4 * Copyright (c) 2003-2012 Michael Niedermayer <michaelni@gmx.at>
6 * This file is part of FFmpeg.
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.
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.
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
25 * FF Video Codec 1 (a lossless codec) encoder
28 #include "libavutil/avassert.h"
29 #include "libavutil/crc.h"
30 #include "libavutil/opt.h"
31 #include "libavutil/imgutils.h"
32 #include "libavutil/pixdesc.h"
33 #include "libavutil/timer.h"
38 #include "rangecoder.h"
43 static const int8_t quant5_10bit[256] = {
44 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1,
45 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
46 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
47 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
48 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
49 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
50 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
51 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
52 -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2,
53 -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2,
54 -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2,
55 -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2,
56 -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -1,
57 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
58 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
59 -1, -1, -1, -1, -1, -1, -0, -0, -0, -0, -0, -0, -0, -0, -0, -0,
62 static const int8_t quant5[256] = {
63 0, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
64 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
65 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
66 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
67 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
68 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
69 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
70 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
71 -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2,
72 -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2,
73 -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2,
74 -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2,
75 -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2,
76 -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2,
77 -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2,
78 -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2, -1, -1, -1,
81 static const int8_t quant9_10bit[256] = {
82 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2,
83 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3,
84 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,
85 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4,
86 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
87 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
88 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
89 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
90 -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4,
91 -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4,
92 -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4,
93 -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4,
94 -4, -4, -4, -4, -4, -4, -4, -4, -4, -3, -3, -3, -3, -3, -3, -3,
95 -3, -3, -3, -3, -3, -3, -3, -3, -3, -3, -3, -3, -3, -3, -3, -3,
96 -3, -3, -3, -3, -3, -3, -2, -2, -2, -2, -2, -2, -2, -2, -2, -2,
97 -2, -2, -2, -2, -1, -1, -1, -1, -1, -1, -1, -1, -0, -0, -0, -0,
100 static const int8_t quant11[256] = {
101 0, 1, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4,
102 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
103 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5,
104 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5,
105 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5,
106 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5,
107 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5,
108 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5,
109 -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5,
110 -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5,
111 -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5,
112 -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5,
113 -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5,
114 -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -5, -4, -4,
115 -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4, -4,
116 -4, -4, -4, -4, -4, -3, -3, -3, -3, -3, -3, -3, -2, -2, -2, -1,
119 static const uint8_t ver2_state[256] = {
120 0, 10, 10, 10, 10, 16, 16, 16, 28, 16, 16, 29, 42, 49, 20, 49,
121 59, 25, 26, 26, 27, 31, 33, 33, 33, 34, 34, 37, 67, 38, 39, 39,
122 40, 40, 41, 79, 43, 44, 45, 45, 48, 48, 64, 50, 51, 52, 88, 52,
123 53, 74, 55, 57, 58, 58, 74, 60, 101, 61, 62, 84, 66, 66, 68, 69,
124 87, 82, 71, 97, 73, 73, 82, 75, 111, 77, 94, 78, 87, 81, 83, 97,
125 85, 83, 94, 86, 99, 89, 90, 99, 111, 92, 93, 134, 95, 98, 105, 98,
126 105, 110, 102, 108, 102, 118, 103, 106, 106, 113, 109, 112, 114, 112, 116, 125,
127 115, 116, 117, 117, 126, 119, 125, 121, 121, 123, 145, 124, 126, 131, 127, 129,
128 165, 130, 132, 138, 133, 135, 145, 136, 137, 139, 146, 141, 143, 142, 144, 148,
129 147, 155, 151, 149, 151, 150, 152, 157, 153, 154, 156, 168, 158, 162, 161, 160,
130 172, 163, 169, 164, 166, 184, 167, 170, 177, 174, 171, 173, 182, 176, 180, 178,
131 175, 189, 179, 181, 186, 183, 192, 185, 200, 187, 191, 188, 190, 197, 193, 196,
132 197, 194, 195, 196, 198, 202, 199, 201, 210, 203, 207, 204, 205, 206, 208, 214,
133 209, 211, 221, 212, 213, 215, 224, 216, 217, 218, 219, 220, 222, 228, 223, 225,
134 226, 224, 227, 229, 240, 230, 231, 232, 233, 234, 235, 236, 238, 239, 237, 242,
135 241, 243, 242, 244, 245, 246, 247, 248, 249, 250, 251, 252, 252, 253, 254, 255,
138 static void find_best_state(uint8_t best_state[256][256],
139 const uint8_t one_state[256])
144 for (i = 1; i < 256; i++)
145 l2tab[i] = log2(i / 256.0);
147 for (i = 0; i < 256; i++) {
148 double best_len[256];
149 double p = i / 256.0;
151 for (j = 0; j < 256; j++)
152 best_len[j] = 1 << 30;
154 for (j = FFMAX(i - 10, 1); j < FFMIN(i + 11, 256); j++) {
155 double occ[256] = { 0 };
158 for (k = 0; k < 256; k++) {
159 double newocc[256] = { 0 };
160 for (m = 1; m < 256; m++)
162 len -=occ[m]*( p *l2tab[ m]
163 + (1-p)*l2tab[256-m]);
165 if (len < best_len[k]) {
167 best_state[i][k] = j;
169 for (m = 0; m < 256; m++)
171 newocc[ one_state[ m]] += occ[m] * p;
172 newocc[256 - one_state[256 - m]] += occ[m] * (1 - p);
174 memcpy(occ, newocc, sizeof(occ));
180 static av_always_inline av_flatten void put_symbol_inline(RangeCoder *c,
181 uint8_t *state, int v,
183 uint64_t rc_stat[256][2],
184 uint64_t rc_stat2[32][2])
188 #define put_rac(C, S, B) \
191 rc_stat[*(S)][B]++; \
192 rc_stat2[(S) - state][B]++; \
198 const int a = FFABS(v);
199 const int e = av_log2(a);
200 put_rac(c, state + 0, 0);
202 for (i = 0; i < e; i++)
203 put_rac(c, state + 1 + i, 1); // 1..10
204 put_rac(c, state + 1 + i, 0);
206 for (i = e - 1; i >= 0; i--)
207 put_rac(c, state + 22 + i, (a >> i) & 1); // 22..31
210 put_rac(c, state + 11 + e, v < 0); // 11..21
212 for (i = 0; i < e; i++)
213 put_rac(c, state + 1 + FFMIN(i, 9), 1); // 1..10
214 put_rac(c, state + 1 + 9, 0);
216 for (i = e - 1; i >= 0; i--)
217 put_rac(c, state + 22 + FFMIN(i, 9), (a >> i) & 1); // 22..31
220 put_rac(c, state + 11 + 10, v < 0); // 11..21
223 put_rac(c, state + 0, 1);
228 static av_noinline void put_symbol(RangeCoder *c, uint8_t *state,
229 int v, int is_signed)
231 put_symbol_inline(c, state, v, is_signed, NULL, NULL);
235 static inline void put_vlc_symbol(PutBitContext *pb, VlcState *const state,
239 v = fold(v - state->bias, bits);
243 while (i < state->error_sum) { // FIXME: optimize
251 if (k == 0 && 2 * state->drift <= -state->count)
256 code = v ^ ((2 * state->drift + state->count) >> 31);
259 av_dlog(NULL, "v:%d/%d bias:%d error:%d drift:%d count:%d k:%d\n", v, code,
260 state->bias, state->error_sum, state->drift, state->count, k);
261 set_sr_golomb(pb, code, k, 12, bits);
263 update_vlc_state(state, v);
266 static av_always_inline int encode_line(FFV1Context *s, int w,
268 int plane_index, int bits)
270 PlaneContext *const p = &s->plane[plane_index];
271 RangeCoder *const c = &s->c;
273 int run_index = s->run_index;
278 if (c->bytestream_end - c->bytestream < w * 20) {
279 av_log(s->avctx, AV_LOG_ERROR, "encoded frame too large\n");
280 return AVERROR_INVALIDDATA;
283 if (s->pb.buf_end - s->pb.buf - (put_bits_count(&s->pb) >> 3) < w * 4) {
284 av_log(s->avctx, AV_LOG_ERROR, "encoded frame too large\n");
285 return AVERROR_INVALIDDATA;
289 for (x = 0; x < w; x++) {
292 context = get_context(p, sample[0] + x, sample[1] + x, sample[2] + x);
293 diff = sample[0][x] - predict(sample[0] + x, sample[1] + x);
300 diff = fold(diff, bits);
303 if (s->flags & CODEC_FLAG_PASS1) {
304 put_symbol_inline(c, p->state[context], diff, 1, s->rc_stat,
305 s->rc_stat2[p->quant_table_index][context]);
307 put_symbol_inline(c, p->state[context], diff, 1, NULL, NULL);
315 while (run_count >= 1 << ff_log2_run[run_index]) {
316 run_count -= 1 << ff_log2_run[run_index];
318 put_bits(&s->pb, 1, 1);
321 put_bits(&s->pb, 1 + ff_log2_run[run_index], run_count);
333 av_dlog(s->avctx, "count:%d index:%d, mode:%d, x:%d pos:%d\n",
334 run_count, run_index, run_mode, x,
335 (int)put_bits_count(&s->pb));
338 put_vlc_symbol(&s->pb, &p->vlc_state[context], diff, bits);
342 while (run_count >= 1 << ff_log2_run[run_index]) {
343 run_count -= 1 << ff_log2_run[run_index];
345 put_bits(&s->pb, 1, 1);
349 put_bits(&s->pb, 1, 1);
351 s->run_index = run_index;
356 static void encode_plane(FFV1Context *s, uint8_t *src, int w, int h,
357 int stride, int plane_index)
360 const int ring_size = s->avctx->context_model ? 3 : 2;
364 memset(s->sample_buffer, 0, ring_size * (w + 6) * sizeof(*s->sample_buffer));
366 for (y = 0; y < h; y++) {
367 for (i = 0; i < ring_size; i++)
368 sample[i] = s->sample_buffer + (w + 6) * ((h + i - y) % ring_size) + 3;
370 sample[0][-1]= sample[1][0 ];
371 sample[1][ w]= sample[1][w-1];
373 if (s->bits_per_raw_sample <= 8) {
374 for (x = 0; x < w; x++)
375 sample[0][x] = src[x + stride * y];
376 encode_line(s, w, sample, plane_index, 8);
378 if (s->packed_at_lsb) {
379 for (x = 0; x < w; x++) {
380 sample[0][x] = ((uint16_t*)(src + stride*y))[x];
383 for (x = 0; x < w; x++) {
384 sample[0][x] = ((uint16_t*)(src + stride*y))[x] >> (16 - s->bits_per_raw_sample);
387 encode_line(s, w, sample, plane_index, s->bits_per_raw_sample);
389 // STOP_TIMER("encode line") }
393 static void encode_rgb_frame(FFV1Context *s, uint8_t *src[3], int w, int h, int stride[3])
396 const int ring_size = s->avctx->context_model ? 3 : 2;
397 int16_t *sample[4][3];
398 int lbd = s->avctx->bits_per_raw_sample <= 8;
399 int bits = s->avctx->bits_per_raw_sample > 0 ? s->avctx->bits_per_raw_sample : 8;
400 int offset = 1 << bits;
404 memset(s->sample_buffer, 0, ring_size * MAX_PLANES *
405 (w + 6) * sizeof(*s->sample_buffer));
407 for (y = 0; y < h; y++) {
408 for (i = 0; i < ring_size; i++)
409 for (p = 0; p < MAX_PLANES; p++)
410 sample[p][i]= s->sample_buffer + p*ring_size*(w+6) + ((h+i-y)%ring_size)*(w+6) + 3;
412 for (x = 0; x < w; x++) {
413 int b, g, r, av_uninit(a);
415 unsigned v = *((uint32_t*)(src[0] + x*4 + stride[0]*y));
418 r = (v >> 16) & 0xFF;
421 b = *((uint16_t*)(src[0] + x*2 + stride[0]*y));
422 g = *((uint16_t*)(src[1] + x*2 + stride[1]*y));
423 r = *((uint16_t*)(src[2] + x*2 + stride[2]*y));
437 for (p = 0; p < 3 + s->transparency; p++) {
438 sample[p][0][-1] = sample[p][1][0 ];
439 sample[p][1][ w] = sample[p][1][w-1];
441 encode_line(s, w, sample[p], (p + 1) / 2, 9);
443 encode_line(s, w, sample[p], (p + 1) / 2, bits + 1);
448 static void write_quant_table(RangeCoder *c, int16_t *quant_table)
452 uint8_t state[CONTEXT_SIZE];
453 memset(state, 128, sizeof(state));
455 for (i = 1; i < 128; i++)
456 if (quant_table[i] != quant_table[i - 1]) {
457 put_symbol(c, state, i - last - 1, 0);
460 put_symbol(c, state, i - last - 1, 0);
463 static void write_quant_tables(RangeCoder *c,
464 int16_t quant_table[MAX_CONTEXT_INPUTS][256])
467 for (i = 0; i < 5; i++)
468 write_quant_table(c, quant_table[i]);
471 static void write_header(FFV1Context *f)
473 uint8_t state[CONTEXT_SIZE];
475 RangeCoder *const c = &f->slice_context[0]->c;
477 memset(state, 128, sizeof(state));
479 if (f->version < 2) {
480 put_symbol(c, state, f->version, 0);
481 put_symbol(c, state, f->ac, 0);
483 for (i = 1; i < 256; i++)
485 f->state_transition[i] - c->one_state[i], 1);
487 put_symbol(c, state, f->colorspace, 0); //YUV cs type
489 put_symbol(c, state, f->bits_per_raw_sample, 0);
490 put_rac(c, state, f->chroma_planes);
491 put_symbol(c, state, f->chroma_h_shift, 0);
492 put_symbol(c, state, f->chroma_v_shift, 0);
493 put_rac(c, state, f->transparency);
495 write_quant_tables(c, f->quant_table);
496 } else if (f->version < 3) {
497 put_symbol(c, state, f->slice_count, 0);
498 for (i = 0; i < f->slice_count; i++) {
499 FFV1Context *fs = f->slice_context[i];
501 (fs->slice_x + 1) * f->num_h_slices / f->width, 0);
503 (fs->slice_y + 1) * f->num_v_slices / f->height, 0);
505 (fs->slice_width + 1) * f->num_h_slices / f->width - 1,
508 (fs->slice_height + 1) * f->num_v_slices / f->height - 1,
510 for (j = 0; j < f->plane_count; j++) {
511 put_symbol(c, state, f->plane[j].quant_table_index, 0);
512 av_assert0(f->plane[j].quant_table_index == f->avctx->context_model);
518 static int write_extradata(FFV1Context *f)
520 RangeCoder *const c = &f->c;
521 uint8_t state[CONTEXT_SIZE];
523 uint8_t state2[32][CONTEXT_SIZE];
526 memset(state2, 128, sizeof(state2));
527 memset(state, 128, sizeof(state));
529 f->avctx->extradata_size = 10000 + 4 +
530 (11 * 11 * 5 * 5 * 5 + 11 * 11 * 11) * 32;
531 f->avctx->extradata = av_malloc(f->avctx->extradata_size);
532 ff_init_range_encoder(c, f->avctx->extradata, f->avctx->extradata_size);
533 ff_build_rac_states(c, 0.05 * (1LL << 32), 256 - 8);
535 put_symbol(c, state, f->version, 0);
536 if (f->version > 2) {
538 f->minor_version = 2;
539 put_symbol(c, state, f->minor_version, 0);
542 put_symbol(c, state, f->ac, 0);
544 for (i = 1; i < 256; i++)
545 put_symbol(c, state, f->state_transition[i] - c->one_state[i], 1);
547 put_symbol(c, state, f->colorspace, 0); // YUV cs type
548 put_symbol(c, state, f->bits_per_raw_sample, 0);
549 put_rac(c, state, f->chroma_planes);
550 put_symbol(c, state, f->chroma_h_shift, 0);
551 put_symbol(c, state, f->chroma_v_shift, 0);
552 put_rac(c, state, f->transparency);
553 put_symbol(c, state, f->num_h_slices - 1, 0);
554 put_symbol(c, state, f->num_v_slices - 1, 0);
556 put_symbol(c, state, f->quant_table_count, 0);
557 for (i = 0; i < f->quant_table_count; i++)
558 write_quant_tables(c, f->quant_tables[i]);
560 for (i = 0; i < f->quant_table_count; i++) {
561 for (j = 0; j < f->context_count[i] * CONTEXT_SIZE; j++)
562 if (f->initial_states[i] && f->initial_states[i][0][j] != 128)
564 if (j < f->context_count[i] * CONTEXT_SIZE) {
565 put_rac(c, state, 1);
566 for (j = 0; j < f->context_count[i]; j++)
567 for (k = 0; k < CONTEXT_SIZE; k++) {
568 int pred = j ? f->initial_states[i][j - 1][k] : 128;
569 put_symbol(c, state2[k],
570 (int8_t)(f->initial_states[i][j][k] - pred), 1);
573 put_rac(c, state, 0);
577 if (f->version > 2) {
578 put_symbol(c, state, f->ec, 0);
581 f->avctx->extradata_size = ff_rac_terminate(c);
582 v = av_crc(av_crc_get_table(AV_CRC_32_IEEE), 0, f->avctx->extradata, f->avctx->extradata_size);
583 AV_WL32(f->avctx->extradata + f->avctx->extradata_size, v);
584 f->avctx->extradata_size += 4;
589 static int sort_stt(FFV1Context *s, uint8_t stt[256])
591 int i, i2, changed, print = 0;
595 for (i = 12; i < 244; i++) {
596 for (i2 = i + 1; i2 < 245 && i2 < i + 4; i2++) {
598 #define COST(old, new) \
599 s->rc_stat[old][0] * -log2((256 - (new)) / 256.0) + \
600 s->rc_stat[old][1] * -log2((new) / 256.0)
602 #define COST2(old, new) \
603 COST(old, new) + COST(256 - (old), 256 - (new))
605 double size0 = COST2(i, i) + COST2(i2, i2);
606 double sizeX = COST2(i, i2) + COST2(i2, i);
607 if (size0 - sizeX > size0*(1e-14) && i != 128 && i2 != 128) {
609 FFSWAP(int, stt[i], stt[i2]);
610 FFSWAP(int, s->rc_stat[i][0], s->rc_stat[i2][0]);
611 FFSWAP(int, s->rc_stat[i][1], s->rc_stat[i2][1]);
613 FFSWAP(int, stt[256 - i], stt[256 - i2]);
614 FFSWAP(int, s->rc_stat[256 - i][0], s->rc_stat[256 - i2][0]);
615 FFSWAP(int, s->rc_stat[256 - i][1], s->rc_stat[256 - i2][1]);
617 for (j = 1; j < 256; j++) {
620 else if (stt[j] == i2)
623 if (stt[256 - j] == 256 - i)
624 stt[256 - j] = 256 - i2;
625 else if (stt[256 - j] == 256 - i2)
626 stt[256 - j] = 256 - i;
637 static av_cold int encode_init(AVCodecContext *avctx)
639 FFV1Context *s = avctx->priv_data;
640 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(avctx->pix_fmt);
643 ffv1_common_init(avctx);
647 if ((avctx->flags & (CODEC_FLAG_PASS1|CODEC_FLAG_PASS2)) || avctx->slices>1)
648 s->version = FFMAX(s->version, 2);
650 if (avctx->level == 3) {
655 s->ec = (s->version >= 3);
658 if (s->version >= 2 && avctx->strict_std_compliance > FF_COMPLIANCE_EXPERIMENTAL) {
659 av_log(avctx, AV_LOG_ERROR, "Version 2 needed for requested features but version 2 is experimental and not enabled\n");
660 return AVERROR_INVALIDDATA;
663 s->ac = avctx->coder_type > 0 ? 2 : 0;
666 switch(avctx->pix_fmt) {
667 case AV_PIX_FMT_YUV444P9:
668 case AV_PIX_FMT_YUV422P9:
669 case AV_PIX_FMT_YUV420P9:
670 if (!avctx->bits_per_raw_sample)
671 s->bits_per_raw_sample = 9;
672 case AV_PIX_FMT_YUV444P10:
673 case AV_PIX_FMT_YUV420P10:
674 case AV_PIX_FMT_YUV422P10:
675 s->packed_at_lsb = 1;
676 if (!avctx->bits_per_raw_sample && !s->bits_per_raw_sample)
677 s->bits_per_raw_sample = 10;
678 case AV_PIX_FMT_GRAY16:
679 case AV_PIX_FMT_YUV444P16:
680 case AV_PIX_FMT_YUV422P16:
681 case AV_PIX_FMT_YUV420P16:
682 if (!avctx->bits_per_raw_sample && !s->bits_per_raw_sample) {
683 s->bits_per_raw_sample = 16;
684 } else if (!s->bits_per_raw_sample) {
685 s->bits_per_raw_sample = avctx->bits_per_raw_sample;
687 if (s->bits_per_raw_sample <= 8) {
688 av_log(avctx, AV_LOG_ERROR, "bits_per_raw_sample invalid\n");
689 return AVERROR_INVALIDDATA;
691 if (!s->ac && avctx->coder_type == -1) {
692 av_log(avctx, AV_LOG_INFO, "bits_per_raw_sample > 8, forcing coder 1\n");
696 av_log(avctx, AV_LOG_ERROR, "bits_per_raw_sample of more than 8 needs -coder 1 currently\n");
697 return AVERROR(ENOSYS);
699 s->version = FFMAX(s->version, 1);
700 case AV_PIX_FMT_GRAY8:
701 case AV_PIX_FMT_YUV444P:
702 case AV_PIX_FMT_YUV440P:
703 case AV_PIX_FMT_YUV422P:
704 case AV_PIX_FMT_YUV420P:
705 case AV_PIX_FMT_YUV411P:
706 case AV_PIX_FMT_YUV410P:
707 s->chroma_planes = desc->nb_components < 3 ? 0 : 1;
710 case AV_PIX_FMT_YUVA444P:
711 case AV_PIX_FMT_YUVA422P:
712 case AV_PIX_FMT_YUVA420P:
713 s->chroma_planes = 1;
717 case AV_PIX_FMT_RGB32:
721 case AV_PIX_FMT_0RGB32:
724 case AV_PIX_FMT_GBRP9:
725 if (!avctx->bits_per_raw_sample)
726 s->bits_per_raw_sample = 9;
727 case AV_PIX_FMT_GBRP10:
728 if (!avctx->bits_per_raw_sample && !s->bits_per_raw_sample)
729 s->bits_per_raw_sample = 10;
730 case AV_PIX_FMT_GBRP12:
731 if (!avctx->bits_per_raw_sample && !s->bits_per_raw_sample)
732 s->bits_per_raw_sample = 12;
733 case AV_PIX_FMT_GBRP14:
734 if (!avctx->bits_per_raw_sample && !s->bits_per_raw_sample)
735 s->bits_per_raw_sample = 14;
736 else if (!s->bits_per_raw_sample)
737 s->bits_per_raw_sample = avctx->bits_per_raw_sample;
739 s->chroma_planes = 1;
740 s->version = FFMAX(s->version, 1);
743 av_log(avctx, AV_LOG_ERROR, "format not supported\n");
744 return AVERROR(ENOSYS);
746 if (s->transparency) {
747 av_log(avctx, AV_LOG_WARNING, "Storing alpha plane, this will require a recent FFV1 decoder to playback!\n");
749 if (avctx->context_model > 1U) {
750 av_log(avctx, AV_LOG_ERROR, "Invalid context model %d, valid values are 0 and 1\n", avctx->context_model);
751 return AVERROR(EINVAL);
755 for (i = 1; i < 256; i++)
756 s->state_transition[i] = ver2_state[i];
758 for (i = 0; i < 256; i++) {
759 s->quant_table_count = 2;
760 if (s->bits_per_raw_sample <= 8) {
761 s->quant_tables[0][0][i]= quant11[i];
762 s->quant_tables[0][1][i]= 11*quant11[i];
763 s->quant_tables[0][2][i]= 11*11*quant11[i];
764 s->quant_tables[1][0][i]= quant11[i];
765 s->quant_tables[1][1][i]= 11*quant11[i];
766 s->quant_tables[1][2][i]= 11*11*quant5 [i];
767 s->quant_tables[1][3][i]= 5*11*11*quant5 [i];
768 s->quant_tables[1][4][i]= 5*5*11*11*quant5 [i];
770 s->quant_tables[0][0][i]= quant9_10bit[i];
771 s->quant_tables[0][1][i]= 11*quant9_10bit[i];
772 s->quant_tables[0][2][i]= 11*11*quant9_10bit[i];
773 s->quant_tables[1][0][i]= quant9_10bit[i];
774 s->quant_tables[1][1][i]= 11*quant9_10bit[i];
775 s->quant_tables[1][2][i]= 11*11*quant5_10bit[i];
776 s->quant_tables[1][3][i]= 5*11*11*quant5_10bit[i];
777 s->quant_tables[1][4][i]= 5*5*11*11*quant5_10bit[i];
780 s->context_count[0] = (11 * 11 * 11 + 1) / 2;
781 s->context_count[1] = (11 * 11 * 5 * 5 * 5 + 1) / 2;
782 memcpy(s->quant_table, s->quant_tables[avctx->context_model],
783 sizeof(s->quant_table));
785 for (i = 0; i < s->plane_count; i++) {
786 PlaneContext *const p = &s->plane[i];
788 memcpy(p->quant_table, s->quant_table, sizeof(p->quant_table));
789 p->quant_table_index = avctx->context_model;
790 p->context_count = s->context_count[p->quant_table_index];
793 if ((ret = ffv1_allocate_initial_states(s)) < 0)
796 avctx->coded_frame = &s->picture;
797 if (!s->transparency)
799 avcodec_get_chroma_sub_sample(avctx->pix_fmt, &s->chroma_h_shift, &s->chroma_v_shift);
800 s->picture_number = 0;
802 if (avctx->flags & (CODEC_FLAG_PASS1 | CODEC_FLAG_PASS2)) {
803 for (i = 0; i < s->quant_table_count; i++) {
804 s->rc_stat2[i] = av_mallocz(s->context_count[i] *
805 sizeof(*s->rc_stat2[i]));
807 return AVERROR(ENOMEM);
810 if (avctx->stats_in) {
811 char *p = avctx->stats_in;
812 uint8_t best_state[256][256];
816 av_assert0(s->version >= 2);
819 for (j = 0; j < 256; j++)
820 for (i = 0; i < 2; i++) {
821 s->rc_stat[j][i] = strtol(p, &next, 0);
823 av_log(avctx, AV_LOG_ERROR,
824 "2Pass file invalid at %d %d [%s]\n", j, i, p);
825 return AVERROR_INVALIDDATA;
829 for (i = 0; i < s->quant_table_count; i++)
830 for (j = 0; j < s->context_count[i]; j++) {
831 for (k = 0; k < 32; k++)
832 for (m = 0; m < 2; m++) {
833 s->rc_stat2[i][j][k][m] = strtol(p, &next, 0);
835 av_log(avctx, AV_LOG_ERROR,
836 "2Pass file invalid at %d %d %d %d [%s]\n",
838 return AVERROR_INVALIDDATA;
843 gob_count = strtol(p, &next, 0);
844 if (next == p || gob_count <= 0) {
845 av_log(avctx, AV_LOG_ERROR, "2Pass file invalid\n");
846 return AVERROR_INVALIDDATA;
849 while (*p == '\n' || *p == ' ')
854 sort_stt(s, s->state_transition);
856 find_best_state(best_state, s->state_transition);
858 for (i = 0; i < s->quant_table_count; i++) {
859 for (j = 0; j < s->context_count[i]; j++)
860 for (k = 0; k < 32; k++) {
862 if (s->rc_stat2[i][j][k][0] + s->rc_stat2[i][j][k][1]) {
863 p = 256.0 * s->rc_stat2[i][j][k][1] /
864 (s->rc_stat2[i][j][k][0] + s->rc_stat2[i][j][k][1]);
866 s->initial_states[i][j][k] =
867 best_state[av_clip(round(p), 1, 255)][av_clip((s->rc_stat2[i][j][k][0] +
868 s->rc_stat2[i][j][k][1]) /
874 if (s->version > 1) {
875 for (s->num_v_slices = 2; s->num_v_slices < 9; s->num_v_slices++) {
876 for (s->num_h_slices = s->num_v_slices; s->num_h_slices < 2*s->num_v_slices; s->num_h_slices++) {
877 if (avctx->slices == s->num_h_slices * s->num_v_slices && avctx->slices <= 64 || !avctx->slices)
881 av_log(avctx, AV_LOG_ERROR,
882 "Unsupported number %d of slices requested, please specify a "
883 "supported number with -slices (ex:4,6,9,12,16, ...)\n",
885 return AVERROR(ENOSYS);
890 if ((ret = ffv1_init_slice_contexts(s)) < 0)
892 if ((ret = ffv1_init_slices_state(s)) < 0)
895 #define STATS_OUT_SIZE 1024 * 1024 * 6
896 if (avctx->flags & CODEC_FLAG_PASS1) {
897 avctx->stats_out = av_mallocz(STATS_OUT_SIZE);
898 for (i = 0; i < s->quant_table_count; i++)
899 for (j = 0; j < s->slice_count; j++) {
900 FFV1Context *sf = s->slice_context[j];
901 av_assert0(!sf->rc_stat2[i]);
902 sf->rc_stat2[i] = av_mallocz(s->context_count[i] *
903 sizeof(*sf->rc_stat2[i]));
904 if (!sf->rc_stat2[i])
905 return AVERROR(ENOMEM);
912 static void encode_slice_header(FFV1Context *f, FFV1Context *fs)
914 RangeCoder *c = &fs->c;
915 uint8_t state[CONTEXT_SIZE];
917 memset(state, 128, sizeof(state));
919 put_symbol(c, state, (fs->slice_x +1)*f->num_h_slices / f->width , 0);
920 put_symbol(c, state, (fs->slice_y +1)*f->num_v_slices / f->height , 0);
921 put_symbol(c, state, (fs->slice_width +1)*f->num_h_slices / f->width -1, 0);
922 put_symbol(c, state, (fs->slice_height+1)*f->num_v_slices / f->height-1, 0);
923 for (j=0; j<f->plane_count; j++) {
924 put_symbol(c, state, f->plane[j].quant_table_index, 0);
925 av_assert0(f->plane[j].quant_table_index == f->avctx->context_model);
927 if (!f->picture.interlaced_frame)
928 put_symbol(c, state, 3, 0);
930 put_symbol(c, state, 1 + !f->picture.top_field_first, 0);
931 put_symbol(c, state, f->picture.sample_aspect_ratio.num, 0);
932 put_symbol(c, state, f->picture.sample_aspect_ratio.den, 0);
935 static int encode_slice(AVCodecContext *c, void *arg)
937 FFV1Context *fs = *(void **)arg;
938 FFV1Context *f = fs->avctx->priv_data;
939 int width = fs->slice_width;
940 int height = fs->slice_height;
943 AVFrame *const p = &f->picture;
944 const int ps = av_pix_fmt_desc_get(c->pix_fmt)->comp[0].step_minus1 + 1;
947 ffv1_clear_slice_state(f, fs);
948 if (f->version > 2) {
949 encode_slice_header(f, fs);
953 put_rac(&fs->c, (uint8_t[]) { 129 }, 0);
954 fs->ac_byte_count = f->version > 2 || (!x && !y) ? ff_rac_terminate(&fs->c) : 0;
955 init_put_bits(&fs->pb,
956 fs->c.bytestream_start + fs->ac_byte_count,
957 fs->c.bytestream_end - fs->c.bytestream_start - fs->ac_byte_count);
960 if (f->colorspace == 0) {
961 const int chroma_width = -((-width) >> f->chroma_h_shift);
962 const int chroma_height = -((-height) >> f->chroma_v_shift);
963 const int cx = x >> f->chroma_h_shift;
964 const int cy = y >> f->chroma_v_shift;
966 encode_plane(fs, p->data[0] + ps*x + y*p->linesize[0], width, height, p->linesize[0], 0);
968 if (f->chroma_planes) {
969 encode_plane(fs, p->data[1] + ps*cx+cy*p->linesize[1], chroma_width, chroma_height, p->linesize[1], 1);
970 encode_plane(fs, p->data[2] + ps*cx+cy*p->linesize[2], chroma_width, chroma_height, p->linesize[2], 1);
972 if (fs->transparency)
973 encode_plane(fs, p->data[3] + ps*x + y*p->linesize[3], width, height, p->linesize[3], 2);
975 uint8_t *planes[3] = {p->data[0] + ps*x + y*p->linesize[0],
976 p->data[1] + ps*x + y*p->linesize[1],
977 p->data[2] + ps*x + y*p->linesize[2]};
978 encode_rgb_frame(fs, planes, width, height, p->linesize);
985 static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
986 const AVFrame *pict, int *got_packet)
988 FFV1Context *f = avctx->priv_data;
989 RangeCoder *const c = &f->slice_context[0]->c;
990 AVFrame *const p = &f->picture;
992 uint8_t keystate = 128;
996 if ((ret = ff_alloc_packet2(avctx, pkt, avctx->width*avctx->height*((8*2+1+1)*4)/8
997 + FF_MIN_BUFFER_SIZE)) < 0)
1000 ff_init_range_encoder(c, pkt->data, pkt->size);
1001 ff_build_rac_states(c, 0.05 * (1LL << 32), 256 - 8);
1004 p->pict_type = AV_PICTURE_TYPE_I;
1006 if (avctx->gop_size == 0 || f->picture_number % avctx->gop_size == 0) {
1007 put_rac(c, &keystate, 1);
1012 put_rac(c, &keystate, 0);
1018 for (i = 1; i < 256; i++) {
1019 c->one_state[i] = f->state_transition[i];
1020 c->zero_state[256 - i] = 256 - c->one_state[i];
1024 for (i = 1; i < f->slice_count; i++) {
1025 FFV1Context *fs = f->slice_context[i];
1026 uint8_t *start = pkt->data + (pkt->size - used_count) * (int64_t)i / f->slice_count;
1027 int len = pkt->size / f->slice_count;
1028 ff_init_range_encoder(&fs->c, start, len);
1030 avctx->execute(avctx, encode_slice, &f->slice_context[0], NULL,
1031 f->slice_count, sizeof(void *));
1034 for (i = 0; i < f->slice_count; i++) {
1035 FFV1Context *fs = f->slice_context[i];
1039 uint8_t state = 129;
1040 put_rac(&fs->c, &state, 0);
1041 bytes = ff_rac_terminate(&fs->c);
1043 flush_put_bits(&fs->pb); // FIXME: nicer padding
1044 bytes = fs->ac_byte_count + (put_bits_count(&fs->pb) + 7) / 8;
1046 if (i > 0 || f->version > 2) {
1047 av_assert0(bytes < pkt->size / f->slice_count);
1048 memmove(buf_p, fs->c.bytestream_start, bytes);
1049 av_assert0(bytes < (1 << 24));
1050 AV_WB24(buf_p + bytes, bytes);
1056 v = av_crc(av_crc_get_table(AV_CRC_32_IEEE), 0, buf_p, bytes);
1057 AV_WL32(buf_p + bytes, v);
1063 if ((avctx->flags & CODEC_FLAG_PASS1) && (f->picture_number & 31) == 0) {
1065 char *p = avctx->stats_out;
1066 char *end = p + STATS_OUT_SIZE;
1068 memset(f->rc_stat, 0, sizeof(f->rc_stat));
1069 for (i = 0; i < f->quant_table_count; i++)
1070 memset(f->rc_stat2[i], 0, f->context_count[i] * sizeof(*f->rc_stat2[i]));
1072 for (j = 0; j < f->slice_count; j++) {
1073 FFV1Context *fs = f->slice_context[j];
1074 for (i = 0; i < 256; i++) {
1075 f->rc_stat[i][0] += fs->rc_stat[i][0];
1076 f->rc_stat[i][1] += fs->rc_stat[i][1];
1078 for (i = 0; i < f->quant_table_count; i++) {
1079 for (k = 0; k < f->context_count[i]; k++)
1080 for (m = 0; m < 32; m++) {
1081 f->rc_stat2[i][k][m][0] += fs->rc_stat2[i][k][m][0];
1082 f->rc_stat2[i][k][m][1] += fs->rc_stat2[i][k][m][1];
1087 for (j = 0; j < 256; j++) {
1088 snprintf(p, end - p, "%" PRIu64 " %" PRIu64 " ",
1089 f->rc_stat[j][0], f->rc_stat[j][1]);
1092 snprintf(p, end - p, "\n");
1094 for (i = 0; i < f->quant_table_count; i++) {
1095 for (j = 0; j < f->context_count[i]; j++)
1096 for (m = 0; m < 32; m++) {
1097 snprintf(p, end - p, "%" PRIu64 " %" PRIu64 " ",
1098 f->rc_stat2[i][j][m][0], f->rc_stat2[i][j][m][1]);
1102 snprintf(p, end - p, "%d\n", f->gob_count);
1103 } else if (avctx->flags & CODEC_FLAG_PASS1)
1104 avctx->stats_out[0] = '\0';
1106 f->picture_number++;
1107 pkt->size = buf_p - pkt->data;
1108 pkt->flags |= AV_PKT_FLAG_KEY * p->key_frame;
1114 #define OFFSET(x) offsetof(FFV1Context, x)
1115 #define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
1116 static const AVOption options[] = {
1117 { "slicecrc", "Protect slices with CRCs", OFFSET(ec), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 1, VE },
1121 static const AVClass class = {
1122 .class_name = "ffv1 encoder",
1123 .item_name = av_default_item_name,
1125 .version = LIBAVUTIL_VERSION_INT,
1128 static const AVCodecDefault ffv1_defaults[] = {
1133 AVCodec ff_ffv1_encoder = {
1135 .type = AVMEDIA_TYPE_VIDEO,
1136 .id = AV_CODEC_ID_FFV1,
1137 .priv_data_size = sizeof(FFV1Context),
1138 .init = encode_init,
1139 .encode2 = encode_frame,
1140 .close = ffv1_close,
1141 .capabilities = CODEC_CAP_SLICE_THREADS,
1142 .pix_fmts = (const enum AVPixelFormat[]) {
1143 AV_PIX_FMT_YUV420P, AV_PIX_FMT_YUVA420P, AV_PIX_FMT_YUVA422P, AV_PIX_FMT_YUV444P,
1144 AV_PIX_FMT_YUVA444P, AV_PIX_FMT_YUV440P, AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUV411P,
1145 AV_PIX_FMT_YUV410P, AV_PIX_FMT_0RGB32, AV_PIX_FMT_RGB32, AV_PIX_FMT_YUV420P16,
1146 AV_PIX_FMT_YUV422P16, AV_PIX_FMT_YUV444P16, AV_PIX_FMT_YUV444P9, AV_PIX_FMT_YUV422P9,
1147 AV_PIX_FMT_YUV420P9, AV_PIX_FMT_YUV420P10, AV_PIX_FMT_YUV422P10, AV_PIX_FMT_YUV444P10,
1148 AV_PIX_FMT_GRAY16, AV_PIX_FMT_GRAY8, AV_PIX_FMT_GBRP9, AV_PIX_FMT_GBRP10,
1149 AV_PIX_FMT_GBRP12, AV_PIX_FMT_GBRP14,
1153 .long_name = NULL_IF_CONFIG_SMALL("FFmpeg video codec #1"),
1154 .defaults = ffv1_defaults,
1155 .priv_class = &class,