroma_h_shift, 0); | |
3475 put_symbol(&s->c, s->header_state, s->chroma_v_shift, 0); | |
2335 | 3476 put_rac(&s->c, s->header_state, s->spatial_scalability); |
3477 // put_rac(&s->c, s->header_state, s->rate_scalability); | |
3314 | 3478 put_symbol(&s->c, s->header_state, s->max_ref_frames-1, 0); |
2138 | 3479 |
5666
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3480 encode_qlogs(s); |
2138 | 3481 } |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3482 |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3483 if(!s->keyframe){ |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3484 int update_mc=0; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3485 for(plane_index=0; plane_index<2; plane_index++){ |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3486 Plane *p= &s->plane[plane_index]; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3487 update_mc |= p->last_htaps != p->htaps; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3488 update_mc |= p->last_diag_mc != p->diag_mc; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3489 update_mc |= !!memcmp(p->last_hcoeff, p->hcoeff, sizeof(p->hcoeff)); |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3490 } |
5667
9242e125395f
do not force the halfpel filter coeffs to be retransmitted on every frame
michael
parents:
5666
diff
changeset
|
3491 put_rac(&s->c, s->header_state, update_mc); |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3492 if(update_mc){ |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3493 for(plane_index=0; plane_index<2; plane_index++){ |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3494 Plane *p= &s->plane[plane_index]; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3495 put_rac(&s->c, s->header_state, p->diag_mc); |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3496 put_symbol(&s->c, s->header_state, p->htaps/2-1, 0); |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3497 for(i= p->htaps/2; i; i--) |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3498 put_symbol(&s->c, s->header_state, FFABS(p->hcoeff[i]), 0); |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3499 } |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3500 } |
5670
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
3501 if(s->last_spatial_decomposition_count != s->spatial_decomposition_count){ |
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
3502 put_rac(&s->c, s->header_state, 1); |
5666
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3503 put_symbol(&s->c, s->header_state, s->spatial_decomposition_count, 0); |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3504 encode_qlogs(s); |
5670
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
3505 }else |
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
3506 put_rac(&s->c, s->header_state, 0); |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3507 } |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3508 |
4331
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3509 put_symbol(&s->c, s->header_state, s->spatial_decomposition_type - s->last_spatial_decomposition_type, 1); |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3510 put_symbol(&s->c, s->header_state, s->qlog - s->last_qlog , 1); |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3511 put_symbol(&s->c, s->header_state, s->mv_scale - s->last_mv_scale, 1); |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3512 put_symbol(&s->c, s->header_state, s->qbias - s->last_qbias , 1); |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3513 put_symbol(&s->c, s->header_state, s->block_max_depth - s->last_block_max_depth, 1); |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3514 |
6144 | 3515 } |
3516 | |
3517 static void update_last_header_values(SnowContext *s){ | |
3518 int plane_index; | |
3519 | |
3520 if(!s->keyframe){ | |
3521 for(plane_index=0; plane_index<2; plane_index++){ | |
3522 Plane *p= &s->plane[plane_index]; | |
3523 p->last_diag_mc= p->diag_mc; | |
3524 p->last_htaps = p->htaps; | |
3525 memcpy(p->last_hcoeff, p->hcoeff, sizeof(p->hcoeff)); | |
3526 } | |
3527 } | |
3528 | |
6414 | 3529 s->last_spatial_decomposition_type = s->spatial_decomposition_type; |
3530 s->last_qlog = s->qlog; | |
3531 s->last_qbias = s->qbias; | |
3532 s->last_mv_scale = s->mv_scale; | |
3533 s->last_block_max_depth = s->block_max_depth; | |
3534 s->last_spatial_decomposition_count = s->spatial_decomposition_count; | |
2138 | 3535 } |
3536 | |
5666
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3537 static void decode_qlogs(SnowContext *s){ |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3538 int plane_index, level, orientation; |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3539 |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3540 for(plane_index=0; plane_index<3; plane_index++){ |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3541 for(level=0; level<s->spatial_decomposition_count; level++){ |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3542 for(orientation=level ? 1:0; orientation<4; orientation++){ |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3543 int q; |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3544 if (plane_index==2) q= s->plane[1].band[level][orientation].qlog; |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3545 else if(orientation==2) q= s->plane[plane_index].band[level][1].qlog; |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3546 else q= get_symbol(&s->c, s->header_state, 1); |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3547 s->plane[plane_index].band[level][orientation].qlog= q; |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3548 } |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3549 } |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3550 } |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3551 } |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3552 |
2138 | 3553 static int decode_header(SnowContext *s){ |
5666
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3554 int plane_index; |
2335 | 3555 uint8_t kstate[32]; |
2138 | 3556 |
2967 | 3557 memset(kstate, MID_STATE, sizeof(kstate)); |
2335 | 3558 |
3559 s->keyframe= get_rac(&s->c, kstate); | |
4331
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3560 if(s->keyframe || s->always_reset){ |
2199 | 3561 reset_contexts(s); |
4331
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3562 s->spatial_decomposition_type= |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3563 s->qlog= |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3564 s->qbias= |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3565 s->mv_scale= |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3566 s->block_max_depth= 0; |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3567 } |
2138 | 3568 if(s->keyframe){ |
3569 s->version= get_symbol(&s->c, s->header_state, 0); | |
3570 if(s->version>0){ | |
3571 av_log(s->avctx, AV_LOG_ERROR, "version %d not supported", s->version); | |
3572 return -1; | |
3573 } | |
2335 | 3574 s->always_reset= get_rac(&s->c, s->header_state); |
2138 | 3575 s->temporal_decomposition_type= get_symbol(&s->c, s->header_state, 0); |
3576 s->temporal_decomposition_count= get_symbol(&s->c, s->header_state, 0); | |
3577 s->spatial_decomposition_count= get_symbol(&s->c, s->header_state, 0); | |
3578 s->colorspace_type= get_symbol(&s->c, s->header_state, 0); | |
3579 s->chroma_h_shift= get_symbol(&s->c, s->header_state, 0); | |
3580 s->chroma_v_shift= get_symbol(&s->c, s->header_state, 0); | |
2335 | 3581 s->spatial_scalability= get_rac(&s->c, s->header_state); |
3582 // s->rate_scalability= get_rac(&s->c, s->header_state); | |
3314 | 3583 s->max_ref_frames= get_symbol(&s->c, s->header_state, 0)+1; |
2138 | 3584 |
5666
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3585 decode_qlogs(s); |
2138 | 3586 } |
2967 | 3587 |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3588 if(!s->keyframe){ |
5667
9242e125395f
do not force the halfpel filter coeffs to be retransmitted on every frame
michael
parents:
5666
diff
changeset
|
3589 if(get_rac(&s->c, s->header_state)){ |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3590 for(plane_index=0; plane_index<2; plane_index++){ |
5807 | 3591 int htaps, i, sum=0; |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3592 Plane *p= &s->plane[plane_index]; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3593 p->diag_mc= get_rac(&s->c, s->header_state); |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3594 htaps= get_symbol(&s->c, s->header_state, 0)*2 + 2; |
5652
941e5deeb2a4
rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
michael
parents:
5651
diff
changeset
|
3595 if((unsigned)htaps > HTAPS_MAX || htaps==0) |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3596 return -1; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3597 p->htaps= htaps; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3598 for(i= htaps/2; i; i--){ |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3599 p->hcoeff[i]= get_symbol(&s->c, s->header_state, 0) * (1-2*(i&1)); |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3600 sum += p->hcoeff[i]; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3601 } |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3602 p->hcoeff[0]= 32-sum; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3603 } |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3604 s->plane[2].diag_mc= s->plane[1].diag_mc; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3605 s->plane[2].htaps = s->plane[1].htaps; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3606 memcpy(s->plane[2].hcoeff, s->plane[1].hcoeff, sizeof(s->plane[1].hcoeff)); |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3607 } |
5666
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3608 if(get_rac(&s->c, s->header_state)){ |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3609 s->spatial_decomposition_count= get_symbol(&s->c, s->header_state, 0); |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3610 decode_qlogs(s); |
b5c137f3f53a
allow changing of the spatial_decomposition_count and quantization tables
michael
parents:
5665
diff
changeset
|
3611 } |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3612 } |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3613 |
4331
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3614 s->spatial_decomposition_type+= get_symbol(&s->c, s->header_state, 1); |
5588
effa59ca89b3
we only have 2 wavelets, the 3rd was just for experimentation ...
michael
parents:
5587
diff
changeset
|
3615 if(s->spatial_decomposition_type > 1){ |
2138 | 3616 av_log(s->avctx, AV_LOG_ERROR, "spatial_decomposition_type %d not supported", s->spatial_decomposition_type); |
3617 return -1; | |
3618 } | |
2967 | 3619 |
4331
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3620 s->qlog += get_symbol(&s->c, s->header_state, 1); |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3621 s->mv_scale += get_symbol(&s->c, s->header_state, 1); |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3622 s->qbias += get_symbol(&s->c, s->header_state, 1); |
e571dfe677be
store a few values in the header as difference to the last
michael
parents:
4283
diff
changeset
|
3623 s->block_max_depth+= get_symbol(&s->c, s->header_state, 1); |
3303
68721b62a528
sanity checks, some might have been exploitable ...
michael
parents:
3206
diff
changeset
|
3624 if(s->block_max_depth > 1 || s->block_max_depth < 0){ |
2952 | 3625 av_log(s->avctx, AV_LOG_ERROR, "block_max_depth= %d is too large", s->block_max_depth); |
3626 s->block_max_depth= 0; | |
3627 return -1; | |
3628 } | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
3629 |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
3630 return 0; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
3631 } |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
3632 |
3075 | 3633 static void init_qexp(void){ |
2600 | 3634 int i; |
3635 double v=128; | |
3636 | |
3637 for(i=0; i<QROOT; i++){ | |
3638 qexp[i]= lrintf(v); | |
2967 | 3639 v *= pow(2, 1.0 / QROOT); |
2600 | 3640 } |
3641 } | |
3642 | |
6517
48759bfbd073
Apply 'cold' attribute to init/uninit functions in libavcodec
zuxy
parents:
6481
diff
changeset
|
3643 static av_cold int common_init(AVCodecContext *avctx){ |
2138 | 3644 SnowContext *s = avctx->priv_data; |
3645 int width, height; | |
3325
c2a017de6bea
Snow: scale predicted mv based on which reference frame the neighbors used.
lorenm
parents:
3324
diff
changeset
|
3646 int i, j; |
2138 | 3647 |
3648 s->avctx= avctx; | |
2967 | 3649 |
2138 | 3650 dsputil_init(&s->dsp, avctx); |
3651 | |
3652 #define mcf(dx,dy)\ | |
3653 s->dsp.put_qpel_pixels_tab [0][dy+dx/4]=\ | |
3654 s->dsp.put_no_rnd_qpel_pixels_tab[0][dy+dx/4]=\ | |
2224
11d54cb7ac4e
100l (forgot to set the 8x8 mc functions, so ME did use the mpeg ones ...)
michael
parents:
2223
diff
changeset
|
3655 s->dsp.put_h264_qpel_pixels_tab[0][dy+dx/4];\ |
11d54cb7ac4e
100l (forgot to set the 8x8 mc functions, so ME did use the mpeg ones ...)
michael
parents:
2223
diff
changeset
|
3656 s->dsp.put_qpel_pixels_tab [1][dy+dx/4]=\ |
11d54cb7ac4e
100l (forgot to set the 8x8 mc functions, so ME did use the mpeg ones ...)
michael
parents:
2223
diff
changeset
|
3657 s->dsp.put_no_rnd_qpel_pixels_tab[1][dy+dx/4]=\ |
11d54cb7ac4e
100l (forgot to set the 8x8 mc functions, so ME did use the mpeg ones ...)
michael
parents:
2223
diff
changeset
|
3658 s->dsp.put_h264_qpel_pixels_tab[1][dy+dx/4]; |
2138 | 3659 |
3660 mcf( 0, 0) | |
3661 mcf( 4, 0) | |
3662 mcf( 8, 0) | |
3663 mcf(12, 0) | |
3664 mcf( 0, 4) | |
3665 mcf( 4, 4) | |
3666 mcf( 8, 4) | |
3667 mcf(12, 4) | |
3668 mcf( 0, 8) | |
3669 mcf( 4, 8) | |
3670 mcf( 8, 8) | |
3671 mcf(12, 8) | |
3672 mcf( 0,12) | |
3673 mcf( 4,12) | |
3674 mcf( 8,12) | |
3675 mcf(12,12) | |
3676 | |
3677 #define mcfh(dx,dy)\ | |
3678 s->dsp.put_pixels_tab [0][dy/4+dx/8]=\ | |
3679 s->dsp.put_no_rnd_pixels_tab[0][dy/4+dx/8]=\ | |
2224
11d54cb7ac4e
100l (forgot to set the 8x8 mc functions, so ME did use the mpeg ones ...)
michael
parents:
2223
diff
changeset
|
3680 mc_block_hpel ## dx ## dy ## 16;\ |
11d54cb7ac4e
100l (forgot to set the 8x8 mc functions, so ME did use the mpeg ones ...)
michael
parents:
2223
diff
changeset
|
3681 s->dsp.put_pixels_tab [1][dy/4+dx/8]=\ |
11d54cb7ac4e
100l (forgot to set the 8x8 mc functions, so ME did use the mpeg ones ...)
michael
parents:
2223
diff
changeset
|
3682 s->dsp.put_no_rnd_pixels_tab[1][dy/4+dx/8]=\ |
11d54cb7ac4e
100l (forgot to set the 8x8 mc functions, so ME did use the mpeg ones ...)
michael
parents:
2223
diff
changeset
|
3683 mc_block_hpel ## dx ## dy ## 8; |
2138 | 3684 |
3685 mcfh(0, 0) | |
3686 mcfh(8, 0) | |
3687 mcfh(0, 8) | |
3688 mcfh(8, 8) | |
2600 | 3689 |
3690 if(!qexp[0]) | |
3691 init_qexp(); | |
3692 | |
2138 | 3693 // dec += FFMAX(s->chroma_h_shift, s->chroma_v_shift); |
2967 | 3694 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
3695 width= s->avctx->width; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
3696 height= s->avctx->height; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
3697 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3698 s->spatial_idwt_buffer= av_mallocz(width*height*sizeof(IDWTELEM)); |
6412 | 3699 s->spatial_dwt_buffer= av_mallocz(width*height*sizeof(DWTELEM)); //FIXME this does not belong here |
2967 | 3700 |
5665
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3701 for(i=0; i<MAX_REF_FRAMES; i++) |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3702 for(j=0; j<MAX_REF_FRAMES; j++) |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3703 scale_mv_ref[i][j] = 256*(i+1)/(j+1); |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3704 |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3705 s->avctx->get_buffer(s->avctx, &s->mconly_picture); |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3706 |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3707 return 0; |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3708 } |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3709 |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3710 static int common_init_after_header(AVCodecContext *avctx){ |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3711 SnowContext *s = avctx->priv_data; |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3712 int plane_index, level, orientation; |
2967 | 3713 |
3714 for(plane_index=0; plane_index<3; plane_index++){ | |
2138 | 3715 int w= s->avctx->width; |
3716 int h= s->avctx->height; | |
3717 | |
3718 if(plane_index){ | |
3719 w>>= s->chroma_h_shift; | |
3720 h>>= s->chroma_v_shift; | |
3721 } | |
3722 s->plane[plane_index].width = w; | |
3723 s->plane[plane_index].height= h; | |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
3724 |
2138 | 3725 for(level=s->spatial_decomposition_count-1; level>=0; level--){ |
3726 for(orientation=level ? 1 : 0; orientation<4; orientation++){ | |
3727 SubBand *b= &s->plane[plane_index].band[level][orientation]; | |
2967 | 3728 |
2138 | 3729 b->buf= s->spatial_dwt_buffer; |
3730 b->level= level; | |
3731 b->stride= s->plane[plane_index].width << (s->spatial_decomposition_count - level); | |
3732 b->width = (w + !(orientation&1))>>1; | |
3733 b->height= (h + !(orientation>1))>>1; | |
2967 | 3734 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3735 b->stride_line = 1 << (s->spatial_decomposition_count - level); |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3736 b->buf_x_offset = 0; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3737 b->buf_y_offset = 0; |
2967 | 3738 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3739 if(orientation&1){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3740 b->buf += (w+1)>>1; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3741 b->buf_x_offset = (w+1)>>1; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3742 } |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3743 if(orientation>1){ |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3744 b->buf += b->stride>>1; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3745 b->buf_y_offset = b->stride_line >> 1; |
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3746 } |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3747 b->ibuf= s->spatial_idwt_buffer + (b->buf - s->spatial_dwt_buffer); |
2967 | 3748 |
2138 | 3749 if(level) |
3750 b->parent= &s->plane[plane_index].band[level-1][orientation]; | |
5665
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3751 //FIXME avoid this realloc |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3752 av_freep(&b->x_coeff); |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
3753 b->x_coeff=av_mallocz(((b->width+1) * b->height+1)*sizeof(x_and_coeff)); |
2138 | 3754 } |
3755 w= (w+1)>>1; | |
3756 h= (h+1)>>1; | |
3757 } | |
3758 } | |
2967 | 3759 |
2138 | 3760 return 0; |
3761 } | |
3762 | |
3322
0b4f548dfb44
Snow: 10l. 1pass ratecontrol failed to set snow's internal quant.
lorenm
parents:
3314
diff
changeset
|
3763 static int qscale2qlog(int qscale){ |
0b4f548dfb44
Snow: 10l. 1pass ratecontrol failed to set snow's internal quant.
lorenm
parents:
3314
diff
changeset
|
3764 return rint(QROOT*log(qscale / (float)FF_QP2LAMBDA)/log(2)) |
0b4f548dfb44
Snow: 10l. 1pass ratecontrol failed to set snow's internal quant.
lorenm
parents:
3314
diff
changeset
|
3765 + 61*QROOT/8; //<64 >60 |
0b4f548dfb44
Snow: 10l. 1pass ratecontrol failed to set snow's internal quant.
lorenm
parents:
3314
diff
changeset
|
3766 } |
0b4f548dfb44
Snow: 10l. 1pass ratecontrol failed to set snow's internal quant.
lorenm
parents:
3314
diff
changeset
|
3767 |
3661
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
3768 static int ratecontrol_1pass(SnowContext *s, AVFrame *pict) |
3313 | 3769 { |
6412 | 3770 /* Estimate the frame's complexity as a sum of weighted dwt coefficients. |
3313 | 3771 * FIXME we know exact mv bits at this point, |
3772 * but ratecontrol isn't set up to include them. */ | |
3773 uint32_t coef_sum= 0; | |
3661
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
3774 int level, orientation, delta_qlog; |
3313 | 3775 |
3776 for(level=0; level<s->spatial_decomposition_count; level++){ | |
3777 for(orientation=level ? 1 : 0; orientation<4; orientation++){ | |
3778 SubBand *b= &s->plane[0].band[level][orientation]; | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3779 IDWTELEM *buf= b->ibuf; |
3313 | 3780 const int w= b->width; |
3781 const int h= b->height; | |
3782 const int stride= b->stride; | |
4594 | 3783 const int qlog= av_clip(2*QROOT + b->qlog, 0, QROOT*16); |
3313 | 3784 const int qmul= qexp[qlog&(QROOT-1)]<<(qlog>>QSHIFT); |
3785 const int qdiv= (1<<16)/qmul; | |
3786 int x, y; | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3787 //FIXME this is ugly |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3788 for(y=0; y<h; y++) |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3789 for(x=0; x<w; x++) |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3790 buf[x+y*stride]= b->buf[x+y*stride]; |
3313 | 3791 if(orientation==0) |
3792 decorrelate(s, b, buf, stride, 1, 0); | |
3793 for(y=0; y<h; y++) | |
3794 for(x=0; x<w; x++) | |
3795 coef_sum+= abs(buf[x+y*stride]) * qdiv >> 16; | |
3796 } | |
3797 } | |
3798 | |
3799 /* ugly, ratecontrol just takes a sqrt again */ | |
3800 coef_sum = (uint64_t)coef_sum * coef_sum >> 16; | |
3801 assert(coef_sum < INT_MAX); | |
3802 | |
6481 | 3803 if(pict->pict_type == FF_I_TYPE){ |
3313 | 3804 s->m.current_picture.mb_var_sum= coef_sum; |
3805 s->m.current_picture.mc_mb_var_sum= 0; | |
3806 }else{ | |
3807 s->m.current_picture.mc_mb_var_sum= coef_sum; | |
3808 s->m.current_picture.mb_var_sum= 0; | |
3809 } | |
3810 | |
3811 pict->quality= ff_rate_estimate_qscale(&s->m, 1); | |
3766 | 3812 if (pict->quality < 0) |
4011
5bce97c30a69
-1 is a valid return value in ratecontrol_1pass() -> 100l for takis
michael
parents:
4001
diff
changeset
|
3813 return INT_MIN; |
3313 | 3814 s->lambda= pict->quality * 3/2; |
3661
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
3815 delta_qlog= qscale2qlog(pict->quality) - s->qlog; |
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
3816 s->qlog+= delta_qlog; |
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
3817 return delta_qlog; |
3313 | 3818 } |
2138 | 3819 |
5909 | 3820 static void calculate_visual_weight(SnowContext *s, Plane *p){ |
2138 | 3821 int width = p->width; |
3822 int height= p->height; | |
2198 | 3823 int level, orientation, x, y; |
2138 | 3824 |
3825 for(level=0; level<s->spatial_decomposition_count; level++){ | |
3826 for(orientation=level ? 1 : 0; orientation<4; orientation++){ | |
3827 SubBand *b= &p->band[level][orientation]; | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3828 IDWTELEM *ibuf= b->ibuf; |
2138 | 3829 int64_t error=0; |
2967 | 3830 |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3831 memset(s->spatial_idwt_buffer, 0, sizeof(*s->spatial_idwt_buffer)*width*height); |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3832 ibuf[b->width/2 + b->height/2*b->stride]= 256*16; |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3833 ff_spatial_idwt(s->spatial_idwt_buffer, width, height, width, s->spatial_decomposition_type, s->spatial_decomposition_count); |
2138 | 3834 for(y=0; y<height; y++){ |
3835 for(x=0; x<width; x++){ | |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
3836 int64_t d= s->spatial_idwt_buffer[x + y*width]*16; |
2138 | 3837 error += d*d; |
3838 } | |
3839 } | |
3840 | |
3841 b->qlog= (int)(log(352256.0/sqrt(error)) / log(pow(2.0, 1.0/QROOT))+0.5); | |
3842 } | |
3843 } | |
3844 } | |
3845 | |
5702 | 3846 #define QUANTIZE2 0 |
3847 | |
3848 #if QUANTIZE2==1 | |
3849 #define Q2_STEP 8 | |
3850 | |
3851 static void find_sse(SnowContext *s, Plane *p, int *score, int score_stride, IDWTELEM *r0, IDWTELEM *r1, int level, int orientation){ | |
3852 SubBand *b= &p->band[level][orientation]; | |
3853 int x, y; | |
3854 int xo=0; | |
3855 int yo=0; | |
3856 int step= 1 << (s->spatial_decomposition_count - level); | |
3857 | |
3858 if(orientation&1) | |
3859 xo= step>>1; | |
3860 if(orientation&2) | |
3861 yo= step>>1; | |
3862 | |
6412 | 3863 //FIXME bias for nonzero ? |
5702 | 3864 //FIXME optimize |
3865 memset(score, 0, sizeof(*score)*score_stride*((p->height + Q2_STEP-1)/Q2_STEP)); | |
3866 for(y=0; y<p->height; y++){ | |
3867 for(x=0; x<p->width; x++){ | |
3868 int sx= (x-xo + step/2) / step / Q2_STEP; | |
3869 int sy= (y-yo + step/2) / step / Q2_STEP; | |
3870 int v= r0[x + y*p->width] - r1[x + y*p->width]; | |
3871 assert(sx>=0 && sy>=0 && sx < score_stride); | |
3872 v= ((v+8)>>4)<<4; | |
3873 score[sx + sy*score_stride] += v*v; | |
3874 assert(score[sx + sy*score_stride] >= 0); | |
3875 } | |
3876 } | |
3877 } | |
3878 | |
3879 static void dequantize_all(SnowContext *s, Plane *p, IDWTELEM *buffer, int width, int height){ | |
3880 int level, orientation; | |
3881 | |
3882 for(level=0; level<s->spatial_decomposition_count; level++){ | |
3883 for(orientation=level ? 1 : 0; orientation<4; orientation++){ | |
3884 SubBand *b= &p->band[level][orientation]; | |
3885 IDWTELEM *dst= buffer + (b->ibuf - s->spatial_idwt_buffer); | |
3886 | |
3887 dequantize(s, b, dst, b->stride); | |
3888 } | |
3889 } | |
3890 } | |
3891 | |
3892 static void dwt_quantize(SnowContext *s, Plane *p, DWTELEM *buffer, int width, int height, int stride, int type){ | |
3893 int level, orientation, ys, xs, x, y, pass; | |
3894 IDWTELEM best_dequant[height * stride]; | |
3895 IDWTELEM idwt2_buffer[height * stride]; | |
3896 const int score_stride= (width + 10)/Q2_STEP; | |
3897 int best_score[(width + 10)/Q2_STEP * (height + 10)/Q2_STEP]; //FIXME size | |
3898 int score[(width + 10)/Q2_STEP * (height + 10)/Q2_STEP]; //FIXME size | |
3899 int threshold= (s->m.lambda * s->m.lambda) >> 6; | |
3900 | |
3901 //FIXME pass the copy cleanly ? | |
3902 | |
3903 // memcpy(dwt_buffer, buffer, height * stride * sizeof(DWTELEM)); | |
3904 ff_spatial_dwt(buffer, width, height, stride, type, s->spatial_decomposition_count); | |
3905 | |
3906 for(level=0; level<s->spatial_decomposition_count; level++){ | |
3907 for(orientation=level ? 1 : 0; orientation<4; orientation++){ | |
3908 SubBand *b= &p->band[level][orientation]; | |
3909 IDWTELEM *dst= best_dequant + (b->ibuf - s->spatial_idwt_buffer); | |
3910 DWTELEM *src= buffer + (b-> buf - s->spatial_dwt_buffer); | |
6412 | 3911 assert(src == b->buf); // code does not depend on this but it is true currently |
5702 | 3912 |
3913 quantize(s, b, dst, src, b->stride, s->qbias); | |
3914 } | |
3915 } | |
3916 for(pass=0; pass<1; pass++){ | |
3917 if(s->qbias == 0) //keyframe | |
3918 continue; | |
3919 for(level=0; level<s->spatial_decomposition_count; level++){ | |
3920 for(orientation=level ? 1 : 0; orientation<4; orientation++){ | |
3921 SubBand *b= &p->band[level][orientation]; | |
3922 IDWTELEM *dst= idwt2_buffer + (b->ibuf - s->spatial_idwt_buffer); | |
3923 IDWTELEM *best_dst= best_dequant + (b->ibuf - s->spatial_idwt_buffer); | |
3924 | |
3925 for(ys= 0; ys<Q2_STEP; ys++){ | |
3926 for(xs= 0; xs<Q2_STEP; xs++){ | |
3927 memcpy(idwt2_buffer, best_dequant, height * stride * sizeof(IDWTELEM)); | |
3928 dequantize_all(s, p, idwt2_buffer, width, height); | |
3929 ff_spatial_idwt(idwt2_buffer, width, height, stride, type, s->spatial_decomposition_count); | |
3930 find_sse(s, p, best_score, score_stride, idwt2_buffer, s->spatial_idwt_buffer, level, orientation); | |
3931 memcpy(idwt2_buffer, best_dequant, height * stride * sizeof(IDWTELEM)); | |
3932 for(y=ys; y<b->height; y+= Q2_STEP){ | |
3933 for(x=xs; x<b->width; x+= Q2_STEP){ | |
3934 if(dst[x + y*b->stride]<0) dst[x + y*b->stride]++; | |
3935 if(dst[x + y*b->stride]>0) dst[x + y*b->stride]--; | |
6412 | 3936 //FIXME try more than just -- |
5702 | 3937 } |
3938 } | |
3939 dequantize_all(s, p, idwt2_buffer, width, height); | |
3940 ff_spatial_idwt(idwt2_buffer, width, height, stride, type, s->spatial_decomposition_count); | |
3941 find_sse(s, p, score, score_stride, idwt2_buffer, s->spatial_idwt_buffer, level, orientation); | |
3942 for(y=ys; y<b->height; y+= Q2_STEP){ | |
3943 for(x=xs; x<b->width; x+= Q2_STEP){ | |
3944 int score_idx= x/Q2_STEP + (y/Q2_STEP)*score_stride; | |
3945 if(score[score_idx] <= best_score[score_idx] + threshold){ | |
3946 best_score[score_idx]= score[score_idx]; | |
3947 if(best_dst[x + y*b->stride]<0) best_dst[x + y*b->stride]++; | |
3948 if(best_dst[x + y*b->stride]>0) best_dst[x + y*b->stride]--; | |
3949 //FIXME copy instead | |
3950 } | |
3951 } | |
3952 } | |
3953 } | |
3954 } | |
3955 } | |
3956 } | |
3957 } | |
6412 | 3958 memcpy(s->spatial_idwt_buffer, best_dequant, height * stride * sizeof(IDWTELEM)); //FIXME work with that directly instead of copy at the end |
5702 | 3959 } |
3960 | |
5910 | 3961 #endif /* QUANTIZE2==1 */ |
5702 | 3962 |
6517
48759bfbd073
Apply 'cold' attribute to init/uninit functions in libavcodec
zuxy
parents:
6481
diff
changeset
|
3963 static av_cold int encode_init(AVCodecContext *avctx) |
2138 | 3964 { |
3965 SnowContext *s = avctx->priv_data; | |
2198 | 3966 int plane_index; |
2138 | 3967 |
2658
d1609cfeb1d0
#defines for strict_std_compliance and split between inofficial extensions and non standarized things
michael
parents:
2635
diff
changeset
|
3968 if(avctx->strict_std_compliance > FF_COMPLIANCE_EXPERIMENTAL){ |
6412 | 3969 av_log(avctx, AV_LOG_ERROR, "This codec is under development, files encoded with it may not be decodable with future versions!!!\n" |
3970 "Use vstrict=-2 / -strict -2 to use it anyway.\n"); | |
2151 | 3971 return -1; |
3972 } | |
2967 | 3973 |
3327
955096780e7c
Snow: don't try to encode lossless with 9/7 wavelet, because it isn't lossless.
lorenm
parents:
3326
diff
changeset
|
3974 if(avctx->prediction_method == DWT_97 |
955096780e7c
Snow: don't try to encode lossless with 9/7 wavelet, because it isn't lossless.
lorenm
parents:
3326
diff
changeset
|
3975 && (avctx->flags & CODEC_FLAG_QSCALE) |
955096780e7c
Snow: don't try to encode lossless with 9/7 wavelet, because it isn't lossless.
lorenm
parents:
3326
diff
changeset
|
3976 && avctx->global_quality == 0){ |
6412 | 3977 av_log(avctx, AV_LOG_ERROR, "The 9/7 wavelet is incompatible with lossless mode.\n"); |
3327
955096780e7c
Snow: don't try to encode lossless with 9/7 wavelet, because it isn't lossless.
lorenm
parents:
3326
diff
changeset
|
3978 return -1; |
955096780e7c
Snow: don't try to encode lossless with 9/7 wavelet, because it isn't lossless.
lorenm
parents:
3326
diff
changeset
|
3979 } |
955096780e7c
Snow: don't try to encode lossless with 9/7 wavelet, because it isn't lossless.
lorenm
parents:
3326
diff
changeset
|
3980 |
5665
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3981 s->spatial_decomposition_type= avctx->prediction_method; //FIXME add decorrelator type r transform_type |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3982 |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3983 s->chroma_h_shift= 1; //FIXME XXX |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3984 s->chroma_v_shift= 1; |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3985 |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3986 s->mv_scale = (avctx->flags & CODEC_FLAG_QPEL) ? 2 : 4; |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3987 s->block_max_depth= (avctx->flags & CODEC_FLAG_4MV ) ? 1 : 0; |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3988 |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3989 for(plane_index=0; plane_index<3; plane_index++){ |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3990 s->plane[plane_index].diag_mc= 1; |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3991 s->plane[plane_index].htaps= 6; |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3992 s->plane[plane_index].hcoeff[0]= 40; |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3993 s->plane[plane_index].hcoeff[1]= -10; |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3994 s->plane[plane_index].hcoeff[2]= 2; |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3995 s->plane[plane_index].fast_mc= 1; |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3996 } |
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
3997 |
2138 | 3998 common_init(avctx); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
3999 alloc_blocks(s); |
2967 | 4000 |
2138 | 4001 s->version=0; |
2967 | 4002 |
2608
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4003 s->m.avctx = avctx; |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4004 s->m.flags = avctx->flags; |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4005 s->m.bit_rate= avctx->bit_rate; |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4006 |
2138 | 4007 s->m.me.scratchpad= av_mallocz((avctx->width+64)*2*16*2*sizeof(uint8_t)); |
4008 s->m.me.map = av_mallocz(ME_MAP_SIZE*sizeof(uint32_t)); | |
4009 s->m.me.score_map = av_mallocz(ME_MAP_SIZE*sizeof(uint32_t)); | |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
4010 s->m.obmc_scratchpad= av_mallocz(MB_SIZE*MB_SIZE*12*sizeof(uint32_t)); |
2138 | 4011 h263_encode_init(&s->m); //mv_penalty |
4012 | |
3314 | 4013 s->max_ref_frames = FFMAX(FFMIN(avctx->refs, MAX_REF_FRAMES), 1); |
4014 | |
2608
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4015 if(avctx->flags&CODEC_FLAG_PASS1){ |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4016 if(!avctx->stats_out) |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4017 avctx->stats_out = av_mallocz(256); |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4018 } |
3322
0b4f548dfb44
Snow: 10l. 1pass ratecontrol failed to set snow's internal quant.
lorenm
parents:
3314
diff
changeset
|
4019 if((avctx->flags&CODEC_FLAG_PASS2) || !(avctx->flags&CODEC_FLAG_QSCALE)){ |
2608
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4020 if(ff_rate_control_init(&s->m) < 0) |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4021 return -1; |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4022 } |
3313 | 4023 s->pass1_rc= !(avctx->flags & (CODEC_FLAG_QSCALE|CODEC_FLAG_PASS2)); |
2608
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4024 |
2138 | 4025 avctx->coded_frame= &s->current_picture; |
4026 switch(avctx->pix_fmt){ | |
4027 // case PIX_FMT_YUV444P: | |
4028 // case PIX_FMT_YUV422P: | |
4029 case PIX_FMT_YUV420P: | |
4030 case PIX_FMT_GRAY8: | |
4031 // case PIX_FMT_YUV411P: | |
4032 // case PIX_FMT_YUV410P: | |
4033 s->colorspace_type= 0; | |
4034 break; | |
4494
ce643a22f049
Replace deprecated PIX_FMT names by the newer variants.
diego
parents:
4436
diff
changeset
|
4035 /* case PIX_FMT_RGB32: |
2138 | 4036 s->colorspace= 1; |
4037 break;*/ | |
4038 default: | |
5908 | 4039 av_log(avctx, AV_LOG_ERROR, "pixel format not supported\n"); |
2138 | 4040 return -1; |
4041 } | |
4042 // avcodec_get_chroma_sub_sample(avctx->pix_fmt, &s->chroma_h_shift, &s->chroma_v_shift); | |
4043 s->chroma_h_shift= 1; | |
4044 s->chroma_v_shift= 1; | |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4045 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4046 ff_set_cmp(&s->dsp, s->dsp.me_cmp, s->avctx->me_cmp); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4047 ff_set_cmp(&s->dsp, s->dsp.me_sub_cmp, s->avctx->me_sub_cmp); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4048 |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4049 s->avctx->get_buffer(s->avctx, &s->input_picture); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4050 |
3314 | 4051 if(s->avctx->me_method == ME_ITER){ |
4052 int i; | |
4053 int size= s->b_width * s->b_height << 2*s->block_max_depth; | |
4054 for(i=0; i<s->max_ref_frames; i++){ | |
4055 s->ref_mvs[i]= av_mallocz(size*sizeof(int16_t[2])); | |
4056 s->ref_scores[i]= av_mallocz(size*sizeof(uint32_t)); | |
4057 } | |
4058 } | |
4059 | |
2138 | 4060 return 0; |
4061 } | |
4062 | |
5924 | 4063 #define USE_HALFPEL_PLANE 0 |
4064 | |
5633
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4065 static void halfpel_interpol(SnowContext *s, uint8_t *halfpel[4][4], AVFrame *frame){ |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4066 int p,x,y; |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4067 |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4068 assert(!(s->avctx->flags & CODEC_FLAG_EMU_EDGE)); |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4069 |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4070 for(p=0; p<3; p++){ |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4071 int is_chroma= !!p; |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4072 int w= s->avctx->width >>is_chroma; |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4073 int h= s->avctx->height >>is_chroma; |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4074 int ls= frame->linesize[p]; |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4075 uint8_t *src= frame->data[p]; |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4076 |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4077 halfpel[1][p]= (uint8_t*)av_malloc(ls * (h+2*EDGE_WIDTH)) + EDGE_WIDTH*(1+ls); |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4078 halfpel[2][p]= (uint8_t*)av_malloc(ls * (h+2*EDGE_WIDTH)) + EDGE_WIDTH*(1+ls); |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4079 halfpel[3][p]= (uint8_t*)av_malloc(ls * (h+2*EDGE_WIDTH)) + EDGE_WIDTH*(1+ls); |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4080 |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4081 halfpel[0][p]= src; |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4082 for(y=0; y<h; y++){ |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4083 for(x=0; x<w; x++){ |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4084 int i= y*ls + x; |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4085 |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4086 halfpel[1][p][i]= (20*(src[i] + src[i+1]) - 5*(src[i-1] + src[i+2]) + (src[i-2] + src[i+3]) + 16 )>>5; |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4087 } |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4088 } |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4089 for(y=0; y<h; y++){ |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4090 for(x=0; x<w; x++){ |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4091 int i= y*ls + x; |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4092 |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4093 halfpel[2][p][i]= (20*(src[i] + src[i+ls]) - 5*(src[i-ls] + src[i+2*ls]) + (src[i-2*ls] + src[i+3*ls]) + 16 )>>5; |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4094 } |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4095 } |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4096 src= halfpel[1][p]; |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4097 for(y=0; y<h; y++){ |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4098 for(x=0; x<w; x++){ |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4099 int i= y*ls + x; |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4100 |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4101 halfpel[3][p][i]= (20*(src[i] + src[i+ls]) - 5*(src[i-ls] + src[i+2*ls]) + (src[i-2*ls] + src[i+3*ls]) + 16 )>>5; |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4102 } |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4103 } |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4104 |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4105 //FIXME border! |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4106 } |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4107 } |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4108 |
2138 | 4109 static int frame_start(SnowContext *s){ |
4110 AVFrame tmp; | |
2187 | 4111 int w= s->avctx->width; //FIXME round up to x16 ? |
4112 int h= s->avctx->height; | |
2138 | 4113 |
2187 | 4114 if(s->current_picture.data[0]){ |
6437 | 4115 s->dsp.draw_edges(s->current_picture.data[0], s->current_picture.linesize[0], w , h , EDGE_WIDTH ); |
4116 s->dsp.draw_edges(s->current_picture.data[1], s->current_picture.linesize[1], w>>1, h>>1, EDGE_WIDTH/2); | |
4117 s->dsp.draw_edges(s->current_picture.data[2], s->current_picture.linesize[2], w>>1, h>>1, EDGE_WIDTH/2); | |
2187 | 4118 } |
4119 | |
3314 | 4120 tmp= s->last_picture[s->max_ref_frames-1]; |
4121 memmove(s->last_picture+1, s->last_picture, (s->max_ref_frames-1)*sizeof(AVFrame)); | |
5633
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4122 memmove(s->halfpel_plane+1, s->halfpel_plane, (s->max_ref_frames-1)*sizeof(void*)*4*4); |
5924 | 4123 if(USE_HALFPEL_PLANE && s->current_picture.data[0]) |
5633
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4124 halfpel_interpol(s, s->halfpel_plane[0], &s->current_picture); |
3314 | 4125 s->last_picture[0]= s->current_picture; |
2138 | 4126 s->current_picture= tmp; |
2967 | 4127 |
3314 | 4128 if(s->keyframe){ |
4129 s->ref_frames= 0; | |
4130 }else{ | |
4131 int i; | |
4132 for(i=0; i<s->max_ref_frames && s->last_picture[i].data[0]; i++) | |
4133 if(i && s->last_picture[i-1].key_frame) | |
4134 break; | |
4135 s->ref_frames= i; | |
4136 } | |
4137 | |
2138 | 4138 s->current_picture.reference= 1; |
4139 if(s->avctx->get_buffer(s->avctx, &s->current_picture) < 0){ | |
4140 av_log(s->avctx, AV_LOG_ERROR, "get_buffer() failed\n"); | |
4141 return -1; | |
4142 } | |
2967 | 4143 |
3314 | 4144 s->current_picture.key_frame= s->keyframe; |
4145 | |
2138 | 4146 return 0; |
4147 } | |
4148 | |
4149 static int encode_frame(AVCodecContext *avctx, unsigned char *buf, int buf_size, void *data){ | |
4150 SnowContext *s = avctx->priv_data; | |
2335 | 4151 RangeCoder * const c= &s->c; |
2138 | 4152 AVFrame *pict = data; |
4153 const int width= s->avctx->width; | |
4154 const int height= s->avctx->height; | |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4155 int level, orientation, plane_index, i, y; |
3661
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
4156 uint8_t rc_header_bak[sizeof(s->header_state)]; |
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
4157 uint8_t rc_block_bak[sizeof(s->block_state)]; |
2138 | 4158 |
2335 | 4159 ff_init_range_encoder(c, buf, buf_size); |
4160 ff_build_rac_states(c, 0.05*(1LL<<32), 256-8); | |
2967 | 4161 |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4162 for(i=0; i<3; i++){ |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4163 int shift= !!i; |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4164 for(y=0; y<(height>>shift); y++) |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4165 memcpy(&s->input_picture.data[i][y * s->input_picture.linesize[i]], |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4166 &pict->data[i][y * pict->linesize[i]], |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4167 width>>shift); |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4168 } |
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4169 s->new_picture = *pict; |
2138 | 4170 |
3313 | 4171 s->m.picture_number= avctx->frame_number; |
2608
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4172 if(avctx->flags&CODEC_FLAG_PASS2){ |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4173 s->m.pict_type = |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4174 pict->pict_type= s->m.rc_context.entry[avctx->frame_number].new_pict_type; |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4175 s->keyframe= pict->pict_type==FF_I_TYPE; |
3766 | 4176 if(!(avctx->flags&CODEC_FLAG_QSCALE)) { |
3193 | 4177 pict->quality= ff_rate_estimate_qscale(&s->m, 0); |
3766 | 4178 if (pict->quality < 0) |
4179 return -1; | |
4180 } | |
2608
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4181 }else{ |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4182 s->keyframe= avctx->gop_size==0 || avctx->frame_number % avctx->gop_size == 0; |
3313 | 4183 s->m.pict_type= |
2608
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4184 pict->pict_type= s->keyframe ? FF_I_TYPE : FF_P_TYPE; |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4185 } |
2967 | 4186 |
3313 | 4187 if(s->pass1_rc && avctx->frame_number == 0) |
4188 pict->quality= 2*FF_QP2LAMBDA; | |
2161 | 4189 if(pict->quality){ |
3322
0b4f548dfb44
Snow: 10l. 1pass ratecontrol failed to set snow's internal quant.
lorenm
parents:
3314
diff
changeset
|
4190 s->qlog= qscale2qlog(pict->quality); |
3313 | 4191 s->lambda = pict->quality * 3/2; |
4192 } | |
4193 if(s->qlog < 0 || (!pict->quality && (avctx->flags & CODEC_FLAG_QSCALE))){ | |
2161 | 4194 s->qlog= LOSSLESS_QLOG; |
3313 | 4195 s->lambda = 0; |
6412 | 4196 }//else keep previous frame's qlog until after motion estimation |
2138 | 4197 |
4198 frame_start(s); | |
4199 | |
2608
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4200 s->m.current_picture_ptr= &s->m.current_picture; |
6481 | 4201 if(pict->pict_type == FF_P_TYPE){ |
2138 | 4202 int block_width = (width +15)>>4; |
4203 int block_height= (height+15)>>4; | |
4204 int stride= s->current_picture.linesize[0]; | |
2967 | 4205 |
2138 | 4206 assert(s->current_picture.data[0]); |
3314 | 4207 assert(s->last_picture[0].data[0]); |
2967 | 4208 |
2138 | 4209 s->m.avctx= s->avctx; |
4210 s->m.current_picture.data[0]= s->current_picture.data[0]; | |
3314 | 4211 s->m. last_picture.data[0]= s->last_picture[0].data[0]; |
2138 | 4212 s->m. new_picture.data[0]= s-> input_picture.data[0]; |
4213 s->m. last_picture_ptr= &s->m. last_picture; | |
4214 s->m.linesize= | |
4215 s->m. last_picture.linesize[0]= | |
4216 s->m. new_picture.linesize[0]= | |
4217 s->m.current_picture.linesize[0]= stride; | |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
4218 s->m.uvlinesize= s->current_picture.linesize[1]; |
2138 | 4219 s->m.width = width; |
4220 s->m.height= height; | |
4221 s->m.mb_width = block_width; | |
4222 s->m.mb_height= block_height; | |
4223 s->m.mb_stride= s->m.mb_width+1; | |
4224 s->m.b8_stride= 2*s->m.mb_width+1; | |
4225 s->m.f_code=1; | |
4226 s->m.pict_type= pict->pict_type; | |
4227 s->m.me_method= s->avctx->me_method; | |
4228 s->m.me.scene_change_score=0; | |
4229 s->m.flags= s->avctx->flags; | |
4230 s->m.quarter_sample= (s->avctx->flags & CODEC_FLAG_QPEL)!=0; | |
4231 s->m.out_format= FMT_H263; | |
4232 s->m.unrestricted_mv= 1; | |
4233 | |
3313 | 4234 s->m.lambda = s->lambda; |
2138 | 4235 s->m.qscale= (s->m.lambda*139 + FF_LAMBDA_SCALE*64) >> (FF_LAMBDA_SHIFT + 7); |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
4236 s->lambda2= s->m.lambda2= (s->m.lambda*s->m.lambda + FF_LAMBDA_SCALE/2) >> FF_LAMBDA_SHIFT; |
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
4237 |
2138 | 4238 s->m.dsp= s->dsp; //move |
4239 ff_init_me(&s->m); | |
2993
cb0e26759cca
iterative overlapped block based motion estimation for snow
michael
parents:
2979
diff
changeset
|
4240 s->dsp= s->m.dsp; |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
4241 } |
2967 | 4242 |
3661
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
4243 if(s->pass1_rc){ |
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
4244 memcpy(rc_header_bak, s->header_state, sizeof(s->header_state)); |
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
4245 memcpy(rc_block_bak, s->block_state, sizeof(s->block_state)); |
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
4246 } |
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
4247 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
4248 redo_frame: |
2967 | 4249 |
6481 | 4250 if(pict->pict_type == FF_I_TYPE) |
5670
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
4251 s->spatial_decomposition_count= 5; |
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
4252 else |
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
4253 s->spatial_decomposition_count= 5; |
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
4254 |
3313 | 4255 s->m.pict_type = pict->pict_type; |
6481 | 4256 s->qbias= pict->pict_type == FF_P_TYPE ? 2 : 0; |
2138 | 4257 |
5670
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
4258 common_init_after_header(avctx); |
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
4259 |
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
4260 if(s->last_spatial_decomposition_count != s->spatial_decomposition_count){ |
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
4261 for(plane_index=0; plane_index<3; plane_index++){ |
5909 | 4262 calculate_visual_weight(s, &s->plane[plane_index]); |
5670
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
4263 } |
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
4264 } |
2f3431b34f0f
make changing spatial_decomposition_count per frame work
michael
parents:
5668
diff
changeset
|
4265 |
2138 | 4266 encode_header(s); |
2608
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4267 s->m.misc_bits = 8*(s->c.bytestream - s->c.bytestream_start); |
3661
b4425339894b
fix snow 1pass rc: the qlog used and the one written to the bitstream weren't always the same.
lorenm
parents:
3556
diff
changeset
|
4268 encode_blocks(s, 1); |
2608
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4269 s->m.mv_bits = 8*(s->c.bytestream - s->c.bytestream_start) - s->m.misc_bits; |
2967 | 4270 |
2138 | 4271 for(plane_index=0; plane_index<3; plane_index++){ |
4272 Plane *p= &s->plane[plane_index]; | |
4273 int w= p->width; | |
4274 int h= p->height; | |
4275 int x, y; | |
2198 | 4276 // int bits= put_bits_count(&s->c.pb); |
2138 | 4277 |
6414 | 4278 if(!(avctx->flags2 & CODEC_FLAG2_MEMC_ONLY)){ |
4279 //FIXME optimize | |
4280 if(pict->data[plane_index]) //FIXME gray hack | |
4281 for(y=0; y<h; y++){ | |
4282 for(x=0; x<w; x++){ | |
4283 s->spatial_idwt_buffer[y*w + x]= pict->data[plane_index][y*pict->linesize[plane_index] + x]<<FRAC_BITS; | |
4284 } | |
4285 } | |
4286 predict_plane(s, s->spatial_idwt_buffer, plane_index, 0); | |
4287 | |
4288 if( plane_index==0 | |
6481 | 4289 && pict->pict_type == FF_P_TYPE |
6414 | 4290 && !(avctx->flags&CODEC_FLAG_PASS2) |
4291 && s->m.me.scene_change_score > s->avctx->scenechange_threshold){ | |
4292 ff_init_range_encoder(c, buf, buf_size); | |
4293 ff_build_rac_states(c, 0.05*(1LL<<32), 256-8); | |
4294 pict->pict_type= FF_I_TYPE; | |
4295 s->keyframe=1; | |
4296 s->current_picture.key_frame=1; | |
4297 goto redo_frame; | |
2138 | 4298 } |
6414 | 4299 |
4300 if(s->qlog == LOSSLESS_QLOG){ | |
4301 for(y=0; y<h; y++){ | |
4302 for(x=0; x<w; x++){ | |
4303 s->spatial_dwt_buffer[y*w + x]= (s->spatial_idwt_buffer[y*w + x] + (1<<(FRAC_BITS-1))-1)>>FRAC_BITS; | |
4304 } | |
2161 | 4305 } |
6414 | 4306 }else{ |
4307 for(y=0; y<h; y++){ | |
4308 for(x=0; x<w; x++){ | |
4309 s->spatial_dwt_buffer[y*w + x]=s->spatial_idwt_buffer[y*w + x]<<ENCODER_EXTRA_BITS; | |
4310 } | |
5575 | 4311 } |
4312 } | |
6414 | 4313 |
4314 /* if(QUANTIZE2) | |
4315 dwt_quantize(s, p, s->spatial_dwt_buffer, w, h, w, s->spatial_decomposition_type); | |
4316 else*/ | |
4317 ff_spatial_dwt(s->spatial_dwt_buffer, w, h, w, s->spatial_decomposition_type, s->spatial_decomposition_count); | |
4318 | |
4319 if(s->pass1_rc && plane_index==0){ | |
4320 int delta_qlog = ratecontrol_1pass(s, pict); | |
4321 if (delta_qlog <= INT_MIN) | |
4322 return -1; | |
4323 if(delta_qlog){ | |
4324 //reordering qlog in the bitstream would eliminate this reset | |
4325 ff_init_range_encoder(c, buf, buf_size); | |
4326 memcpy(s->header_state, rc_header_bak, sizeof(s->header_state)); | |
4327 memcpy(s->block_state, rc_block_bak, sizeof(s->block_state)); | |
4328 encode_header(s); | |
4329 encode_blocks(s, 0); | |
2161 | 4330 } |
4331 } | |
6414 | 4332 |
4333 for(level=0; level<s->spatial_decomposition_count; level++){ | |
4334 for(orientation=level ? 1 : 0; orientation<4; orientation++){ | |
4335 SubBand *b= &p->band[level][orientation]; | |
4336 | |
4337 if(!QUANTIZE2) | |
4338 quantize(s, b, b->ibuf, b->buf, b->stride, s->qbias); | |
4339 if(orientation==0) | |
6481 | 4340 decorrelate(s, b, b->ibuf, b->stride, pict->pict_type == FF_P_TYPE, 0); |
6414 | 4341 encode_subband(s, b, b->ibuf, b->parent ? b->parent->ibuf : NULL, b->stride, orientation); |
4342 assert(b->parent==NULL || b->parent->stride == b->stride*2); | |
4343 if(orientation==0) | |
4344 correlate(s, b, b->ibuf, b->stride, 1, 0); | |
4345 } | |
4346 } | |
4347 | |
4348 for(level=0; level<s->spatial_decomposition_count; level++){ | |
4349 for(orientation=level ? 1 : 0; orientation<4; orientation++){ | |
4350 SubBand *b= &p->band[level][orientation]; | |
4351 | |
4352 dequantize(s, b, b->ibuf, b->stride); | |
4353 } | |
4354 } | |
4355 | |
4356 ff_spatial_idwt(s->spatial_idwt_buffer, w, h, w, s->spatial_decomposition_type, s->spatial_decomposition_count); | |
4357 if(s->qlog == LOSSLESS_QLOG){ | |
4358 for(y=0; y<h; y++){ | |
4359 for(x=0; x<w; x++){ | |
4360 s->spatial_idwt_buffer[y*w + x]<<=FRAC_BITS; | |
4361 } | |
4362 } | |
4363 } | |
4364 predict_plane(s, s->spatial_idwt_buffer, plane_index, 1); | |
4365 }else{ | |
3338
937f14bb0f23
support doing motion estimation and compensation without any residual transform or coding
michael
parents:
3327
diff
changeset
|
4366 //ME/MC only |
6481 | 4367 if(pict->pict_type == FF_I_TYPE){ |
3338
937f14bb0f23
support doing motion estimation and compensation without any residual transform or coding
michael
parents:
3327
diff
changeset
|
4368 for(y=0; y<h; y++){ |
937f14bb0f23
support doing motion estimation and compensation without any residual transform or coding
michael
parents:
3327
diff
changeset
|
4369 for(x=0; x<w; x++){ |
937f14bb0f23
support doing motion estimation and compensation without any residual transform or coding
michael
parents:
3327
diff
changeset
|
4370 s->current_picture.data[plane_index][y*s->current_picture.linesize[plane_index] + x]= |
937f14bb0f23
support doing motion estimation and compensation without any residual transform or coding
michael
parents:
3327
diff
changeset
|
4371 pict->data[plane_index][y*pict->linesize[plane_index] + x]; |
937f14bb0f23
support doing motion estimation and compensation without any residual transform or coding
michael
parents:
3327
diff
changeset
|
4372 } |
937f14bb0f23
support doing motion estimation and compensation without any residual transform or coding
michael
parents:
3327
diff
changeset
|
4373 } |
937f14bb0f23
support doing motion estimation and compensation without any residual transform or coding
michael
parents:
3327
diff
changeset
|
4374 }else{ |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
4375 memset(s->spatial_idwt_buffer, 0, sizeof(IDWTELEM)*w*h); |
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
4376 predict_plane(s, s->spatial_idwt_buffer, plane_index, 1); |
3338
937f14bb0f23
support doing motion estimation and compensation without any residual transform or coding
michael
parents:
3327
diff
changeset
|
4377 } |
6414 | 4378 } |
2138 | 4379 if(s->avctx->flags&CODEC_FLAG_PSNR){ |
4380 int64_t error= 0; | |
2967 | 4381 |
6414 | 4382 if(pict->data[plane_index]) //FIXME gray hack |
4383 for(y=0; y<h; y++){ | |
4384 for(x=0; x<w; x++){ | |
4385 int d= s->current_picture.data[plane_index][y*s->current_picture.linesize[plane_index] + x] - pict->data[plane_index][y*pict->linesize[plane_index] + x]; | |
4386 error += d*d; | |
4387 } | |
2138 | 4388 } |
4389 s->avctx->error[plane_index] += error; | |
2232 | 4390 s->current_picture.error[plane_index] = error; |
2138 | 4391 } |
6414 | 4392 |
2138 | 4393 } |
4394 | |
6144 | 4395 update_last_header_values(s); |
4396 | |
5633
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4397 if(s->last_picture[s->max_ref_frames-1].data[0]){ |
3314 | 4398 avctx->release_buffer(avctx, &s->last_picture[s->max_ref_frames-1]); |
5633
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4399 for(i=0; i<9; i++) |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4400 if(s->halfpel_plane[s->max_ref_frames-1][1+i/3][i%3]) |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4401 av_free(s->halfpel_plane[s->max_ref_frames-1][1+i/3][i%3] - EDGE_WIDTH*(1+s->current_picture.linesize[i%3])); |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4402 } |
2138 | 4403 |
2608
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4404 s->current_picture.coded_picture_number = avctx->frame_number; |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4405 s->current_picture.pict_type = pict->pict_type; |
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4406 s->current_picture.quality = pict->quality; |
3313 | 4407 s->m.frame_bits = 8*(s->c.bytestream - s->c.bytestream_start); |
4408 s->m.p_tex_bits = s->m.frame_bits - s->m.misc_bits - s->m.mv_bits; | |
4409 s->m.current_picture.display_picture_number = | |
4410 s->m.current_picture.coded_picture_number = avctx->frame_number; | |
4411 s->m.current_picture.quality = pict->quality; | |
4412 s->m.total_bits += 8*(s->c.bytestream - s->c.bytestream_start); | |
4413 if(s->pass1_rc) | |
3766 | 4414 if (ff_rate_estimate_qscale(&s->m, 0) < 0) |
4415 return -1; | |
3313 | 4416 if(avctx->flags&CODEC_FLAG_PASS1) |
2608
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4417 ff_write_pass1_stats(&s->m); |
3313 | 4418 s->m.last_pict_type = s->m.pict_type; |
4123
5c86acb39889
outputs bit spent on various encoding functions (motion vectors, overhead, etc)
gpoirier
parents:
4122
diff
changeset
|
4419 avctx->frame_bits = s->m.frame_bits; |
5c86acb39889
outputs bit spent on various encoding functions (motion vectors, overhead, etc)
gpoirier
parents:
4122
diff
changeset
|
4420 avctx->mv_bits = s->m.mv_bits; |
5c86acb39889
outputs bit spent on various encoding functions (motion vectors, overhead, etc)
gpoirier
parents:
4122
diff
changeset
|
4421 avctx->misc_bits = s->m.misc_bits; |
5c86acb39889
outputs bit spent on various encoding functions (motion vectors, overhead, etc)
gpoirier
parents:
4122
diff
changeset
|
4422 avctx->p_tex_bits = s->m.p_tex_bits; |
2608
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4423 |
2138 | 4424 emms_c(); |
2967 | 4425 |
2335 | 4426 return ff_rac_terminate(c); |
2138 | 4427 } |
4428 | |
6517
48759bfbd073
Apply 'cold' attribute to init/uninit functions in libavcodec
zuxy
parents:
6481
diff
changeset
|
4429 static av_cold void common_end(SnowContext *s){ |
3314 | 4430 int plane_index, level, orientation, i; |
2192 | 4431 |
2138 | 4432 av_freep(&s->spatial_dwt_buffer); |
5587
3ae03eacbe9f
use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
michael
parents:
5575
diff
changeset
|
4433 av_freep(&s->spatial_idwt_buffer); |
2138 | 4434 |
2967 | 4435 av_freep(&s->m.me.scratchpad); |
2138 | 4436 av_freep(&s->m.me.map); |
4437 av_freep(&s->m.me.score_map); | |
3033
e8599ab02b38
faster iterative_me: avoid duplicate mc of neighboring blocks.
lorenm
parents:
3020
diff
changeset
|
4438 av_freep(&s->m.obmc_scratchpad); |
2967 | 4439 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
4440 av_freep(&s->block); |
2192 | 4441 |
3314 | 4442 for(i=0; i<MAX_REF_FRAMES; i++){ |
4443 av_freep(&s->ref_mvs[i]); | |
4444 av_freep(&s->ref_scores[i]); | |
4445 if(s->last_picture[i].data[0]) | |
4446 s->avctx->release_buffer(s->avctx, &s->last_picture[i]); | |
4447 } | |
4448 | |
2967 | 4449 for(plane_index=0; plane_index<3; plane_index++){ |
2192 | 4450 for(level=s->spatial_decomposition_count-1; level>=0; level--){ |
4451 for(orientation=level ? 1 : 0; orientation<4; orientation++){ | |
4452 SubBand *b= &s->plane[plane_index].band[level][orientation]; | |
2967 | 4453 |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
4454 av_freep(&b->x_coeff); |
2192 | 4455 } |
4456 } | |
4457 } | |
2138 | 4458 } |
4459 | |
6517
48759bfbd073
Apply 'cold' attribute to init/uninit functions in libavcodec
zuxy
parents:
6481
diff
changeset
|
4460 static av_cold int encode_end(AVCodecContext *avctx) |
2138 | 4461 { |
4462 SnowContext *s = avctx->priv_data; | |
4463 | |
4464 common_end(s); | |
2608
4fb7fa34050b
allow 2pass ratecontrol. also fixes psnr displayed by mencoder.
lorenm
parents:
2607
diff
changeset
|
4465 av_free(avctx->stats_out); |
2138 | 4466 |
4467 return 0; | |
4468 } | |
4469 | |
6517
48759bfbd073
Apply 'cold' attribute to init/uninit functions in libavcodec
zuxy
parents:
6481
diff
changeset
|
4470 static av_cold int decode_init(AVCodecContext *avctx) |
2138 | 4471 { |
2635 | 4472 avctx->pix_fmt= PIX_FMT_YUV420P; |
2138 | 4473 |
4474 common_init(avctx); | |
2967 | 4475 |
2138 | 4476 return 0; |
4477 } | |
4478 | |
6255 | 4479 static int decode_frame(AVCodecContext *avctx, void *data, int *data_size, const uint8_t *buf, int buf_size){ |
2138 | 4480 SnowContext *s = avctx->priv_data; |
2335 | 4481 RangeCoder * const c= &s->c; |
2138 | 4482 int bytes_read; |
4483 AVFrame *picture = data; | |
5633
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4484 int level, orientation, plane_index, i; |
2138 | 4485 |
2335 | 4486 ff_init_range_decoder(c, buf, buf_size); |
4487 ff_build_rac_states(c, 0.05*(1LL<<32), 256-8); | |
2138 | 4488 |
4489 s->current_picture.pict_type= FF_I_TYPE; //FIXME I vs. P | |
5668 | 4490 if(decode_header(s)<0) |
4491 return -1; | |
5665
834f359e7257
perform init after reading the values needed for init
michael
parents:
5664
diff
changeset
|
4492 common_init_after_header(avctx); |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
4493 |
5662
77297fe0311c
allocate slice buffer after parsing the header containing needed parameters
michael
parents:
5661
diff
changeset
|
4494 // realloc slice buffer for the case that spatial_decomposition_count changed |
77297fe0311c
allocate slice buffer after parsing the header containing needed parameters
michael
parents:
5661
diff
changeset
|
4495 slice_buffer_destroy(&s->sb); |
77297fe0311c
allocate slice buffer after parsing the header containing needed parameters
michael
parents:
5661
diff
changeset
|
4496 slice_buffer_init(&s->sb, s->plane[0].height, (MB_SIZE >> s->block_max_depth) + s->spatial_decomposition_count * 8 + 1, s->plane[0].width, s->spatial_idwt_buffer); |
77297fe0311c
allocate slice buffer after parsing the header containing needed parameters
michael
parents:
5661
diff
changeset
|
4497 |
5651
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
4498 for(plane_index=0; plane_index<3; plane_index++){ |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
4499 Plane *p= &s->plane[plane_index]; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
4500 p->fast_mc= p->diag_mc && p->htaps==6 && p->hcoeff[0]==40 |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
4501 && p->hcoeff[1]==-10 |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
4502 && p->hcoeff[2]==2; |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
4503 } |
ab023c9f03d0
store halfpel filter coefficients in the header as well as the
michael
parents:
5650
diff
changeset
|
4504 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
4505 if(!s->block) alloc_blocks(s); |
2138 | 4506 |
4507 frame_start(s); | |
6412 | 4508 //keyframe flag duplication mess FIXME |
2138 | 4509 if(avctx->debug&FF_DEBUG_PICT_INFO) |
4510 av_log(avctx, AV_LOG_ERROR, "keyframe:%d qlog:%d\n", s->keyframe, s->qlog); | |
2967 | 4511 |
2189
70b27300a496
quad tree based motion compensation (currently only 16x16 & 8x8 OBMC blocks, but can be extended to other block sizes easily)
michael
parents:
2187
diff
changeset
|
4512 decode_blocks(s); |
2138 | 4513 |
4514 for(plane_index=0; plane_index<3; plane_index++){ | |
4515 Plane *p= &s->plane[plane_index]; | |
4516 int w= p->width; | |
4517 int h= p->height; | |
4518 int x, y; | |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
4519 int decode_state[MAX_DECOMPOSITIONS][4][1]; /* Stored state info for unpack_coeffs. 1 variable per instance. */ |
2967 | 4520 |
6414 | 4521 if(s->avctx->debug&2048){ |
4522 memset(s->spatial_dwt_buffer, 0, sizeof(DWTELEM)*w*h); | |
4523 predict_plane(s, s->spatial_idwt_buffer, plane_index, 1); | |
4524 | |
4525 for(y=0; y<h; y++){ | |
4526 for(x=0; x<w; x++){ | |
4527 int v= s->current_picture.data[plane_index][y*s->current_picture.linesize[plane_index] + x]; | |
4528 s->mconly_picture.data[plane_index][y*s->mconly_picture.linesize[plane_index] + x]= v; | |
4529 } | |
2138 | 4530 } |
4531 } | |
6414 | 4532 |
4533 { | |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
4534 for(level=0; level<s->spatial_decomposition_count; level++){ |
2634
337217ecbb3e
tighter snow slicing patch by (Yartrebo || yartrebo earthlink net)
michael
parents:
2628
diff
changeset
|
4535 for(orientation=level ? 1 : 0; orientation<4; orientation++){ |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
4536 SubBand *b= &p->band[level][orientation]; |
6414 | 4537 unpack_coeffs(s, b, b->parent, orientation); |
4538 } | |
4539 } | |
4540 } | |
4541 | |
4542 { | |
4543 const int mb_h= s->b_height << s->block_max_depth; | |
4544 const int block_size = MB_SIZE >> s->block_max_depth; | |
4545 const int block_w = plane_index ? block_size/2 : block_size; | |
4546 int mb_y; | |
4547 dwt_compose_t cs[MAX_DECOMPOSITIONS]; | |
4548 int yd=0, yq=0; | |
4549 int y; | |
4550 int end_y; | |
4551 | |
4552 ff_spatial_idwt_buffered_init(cs, &s->sb, w, h, 1, s->spatial_decomposition_type, s->spatial_decomposition_count); | |
4553 for(mb_y=0; mb_y<=mb_h; mb_y++){ | |
4554 | |
4555 int slice_starty = block_w*mb_y; | |
4556 int slice_h = block_w*(mb_y+1); | |
4557 if (!(s->keyframe || s->avctx->debug&512)){ | |
4558 slice_starty = FFMAX(0, slice_starty - (block_w >> 1)); | |
4559 slice_h -= (block_w >> 1); | |
4560 } | |
4561 | |
4562 for(level=0; level<s->spatial_decomposition_count; level++){ | |
4563 for(orientation=level ? 1 : 0; orientation<4; orientation++){ | |
4564 SubBand *b= &p->band[level][orientation]; | |
4565 int start_y; | |
4566 int end_y; | |
4567 int our_mb_start = mb_y; | |
4568 int our_mb_end = (mb_y + 1); | |
4569 const int extra= 3; | |
4570 start_y = (mb_y ? ((block_w * our_mb_start) >> (s->spatial_decomposition_count - level)) + s->spatial_decomposition_count - level + extra: 0); | |
4571 end_y = (((block_w * our_mb_end) >> (s->spatial_decomposition_count - level)) + s->spatial_decomposition_count - level + extra); | |
4572 if (!(s->keyframe || s->avctx->debug&512)){ | |
4573 start_y = FFMAX(0, start_y - (block_w >> (1+s->spatial_decomposition_count - level))); | |
4574 end_y = FFMAX(0, end_y - (block_w >> (1+s->spatial_decomposition_count - level))); | |
2634
337217ecbb3e
tighter snow slicing patch by (Yartrebo || yartrebo earthlink net)
michael
parents:
2628
diff
changeset
|
4575 } |
6414 | 4576 start_y = FFMIN(b->height, start_y); |
4577 end_y = FFMIN(b->height, end_y); | |
4578 | |
4579 if (start_y != end_y){ | |
4580 if (orientation == 0){ | |
4581 SubBand * correlate_band = &p->band[0][0]; | |
4582 int correlate_end_y = FFMIN(b->height, end_y + 1); | |
4583 int correlate_start_y = FFMIN(b->height, (start_y ? start_y + 1 : 0)); | |
4584 decode_subband_slice_buffered(s, correlate_band, &s->sb, correlate_start_y, correlate_end_y, decode_state[0][0]); | |
4585 correlate_slice_buffered(s, &s->sb, correlate_band, correlate_band->ibuf, correlate_band->stride, 1, 0, correlate_start_y, correlate_end_y); | |
4586 dequantize_slice_buffered(s, &s->sb, correlate_band, correlate_band->ibuf, correlate_band->stride, start_y, end_y); | |
4587 } | |
4588 else | |
4589 decode_subband_slice_buffered(s, b, &s->sb, start_y, end_y, decode_state[level][orientation]); | |
4590 } | |
2634
337217ecbb3e
tighter snow slicing patch by (Yartrebo || yartrebo earthlink net)
michael
parents:
2628
diff
changeset
|
4591 } |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
4592 } |
6414 | 4593 |
4594 for(; yd<slice_h; yd+=4){ | |
4595 ff_spatial_idwt_buffered_slice(&s->dsp, cs, &s->sb, w, h, 1, s->spatial_decomposition_type, s->spatial_decomposition_count, yd); | |
4596 } | |
4597 | |
4598 if(s->qlog == LOSSLESS_QLOG){ | |
4599 for(; yq<slice_h && yq<h; yq++){ | |
4600 IDWTELEM * line = slice_buffer_get_line(&s->sb, yq); | |
4601 for(x=0; x<w; x++){ | |
4602 line[x] <<= FRAC_BITS; | |
4603 } | |
2161 | 4604 } |
4605 } | |
6414 | 4606 |
4607 predict_slice_buffered(s, &s->sb, s->spatial_idwt_buffer, plane_index, 1, mb_y); | |
4608 | |
4609 y = FFMIN(p->height, slice_starty); | |
4610 end_y = FFMIN(p->height, slice_h); | |
4611 while(y < end_y) | |
4612 slice_buffer_release(&s->sb, y++); | |
2161 | 4613 } |
2562 | 4614 |
6414 | 4615 slice_buffer_flush(&s->sb); |
4616 } | |
4617 | |
2138 | 4618 } |
2967 | 4619 |
2138 | 4620 emms_c(); |
4621 | |
5633
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4622 if(s->last_picture[s->max_ref_frames-1].data[0]){ |
3314 | 4623 avctx->release_buffer(avctx, &s->last_picture[s->max_ref_frames-1]); |
5633
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4624 for(i=0; i<9; i++) |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4625 if(s->halfpel_plane[s->max_ref_frames-1][1+i/3][i%3]) |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4626 av_free(s->halfpel_plane[s->max_ref_frames-1][1+i/3][i%3] - EDGE_WIDTH*(1+s->current_picture.linesize[i%3])); |
873ea64637d9
code to do halfpel interpolation per frame (unfinished and under ifdef but it
michael
parents:
5627
diff
changeset
|
4627 } |
2138 | 4628 |
6414 | 4629 if(!(s->avctx->debug&2048)) |
4630 *picture= s->current_picture; | |
4631 else | |
4632 *picture= s->mconly_picture; | |
2967 | 4633 |
2138 | 4634 *data_size = sizeof(AVFrame); |
2967 | 4635 |
2335 | 4636 bytes_read= c->bytestream - c->bytestream_start; |
4637 if(bytes_read ==0) av_log(s->avctx, AV_LOG_ERROR, "error at end of frame\n"); //FIXME | |
2138 | 4638 |
4639 return bytes_read; | |
4640 } | |
4641 | |
6517
48759bfbd073
Apply 'cold' attribute to init/uninit functions in libavcodec
zuxy
parents:
6481
diff
changeset
|
4642 static av_cold int decode_end(AVCodecContext *avctx) |
2138 | 4643 { |
4644 SnowContext *s = avctx->priv_data; | |
4645 | |
2589
a5a62827f195
Snow Slicing patch by (Yartrebo) yartrebo earthlink net
michael
parents:
2562
diff
changeset
|
4646 slice_buffer_destroy(&s->sb); |
2967 | 4647 |
2138 | 4648 common_end(s); |
4649 | |
4650 return 0; | |
4651 } | |
4652 | |
4653 AVCodec snow_decoder = { | |
4654 "snow", | |
4655 CODEC_TYPE_VIDEO, | |
4656 CODEC_ID_SNOW, | |
4657 sizeof(SnowContext), | |
4658 decode_init, | |
4659 NULL, | |
4660 decode_end, | |
4661 decode_frame, | |
4662 0 /*CODEC_CAP_DR1*/ /*| CODEC_CAP_DRAW_HORIZ_BAND*/, | |
6712 | 4663 NULL, |
7040
e943e1409077
Make AVCodec long_names definition conditional depending on CONFIG_SMALL.
stefano
parents:
6712
diff
changeset
|
4664 .long_name = NULL_IF_CONFIG_SMALL("Snow"), |
2138 | 4665 }; |
4666 | |
5224 | 4667 #ifdef CONFIG_SNOW_ENCODER |
2138 | 4668 AVCodec snow_encoder = { |
4669 "snow", | |
4670 CODEC_TYPE_VIDEO, | |
4671 CODEC_ID_SNOW, | |
4672 sizeof(SnowContext), | |
4673 encode_init, | |
4674 encode_frame, | |
4675 encode_end, | |
7040
e943e1409077
Make AVCodec long_names definition conditional depending on CONFIG_SMALL.
stefano
parents:
6712
diff
changeset
|
4676 .long_name = NULL_IF_CONFIG_SMALL("Snow"), |
2138 | 4677 }; |
2408
a6e4da1c28ee
Disable encoders patch by (Gianluigi Tiesi <mplayer netfarm it>)
michael
parents:
2368
diff
changeset
|
4678 #endif |
2138 | 4679 |
4680 | |
6164
ecaf5226e9b0
Consistently use TEST as the preprocessor condition to enable test code.
diego
parents:
6144
diff
changeset
|
4681 #ifdef TEST |
2138 | 4682 #undef malloc |
4683 #undef free | |
4684 #undef printf | |
5380
389366aa3458
Fix the self tests which are contained in some codecs and are using random().
takis
parents:
5254
diff
changeset
|
4685 #undef random |
2138 | 4686 |
5934 | 4687 int main(void){ |
2138 | 4688 int width=256; |
4689 int height=256; | |
4690 int buffer[2][width*height]; | |
4691 SnowContext s; | |
4692 int i; | |
4693 s.spatial_decomposition_count=6; | |
4694 s.spatial_decomposition_type=1; | |
2967 | 4695 |
2138 | 4696 printf("testing 5/3 DWT\n"); |
4697 for(i=0; i<width*height; i++) | |
4698 buffer[0][i]= buffer[1][i]= random()%54321 - 12345; | |
2967 | 4699 |
2951 | 4700 ff_spatial_dwt(buffer[0], width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count); |
4701 ff_spatial_idwt(buffer[0], width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count); | |
2967 | 4702 |
2138 | 4703 for(i=0; i<width*height; i++) |
4704 if(buffer[0][i]!= buffer[1][i]) printf("fsck: %d %d %d\n",i, buffer[0][i], buffer[1][i]); | |
4705 | |
4706 printf("testing 9/7 DWT\n"); | |
4707 s.spatial_decomposition_type=0; | |
4708 for(i=0; i<width*height; i++) | |
4709 buffer[0][i]= buffer[1][i]= random()%54321 - 12345; | |
2967 | 4710 |
2951 | 4711 ff_spatial_dwt(buffer[0], width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count); |
4712 ff_spatial_idwt(buffer[0], width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count); | |
2967 | 4713 |
2138 | 4714 for(i=0; i<width*height; i++) |
4001 | 4715 if(FFABS(buffer[0][i] - buffer[1][i])>20) printf("fsck: %d %d %d\n",i, buffer[0][i], buffer[1][i]); |
2967 | 4716 |
2951 | 4717 #if 0 |
2138 | 4718 printf("testing AC coder\n"); |
4719 memset(s.header_state, 0, sizeof(s.header_state)); | |
2335 | 4720 ff_init_range_encoder(&s.c, buffer[0], 256*256); |
2138 | 4721 ff_init_cabac_states(&s.c, ff_h264_lps_range, ff_h264_mps_state, ff_h264_lps_state, 64); |
2967 | 4722 |
2138 | 4723 for(i=-256; i<256; i++){ |
4001 | 4724 put_symbol(&s.c, s.header_state, i*i*i/3*FFABS(i), 1); |
2138 | 4725 } |
2335 | 4726 ff_rac_terminate(&s.c); |
2138 | 4727 |
4728 memset(s.header_state, 0, sizeof(s.header_state)); | |
2335 | 4729 ff_init_range_decoder(&s.c, buffer[0], 256*256); |
2138 | 4730 ff_init_cabac_states(&s.c, ff_h264_lps_range, ff_h264_mps_state, ff_h264_lps_state, 64); |
2967 | 4731 |
2138 | 4732 for(i=-256; i<256; i++){ |
4733 int j; | |
4734 j= get_symbol(&s.c, s.header_state, 1); | |
4001 | 4735 if(j!=i*i*i/3*FFABS(i)) printf("fsck: %d != %d\n", i, j); |
2138 | 4736 } |
2951 | 4737 #endif |
6414 | 4738 { |
4739 int level, orientation, x, y; | |
4740 int64_t errors[8][4]; | |
4741 int64_t g=0; | |
4742 | |
4743 memset(errors, 0, sizeof(errors)); | |
4744 s.spatial_decomposition_count=3; | |
4745 s.spatial_decomposition_type=0; | |
4746 for(level=0; level<s.spatial_decomposition_count; level++){ | |
4747 for(orientation=level ? 1 : 0; orientation<4; orientation++){ | |
4748 int w= width >> (s.spatial_decomposition_count-level); | |
4749 int h= height >> (s.spatial_decomposition_count-level); | |
4750 int stride= width << (s.spatial_decomposition_count-level); | |
4751 DWTELEM *buf= buffer[0]; | |
4752 int64_t error=0; | |
4753 | |
4754 if(orientation&1) buf+=w; | |
4755 if(orientation>1) buf+=stride>>1; | |
4756 | |
4757 memset(buffer[0], 0, sizeof(int)*width*height); | |
4758 buf[w/2 + h/2*stride]= 256*256; | |
4759 ff_spatial_idwt(buffer[0], width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count); | |
4760 for(y=0; y<height; y++){ | |
4761 for(x=0; x<width; x++){ | |
4762 int64_t d= buffer[0][x + y*width]; | |
4763 error += d*d; | |
4764 if(FFABS(width/2-x)<9 && FFABS(height/2-y)<9 && level==2) printf("%8"PRId64" ", d); | |
4765 } | |
4766 if(FFABS(height/2-y)<9 && level==2) printf("\n"); | |
2138 | 4767 } |
6414 | 4768 error= (int)(sqrt(error)+0.5); |
4769 errors[level][orientation]= error; | |
4770 if(g) g=ff_gcd(g, error); | |
4771 else g= error; | |
2138 | 4772 } |
4773 } | |
6414 | 4774 printf("static int const visual_weight[][4]={\n"); |
4775 for(level=0; level<s.spatial_decomposition_count; level++){ | |
4776 printf(" {"); | |
4777 for(orientation=0; orientation<4; orientation++){ | |
4778 printf("%8"PRId64",", errors[level][orientation]/g); | |
4779 } | |
4780 printf("},\n"); | |
2138 | 4781 } |
6414 | 4782 printf("};\n"); |
4783 { | |
2138 | 4784 int level=2; |
4785 int w= width >> (s.spatial_decomposition_count-level); | |
6168 | 4786 //int h= height >> (s.spatial_decomposition_count-level); |
2138 | 4787 int stride= width << (s.spatial_decomposition_count-level); |
4788 DWTELEM *buf= buffer[0]; | |
4789 int64_t error=0; | |
4790 | |
4791 buf+=w; | |
4792 buf+=stride>>1; | |
2967 | 4793 |
2138 | 4794 memset(buffer[0], 0, sizeof(int)*width*height); |
4795 #if 1 | |
4796 for(y=0; y<height; y++){ | |
4797 for(x=0; x<width; x++){ | |
4798 int tab[4]={0,2,3,1}; | |
4799 buffer[0][x+width*y]= 256*256*tab[(x&1) + 2*(y&1)]; | |
4800 } | |
4801 } | |
2951 | 4802 ff_spatial_dwt(buffer[0], width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count); |
2138 | 4803 #else |
4804 for(y=0; y<h; y++){ | |
4805 for(x=0; x<w; x++){ | |
4806 buf[x + y*stride ]=169; | |
4807 buf[x + y*stride-w]=64; | |
4808 } | |
4809 } | |
2951 | 4810 ff_spatial_idwt(buffer[0], width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count); |
2138 | 4811 #endif |
4812 for(y=0; y<height; y++){ | |
4813 for(x=0; x<width; x++){ | |
4814 int64_t d= buffer[0][x + y*width]; | |
4815 error += d*d; | |
4122
daae66c03857
Replace most of the %lld and %llx by their (cleaner) PRI*64 counterparts.
diego
parents:
4011
diff
changeset
|
4816 if(FFABS(width/2-x)<9 && FFABS(height/2-y)<9) printf("%8"PRId64" ", d); |
2138 | 4817 } |
4001 | 4818 if(FFABS(height/2-y)<9) printf("\n"); |
2138 | 4819 } |
6414 | 4820 } |
4821 | |
2138 | 4822 } |
4823 return 0; | |
4824 } | |
6164
ecaf5226e9b0
Consistently use TEST as the preprocessor condition to enable test code.
diego
parents:
6144
diff
changeset
|
4825 #endif /* TEST */ |