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 Libav.
7 * Libav 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 * Libav 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 Libav; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
24 * H.264 / AVC / MPEG4 part10 sei decoding.
25 * @author Michael Niedermayer <michaelni@gmx.at>
33 static const uint8_t sei_num_clock_ts_table[9] = {
34 1, 1, 1, 2, 2, 3, 3, 2, 3
37 void ff_h264_reset_sei(H264Context *h)
39 h->sei_recovery_frame_cnt = -1;
40 h->sei_dpb_output_delay = 0;
41 h->sei_cpb_removal_delay = -1;
42 h->sei_buffering_period_present = 0;
43 h->sei_frame_packing_present = 0;
44 h->sei_display_orientation_present = 0;
45 h->sei_reguserdata_afd_present = 0;
47 h->a53_caption_size = 0;
48 av_freep(&h->a53_caption);
51 static int decode_picture_timing(H264Context *h)
53 if (h->sps.nal_hrd_parameters_present_flag ||
54 h->sps.vcl_hrd_parameters_present_flag) {
55 h->sei_cpb_removal_delay = get_bits(&h->gb,
56 h->sps.cpb_removal_delay_length);
57 h->sei_dpb_output_delay = get_bits(&h->gb,
58 h->sps.dpb_output_delay_length);
60 if (h->sps.pic_struct_present_flag) {
61 unsigned int i, num_clock_ts;
63 h->sei_pic_struct = get_bits(&h->gb, 4);
66 if (h->sei_pic_struct > SEI_PIC_STRUCT_FRAME_TRIPLING)
67 return AVERROR_INVALIDDATA;
69 num_clock_ts = sei_num_clock_ts_table[h->sei_pic_struct];
71 for (i = 0; i < num_clock_ts; i++) {
72 if (get_bits(&h->gb, 1)) { /* clock_timestamp_flag */
73 unsigned int full_timestamp_flag;
75 h->sei_ct_type |= 1 << get_bits(&h->gb, 2);
76 skip_bits(&h->gb, 1); /* nuit_field_based_flag */
77 skip_bits(&h->gb, 5); /* counting_type */
78 full_timestamp_flag = get_bits(&h->gb, 1);
79 skip_bits(&h->gb, 1); /* discontinuity_flag */
80 skip_bits(&h->gb, 1); /* cnt_dropped_flag */
81 skip_bits(&h->gb, 8); /* n_frames */
82 if (full_timestamp_flag) {
83 skip_bits(&h->gb, 6); /* seconds_value 0..59 */
84 skip_bits(&h->gb, 6); /* minutes_value 0..59 */
85 skip_bits(&h->gb, 5); /* hours_value 0..23 */
87 if (get_bits(&h->gb, 1)) { /* seconds_flag */
88 skip_bits(&h->gb, 6); /* seconds_value range 0..59 */
89 if (get_bits(&h->gb, 1)) { /* minutes_flag */
90 skip_bits(&h->gb, 6); /* minutes_value 0..59 */
91 if (get_bits(&h->gb, 1)) /* hours_flag */
92 skip_bits(&h->gb, 5); /* hours_value 0..23 */
96 if (h->sps.time_offset_length > 0)
98 h->sps.time_offset_length); /* time_offset */
102 if (h->avctx->debug & FF_DEBUG_PICT_INFO)
103 av_log(h->avctx, AV_LOG_DEBUG, "ct_type:%X pic_struct:%d\n",
104 h->sei_ct_type, h->sei_pic_struct);
109 static int decode_registered_user_data_afd(H264Context *h, int size)
114 return AVERROR_INVALIDDATA;
115 skip_bits(&h->gb, 1); // 0
116 flag = get_bits(&h->gb, 1); // active_format_flag
117 skip_bits(&h->gb, 6); // reserved
121 return AVERROR_INVALIDDATA;
122 skip_bits(&h->gb, 4); // reserved
123 h->active_format_description = get_bits(&h->gb, 4);
124 h->sei_reguserdata_afd_present = 1;
130 static int decode_registered_user_data_closed_caption(H264Context *h, int size)
133 int user_data_type_code;
137 return AVERROR(EINVAL);
139 user_data_type_code = get_bits(&h->gb, 8);
140 if (user_data_type_code == 0x3) {
141 skip_bits(&h->gb, 1); // reserved
143 flag = get_bits(&h->gb, 1); // process_cc_data_flag
145 skip_bits(&h->gb, 1); // zero bit
146 cc_count = get_bits(&h->gb, 5);
147 skip_bits(&h->gb, 8); // reserved
150 if (cc_count && size >= cc_count * 3) {
151 const uint64_t new_size = (h->a53_caption_size + cc_count
155 if (new_size > INT_MAX)
156 return AVERROR(EINVAL);
158 /* Allow merging of the cc data from two fields. */
159 ret = av_reallocp(&h->a53_caption, new_size);
163 for (i = 0; i < cc_count; i++) {
164 h->a53_caption[h->a53_caption_size++] = get_bits(&h->gb, 8);
165 h->a53_caption[h->a53_caption_size++] = get_bits(&h->gb, 8);
166 h->a53_caption[h->a53_caption_size++] = get_bits(&h->gb, 8);
169 skip_bits(&h->gb, 8); // marker_bits
174 avpriv_request_sample("Subtitles with data type 0x%02x",
175 user_data_type_code);
176 for (i = 0; i < size - 1; i++)
177 skip_bits(&h->gb, 8);
183 static int decode_registered_user_data(H264Context *h, int size)
185 uint32_t country_code;
186 uint32_t user_identifier;
189 return AVERROR_INVALIDDATA;
192 country_code = get_bits(&h->gb, 8); // itu_t_t35_country_code
193 if (country_code == 0xFF) {
194 skip_bits(&h->gb, 8); // itu_t_t35_country_code_extension_byte
198 /* itu_t_t35_payload_byte follows */
199 skip_bits(&h->gb, 8); // terminal provider code
200 skip_bits(&h->gb, 8); // terminal provider oriented code
201 user_identifier = get_bits_long(&h->gb, 32);
203 switch (user_identifier) {
204 case MKBETAG('D', 'T', 'G', '1'): // afd_data
205 return decode_registered_user_data_afd(h, size);
206 case MKBETAG('G', 'A', '9', '4'): // closed captions
207 return decode_registered_user_data_closed_caption(h, size);
209 skip_bits(&h->gb, size * 8);
216 static int decode_unregistered_user_data(H264Context *h, int size)
218 uint8_t user_data[16 + 256];
222 return AVERROR_INVALIDDATA;
224 for (i = 0; i < sizeof(user_data) - 1 && i < size; i++)
225 user_data[i] = get_bits(&h->gb, 8);
228 e = sscanf(user_data + 16, "x264 - core %d", &build);
229 if (e == 1 && build > 0)
230 h->x264_build = build;
232 if (h->avctx->debug & FF_DEBUG_BUGS)
233 av_log(h->avctx, AV_LOG_DEBUG, "user data:\"%s\"\n", user_data + 16);
235 for (; i < size; i++)
236 skip_bits(&h->gb, 8);
241 static int decode_recovery_point(H264Context *h)
243 h->sei_recovery_frame_cnt = get_ue_golomb(&h->gb);
245 /* 1b exact_match_flag,
246 * 1b broken_link_flag,
247 * 2b changing_slice_group_idc */
248 skip_bits(&h->gb, 4);
253 static int decode_buffering_period(H264Context *h)
259 sps_id = get_ue_golomb_31(&h->gb);
260 if (sps_id > 31 || !h->sps_buffers[sps_id]) {
261 av_log(h->avctx, AV_LOG_ERROR,
262 "non-existing SPS %d referenced in buffering period\n", sps_id);
263 return AVERROR_INVALIDDATA;
265 sps = h->sps_buffers[sps_id];
267 // NOTE: This is really so duplicated in the standard... See H.264, D.1.1
268 if (sps->nal_hrd_parameters_present_flag) {
269 for (sched_sel_idx = 0; sched_sel_idx < sps->cpb_cnt; sched_sel_idx++) {
270 h->initial_cpb_removal_delay[sched_sel_idx] =
271 get_bits(&h->gb, sps->initial_cpb_removal_delay_length);
272 // initial_cpb_removal_delay_offset
273 skip_bits(&h->gb, sps->initial_cpb_removal_delay_length);
276 if (sps->vcl_hrd_parameters_present_flag) {
277 for (sched_sel_idx = 0; sched_sel_idx < sps->cpb_cnt; sched_sel_idx++) {
278 h->initial_cpb_removal_delay[sched_sel_idx] =
279 get_bits(&h->gb, sps->initial_cpb_removal_delay_length);
280 // initial_cpb_removal_delay_offset
281 skip_bits(&h->gb, sps->initial_cpb_removal_delay_length);
285 h->sei_buffering_period_present = 1;
289 static int decode_frame_packing_arrangement(H264Context *h)
291 get_ue_golomb(&h->gb); // frame_packing_arrangement_id
292 h->sei_frame_packing_present = !get_bits1(&h->gb);
294 if (h->sei_frame_packing_present) {
295 h->frame_packing_arrangement_type = get_bits(&h->gb, 7);
296 h->quincunx_subsampling = get_bits1(&h->gb);
297 h->content_interpretation_type = get_bits(&h->gb, 6);
299 // the following skips: spatial_flipping_flag, frame0_flipped_flag,
300 // field_views_flag, current_frame_is_frame0_flag,
301 // frame0_self_contained_flag, frame1_self_contained_flag
302 skip_bits(&h->gb, 6);
304 if (!h->quincunx_subsampling && h->frame_packing_arrangement_type != 5)
305 skip_bits(&h->gb, 16); // frame[01]_grid_position_[xy]
306 skip_bits(&h->gb, 8); // frame_packing_arrangement_reserved_byte
307 get_ue_golomb(&h->gb); // frame_packing_arrangement_repetition_period
309 skip_bits1(&h->gb); // frame_packing_arrangement_extension_flag
314 static int decode_display_orientation(H264Context *h)
316 h->sei_display_orientation_present = !get_bits1(&h->gb);
318 if (h->sei_display_orientation_present) {
319 h->sei_hflip = get_bits1(&h->gb); // hor_flip
320 h->sei_vflip = get_bits1(&h->gb); // ver_flip
322 h->sei_anticlockwise_rotation = get_bits(&h->gb, 16);
323 get_ue_golomb(&h->gb); // display_orientation_repetition_period
324 skip_bits1(&h->gb); // display_orientation_extension_flag
330 int ff_h264_decode_sei(H264Context *h)
332 while (get_bits_left(&h->gb) > 16) {
338 while (get_bits_left(&h->gb) >= 8 &&
339 (last = get_bits(&h->gb, 8)) == 255) {
345 while (get_bits_left(&h->gb) >= 8 &&
346 (last = get_bits(&h->gb, 8)) == 255) {
351 if (size > get_bits_left(&h->gb) / 8) {
352 av_log(h->avctx, AV_LOG_ERROR, "SEI type %d truncated at %d\n",
353 type, get_bits_left(&h->gb));
354 return AVERROR_INVALIDDATA;
358 case SEI_TYPE_PIC_TIMING: // Picture timing SEI
359 ret = decode_picture_timing(h);
361 case SEI_TYPE_USER_DATA_REGISTERED:
362 ret = decode_registered_user_data(h, size);
364 case SEI_TYPE_USER_DATA_UNREGISTERED:
365 ret = decode_unregistered_user_data(h, size);
367 case SEI_TYPE_RECOVERY_POINT:
368 ret = decode_recovery_point(h);
370 case SEI_TYPE_BUFFERING_PERIOD:
371 ret = decode_buffering_period(h);
373 case SEI_TYPE_FRAME_PACKING:
374 ret = decode_frame_packing_arrangement(h);
376 case SEI_TYPE_DISPLAY_ORIENTATION:
377 ret = decode_display_orientation(h);
380 av_log(h->avctx, AV_LOG_DEBUG, "unknown SEI type %d\n", type);
381 skip_bits(&h->gb, 8 * size);
386 // FIXME check bits here
387 align_get_bits(&h->gb);