Mercurial > libavcodec.hg
annotate dctref.c @ 12266:48d6738904a9 libavcodec
Fix SPLATB_REG mess. Used to be a if/elseif/elseif/elseif spaghetti, so this
splits it into small optimization-specific macros which are selected for each
DSP function. The advantage of this approach is that the sse4 functions now
use the ssse3 codepath also without needing an explicit sse4 codepath.
author | rbultje |
---|---|
date | Sat, 24 Jul 2010 19:33:05 +0000 |
parents | 7dd2a45249a9 |
children |
rev | line source |
---|---|
9308 | 1 /* |
2 * reference discrete cosine transform (double precision) | |
3 * Copyright (C) 2009 Dylan Yudaken | |
4 * | |
5 * This file is part of FFmpeg. | |
6 * | |
7 * FFmpeg is free software; you can redistribute it and/or | |
8 * modify it under the terms of the GNU Lesser General Public | |
9 * License as published by the Free Software Foundation; either | |
10 * version 2.1 of the License, or (at your option) any later version. | |
11 * | |
12 * FFmpeg is distributed in the hope that it will be useful, | |
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
15 * Lesser General Public License for more details. | |
16 * | |
17 * You should have received a copy of the GNU Lesser General Public | |
18 * License along with FFmpeg; if not, write to the Free Software | |
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | |
20 */ | |
21 | |
22 /** | |
11644
7dd2a45249a9
Remove explicit filename from Doxygen @file commands.
diego
parents:
11408
diff
changeset
|
23 * @file |
9308 | 24 * reference discrete cosine transform (double precision) |
25 * | |
26 * @author Dylan Yudaken (dyudaken at gmail) | |
27 * | |
28 * @note This file could be optimized a lot, but is for | |
29 * reference and so readability is better. | |
30 */ | |
31 | |
32 #include "libavutil/mathematics.h" | |
11408 | 33 #include "dctref.h" |
34 | |
9308 | 35 static double coefficients[8 * 8]; |
36 | |
37 /** | |
38 * Initialize the double precision discrete cosine transform | |
39 * functions fdct & idct. | |
40 */ | |
41 av_cold void ff_ref_dct_init(void) | |
42 { | |
43 unsigned int i, j; | |
44 | |
45 for (j = 0; j < 8; ++j) { | |
46 coefficients[j] = sqrt(0.125); | |
47 for (i = 8; i < 64; i += 8) { | |
48 coefficients[i + j] = 0.5 * cos(i * (j + 0.5) * M_PI / 64.0); | |
49 } | |
50 } | |
51 } | |
52 | |
53 /** | |
54 * Transform 8x8 block of data with a double precision forward DCT <br> | |
55 * This is a reference implementation. | |
56 * | |
57 * @param block pointer to 8x8 block of data to transform | |
58 */ | |
59 void ff_ref_fdct(short *block) | |
60 { | |
61 /* implement the equation: block = coefficients * block * coefficients' */ | |
62 | |
63 unsigned int i, j, k; | |
64 double out[8 * 8]; | |
65 | |
66 /* out = coefficients * block */ | |
67 for (i = 0; i < 64; i += 8) { | |
68 for (j = 0; j < 8; ++j) { | |
69 double tmp = 0; | |
70 for (k = 0; k < 8; ++k) { | |
71 tmp += coefficients[i + k] * block[k * 8 + j]; | |
72 } | |
73 out[i + j] = tmp * 8; | |
74 } | |
75 } | |
76 | |
77 /* block = out * (coefficients') */ | |
78 for (j = 0; j < 8; ++j) { | |
79 for (i = 0; i < 64; i += 8) { | |
80 double tmp = 0; | |
81 for (k = 0; k < 8; ++k) { | |
82 tmp += out[i + k] * coefficients[j * 8 + k]; | |
83 } | |
84 block[i + j] = floor(tmp + 0.499999999999); | |
85 } | |
86 } | |
87 } | |
88 | |
89 /** | |
90 * Transform 8x8 block of data with a double precision inverse DCT <br> | |
91 * This is a reference implementation. | |
92 * | |
93 * @param block pointer to 8x8 block of data to transform | |
94 */ | |
95 void ff_ref_idct(short *block) | |
96 { | |
97 /* implement the equation: block = (coefficients') * block * coefficients */ | |
98 | |
99 unsigned int i, j, k; | |
100 double out[8 * 8]; | |
101 | |
102 /* out = block * coefficients */ | |
103 for (i = 0; i < 64; i += 8) { | |
104 for (j = 0; j < 8; ++j) { | |
105 double tmp = 0; | |
106 for (k = 0; k < 8; ++k) { | |
107 tmp += block[i + k] * coefficients[k * 8 + j]; | |
108 } | |
109 out[i + j] = tmp; | |
110 } | |
111 } | |
112 | |
113 /* block = (coefficients') * out */ | |
114 for (i = 0; i < 8; ++i) { | |
115 for (j = 0; j < 8; ++j) { | |
116 double tmp = 0; | |
117 for (k = 0; k < 64; k += 8) { | |
118 tmp += coefficients[k + i] * out[k + j]; | |
119 } | |
120 block[i * 8 + j] = floor(tmp + 0.5); | |
121 } | |
122 } | |
123 } |