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>
35 #define AVERROR_PS_NOT_FOUND FFERRTAG(0xF8,'?','P','S')
37 static const uint8_t sei_num_clock_ts_table[9] = {
38 1, 1, 1, 2, 2, 3, 3, 2, 3
41 void ff_h264_sei_uninit(H264SEIContext *h)
43 h->recovery_point.recovery_frame_cnt = -1;
45 h->picture_timing.dpb_output_delay = 0;
46 h->picture_timing.cpb_removal_delay = -1;
48 h->picture_timing.present = 0;
49 h->buffering_period.present = 0;
50 h->frame_packing.present = 0;
51 h->display_orientation.present = 0;
54 av_buffer_unref(&h->a53_caption.buf_ref);
57 static int decode_picture_timing(H264SEIPictureTiming *h, GetBitContext *gb,
58 const H264ParamSets *ps, void *logctx)
61 const SPS *sps = ps->sps;
63 for (i = 0; i<MAX_SPS_COUNT; i++)
64 if ((!sps || !sps->log2_max_frame_num) && ps->sps_list[i])
65 sps = (const SPS *)ps->sps_list[i]->data;
68 av_log(logctx, AV_LOG_ERROR, "SPS unavailable in decode_picture_timing\n");
69 return AVERROR_PS_NOT_FOUND;
72 if (sps->nal_hrd_parameters_present_flag ||
73 sps->vcl_hrd_parameters_present_flag) {
74 h->cpb_removal_delay = get_bits_long(gb, sps->cpb_removal_delay_length);
75 h->dpb_output_delay = get_bits_long(gb, sps->dpb_output_delay_length);
77 if (sps->pic_struct_present_flag) {
78 unsigned int i, num_clock_ts;
80 h->pic_struct = get_bits(gb, 4);
83 if (h->pic_struct > H264_SEI_PIC_STRUCT_FRAME_TRIPLING)
84 return AVERROR_INVALIDDATA;
86 num_clock_ts = sei_num_clock_ts_table[h->pic_struct];
88 for (i = 0; i < num_clock_ts; i++) {
89 if (get_bits(gb, 1)) { /* clock_timestamp_flag */
90 H264SEITimeCode *tc = &h->timecode[h->timecode_cnt++];
91 unsigned int full_timestamp_flag;
92 unsigned int counting_type, cnt_dropped_flag;
93 h->ct_type |= 1 << get_bits(gb, 2);
94 skip_bits(gb, 1); /* nuit_field_based_flag */
95 counting_type = get_bits(gb, 5); /* counting_type */
96 full_timestamp_flag = get_bits(gb, 1);
97 skip_bits(gb, 1); /* discontinuity_flag */
98 cnt_dropped_flag = get_bits(gb, 1); /* cnt_dropped_flag */
99 if (cnt_dropped_flag && counting_type > 1 && counting_type < 7)
101 tc->frame = get_bits(gb, 8); /* n_frames */
102 if (full_timestamp_flag) {
104 tc->seconds = get_bits(gb, 6); /* seconds_value 0..59 */
105 tc->minutes = get_bits(gb, 6); /* minutes_value 0..59 */
106 tc->hours = get_bits(gb, 5); /* hours_value 0..23 */
108 tc->seconds = tc->minutes = tc->hours = tc->full = 0;
109 if (get_bits(gb, 1)) { /* seconds_flag */
110 tc->seconds = get_bits(gb, 6);
111 if (get_bits(gb, 1)) { /* minutes_flag */
112 tc->minutes = get_bits(gb, 6);
113 if (get_bits(gb, 1)) /* hours_flag */
114 tc->hours = get_bits(gb, 5);
119 if (sps->time_offset_length > 0)
121 sps->time_offset_length); /* time_offset */
125 av_log(logctx, AV_LOG_DEBUG, "ct_type:%X pic_struct:%d\n",
126 h->ct_type, h->pic_struct);
133 static int decode_registered_user_data_afd(H264SEIAFD *h, GetBitContext *gb, int size)
138 return AVERROR_INVALIDDATA;
139 skip_bits(gb, 1); // 0
140 flag = get_bits(gb, 1); // active_format_flag
141 skip_bits(gb, 6); // reserved
145 return AVERROR_INVALIDDATA;
146 skip_bits(gb, 4); // reserved
147 h->active_format_description = get_bits(gb, 4);
154 static int decode_registered_user_data_closed_caption(H264SEIA53Caption *h,
155 GetBitContext *gb, void *logctx,
159 int user_data_type_code;
163 return AVERROR(EINVAL);
165 user_data_type_code = get_bits(gb, 8);
166 if (user_data_type_code == 0x3) {
167 skip_bits(gb, 1); // reserved
169 flag = get_bits(gb, 1); // process_cc_data_flag
171 skip_bits(gb, 1); // zero bit
172 cc_count = get_bits(gb, 5);
173 skip_bits(gb, 8); // reserved
176 if (cc_count && size >= cc_count * 3) {
177 int old_size = h->buf_ref ? h->buf_ref->size : 0;
178 const uint64_t new_size = (old_size + cc_count
182 if (new_size > INT_MAX)
183 return AVERROR(EINVAL);
185 /* Allow merging of the cc data from two fields. */
186 ret = av_buffer_realloc(&h->buf_ref, new_size);
190 /* Use of av_buffer_realloc assumes buffer is writeable */
191 for (i = 0; i < cc_count; i++) {
192 h->buf_ref->data[old_size++] = get_bits(gb, 8);
193 h->buf_ref->data[old_size++] = get_bits(gb, 8);
194 h->buf_ref->data[old_size++] = get_bits(gb, 8);
197 skip_bits(gb, 8); // marker_bits
202 for (i = 0; i < size - 1; i++)
209 static int decode_registered_user_data(H264SEIContext *h, GetBitContext *gb,
210 void *logctx, int size)
212 uint32_t country_code;
213 uint32_t user_identifier;
216 return AVERROR_INVALIDDATA;
219 country_code = get_bits(gb, 8); // itu_t_t35_country_code
220 if (country_code == 0xFF) {
221 skip_bits(gb, 8); // itu_t_t35_country_code_extension_byte
225 /* itu_t_t35_payload_byte follows */
226 skip_bits(gb, 8); // terminal provider code
227 skip_bits(gb, 8); // terminal provider oriented code
228 user_identifier = get_bits_long(gb, 32);
230 switch (user_identifier) {
231 case MKBETAG('D', 'T', 'G', '1'): // afd_data
232 return decode_registered_user_data_afd(&h->afd, gb, size);
233 case MKBETAG('G', 'A', '9', '4'): // closed captions
234 return decode_registered_user_data_closed_caption(&h->a53_caption, gb,
237 skip_bits(gb, size * 8);
244 static int decode_unregistered_user_data(H264SEIUnregistered *h, GetBitContext *gb,
245 void *logctx, int size)
250 if (size < 16 || size >= INT_MAX - 16)
251 return AVERROR_INVALIDDATA;
253 user_data = av_malloc(16 + size + 1);
255 return AVERROR(ENOMEM);
257 for (i = 0; i < size + 16; i++)
258 user_data[i] = get_bits(gb, 8);
261 e = sscanf(user_data + 16, "x264 - core %d", &build);
262 if (e == 1 && build > 0)
263 h->x264_build = build;
264 if (e == 1 && build == 1 && !strncmp(user_data+16, "x264 - core 0000", 16))
271 static int decode_recovery_point(H264SEIRecoveryPoint *h, GetBitContext *gb, void *logctx)
273 unsigned recovery_frame_cnt = get_ue_golomb_long(gb);
275 if (recovery_frame_cnt >= (1<<MAX_LOG2_MAX_FRAME_NUM)) {
276 av_log(logctx, AV_LOG_ERROR, "recovery_frame_cnt %u is out of range\n", recovery_frame_cnt);
277 return AVERROR_INVALIDDATA;
280 h->recovery_frame_cnt = recovery_frame_cnt;
281 /* 1b exact_match_flag,
282 * 1b broken_link_flag,
283 * 2b changing_slice_group_idc */
289 static int decode_buffering_period(H264SEIBufferingPeriod *h, GetBitContext *gb,
290 const H264ParamSets *ps, void *logctx)
296 sps_id = get_ue_golomb_31(gb);
297 if (sps_id > 31 || !ps->sps_list[sps_id]) {
298 av_log(logctx, AV_LOG_ERROR,
299 "non-existing SPS %d referenced in buffering period\n", sps_id);
300 return sps_id > 31 ? AVERROR_INVALIDDATA : AVERROR_PS_NOT_FOUND;
302 sps = (const SPS*)ps->sps_list[sps_id]->data;
304 // NOTE: This is really so duplicated in the standard... See H.264, D.1.1
305 if (sps->nal_hrd_parameters_present_flag) {
306 for (sched_sel_idx = 0; sched_sel_idx < sps->cpb_cnt; sched_sel_idx++) {
307 h->initial_cpb_removal_delay[sched_sel_idx] =
308 get_bits_long(gb, sps->initial_cpb_removal_delay_length);
309 // initial_cpb_removal_delay_offset
310 skip_bits(gb, sps->initial_cpb_removal_delay_length);
313 if (sps->vcl_hrd_parameters_present_flag) {
314 for (sched_sel_idx = 0; sched_sel_idx < sps->cpb_cnt; sched_sel_idx++) {
315 h->initial_cpb_removal_delay[sched_sel_idx] =
316 get_bits_long(gb, sps->initial_cpb_removal_delay_length);
317 // initial_cpb_removal_delay_offset
318 skip_bits(gb, sps->initial_cpb_removal_delay_length);
326 static int decode_frame_packing_arrangement(H264SEIFramePacking *h,
329 h->arrangement_id = get_ue_golomb_long(gb);
330 h->arrangement_cancel_flag = get_bits1(gb);
331 h->present = !h->arrangement_cancel_flag;
334 h->arrangement_type = get_bits(gb, 7);
335 h->quincunx_sampling_flag = get_bits1(gb);
336 h->content_interpretation_type = get_bits(gb, 6);
338 // spatial_flipping_flag, frame0_flipped_flag, field_views_flag
340 h->current_frame_is_frame0_flag = get_bits1(gb);
341 // frame0_self_contained_flag, frame1_self_contained_flag
344 if (!h->quincunx_sampling_flag && h->arrangement_type != 5)
345 skip_bits(gb, 16); // frame[01]_grid_position_[xy]
346 skip_bits(gb, 8); // frame_packing_arrangement_reserved_byte
347 h->arrangement_repetition_period = get_ue_golomb_long(gb);
349 skip_bits1(gb); // frame_packing_arrangement_extension_flag
354 static int decode_display_orientation(H264SEIDisplayOrientation *h,
357 h->present = !get_bits1(gb);
360 h->hflip = get_bits1(gb); // hor_flip
361 h->vflip = get_bits1(gb); // ver_flip
363 h->anticlockwise_rotation = get_bits(gb, 16);
364 get_ue_golomb_long(gb); // display_orientation_repetition_period
365 skip_bits1(gb); // display_orientation_extension_flag
371 static int decode_green_metadata(H264SEIGreenMetaData *h, GetBitContext *gb)
373 h->green_metadata_type = get_bits(gb, 8);
375 if (h->green_metadata_type == 0) {
376 h->period_type = get_bits(gb, 8);
378 if (h->period_type == 2)
379 h->num_seconds = get_bits(gb, 16);
380 else if (h->period_type == 3)
381 h->num_pictures = get_bits(gb, 16);
383 h->percent_non_zero_macroblocks = get_bits(gb, 8);
384 h->percent_intra_coded_macroblocks = get_bits(gb, 8);
385 h->percent_six_tap_filtering = get_bits(gb, 8);
386 h->percent_alpha_point_deblocking_instance = get_bits(gb, 8);
388 } else if (h->green_metadata_type == 1) {
389 h->xsd_metric_type = get_bits(gb, 8);
390 h->xsd_metric_value = get_bits(gb, 16);
396 static int decode_alternative_transfer(H264SEIAlternativeTransfer *h,
400 h->preferred_transfer_characteristics = get_bits(gb, 8);
404 int ff_h264_sei_decode(H264SEIContext *h, GetBitContext *gb,
405 const H264ParamSets *ps, void *logctx)
409 while (get_bits_left(gb) > 16 && show_bits(gb, 16)) {
416 if (get_bits_left(gb) < 8)
417 return AVERROR_INVALIDDATA;
418 type += show_bits(gb, 8);
419 } while (get_bits(gb, 8) == 255);
422 if (get_bits_left(gb) < 8)
423 return AVERROR_INVALIDDATA;
424 size += show_bits(gb, 8);
425 } while (get_bits(gb, 8) == 255);
427 if (size > get_bits_left(gb) / 8) {
428 av_log(logctx, AV_LOG_ERROR, "SEI type %d size %d truncated at %d\n",
429 type, 8*size, get_bits_left(gb));
430 return AVERROR_INVALIDDATA;
432 next = get_bits_count(gb) + 8 * size;
435 case H264_SEI_TYPE_PIC_TIMING: // Picture timing SEI
436 ret = decode_picture_timing(&h->picture_timing, gb, ps, logctx);
438 case H264_SEI_TYPE_USER_DATA_REGISTERED:
439 ret = decode_registered_user_data(h, gb, logctx, size);
441 case H264_SEI_TYPE_USER_DATA_UNREGISTERED:
442 ret = decode_unregistered_user_data(&h->unregistered, gb, logctx, size);
444 case H264_SEI_TYPE_RECOVERY_POINT:
445 ret = decode_recovery_point(&h->recovery_point, gb, logctx);
447 case H264_SEI_TYPE_BUFFERING_PERIOD:
448 ret = decode_buffering_period(&h->buffering_period, gb, ps, logctx);
450 case H264_SEI_TYPE_FRAME_PACKING:
451 ret = decode_frame_packing_arrangement(&h->frame_packing, gb);
453 case H264_SEI_TYPE_DISPLAY_ORIENTATION:
454 ret = decode_display_orientation(&h->display_orientation, gb);
456 case H264_SEI_TYPE_GREEN_METADATA:
457 ret = decode_green_metadata(&h->green_metadata, gb);
459 case H264_SEI_TYPE_ALTERNATIVE_TRANSFER:
460 ret = decode_alternative_transfer(&h->alternative_transfer, gb);
463 av_log(logctx, AV_LOG_DEBUG, "unknown SEI type %d\n", type);
465 if (ret < 0 && ret != AVERROR_PS_NOT_FOUND)
470 skip_bits_long(gb, next - get_bits_count(gb));
472 // FIXME check bits here
479 const char *ff_h264_sei_stereo_mode(const H264SEIFramePacking *h)
481 if (h->arrangement_cancel_flag == 0) {
482 switch (h->arrangement_type) {
483 case H264_SEI_FPA_TYPE_CHECKERBOARD:
484 if (h->content_interpretation_type == 2)
485 return "checkerboard_rl";
487 return "checkerboard_lr";
488 case H264_SEI_FPA_TYPE_INTERLEAVE_COLUMN:
489 if (h->content_interpretation_type == 2)
490 return "col_interleaved_rl";
492 return "col_interleaved_lr";
493 case H264_SEI_FPA_TYPE_INTERLEAVE_ROW:
494 if (h->content_interpretation_type == 2)
495 return "row_interleaved_rl";
497 return "row_interleaved_lr";
498 case H264_SEI_FPA_TYPE_SIDE_BY_SIDE:
499 if (h->content_interpretation_type == 2)
503 case H264_SEI_FPA_TYPE_TOP_BOTTOM:
504 if (h->content_interpretation_type == 2)
508 case H264_SEI_FPA_TYPE_INTERLEAVE_TEMPORAL:
509 if (h->content_interpretation_type == 2)
513 case H264_SEI_FPA_TYPE_2D:
517 } else if (h->arrangement_cancel_flag == 1) {