2 * H.26L/H.264/AVC/JVT/14496-10/... SEI decoding
3 * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
5 * This file is part of FFmpeg.
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.
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.
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
24 * H.264 / AVC / MPEG-4 part10 SEI decoding.
25 * @author Michael Niedermayer <michaelni@gmx.at>
36 #define AVERROR_PS_NOT_FOUND FFERRTAG(0xF8,'?','P','S')
38 static const uint8_t sei_num_clock_ts_table[9] = {
39 1, 1, 1, 2, 2, 3, 3, 2, 3
42 void ff_h264_sei_uninit(H264SEIContext *h)
44 h->recovery_point.recovery_frame_cnt = -1;
46 h->picture_timing.dpb_output_delay = 0;
47 h->picture_timing.cpb_removal_delay = -1;
49 h->picture_timing.present = 0;
50 h->buffering_period.present = 0;
51 h->frame_packing.present = 0;
52 h->display_orientation.present = 0;
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);
62 int ff_h264_sei_process_picture_timing(H264SEIPictureTiming *h, const SPS *sps,
67 init_get_bits(&gb, h->payload, h->payload_size_bits);
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);
74 if (sps->pic_struct_present_flag) {
75 unsigned int i, num_clock_ts;
77 h->pic_struct = get_bits(&gb, 4);
80 if (h->pic_struct > H264_SEI_PIC_STRUCT_FRAME_TRIPLING)
81 return AVERROR_INVALIDDATA;
83 num_clock_ts = sei_num_clock_ts_table[h->pic_struct];
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)
98 tc->frame = get_bits(&gb, 8); /* n_frames */
99 if (full_timestamp_flag) {
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 */
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);
116 if (sps->time_offset_length > 0)
118 sps->time_offset_length); /* time_offset */
122 av_log(logctx, AV_LOG_DEBUG, "ct_type:%X pic_struct:%d\n",
123 h->ct_type, h->pic_struct);
129 static int decode_picture_timing(H264SEIPictureTiming *h, GetBitContext *gb,
132 int index = get_bits_count(gb);
133 int size_bits = get_bits_left(gb);
134 int size = (size_bits + 7) / 8;
137 av_log(logctx, AV_LOG_ERROR, "Unaligned SEI payload\n");
138 return AVERROR_INVALIDDATA;
140 if (size > sizeof(h->payload)) {
141 av_log(logctx, AV_LOG_ERROR, "Picture timing SEI payload too large\n");
142 return AVERROR_INVALIDDATA;
144 memcpy(h->payload, gb->buffer + index / 8, size);
146 h->payload_size_bits = size_bits;
152 static int decode_registered_user_data_afd(H264SEIAFD *h, GetBitContext *gb, int size)
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
164 return AVERROR_INVALIDDATA;
165 skip_bits(gb, 4); // reserved
166 h->active_format_description = get_bits(gb, 4);
173 static int decode_registered_user_data_closed_caption(H264SEIA53Caption *h,
174 GetBitContext *gb, void *logctx,
178 return AVERROR(EINVAL);
180 return ff_parse_a53_cc(&h->buf_ref, gb->buffer + get_bits_count(gb) / 8, size);
183 static int decode_registered_user_data(H264SEIContext *h, GetBitContext *gb,
184 void *logctx, int size)
186 uint32_t country_code;
187 uint32_t user_identifier;
190 return AVERROR_INVALIDDATA;
193 country_code = get_bits(gb, 8); // itu_t_t35_country_code
194 if (country_code == 0xFF) {
195 skip_bits(gb, 8); // itu_t_t35_country_code_extension_byte
199 /* itu_t_t35_payload_byte follows */
200 skip_bits(gb, 8); // terminal provider code
201 skip_bits(gb, 8); // terminal provider oriented code
202 user_identifier = get_bits_long(gb, 32);
204 switch (user_identifier) {
205 case MKBETAG('D', 'T', 'G', '1'): // afd_data
206 return decode_registered_user_data_afd(&h->afd, gb, size);
207 case MKBETAG('G', 'A', '9', '4'): // closed captions
208 return decode_registered_user_data_closed_caption(&h->a53_caption, gb,
211 skip_bits(gb, size * 8);
218 static int decode_unregistered_user_data(H264SEIUnregistered *h, GetBitContext *gb,
219 void *logctx, int size)
223 AVBufferRef *buf_ref, **tmp;
225 if (size < 16 || size >= INT_MAX - 1)
226 return AVERROR_INVALIDDATA;
228 tmp = av_realloc_array(h->buf_ref, h->nb_buf_ref + 1, sizeof(*h->buf_ref));
230 return AVERROR(ENOMEM);
233 buf_ref = av_buffer_alloc(size + 1);
235 return AVERROR(ENOMEM);
236 user_data = buf_ref->data;
238 for (i = 0; i < size; i++)
239 user_data[i] = get_bits(gb, 8);
242 buf_ref->size = size;
243 h->buf_ref[h->nb_buf_ref++] = buf_ref;
245 e = sscanf(user_data + 16, "x264 - core %d", &build);
246 if (e == 1 && build > 0)
247 h->x264_build = build;
248 if (e == 1 && build == 1 && !strncmp(user_data+16, "x264 - core 0000", 16))
254 static int decode_recovery_point(H264SEIRecoveryPoint *h, GetBitContext *gb, void *logctx)
256 unsigned recovery_frame_cnt = get_ue_golomb_long(gb);
258 if (recovery_frame_cnt >= (1<<MAX_LOG2_MAX_FRAME_NUM)) {
259 av_log(logctx, AV_LOG_ERROR, "recovery_frame_cnt %u is out of range\n", recovery_frame_cnt);
260 return AVERROR_INVALIDDATA;
263 h->recovery_frame_cnt = recovery_frame_cnt;
264 /* 1b exact_match_flag,
265 * 1b broken_link_flag,
266 * 2b changing_slice_group_idc */
272 static int decode_buffering_period(H264SEIBufferingPeriod *h, GetBitContext *gb,
273 const H264ParamSets *ps, void *logctx)
279 sps_id = get_ue_golomb_31(gb);
280 if (sps_id > 31 || !ps->sps_list[sps_id]) {
281 av_log(logctx, AV_LOG_ERROR,
282 "non-existing SPS %d referenced in buffering period\n", sps_id);
283 return sps_id > 31 ? AVERROR_INVALIDDATA : AVERROR_PS_NOT_FOUND;
285 sps = (const SPS*)ps->sps_list[sps_id]->data;
287 // NOTE: This is really so duplicated in the standard... See H.264, D.1.1
288 if (sps->nal_hrd_parameters_present_flag) {
289 for (sched_sel_idx = 0; sched_sel_idx < sps->cpb_cnt; sched_sel_idx++) {
290 h->initial_cpb_removal_delay[sched_sel_idx] =
291 get_bits_long(gb, sps->initial_cpb_removal_delay_length);
292 // initial_cpb_removal_delay_offset
293 skip_bits(gb, sps->initial_cpb_removal_delay_length);
296 if (sps->vcl_hrd_parameters_present_flag) {
297 for (sched_sel_idx = 0; sched_sel_idx < sps->cpb_cnt; sched_sel_idx++) {
298 h->initial_cpb_removal_delay[sched_sel_idx] =
299 get_bits_long(gb, sps->initial_cpb_removal_delay_length);
300 // initial_cpb_removal_delay_offset
301 skip_bits(gb, sps->initial_cpb_removal_delay_length);
309 static int decode_frame_packing_arrangement(H264SEIFramePacking *h,
312 h->arrangement_id = get_ue_golomb_long(gb);
313 h->arrangement_cancel_flag = get_bits1(gb);
314 h->present = !h->arrangement_cancel_flag;
317 h->arrangement_type = get_bits(gb, 7);
318 h->quincunx_sampling_flag = get_bits1(gb);
319 h->content_interpretation_type = get_bits(gb, 6);
321 // spatial_flipping_flag, frame0_flipped_flag, field_views_flag
323 h->current_frame_is_frame0_flag = get_bits1(gb);
324 // frame0_self_contained_flag, frame1_self_contained_flag
327 if (!h->quincunx_sampling_flag && h->arrangement_type != 5)
328 skip_bits(gb, 16); // frame[01]_grid_position_[xy]
329 skip_bits(gb, 8); // frame_packing_arrangement_reserved_byte
330 h->arrangement_repetition_period = get_ue_golomb_long(gb);
332 skip_bits1(gb); // frame_packing_arrangement_extension_flag
337 static int decode_display_orientation(H264SEIDisplayOrientation *h,
340 h->present = !get_bits1(gb);
343 h->hflip = get_bits1(gb); // hor_flip
344 h->vflip = get_bits1(gb); // ver_flip
346 h->anticlockwise_rotation = get_bits(gb, 16);
347 get_ue_golomb_long(gb); // display_orientation_repetition_period
348 skip_bits1(gb); // display_orientation_extension_flag
354 static int decode_green_metadata(H264SEIGreenMetaData *h, GetBitContext *gb)
356 h->green_metadata_type = get_bits(gb, 8);
358 if (h->green_metadata_type == 0) {
359 h->period_type = get_bits(gb, 8);
361 if (h->period_type == 2)
362 h->num_seconds = get_bits(gb, 16);
363 else if (h->period_type == 3)
364 h->num_pictures = get_bits(gb, 16);
366 h->percent_non_zero_macroblocks = get_bits(gb, 8);
367 h->percent_intra_coded_macroblocks = get_bits(gb, 8);
368 h->percent_six_tap_filtering = get_bits(gb, 8);
369 h->percent_alpha_point_deblocking_instance = get_bits(gb, 8);
371 } else if (h->green_metadata_type == 1) {
372 h->xsd_metric_type = get_bits(gb, 8);
373 h->xsd_metric_value = get_bits(gb, 16);
379 static int decode_alternative_transfer(H264SEIAlternativeTransfer *h,
383 h->preferred_transfer_characteristics = get_bits(gb, 8);
387 int ff_h264_sei_decode(H264SEIContext *h, GetBitContext *gb,
388 const H264ParamSets *ps, void *logctx)
392 while (get_bits_left(gb) > 16 && show_bits(gb, 16)) {
393 GetBitContext gb_payload;
399 if (get_bits_left(gb) < 8)
400 return AVERROR_INVALIDDATA;
401 type += show_bits(gb, 8);
402 } while (get_bits(gb, 8) == 255);
405 if (get_bits_left(gb) < 8)
406 return AVERROR_INVALIDDATA;
407 size += show_bits(gb, 8);
408 } while (get_bits(gb, 8) == 255);
410 if (size > get_bits_left(gb) / 8) {
411 av_log(logctx, AV_LOG_ERROR, "SEI type %d size %d truncated at %d\n",
412 type, 8*size, get_bits_left(gb));
413 return AVERROR_INVALIDDATA;
416 ret = init_get_bits8(&gb_payload, gb->buffer + get_bits_count(gb) / 8, size);
421 case H264_SEI_TYPE_PIC_TIMING: // Picture timing SEI
422 ret = decode_picture_timing(&h->picture_timing, &gb_payload, logctx);
424 case H264_SEI_TYPE_USER_DATA_REGISTERED:
425 ret = decode_registered_user_data(h, &gb_payload, logctx, size);
427 case H264_SEI_TYPE_USER_DATA_UNREGISTERED:
428 ret = decode_unregistered_user_data(&h->unregistered, &gb_payload, logctx, size);
430 case H264_SEI_TYPE_RECOVERY_POINT:
431 ret = decode_recovery_point(&h->recovery_point, &gb_payload, logctx);
433 case H264_SEI_TYPE_BUFFERING_PERIOD:
434 ret = decode_buffering_period(&h->buffering_period, &gb_payload, ps, logctx);
436 case H264_SEI_TYPE_FRAME_PACKING:
437 ret = decode_frame_packing_arrangement(&h->frame_packing, &gb_payload);
439 case H264_SEI_TYPE_DISPLAY_ORIENTATION:
440 ret = decode_display_orientation(&h->display_orientation, &gb_payload);
442 case H264_SEI_TYPE_GREEN_METADATA:
443 ret = decode_green_metadata(&h->green_metadata, &gb_payload);
445 case H264_SEI_TYPE_ALTERNATIVE_TRANSFER:
446 ret = decode_alternative_transfer(&h->alternative_transfer, &gb_payload);
449 av_log(logctx, AV_LOG_DEBUG, "unknown SEI type %d\n", type);
451 if (ret < 0 && ret != AVERROR_PS_NOT_FOUND)
456 if (get_bits_left(&gb_payload) < 0) {
457 av_log(logctx, AV_LOG_WARNING, "SEI type %d overread by %d bits\n",
458 type, -get_bits_left(&gb_payload));
461 skip_bits_long(gb, 8 * size);
467 const char *ff_h264_sei_stereo_mode(const H264SEIFramePacking *h)
469 if (h->arrangement_cancel_flag == 0) {
470 switch (h->arrangement_type) {
471 case H264_SEI_FPA_TYPE_CHECKERBOARD:
472 if (h->content_interpretation_type == 2)
473 return "checkerboard_rl";
475 return "checkerboard_lr";
476 case H264_SEI_FPA_TYPE_INTERLEAVE_COLUMN:
477 if (h->content_interpretation_type == 2)
478 return "col_interleaved_rl";
480 return "col_interleaved_lr";
481 case H264_SEI_FPA_TYPE_INTERLEAVE_ROW:
482 if (h->content_interpretation_type == 2)
483 return "row_interleaved_rl";
485 return "row_interleaved_lr";
486 case H264_SEI_FPA_TYPE_SIDE_BY_SIDE:
487 if (h->content_interpretation_type == 2)
491 case H264_SEI_FPA_TYPE_TOP_BOTTOM:
492 if (h->content_interpretation_type == 2)
496 case H264_SEI_FPA_TYPE_INTERLEAVE_TEMPORAL:
497 if (h->content_interpretation_type == 2)
501 case H264_SEI_FPA_TYPE_2D:
505 } else if (h->arrangement_cancel_flag == 1) {