FFmpeg coverage


Directory: ../../../ffmpeg/
File: src/libavcodec/proresdsp.c
Date: 2026-08-26 01:27:24
Exec Total Coverage
Lines: 51 75 68.0%
Functions: 8 11 72.7%
Branches: 20 32 62.5%

Line Branch Exec Source
1 /*
2 * Apple ProRes compatible decoder
3 *
4 * Copyright (c) 2010-2011 Maxim Poliakovski
5 *
6 * This file is part of FFmpeg.
7 *
8 * FFmpeg is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
12 *
13 * FFmpeg is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
17 *
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with FFmpeg; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21 */
22
23 #include "config.h"
24 #include "libavutil/attributes.h"
25 #include "libavutil/avassert.h"
26 #include "libavutil/common.h"
27 #include "libavutil/intreadwrite.h"
28 #include "idctdsp.h"
29 #include "proresdsp.h"
30
31 #define IN_IDCT_DEPTH 16
32 #define PRORES_ONLY
33
34 #define BIT_DEPTH 10
35 #define EXTRA_SHIFT
36 #include "simple_idct_template.c"
37 #undef BIT_DEPTH
38 #undef EXTRA_SHIFT
39
40 #define BIT_DEPTH 12
41 #include "simple_idct_template.c"
42 #undef BIT_DEPTH
43 #undef IN_IDCT_DEPTH
44
45 /* 32bit iDCT for the ProRes RAW */
46 #define IN_IDCT_DEPTH 32
47 #define BIT_DEPTH 12
48 #include "simple_idct_template.c"
49 #undef BIT_DEPTH
50 #undef IN_IDCT_DEPTH
51
52 /**
53 * Special version of ff_simple_idct_int16_10bit() which does dequantization
54 * and scales by a factor of 2 more between the two IDCTs to account
55 * for larger scale of input coefficients.
56 */
57 2189760 static void prores_idct_10(int16_t *restrict block, const int16_t *restrict qmat)
58 {
59
2/2
✓ Branch 0 taken 140144640 times.
✓ Branch 1 taken 2189760 times.
142334400 for (int i = 0; i < 64; i++)
60 140144640 block[i] *= qmat[i];
61
62
2/2
✓ Branch 0 taken 17518080 times.
✓ Branch 1 taken 2189760 times.
19707840 for (int i = 0; i < 8; i++)
63 17518080 idctRowCondDC_extrashift_10(block + i*8, 2);
64
65
2/2
✓ Branch 0 taken 17518080 times.
✓ Branch 1 taken 2189760 times.
19707840 for (int i = 0; i < 8; i++) {
66 17518080 block[i] += 8192;
67 17518080 idctSparseCol_extrashift_10(block + i);
68 }
69 2189760 }
70
71 2025720 static void prores_idct_12(int16_t *restrict block, const int16_t *restrict qmat)
72 {
73
2/2
✓ Branch 0 taken 129646080 times.
✓ Branch 1 taken 2025720 times.
131671800 for (int i = 0; i < 64; i++)
74 129646080 block[i] *= qmat[i];
75
76
2/2
✓ Branch 0 taken 16205760 times.
✓ Branch 1 taken 2025720 times.
18231480 for (int i = 0; i < 8; i++)
77 16205760 idctRowCondDC_int16_12bit(block + i*8, 0);
78
79
2/2
✓ Branch 0 taken 16205760 times.
✓ Branch 1 taken 2025720 times.
18231480 for (int i = 0; i < 8; i++) {
80 16205760 block[i] += 8192;
81 16205760 idctSparseCol_int16_12bit(block + i);
82 }
83 2025720 }
84
85 /*
86 * 32-bit iDCT for the ProRes RAW
87 * qmat must be s->qmat[i] * scale
88 */
89 static void prores_idct_bayer_32(int32_t *restrict block, const int16_t *restrict qmat)
90 {
91 for (int i = 0; i < 64; i++)
92 block[i] = (block[i] * qmat[i]) >> 1;
93
94 for (int i = 0; i < 8; i++)
95 idctRowCondDC_int32_12bit(block + i*8, 0);
96
97 for (int i = 0; i < 8; i++) {
98 block[i] += 8192;
99 idctSparseCol_int32_12bit(block + i);
100 }
101 }
102
103 #define CLIP_MIN (1 << 2) ///< minimum value for clipping resulting pixels
104 #define CLIP_MAX_10 (1 << 10) - CLIP_MIN - 1 ///< maximum value for clipping resulting pixels
105 #define CLIP_MAX_12 (1 << 12) - CLIP_MIN - 1 ///< maximum value for clipping resulting pixels
106
107 #define CLIP_10(x) (av_clip((x), CLIP_MIN, CLIP_MAX_10))
108 #define CLIP_12(x) (av_clip((x), CLIP_MIN, CLIP_MAX_12))
109
110 /**
111 * Add bias value, clamp and output pixels of a slice
112 */
113
114 4155480 static inline void put_pixel(uint16_t *dst, ptrdiff_t linesize, const int16_t *in, int bits_per_raw_sample) {
115
2/2
✓ Branch 0 taken 33243840 times.
✓ Branch 1 taken 4155480 times.
37399320 for (int y = 0; y < 8; y++, dst += linesize) {
116
2/2
✓ Branch 0 taken 265950720 times.
✓ Branch 1 taken 33243840 times.
299194560 for (int x = 0; x < 8; x++) {
117 265950720 int src_offset = (y << 3) + x;
118
119
2/2
✓ Branch 0 taken 136304640 times.
✓ Branch 1 taken 129646080 times.
265950720 if (bits_per_raw_sample == 10) {
120 136304640 dst[x] = CLIP_10(in[src_offset]);
121 } else {//12b
122 129646080 dst[x] = CLIP_12(in[src_offset]);
123 }
124 }
125 }
126 4155480 }
127
128 /* Apply the 8-point combined linearization curve (inv. transfer fn + encoder shaping) */
129 static inline void put_pixel_bayer_lin_curve_12(uint16_t *dst, ptrdiff_t linesize,
130 const int32_t *in, const uint16_t *lin_curve)
131 {
132 for (int y = 0; y < 8; y++, dst += linesize) {
133 for (int x = 0; x < 8; x++) {
134 /* Convert the 32-bit input into 16-bits (lrintf(x*16 - 15.5f) = 16) */
135 int u = av_clip_uint16(in[(y << 3) + x]*16 - 16);
136 uint32_t seg = (uint32_t)u >> 13;
137 uint32_t frac = (uint32_t)u & 0x1FFF;
138 uint32_t cp0 = lin_curve[seg];
139 uint32_t cp1 = seg < 7 ? lin_curve[seg + 1] : 0;
140 uint32_t o = (cp0 * 8192 + ((cp1 - cp0) & 0xFFFF) * frac + 4096) >> 13;
141 dst[x*2] = FFMIN(o, 0xFFFF);
142 }
143 }
144 }
145
146 2129760 static void put_pixels_10(uint16_t *dst, ptrdiff_t linesize, const int16_t *in)
147 {
148 2129760 put_pixel(dst, linesize, in, 10);
149 2129760 }
150
151 2025720 static void put_pixels_12(uint16_t *dst, ptrdiff_t linesize, const int16_t *in)
152 {
153 2025720 put_pixel(dst, linesize, in, 12);
154 2025720 }
155
156 2129760 static void prores_idct_put_10_c(uint16_t *out, ptrdiff_t linesize, int16_t *block, const int16_t *qmat)
157 {
158 2129760 prores_idct_10(block, qmat);
159 2129760 put_pixels_10(out, linesize >> 1, block);
160 2129760 }
161
162 2025720 static void prores_idct_put_12_c(uint16_t *out, ptrdiff_t linesize, int16_t *block, const int16_t *qmat)
163 {
164 2025720 prores_idct_12(block, qmat);
165 2025720 put_pixels_12(out, linesize >> 1, block);
166 2025720 }
167
168 static void prores_idct_put_bayer_12_c(uint16_t *out, ptrdiff_t linesize,
169 int32_t *block, const int16_t *qmat,
170 const uint16_t *lin_curve)
171 {
172 prores_idct_bayer_32(block, qmat);
173 put_pixel_bayer_lin_curve_12(out, linesize << 1, block, lin_curve);
174 }
175
176 76 av_cold void ff_proresdsp_init(ProresDSPContext *dsp, int bits_per_raw_sample)
177 {
178
2/2
✓ Branch 0 taken 52 times.
✓ Branch 1 taken 24 times.
76 if (bits_per_raw_sample == 10) {
179 52 dsp->idct_put = prores_idct_put_10_c;
180 52 dsp->idct_permutation_type = FF_IDCT_PERM_NONE;
181 } else {
182 av_assert1(bits_per_raw_sample == 12);
183 24 dsp->idct_put = prores_idct_put_12_c;
184 24 dsp->idct_put_bayer = prores_idct_put_bayer_12_c;
185 24 dsp->idct_permutation_type = FF_IDCT_PERM_NONE;
186 }
187
188 #if ARCH_X86 && HAVE_X86ASM
189 76 ff_proresdsp_init_x86(dsp, bits_per_raw_sample);
190 #endif
191
192 76 ff_init_scantable_permutation(dsp->idct_permutation,
193 76 dsp->idct_permutation_type);
194 76 }
195