Mercurial > libavcodec.hg
comparison dctref.c @ 9308:aad816bc3d54 libavcodec
Replacement reference DCT implementation.
patch by Dylan Yudaken, dyudaken gmail com
author | diego |
---|---|
date | Tue, 31 Mar 2009 15:48:47 +0000 |
parents | |
children | 4b6b3ffbaee3 |
comparison
equal
deleted
inserted
replaced
9307:851ca5db8327 | 9308:aad816bc3d54 |
---|---|
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 /** | |
23 * @file libavcodec/dctref.c | |
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" | |
33 static double coefficients[8 * 8]; | |
34 | |
35 /** | |
36 * Initialize the double precision discrete cosine transform | |
37 * functions fdct & idct. | |
38 */ | |
39 av_cold void ff_ref_dct_init(void) | |
40 { | |
41 unsigned int i, j; | |
42 | |
43 for (j = 0; j < 8; ++j) { | |
44 coefficients[j] = sqrt(0.125); | |
45 for (i = 8; i < 64; i += 8) { | |
46 coefficients[i + j] = 0.5 * cos(i * (j + 0.5) * M_PI / 64.0); | |
47 } | |
48 } | |
49 } | |
50 | |
51 /** | |
52 * Transform 8x8 block of data with a double precision forward DCT <br> | |
53 * This is a reference implementation. | |
54 * | |
55 * @param block pointer to 8x8 block of data to transform | |
56 */ | |
57 void ff_ref_fdct(short *block) | |
58 { | |
59 /* implement the equation: block = coefficients * block * coefficients' */ | |
60 | |
61 unsigned int i, j, k; | |
62 double out[8 * 8]; | |
63 | |
64 /* out = coefficients * block */ | |
65 for (i = 0; i < 64; i += 8) { | |
66 for (j = 0; j < 8; ++j) { | |
67 double tmp = 0; | |
68 for (k = 0; k < 8; ++k) { | |
69 tmp += coefficients[i + k] * block[k * 8 + j]; | |
70 } | |
71 out[i + j] = tmp * 8; | |
72 } | |
73 } | |
74 | |
75 /* block = out * (coefficients') */ | |
76 for (j = 0; j < 8; ++j) { | |
77 for (i = 0; i < 64; i += 8) { | |
78 double tmp = 0; | |
79 for (k = 0; k < 8; ++k) { | |
80 tmp += out[i + k] * coefficients[j * 8 + k]; | |
81 } | |
82 block[i + j] = floor(tmp + 0.499999999999); | |
83 } | |
84 } | |
85 } | |
86 | |
87 /** | |
88 * Transform 8x8 block of data with a double precision inverse DCT <br> | |
89 * This is a reference implementation. | |
90 * | |
91 * @param block pointer to 8x8 block of data to transform | |
92 */ | |
93 void ff_ref_idct(short *block) | |
94 { | |
95 /* implement the equation: block = (coefficients') * block * coefficients */ | |
96 | |
97 unsigned int i, j, k; | |
98 double out[8 * 8]; | |
99 | |
100 /* out = block * coefficients */ | |
101 for (i = 0; i < 64; i += 8) { | |
102 for (j = 0; j < 8; ++j) { | |
103 double tmp = 0; | |
104 for (k = 0; k < 8; ++k) { | |
105 tmp += block[i + k] * coefficients[k * 8 + j]; | |
106 } | |
107 out[i + j] = tmp; | |
108 } | |
109 } | |
110 | |
111 /* block = (coefficients') * out */ | |
112 for (i = 0; i < 8; ++i) { | |
113 for (j = 0; j < 8; ++j) { | |
114 double tmp = 0; | |
115 for (k = 0; k < 64; k += 8) { | |
116 tmp += coefficients[k + i] * out[k + j]; | |
117 } | |
118 block[i * 8 + j] = floor(tmp + 0.5); | |
119 } | |
120 } | |
121 } |