Line | Branch | Exec | Source |
---|---|---|---|
1 | /* | ||
2 | * This file is part of FFmpeg. | ||
3 | * | ||
4 | * FFmpeg is free software; you can redistribute it and/or | ||
5 | * modify it under the terms of the GNU Lesser General Public | ||
6 | * License as published by the Free Software Foundation; either | ||
7 | * version 2.1 of the License, or (at your option) any later version. | ||
8 | * | ||
9 | * FFmpeg is distributed in the hope that it will be useful, | ||
10 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
11 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | ||
12 | * Lesser General Public License for more details. | ||
13 | * | ||
14 | * You should have received a copy of the GNU Lesser General Public | ||
15 | * License along with FFmpeg; if not, write to the Free Software | ||
16 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | ||
17 | */ | ||
18 | |||
19 | #include <stdint.h> | ||
20 | |||
21 | #include "config.h" | ||
22 | #include "libavutil/attributes.h" | ||
23 | #include "libavutil/common.h" | ||
24 | |||
25 | #include "apv.h" | ||
26 | #include "apv_dsp.h" | ||
27 | |||
28 | |||
29 | static const int8_t apv_trans_matrix[8][8] = { | ||
30 | { 64, 64, 64, 64, 64, 64, 64, 64 }, | ||
31 | { 89, 75, 50, 18, -18, -50, -75, -89 }, | ||
32 | { 84, 35, -35, -84, -84, -35, 35, 84 }, | ||
33 | { 75, -18, -89, -50, 50, 89, 18, -75 }, | ||
34 | { 64, -64, -64, 64, 64, -64, -64, 64 }, | ||
35 | { 50, -89, 18, 75, -75, -18, 89, -50 }, | ||
36 | { 35, -84, 84, -35, -35, 84, -84, 35 }, | ||
37 | { 18, -50, 75, -89, 89, -75, 50, -18 }, | ||
38 | }; | ||
39 | |||
40 | 9606 | static void apv_decode_transquant_c(void *output, | |
41 | ptrdiff_t pitch, | ||
42 | const int16_t *input_flat, | ||
43 | const int16_t *qmatrix_flat, | ||
44 | int bit_depth, | ||
45 | int qp_shift) | ||
46 | { | ||
47 | 9606 | const int16_t (*input)[8] = (const int16_t(*)[8])input_flat; | |
48 | 9606 | const int16_t (*qmatrix)[8] = (const int16_t(*)[8])qmatrix_flat; | |
49 | |||
50 | int16_t scaled_coeff[8][8]; | ||
51 | int32_t recon_sample[8][8]; | ||
52 | |||
53 | // Dequant. | ||
54 | { | ||
55 | // Note that level_scale was already combined into qmatrix | ||
56 | // before we got here. | ||
57 | 9606 | int bd_shift = bit_depth + 3 - 5; | |
58 | |||
59 |
2/2✓ Branch 0 taken 76848 times.
✓ Branch 1 taken 9606 times.
|
86454 | for (int y = 0; y < 8; y++) { |
60 |
2/2✓ Branch 0 taken 614784 times.
✓ Branch 1 taken 76848 times.
|
691632 | for (int x = 0; x < 8; x++) { |
61 | 614784 | int coeff = (input[y][x] * qmatrix[y][x] * (1 << qp_shift) + | |
62 | 614784 | (1 << (bd_shift - 1))) >> bd_shift; | |
63 | |||
64 | 614784 | scaled_coeff[y][x] = | |
65 | 614784 | av_clip(coeff, APV_MIN_TRANS_COEFF, | |
66 | APV_MAX_TRANS_COEFF); | ||
67 | } | ||
68 | } | ||
69 | } | ||
70 | |||
71 | // Transform. | ||
72 | { | ||
73 | int32_t tmp[8][8]; | ||
74 | |||
75 | // Vertical transform of columns. | ||
76 |
2/2✓ Branch 0 taken 76848 times.
✓ Branch 1 taken 9606 times.
|
86454 | for (int x = 0; x < 8; x++) { |
77 |
2/2✓ Branch 0 taken 614784 times.
✓ Branch 1 taken 76848 times.
|
691632 | for (int i = 0; i < 8; i++) { |
78 | 614784 | int sum = 0; | |
79 |
2/2✓ Branch 0 taken 4918272 times.
✓ Branch 1 taken 614784 times.
|
5533056 | for (int j = 0; j < 8; j++) |
80 | 4918272 | sum += apv_trans_matrix[j][i] * scaled_coeff[j][x]; | |
81 | 614784 | tmp[i][x] = sum; | |
82 | } | ||
83 | } | ||
84 | |||
85 | // Renormalise. | ||
86 |
2/2✓ Branch 0 taken 76848 times.
✓ Branch 1 taken 9606 times.
|
86454 | for (int x = 0; x < 8; x++) { |
87 |
2/2✓ Branch 0 taken 614784 times.
✓ Branch 1 taken 76848 times.
|
691632 | for (int y = 0; y < 8; y++) |
88 | 614784 | tmp[y][x] = (tmp[y][x] + 64) >> 7; | |
89 | } | ||
90 | |||
91 | // Horizontal transform of rows. | ||
92 |
2/2✓ Branch 0 taken 76848 times.
✓ Branch 1 taken 9606 times.
|
86454 | for (int y = 0; y < 8; y++) { |
93 |
2/2✓ Branch 0 taken 614784 times.
✓ Branch 1 taken 76848 times.
|
691632 | for (int i = 0; i < 8; i++) { |
94 | 614784 | int sum = 0; | |
95 |
2/2✓ Branch 0 taken 4918272 times.
✓ Branch 1 taken 614784 times.
|
5533056 | for (int j = 0; j < 8; j++) |
96 | 4918272 | sum += apv_trans_matrix[j][i] * tmp[y][j]; | |
97 | 614784 | recon_sample[y][i] = sum; | |
98 | } | ||
99 | } | ||
100 | } | ||
101 | |||
102 | // Output. | ||
103 |
2/2✓ Branch 0 taken 3 times.
✓ Branch 1 taken 9603 times.
|
9606 | if (bit_depth == 8) { |
104 | 3 | uint8_t *ptr = output; | |
105 | 3 | int bd_shift = 20 - bit_depth; | |
106 | |||
107 |
2/2✓ Branch 0 taken 24 times.
✓ Branch 1 taken 3 times.
|
27 | for (int y = 0; y < 8; y++) { |
108 |
2/2✓ Branch 0 taken 192 times.
✓ Branch 1 taken 24 times.
|
216 | for (int x = 0; x < 8; x++) { |
109 | 192 | int sample = ((recon_sample[y][x] + | |
110 | 192 | (1 << (bd_shift - 1))) >> bd_shift) + | |
111 | 192 | (1 << (bit_depth - 1)); | |
112 | 192 | ptr[x] = av_clip_uintp2(sample, bit_depth); | |
113 | } | ||
114 | 24 | ptr += pitch; | |
115 | } | ||
116 | } else { | ||
117 | 9603 | uint16_t *ptr = output; | |
118 | 9603 | int bd_shift = 20 - bit_depth; | |
119 | 9603 | pitch /= 2; // Pitch was in bytes, 2 bytes per sample. | |
120 | |||
121 |
2/2✓ Branch 0 taken 76824 times.
✓ Branch 1 taken 9603 times.
|
86427 | for (int y = 0; y < 8; y++) { |
122 |
2/2✓ Branch 0 taken 614592 times.
✓ Branch 1 taken 76824 times.
|
691416 | for (int x = 0; x < 8; x++) { |
123 | 614592 | int sample = ((recon_sample[y][x] + | |
124 | 614592 | (1 << (bd_shift - 1))) >> bd_shift) + | |
125 | 614592 | (1 << (bit_depth - 1)); | |
126 | 614592 | ptr[x] = av_clip_uintp2(sample, bit_depth); | |
127 | } | ||
128 | 76824 | ptr += pitch; | |
129 | } | ||
130 | } | ||
131 | 9606 | } | |
132 | |||
133 | 16 | av_cold void ff_apv_dsp_init(APVDSPContext *dsp) | |
134 | { | ||
135 | 16 | dsp->decode_transquant = apv_decode_transquant_c; | |
136 | |||
137 | #if ARCH_X86_64 | ||
138 | 16 | ff_apv_dsp_init_x86_64(dsp); | |
139 | #endif | ||
140 | 16 | } | |
141 |