1 /*
2 * H.26L/H.264/AVC/JVT/14496-10/... SEI decoding
3 * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
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
24 * H.264 / AVC / MPEG-4 part10 SEI decoding.
25 * @author Michael Niedermayer <michaelni@gmx.at>
26 */
27
28 #include "atsc_a53.h"
29 #include "avcodec.h"
30 #include "get_bits.h"
31 #include "golomb.h"
32 #include "h264_ps.h"
33 #include "h264_sei.h"
34 #include "internal.h"
35
36 #define AVERROR_PS_NOT_FOUND FFERRTAG(0xF8,'?','P','S')
37
38 static const uint8_t sei_num_clock_ts_table[9] = {
39 1, 1, 1, 2, 2, 3, 3, 2, 3
40 };
41
ff_h264_sei_uninit(H264SEIContext * h)42 void ff_h264_sei_uninit(H264SEIContext *h)
43 {
44 h->recovery_point.recovery_frame_cnt = -1;
45
46 h->picture_timing.dpb_output_delay = 0;
47 h->picture_timing.cpb_removal_delay = -1;
48
49 h->picture_timing.present = 0;
50 h->buffering_period.present = 0;
51 h->frame_packing.present = 0;
52 h->display_orientation.present = 0;
53 h->afd.present = 0;
54
55 av_buffer_unref(&h->a53_caption.buf_ref);
56 for (int i = 0; i < h->unregistered.nb_buf_ref; i++)
57 av_buffer_unref(&h->unregistered.buf_ref[i]);
58 h->unregistered.nb_buf_ref = 0;
59 av_freep(&h->unregistered.buf_ref);
60 }
61
ff_h264_sei_process_picture_timing(H264SEIPictureTiming * h,const SPS * sps,void * logctx)62 int ff_h264_sei_process_picture_timing(H264SEIPictureTiming *h, const SPS *sps,
63 void *logctx)
64 {
65 GetBitContext gb;
66
67 init_get_bits(&gb, h->payload, h->payload_size_bits);
68
69 if (sps->nal_hrd_parameters_present_flag ||
70 sps->vcl_hrd_parameters_present_flag) {
71 h->cpb_removal_delay = get_bits_long(&gb, sps->cpb_removal_delay_length);
72 h->dpb_output_delay = get_bits_long(&gb, sps->dpb_output_delay_length);
73 }
74 if (sps->pic_struct_present_flag) {
75 unsigned int i, num_clock_ts;
76
77 h->pic_struct = get_bits(&gb, 4);
78 h->ct_type = 0;
79
80 if (h->pic_struct > H264_SEI_PIC_STRUCT_FRAME_TRIPLING)
81 return AVERROR_INVALIDDATA;
82
83 num_clock_ts = sei_num_clock_ts_table[h->pic_struct];
84 h->timecode_cnt = 0;
85 for (i = 0; i < num_clock_ts; i++) {
86 if (get_bits(&gb, 1)) { /* clock_timestamp_flag */
87 H264SEITimeCode *tc = &h->timecode[h->timecode_cnt++];
88 unsigned int full_timestamp_flag;
89 unsigned int counting_type, cnt_dropped_flag;
90 h->ct_type |= 1 << get_bits(&gb, 2);
91 skip_bits(&gb, 1); /* nuit_field_based_flag */
92 counting_type = get_bits(&gb, 5); /* counting_type */
93 full_timestamp_flag = get_bits(&gb, 1);
94 skip_bits(&gb, 1); /* discontinuity_flag */
95 cnt_dropped_flag = get_bits(&gb, 1); /* cnt_dropped_flag */
96 if (cnt_dropped_flag && counting_type > 1 && counting_type < 7)
97 tc->dropframe = 1;
98 tc->frame = get_bits(&gb, 8); /* n_frames */
99 if (full_timestamp_flag) {
100 tc->full = 1;
101 tc->seconds = get_bits(&gb, 6); /* seconds_value 0..59 */
102 tc->minutes = get_bits(&gb, 6); /* minutes_value 0..59 */
103 tc->hours = get_bits(&gb, 5); /* hours_value 0..23 */
104 } else {
105 tc->seconds = tc->minutes = tc->hours = tc->full = 0;
106 if (get_bits(&gb, 1)) { /* seconds_flag */
107 tc->seconds = get_bits(&gb, 6);
108 if (get_bits(&gb, 1)) { /* minutes_flag */
109 tc->minutes = get_bits(&gb, 6);
110 if (get_bits(&gb, 1)) /* hours_flag */
111 tc->hours = get_bits(&gb, 5);
112 }
113 }
114 }
115
116 if (sps->time_offset_length > 0)
117 skip_bits(&gb,
118 sps->time_offset_length); /* time_offset */
119 }
120 }
121
122 av_log(logctx, AV_LOG_DEBUG, "ct_type:%X pic_struct:%d\n",
123 h->ct_type, h->pic_struct);
124 }
125
126 return 0;
127 }
128
decode_picture_timing(H264SEIPictureTiming * h,GetBitContext * gb,void * logctx)129 static int decode_picture_timing(H264SEIPictureTiming *h, GetBitContext *gb,
130 void *logctx)
131 {
132 int index = get_bits_count(gb);
133 int size_bits = get_bits_left(gb);
134 int size = (size_bits + 7) / 8;
135
136 if (index & 7) {
137 av_log(logctx, AV_LOG_ERROR, "Unaligned SEI payload\n");
138 return AVERROR_INVALIDDATA;
139 }
140 if (size > sizeof(h->payload)) {
141 av_log(logctx, AV_LOG_ERROR, "Picture timing SEI payload too large\n");
142 return AVERROR_INVALIDDATA;
143 }
144 memcpy(h->payload, gb->buffer + index / 8, size);
145
146 h->payload_size_bits = size_bits;
147
148 h->present = 1;
149 return 0;
150 }
151
decode_registered_user_data_afd(H264SEIAFD * h,GetBitContext * gb,int size)152 static int decode_registered_user_data_afd(H264SEIAFD *h, GetBitContext *gb, int size)
153 {
154 int flag;
155
156 if (size-- < 1)
157 return AVERROR_INVALIDDATA;
158 skip_bits(gb, 1); // 0
159 flag = get_bits(gb, 1); // active_format_flag
160 skip_bits(gb, 6); // reserved
161
162 if (flag) {
163 if (size-- < 1)
164 return AVERROR_INVALIDDATA;
165 skip_bits(gb, 4); // reserved
166 h->active_format_description = get_bits(gb, 4);
167 h->present = 1;
168 }
169
170 return 0;
171 }
172
decode_registered_user_data_closed_caption(H264SEIA53Caption * h,GetBitContext * gb,void * logctx,int size)173 static int decode_registered_user_data_closed_caption(H264SEIA53Caption *h,
174 GetBitContext *gb, void *logctx,
175 int size)
176 {
177 if (size < 3)
178 return AVERROR(EINVAL);
179
180 return ff_parse_a53_cc(&h->buf_ref, gb->buffer + get_bits_count(gb) / 8, size);
181 }
182
decode_registered_user_data(H264SEIContext * h,GetBitContext * gb,void * logctx,int size)183 static int decode_registered_user_data(H264SEIContext *h, GetBitContext *gb,
184 void *logctx, int size)
185 {
186 int country_code, provider_code;
187
188 if (size < 3)
189 return AVERROR_INVALIDDATA;
190 size -= 3;
191
192 country_code = get_bits(gb, 8); // itu_t_t35_country_code
193 if (country_code == 0xFF) {
194 if (size < 1)
195 return AVERROR_INVALIDDATA;
196
197 skip_bits(gb, 8); // itu_t_t35_country_code_extension_byte
198 size--;
199 }
200
201 if (country_code != 0xB5) { // usa_country_code
202 av_log(logctx, AV_LOG_VERBOSE,
203 "Unsupported User Data Registered ITU-T T35 SEI message (country_code = %d)\n",
204 country_code);
205 return 0;
206 }
207
208 /* itu_t_t35_payload_byte follows */
209 provider_code = get_bits(gb, 16);
210
211 switch (provider_code) {
212 case 0x31: { // atsc_provider_code
213 uint32_t user_identifier;
214
215 if (size < 4)
216 return AVERROR_INVALIDDATA;
217 size -= 4;
218
219 user_identifier = get_bits_long(gb, 32);
220 switch (user_identifier) {
221 case MKBETAG('D', 'T', 'G', '1'): // afd_data
222 return decode_registered_user_data_afd(&h->afd, gb, size);
223 case MKBETAG('G', 'A', '9', '4'): // closed captions
224 return decode_registered_user_data_closed_caption(&h->a53_caption, gb,
225 logctx, size);
226 default:
227 av_log(logctx, AV_LOG_VERBOSE,
228 "Unsupported User Data Registered ITU-T T35 SEI message (atsc user_identifier = 0x%04x)\n",
229 user_identifier);
230 break;
231 }
232 break;
233 }
234 default:
235 av_log(logctx, AV_LOG_VERBOSE,
236 "Unsupported User Data Registered ITU-T T35 SEI message (provider_code = %d)\n",
237 provider_code);
238 break;
239 }
240
241 return 0;
242 }
243
decode_unregistered_user_data(H264SEIUnregistered * h,GetBitContext * gb,void * logctx,int size)244 static int decode_unregistered_user_data(H264SEIUnregistered *h, GetBitContext *gb,
245 void *logctx, int size)
246 {
247 uint8_t *user_data;
248 int e, build, i;
249 AVBufferRef *buf_ref, **tmp;
250
251 if (size < 16 || size >= INT_MAX - 1)
252 return AVERROR_INVALIDDATA;
253
254 tmp = av_realloc_array(h->buf_ref, h->nb_buf_ref + 1, sizeof(*h->buf_ref));
255 if (!tmp)
256 return AVERROR(ENOMEM);
257 h->buf_ref = tmp;
258
259 buf_ref = av_buffer_alloc(size + 1);
260 if (!buf_ref)
261 return AVERROR(ENOMEM);
262 user_data = buf_ref->data;
263
264 for (i = 0; i < size; i++)
265 user_data[i] = get_bits(gb, 8);
266
267 user_data[i] = 0;
268 buf_ref->size = size;
269 h->buf_ref[h->nb_buf_ref++] = buf_ref;
270
271 e = sscanf(user_data + 16, "x264 - core %d", &build);
272 if (e == 1 && build > 0)
273 h->x264_build = build;
274 if (e == 1 && build == 1 && !strncmp(user_data+16, "x264 - core 0000", 16))
275 h->x264_build = 67;
276
277 return 0;
278 }
279
decode_recovery_point(H264SEIRecoveryPoint * h,GetBitContext * gb,void * logctx)280 static int decode_recovery_point(H264SEIRecoveryPoint *h, GetBitContext *gb, void *logctx)
281 {
282 unsigned recovery_frame_cnt = get_ue_golomb_long(gb);
283
284 if (recovery_frame_cnt >= (1<<MAX_LOG2_MAX_FRAME_NUM)) {
285 av_log(logctx, AV_LOG_ERROR, "recovery_frame_cnt %u is out of range\n", recovery_frame_cnt);
286 return AVERROR_INVALIDDATA;
287 }
288
289 h->recovery_frame_cnt = recovery_frame_cnt;
290 /* 1b exact_match_flag,
291 * 1b broken_link_flag,
292 * 2b changing_slice_group_idc */
293 skip_bits(gb, 4);
294
295 return 0;
296 }
297
decode_buffering_period(H264SEIBufferingPeriod * h,GetBitContext * gb,const H264ParamSets * ps,void * logctx)298 static int decode_buffering_period(H264SEIBufferingPeriod *h, GetBitContext *gb,
299 const H264ParamSets *ps, void *logctx)
300 {
301 unsigned int sps_id;
302 int sched_sel_idx;
303 const SPS *sps;
304
305 sps_id = get_ue_golomb_31(gb);
306 if (sps_id > 31 || !ps->sps_list[sps_id]) {
307 av_log(logctx, AV_LOG_ERROR,
308 "non-existing SPS %d referenced in buffering period\n", sps_id);
309 return sps_id > 31 ? AVERROR_INVALIDDATA : AVERROR_PS_NOT_FOUND;
310 }
311 sps = (const SPS*)ps->sps_list[sps_id]->data;
312
313 // NOTE: This is really so duplicated in the standard... See H.264, D.1.1
314 if (sps->nal_hrd_parameters_present_flag) {
315 for (sched_sel_idx = 0; sched_sel_idx < sps->cpb_cnt; sched_sel_idx++) {
316 h->initial_cpb_removal_delay[sched_sel_idx] =
317 get_bits_long(gb, sps->initial_cpb_removal_delay_length);
318 // initial_cpb_removal_delay_offset
319 skip_bits(gb, sps->initial_cpb_removal_delay_length);
320 }
321 }
322 if (sps->vcl_hrd_parameters_present_flag) {
323 for (sched_sel_idx = 0; sched_sel_idx < sps->cpb_cnt; sched_sel_idx++) {
324 h->initial_cpb_removal_delay[sched_sel_idx] =
325 get_bits_long(gb, sps->initial_cpb_removal_delay_length);
326 // initial_cpb_removal_delay_offset
327 skip_bits(gb, sps->initial_cpb_removal_delay_length);
328 }
329 }
330
331 h->present = 1;
332 return 0;
333 }
334
decode_frame_packing_arrangement(H264SEIFramePacking * h,GetBitContext * gb)335 static int decode_frame_packing_arrangement(H264SEIFramePacking *h,
336 GetBitContext *gb)
337 {
338 h->arrangement_id = get_ue_golomb_long(gb);
339 h->arrangement_cancel_flag = get_bits1(gb);
340 h->present = !h->arrangement_cancel_flag;
341
342 if (h->present) {
343 h->arrangement_type = get_bits(gb, 7);
344 h->quincunx_sampling_flag = get_bits1(gb);
345 h->content_interpretation_type = get_bits(gb, 6);
346
347 // spatial_flipping_flag, frame0_flipped_flag, field_views_flag
348 skip_bits(gb, 3);
349 h->current_frame_is_frame0_flag = get_bits1(gb);
350 // frame0_self_contained_flag, frame1_self_contained_flag
351 skip_bits(gb, 2);
352
353 if (!h->quincunx_sampling_flag && h->arrangement_type != 5)
354 skip_bits(gb, 16); // frame[01]_grid_position_[xy]
355 skip_bits(gb, 8); // frame_packing_arrangement_reserved_byte
356 h->arrangement_repetition_period = get_ue_golomb_long(gb);
357 }
358 skip_bits1(gb); // frame_packing_arrangement_extension_flag
359
360 return 0;
361 }
362
decode_display_orientation(H264SEIDisplayOrientation * h,GetBitContext * gb)363 static int decode_display_orientation(H264SEIDisplayOrientation *h,
364 GetBitContext *gb)
365 {
366 h->present = !get_bits1(gb);
367
368 if (h->present) {
369 h->hflip = get_bits1(gb); // hor_flip
370 h->vflip = get_bits1(gb); // ver_flip
371
372 h->anticlockwise_rotation = get_bits(gb, 16);
373 get_ue_golomb_long(gb); // display_orientation_repetition_period
374 skip_bits1(gb); // display_orientation_extension_flag
375 }
376
377 return 0;
378 }
379
decode_green_metadata(H264SEIGreenMetaData * h,GetBitContext * gb)380 static int decode_green_metadata(H264SEIGreenMetaData *h, GetBitContext *gb)
381 {
382 h->green_metadata_type = get_bits(gb, 8);
383
384 if (h->green_metadata_type == 0) {
385 h->period_type = get_bits(gb, 8);
386
387 if (h->period_type == 2)
388 h->num_seconds = get_bits(gb, 16);
389 else if (h->period_type == 3)
390 h->num_pictures = get_bits(gb, 16);
391
392 h->percent_non_zero_macroblocks = get_bits(gb, 8);
393 h->percent_intra_coded_macroblocks = get_bits(gb, 8);
394 h->percent_six_tap_filtering = get_bits(gb, 8);
395 h->percent_alpha_point_deblocking_instance = get_bits(gb, 8);
396
397 } else if (h->green_metadata_type == 1) {
398 h->xsd_metric_type = get_bits(gb, 8);
399 h->xsd_metric_value = get_bits(gb, 16);
400 }
401
402 return 0;
403 }
404
decode_alternative_transfer(H264SEIAlternativeTransfer * h,GetBitContext * gb)405 static int decode_alternative_transfer(H264SEIAlternativeTransfer *h,
406 GetBitContext *gb)
407 {
408 h->present = 1;
409 h->preferred_transfer_characteristics = get_bits(gb, 8);
410 return 0;
411 }
412
ff_h264_sei_decode(H264SEIContext * h,GetBitContext * gb,const H264ParamSets * ps,void * logctx)413 int ff_h264_sei_decode(H264SEIContext *h, GetBitContext *gb,
414 const H264ParamSets *ps, void *logctx)
415 {
416 int master_ret = 0;
417
418 while (get_bits_left(gb) > 16 && show_bits(gb, 16)) {
419 GetBitContext gb_payload;
420 int type = 0;
421 unsigned size = 0;
422 int ret = 0;
423
424 do {
425 if (get_bits_left(gb) < 8)
426 return AVERROR_INVALIDDATA;
427 type += show_bits(gb, 8);
428 } while (get_bits(gb, 8) == 255);
429
430 do {
431 if (get_bits_left(gb) < 8)
432 return AVERROR_INVALIDDATA;
433 size += show_bits(gb, 8);
434 } while (get_bits(gb, 8) == 255);
435
436 if (size > get_bits_left(gb) / 8) {
437 av_log(logctx, AV_LOG_ERROR, "SEI type %d size %d truncated at %d\n",
438 type, 8*size, get_bits_left(gb));
439 return AVERROR_INVALIDDATA;
440 }
441
442 ret = init_get_bits8(&gb_payload, gb->buffer + get_bits_count(gb) / 8, size);
443 if (ret < 0)
444 return ret;
445
446 switch (type) {
447 case SEI_TYPE_PIC_TIMING: // Picture timing SEI
448 ret = decode_picture_timing(&h->picture_timing, &gb_payload, logctx);
449 break;
450 case SEI_TYPE_USER_DATA_REGISTERED_ITU_T_T35:
451 ret = decode_registered_user_data(h, &gb_payload, logctx, size);
452 break;
453 case SEI_TYPE_USER_DATA_UNREGISTERED:
454 ret = decode_unregistered_user_data(&h->unregistered, &gb_payload, logctx, size);
455 break;
456 case SEI_TYPE_RECOVERY_POINT:
457 ret = decode_recovery_point(&h->recovery_point, &gb_payload, logctx);
458 break;
459 case SEI_TYPE_BUFFERING_PERIOD:
460 ret = decode_buffering_period(&h->buffering_period, &gb_payload, ps, logctx);
461 break;
462 case SEI_TYPE_FRAME_PACKING_ARRANGEMENT:
463 ret = decode_frame_packing_arrangement(&h->frame_packing, &gb_payload);
464 break;
465 case SEI_TYPE_DISPLAY_ORIENTATION:
466 ret = decode_display_orientation(&h->display_orientation, &gb_payload);
467 break;
468 case SEI_TYPE_GREEN_METADATA:
469 ret = decode_green_metadata(&h->green_metadata, &gb_payload);
470 break;
471 case SEI_TYPE_ALTERNATIVE_TRANSFER_CHARACTERISTICS:
472 ret = decode_alternative_transfer(&h->alternative_transfer, &gb_payload);
473 break;
474 default:
475 av_log(logctx, AV_LOG_DEBUG, "unknown SEI type %d\n", type);
476 }
477 if (ret < 0 && ret != AVERROR_PS_NOT_FOUND)
478 return ret;
479 if (ret < 0)
480 master_ret = ret;
481
482 if (get_bits_left(&gb_payload) < 0) {
483 av_log(logctx, AV_LOG_WARNING, "SEI type %d overread by %d bits\n",
484 type, -get_bits_left(&gb_payload));
485 }
486
487 skip_bits_long(gb, 8 * size);
488 }
489
490 return master_ret;
491 }
492
ff_h264_sei_stereo_mode(const H264SEIFramePacking * h)493 const char *ff_h264_sei_stereo_mode(const H264SEIFramePacking *h)
494 {
495 if (h->arrangement_cancel_flag == 0) {
496 switch (h->arrangement_type) {
497 case H264_SEI_FPA_TYPE_CHECKERBOARD:
498 if (h->content_interpretation_type == 2)
499 return "checkerboard_rl";
500 else
501 return "checkerboard_lr";
502 case H264_SEI_FPA_TYPE_INTERLEAVE_COLUMN:
503 if (h->content_interpretation_type == 2)
504 return "col_interleaved_rl";
505 else
506 return "col_interleaved_lr";
507 case H264_SEI_FPA_TYPE_INTERLEAVE_ROW:
508 if (h->content_interpretation_type == 2)
509 return "row_interleaved_rl";
510 else
511 return "row_interleaved_lr";
512 case H264_SEI_FPA_TYPE_SIDE_BY_SIDE:
513 if (h->content_interpretation_type == 2)
514 return "right_left";
515 else
516 return "left_right";
517 case H264_SEI_FPA_TYPE_TOP_BOTTOM:
518 if (h->content_interpretation_type == 2)
519 return "bottom_top";
520 else
521 return "top_bottom";
522 case H264_SEI_FPA_TYPE_INTERLEAVE_TEMPORAL:
523 if (h->content_interpretation_type == 2)
524 return "block_rl";
525 else
526 return "block_lr";
527 case H264_SEI_FPA_TYPE_2D:
528 default:
529 return "mono";
530 }
531 } else if (h->arrangement_cancel_flag == 1) {
532 return "mono";
533 } else {
534 return NULL;
535 }
536 }
537