4bf001ae07420827263c6a1fe6e8218a5268908e
[libav.git] / libavcodec / h264_sei.c
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 Libav.
6 *
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.
11 *
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.
16 *
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
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 "avcodec.h"
29 #include "get_bits.h"
30 #include "golomb_legacy.h"
31 #include "h264_ps.h"
32 #include "h264_sei.h"
33 #include "internal.h"
34
35 static const uint8_t sei_num_clock_ts_table[9] = {
36 1, 1, 1, 2, 2, 3, 3, 2, 3
37 };
38
39 void ff_h264_sei_uninit(H264SEIContext *h)
40 {
41 h->unregistered.x264_build = -1;
42 h->recovery_point.recovery_frame_cnt = -1;
43
44 h->picture_timing.dpb_output_delay = 0;
45 h->picture_timing.cpb_removal_delay = -1;
46
47 h->buffering_period.present = 0;
48 h->frame_packing.present = 0;
49 h->display_orientation.present = 0;
50 h->afd.present = 0;
51
52 h->a53_caption.a53_caption_size = 0;
53 av_freep(&h->a53_caption.a53_caption);
54 }
55
56 static int decode_picture_timing(H264SEIPictureTiming *h, GetBitContext *gb,
57 const SPS *sps, void *logctx)
58 {
59 if (!sps)
60 return AVERROR_INVALIDDATA;
61
62 if (sps->nal_hrd_parameters_present_flag ||
63 sps->vcl_hrd_parameters_present_flag) {
64 h->cpb_removal_delay = get_bits(gb, sps->cpb_removal_delay_length);
65 h->dpb_output_delay = get_bits(gb, sps->dpb_output_delay_length);
66 }
67 if (sps->pic_struct_present_flag) {
68 unsigned int i, num_clock_ts;
69
70 h->pic_struct = get_bits(gb, 4);
71 h->ct_type = 0;
72
73 if (h->pic_struct > SEI_PIC_STRUCT_FRAME_TRIPLING)
74 return AVERROR_INVALIDDATA;
75
76 num_clock_ts = sei_num_clock_ts_table[h->pic_struct];
77
78 for (i = 0; i < num_clock_ts; i++) {
79 if (get_bits(gb, 1)) { /* clock_timestamp_flag */
80 unsigned int full_timestamp_flag;
81
82 h->ct_type |= 1 << get_bits(gb, 2);
83 skip_bits(gb, 1); /* nuit_field_based_flag */
84 skip_bits(gb, 5); /* counting_type */
85 full_timestamp_flag = get_bits(gb, 1);
86 skip_bits(gb, 1); /* discontinuity_flag */
87 skip_bits(gb, 1); /* cnt_dropped_flag */
88 skip_bits(gb, 8); /* n_frames */
89 if (full_timestamp_flag) {
90 skip_bits(gb, 6); /* seconds_value 0..59 */
91 skip_bits(gb, 6); /* minutes_value 0..59 */
92 skip_bits(gb, 5); /* hours_value 0..23 */
93 } else {
94 if (get_bits(gb, 1)) { /* seconds_flag */
95 skip_bits(gb, 6); /* seconds_value range 0..59 */
96 if (get_bits(gb, 1)) { /* minutes_flag */
97 skip_bits(gb, 6); /* minutes_value 0..59 */
98 if (get_bits(gb, 1)) /* hours_flag */
99 skip_bits(gb, 5); /* hours_value 0..23 */
100 }
101 }
102 }
103 if (sps->time_offset_length > 0)
104 skip_bits(gb,
105 sps->time_offset_length); /* time_offset */
106 }
107 }
108
109 av_log(logctx, AV_LOG_DEBUG, "ct_type:%X pic_struct:%d\n",
110 h->ct_type, h->pic_struct);
111 }
112 return 0;
113 }
114
115 static int decode_registered_user_data_afd(H264SEIAFD *h, GetBitContext *gb, int size)
116 {
117 int flag;
118
119 if (size-- < 1)
120 return AVERROR_INVALIDDATA;
121 skip_bits(gb, 1); // 0
122 flag = get_bits(gb, 1); // active_format_flag
123 skip_bits(gb, 6); // reserved
124
125 if (flag) {
126 if (size-- < 1)
127 return AVERROR_INVALIDDATA;
128 skip_bits(gb, 4); // reserved
129 h->active_format_description = get_bits(gb, 4);
130 h->present = 1;
131 }
132
133 return 0;
134 }
135
136 static int decode_registered_user_data_closed_caption(H264SEIA53Caption *h,
137 GetBitContext *gb, void *logctx,
138 int size)
139 {
140 int flag;
141 int user_data_type_code;
142 int cc_count;
143
144 if (size < 3)
145 return AVERROR(EINVAL);
146
147 user_data_type_code = get_bits(gb, 8);
148 if (user_data_type_code == 0x3) {
149 skip_bits(gb, 1); // reserved
150
151 flag = get_bits(gb, 1); // process_cc_data_flag
152 if (flag) {
153 skip_bits(gb, 1); // zero bit
154 cc_count = get_bits(gb, 5);
155 skip_bits(gb, 8); // reserved
156 size -= 2;
157
158 if (cc_count && size >= cc_count * 3) {
159 const uint64_t new_size = (h->a53_caption_size + cc_count
160 * UINT64_C(3));
161 int i, ret;
162
163 if (new_size > INT_MAX)
164 return AVERROR(EINVAL);
165
166 /* Allow merging of the cc data from two fields. */
167 ret = av_reallocp(&h->a53_caption, new_size);
168 if (ret < 0)
169 return ret;
170
171 for (i = 0; i < cc_count; i++) {
172 h->a53_caption[h->a53_caption_size++] = get_bits(gb, 8);
173 h->a53_caption[h->a53_caption_size++] = get_bits(gb, 8);
174 h->a53_caption[h->a53_caption_size++] = get_bits(gb, 8);
175 }
176
177 skip_bits(gb, 8); // marker_bits
178 }
179 }
180 } else {
181 int i;
182 avpriv_request_sample(logctx, "Subtitles with data type 0x%02x",
183 user_data_type_code);
184 for (i = 0; i < size - 1; i++)
185 skip_bits(gb, 8);
186 }
187
188 return 0;
189 }
190
191 static int decode_registered_user_data(H264SEIContext *h, GetBitContext *gb,
192 void *logctx, int size)
193 {
194 uint32_t country_code;
195 uint32_t user_identifier;
196
197 if (size < 7)
198 return AVERROR_INVALIDDATA;
199 size -= 7;
200
201 country_code = get_bits(gb, 8); // itu_t_t35_country_code
202 if (country_code == 0xFF) {
203 skip_bits(gb, 8); // itu_t_t35_country_code_extension_byte
204 size--;
205 }
206
207 /* itu_t_t35_payload_byte follows */
208 skip_bits(gb, 8); // terminal provider code
209 skip_bits(gb, 8); // terminal provider oriented code
210 user_identifier = get_bits_long(gb, 32);
211
212 switch (user_identifier) {
213 case MKBETAG('D', 'T', 'G', '1'): // afd_data
214 return decode_registered_user_data_afd(&h->afd, gb, size);
215 case MKBETAG('G', 'A', '9', '4'): // closed captions
216 return decode_registered_user_data_closed_caption(&h->a53_caption, gb,
217 logctx, size);
218 default:
219 skip_bits(gb, size * 8);
220 break;
221 }
222
223 return 0;
224 }
225
226 static int decode_unregistered_user_data(H264SEIUnregistered *h, GetBitContext *gb,
227 void *logctx, int size)
228 {
229 uint8_t *user_data;
230 int e, build, i;
231
232 if (size < 16 || size >= INT_MAX - 16)
233 return AVERROR_INVALIDDATA;
234
235 user_data = av_malloc(16 + size + 1);
236 if (!user_data)
237 return AVERROR(ENOMEM);
238
239 for (i = 0; i < size + 16; i++)
240 user_data[i] = get_bits(gb, 8);
241
242 user_data[i] = 0;
243 e = sscanf(user_data + 16, "x264 - core %d", &build);
244 if (e == 1 && build > 0)
245 h->x264_build = build;
246
247 if (strlen(user_data + 16) > 0)
248 av_log(logctx, AV_LOG_DEBUG, "user data:\"%s\"\n", user_data + 16);
249
250 av_free(user_data);
251 return 0;
252 }
253
254 static int decode_recovery_point(H264SEIRecoveryPoint *h, GetBitContext *gb)
255 {
256 h->recovery_frame_cnt = get_ue_golomb(gb);
257
258 /* 1b exact_match_flag,
259 * 1b broken_link_flag,
260 * 2b changing_slice_group_idc */
261 skip_bits(gb, 4);
262
263 return 0;
264 }
265
266 static int decode_buffering_period(H264SEIBufferingPeriod *h, GetBitContext *gb,
267 const H264ParamSets *ps, void *logctx)
268 {
269 unsigned int sps_id;
270 int sched_sel_idx;
271 SPS *sps;
272
273 sps_id = get_ue_golomb_31(gb);
274 if (sps_id > 31 || !ps->sps_list[sps_id]) {
275 av_log(logctx, AV_LOG_ERROR,
276 "non-existing SPS %d referenced in buffering period\n", sps_id);
277 return AVERROR_INVALIDDATA;
278 }
279 sps = (SPS*)ps->sps_list[sps_id]->data;
280
281 // NOTE: This is really so duplicated in the standard... See H.264, D.1.1
282 if (sps->nal_hrd_parameters_present_flag) {
283 for (sched_sel_idx = 0; sched_sel_idx < sps->cpb_cnt; sched_sel_idx++) {
284 h->initial_cpb_removal_delay[sched_sel_idx] =
285 get_bits(gb, sps->initial_cpb_removal_delay_length);
286 // initial_cpb_removal_delay_offset
287 skip_bits(gb, sps->initial_cpb_removal_delay_length);
288 }
289 }
290 if (sps->vcl_hrd_parameters_present_flag) {
291 for (sched_sel_idx = 0; sched_sel_idx < sps->cpb_cnt; sched_sel_idx++) {
292 h->initial_cpb_removal_delay[sched_sel_idx] =
293 get_bits(gb, sps->initial_cpb_removal_delay_length);
294 // initial_cpb_removal_delay_offset
295 skip_bits(gb, sps->initial_cpb_removal_delay_length);
296 }
297 }
298
299 h->present = 1;
300 return 0;
301 }
302
303 static int decode_frame_packing_arrangement(H264SEIFramePacking *h,
304 GetBitContext *gb)
305 {
306 get_ue_golomb(gb); // frame_packing_arrangement_id
307 h->present = !get_bits1(gb);
308
309 if (h->present) {
310 h->arrangement_type = get_bits(gb, 7);
311 h->quincunx_subsampling = get_bits1(gb);
312 h->content_interpretation_type = get_bits(gb, 6);
313
314 // the following skips: spatial_flipping_flag, frame0_flipped_flag,
315 // field_views_flag, current_frame_is_frame0_flag,
316 // frame0_self_contained_flag, frame1_self_contained_flag
317 skip_bits(gb, 6);
318
319 if (!h->quincunx_subsampling && h->arrangement_type != 5)
320 skip_bits(gb, 16); // frame[01]_grid_position_[xy]
321 skip_bits(gb, 8); // frame_packing_arrangement_reserved_byte
322 get_ue_golomb(gb); // frame_packing_arrangement_repetition_period
323 }
324 skip_bits1(gb); // frame_packing_arrangement_extension_flag
325
326 return 0;
327 }
328
329 static int decode_display_orientation(H264SEIDisplayOrientation *h,
330 GetBitContext *gb)
331 {
332 h->present = !get_bits1(gb);
333
334 if (h->present) {
335 h->hflip = get_bits1(gb); // hor_flip
336 h->vflip = get_bits1(gb); // ver_flip
337
338 h->anticlockwise_rotation = get_bits(gb, 16);
339 get_ue_golomb(gb); // display_orientation_repetition_period
340 skip_bits1(gb); // display_orientation_extension_flag
341 }
342
343 return 0;
344 }
345
346 int ff_h264_sei_decode(H264SEIContext *h, GetBitContext *gb,
347 const H264ParamSets *ps, void *logctx)
348 {
349 while (get_bits_left(gb) > 16) {
350 int size = 0;
351 int type = 0;
352 int ret = 0;
353 int last = 0;
354
355 while (get_bits_left(gb) >= 8 &&
356 (last = get_bits(gb, 8)) == 255) {
357 type += 255;
358 }
359 type += last;
360
361 last = 0;
362 while (get_bits_left(gb) >= 8 &&
363 (last = get_bits(gb, 8)) == 255) {
364 size += 255;
365 }
366 size += last;
367
368 if (size > get_bits_left(gb) / 8) {
369 av_log(logctx, AV_LOG_ERROR, "SEI type %d truncated at %d\n",
370 type, get_bits_left(gb));
371 return AVERROR_INVALIDDATA;
372 }
373
374 switch (type) {
375 case SEI_TYPE_PIC_TIMING: // Picture timing SEI
376 ret = decode_picture_timing(&h->picture_timing, gb, ps->sps, logctx);
377 break;
378 case SEI_TYPE_USER_DATA_REGISTERED:
379 ret = decode_registered_user_data(h, gb, logctx, size);
380 break;
381 case SEI_TYPE_USER_DATA_UNREGISTERED:
382 ret = decode_unregistered_user_data(&h->unregistered, gb, logctx, size);
383 break;
384 case SEI_TYPE_RECOVERY_POINT:
385 ret = decode_recovery_point(&h->recovery_point, gb);
386 break;
387 case SEI_TYPE_BUFFERING_PERIOD:
388 ret = decode_buffering_period(&h->buffering_period, gb, ps, logctx);
389 break;
390 case SEI_TYPE_FRAME_PACKING:
391 ret = decode_frame_packing_arrangement(&h->frame_packing, gb);
392 break;
393 case SEI_TYPE_DISPLAY_ORIENTATION:
394 ret = decode_display_orientation(&h->display_orientation, gb);
395 break;
396 default:
397 av_log(logctx, AV_LOG_DEBUG, "unknown SEI type %d\n", type);
398 skip_bits(gb, 8 * size);
399 }
400 if (ret < 0)
401 return ret;
402
403 // FIXME check bits here
404 align_get_bits(gb);
405 }
406
407 return 0;
408 }