File: prores_raw_decode.comp

package info (click to toggle)
chromium 145.0.7632.159-1
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid
  • size: 5,976,224 kB
  • sloc: cpp: 36,198,469; ansic: 7,634,080; javascript: 3,564,060; python: 1,649,622; xml: 838,470; asm: 717,087; pascal: 185,708; sh: 88,786; perl: 88,718; objc: 79,984; sql: 59,811; cs: 42,452; fortran: 24,101; makefile: 21,144; tcl: 15,277; php: 14,022; yacc: 9,066; ruby: 7,553; awk: 3,720; lisp: 3,233; lex: 1,328; ada: 727; jsp: 228; sed: 36
file content (231 lines) | stat: -rw-r--r-- 7,413 bytes parent folder | download | duplicates (3)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
/*
 * ProRes RAW decoder
 *
 * Copyright (c) 2025 Lynne <dev@lynne.ee>
 *
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * FFmpeg is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with FFmpeg; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#define COMP_ID (gl_LocalInvocationID.x)

GetBitContext gb;

#define DC_CB_MAX 12
const uint8_t dc_cb[DC_CB_MAX + 1] = {
    U8(16), U8(33), U8(50), U8(51), U8(51), U8(51),
    U8(68), U8(68), U8(68), U8(68), U8(68), U8(68), U8(118)
};

#define AC_CB_MAX 94
const int16_t ac_cb[AC_CB_MAX + 1] = {
    I16(  0), I16(529), I16(273), I16(273), I16(546), I16(546),
    I16(546), I16(290), I16(290), I16(290), I16(563), I16(563),
    I16(563), I16(563), I16(563), I16(563), I16(563), I16(563),
    I16(307), I16(307), I16(580), I16(580), I16(580), I16(580),
    I16(580), I16(580), I16(580), I16(580), I16(580), I16(580),
    I16(580), I16(580), I16(580), I16(580), I16(580), I16(580),
    I16(580), I16(580), I16(580), I16(580), I16(580), I16(580),
    I16(853), I16(853), I16(853), I16(853), I16(853), I16(853),
    I16(853), I16(853), I16(853), I16(853), I16(853), I16(853),
    I16(853), I16(853), I16(853), I16(853), I16(853), I16(853),
    I16(853), I16(853), I16(853), I16(853), I16(853), I16(853),
    I16(853), I16(853), I16(853), I16(853), I16(853), I16(853),
    I16(853), I16(853), I16(853), I16(853), I16(853), I16(853),
    I16(853), I16(853), I16(853), I16(853), I16(853), I16(853),
    I16(853), I16(853), I16(853), I16(853), I16(853), I16(853),
    I16(853), I16(853), I16(853), I16(853), I16(358)
};

#define RN_CB_MAX 27
const int16_t rn_cb[RN_CB_MAX + 1] = {
    I16(512), I16(256), I16(  0), I16(  0), I16(529), I16(529), I16(273),
    I16(273), I16( 17), I16( 17), I16( 33), I16( 33), I16(546), I16( 34),
    I16( 34), I16( 34), I16( 34), I16( 34), I16( 34), I16( 34), I16( 34),
    I16( 34), I16( 34), I16( 34), I16( 34), I16( 50), I16( 50), I16( 68),
};

#define LN_CB_MAX 14
const int16_t ln_cb[LN_CB_MAX + 1] = {
    I16( 256), I16( 273), I16( 546), I16( 546), I16( 290), I16( 290), I16( 1075),
    I16(1075), I16( 563), I16( 563), I16( 563), I16( 563), I16( 563), I16( 563),
    I16( 51)
};

int16_t get_value(int16_t codebook)
{
    const int16_t switch_bits = codebook >> 8;
    const int16_t rice_order  = codebook & I16(0xf);
    const int16_t exp_order   = (codebook >> 4) & I16(0xf);

    uint32_t b = show_bits(gb, 32);
    if (expectEXT(b == 0, false))
        return I16(0);
    int16_t q = I16(31) - I16(findMSB(b));

    if ((b & 0x80000000) != 0) {
        skip_bits(gb, 1 + rice_order);
        return I16((b & 0x7FFFFFFF) >> (31 - rice_order));
    }

    if (q <= switch_bits) {
        skip_bits(gb, q + rice_order + 1);
        return I16((q << rice_order) +
                   (((b << (q + 1)) >> 1) >> (31 - rice_order)));
    }

    int16_t bits = exp_order + (q << 1) - switch_bits;
    skip_bits(gb, bits);
    return I16((b >> (32 - bits)) +
               ((switch_bits + 1) << rice_order) -
               (1 << exp_order));
}

#define TODCCODEBOOK(x) ((x + 1) >> 1)

void store_val(ivec2 offs, int blk, int c, int16_t v)
{
    imageStore(dst, offs + 2*ivec2(blk*8 + (c & 7), c >> 3),
               ivec4(v & 0xFFFF));
}

void read_dc_vals(ivec2 offs, int nb_blocks)
{
    int16_t dc, dc_add;
    int16_t prev_dc = I16(0), sign = I16(0);

    /* Special handling for first block */
    dc = get_value(I16(700));
    prev_dc = (dc >> 1) ^ -(dc & I16(1));
    store_val(offs, 0, 0, prev_dc);

    for (int n = 1; n < nb_blocks; n++) {
        if (expectEXT(left_bits(gb) <= 0, false))
            break;

        uint8_t dc_codebook;
        if ((n & 15) == 1)
            dc_codebook = uint8_t(100);
        else
            dc_codebook = dc_cb[min(TODCCODEBOOK(dc), 13 - 1)];

        dc = get_value(dc_codebook);

        sign = sign ^ dc & int16_t(1);
        dc_add = (-sign ^ I16(TODCCODEBOOK(dc))) + sign;
        sign = I16(dc_add < 0);
        prev_dc += dc_add;

        store_val(offs, n, 0, prev_dc);
    }
}

void read_ac_vals(ivec2 offs, int nb_blocks)
{
    const int nb_codes = nb_blocks << 6;
    const int log2_nb_blocks = findMSB(nb_blocks);
    const int block_mask = (1 << log2_nb_blocks) - 1;

    int16_t ac, rn, ln;
    int16_t ac_codebook = I16(49);
    int16_t rn_codebook = I16( 0);
    int16_t ln_codebook = I16(66);
    int16_t sign;
    int16_t val;

    for (int n = nb_blocks; n <= nb_codes;) {
        if (expectEXT(left_bits(gb) <= 0, false))
            break;

        ln = get_value(ln_codebook);
        for (int i = 0; i < ln; i++) {
            if (expectEXT(left_bits(gb) <= 0, false))
                break;

            if (expectEXT(n >= nb_codes, false))
                break;

            ac = get_value(ac_codebook);
            ac_codebook = ac_cb[min(ac, 95 - 1)];
            sign = -int16_t(get_bit(gb));

            val = ((ac + I16(1)) ^ sign) - sign;
            store_val(offs, n & block_mask, n >> log2_nb_blocks, val);

            n++;
        }

        if (expectEXT(n >= nb_codes, false))
            break;

        rn = get_value(rn_codebook);
        rn_codebook = rn_cb[min(rn, 28 - 1)];

        n += rn + 1;
        if (expectEXT(n >= nb_codes, false))
            break;

        if (expectEXT(left_bits(gb) <= 0, false))
            break;

        ac = get_value(ac_codebook);
        sign = -int16_t(get_bit(gb));

        val = ((ac + I16(1)) ^ sign) - sign;
        store_val(offs, n & block_mask, n >> log2_nb_blocks, val);

        ac_codebook = ac_cb[min(ac, 95 - 1)];
        ln_codebook = ln_cb[min(ac, 15 - 1)];

        n++;
    }
}

void main(void)
{
    const uint tile_idx = gl_WorkGroupID.y*gl_NumWorkGroups.x + gl_WorkGroupID.x;
    TileData td = tile_data[tile_idx];

    if (expectEXT(td.pos.x >= frame_size.x, false))
        return;

    uint64_t pkt_offset = uint64_t(pkt_data) + td.offset;
    u8vec2buf hdr_data = u8vec2buf(pkt_offset);
    int header_len = hdr_data[0].v.x >> 3;

    ivec4 size = ivec4(td.size,
                       pack16(hdr_data[2].v.yx),
                       pack16(hdr_data[1].v.yx),
                       pack16(hdr_data[3].v.yx));
    size[0] = size[0] - size[1] - size[2] - size[3] - header_len;
    if (expectEXT(size[0] < 0, false))
        return;

    const ivec2 offs = td.pos + ivec2(COMP_ID & 1, COMP_ID >> 1);
    const int w = min(tile_size.x, frame_size.x - td.pos.x) / 2;
    const int nb_blocks = w / 8;

    const ivec4 comp_offset = ivec4(size[2] + size[1] + size[3],
                                    size[2],
                                    0,
                                    size[2] + size[1]);

    init_get_bits(gb, u8buf(pkt_offset + header_len + comp_offset[COMP_ID]),
                  size[COMP_ID]);

    read_dc_vals(offs, nb_blocks);
    read_ac_vals(offs, nb_blocks);
}