blob: 602ce975302e11bcb69cb98db7b9f4a8b05149e1
1 | /* |
2 | * WAV demuxer |
3 | * Copyright (c) 2001, 2002 Fabrice Bellard |
4 | * |
5 | * Sony Wave64 demuxer |
6 | * RF64 demuxer |
7 | * Copyright (c) 2009 Daniel Verkamp |
8 | * |
9 | * This file is part of FFmpeg. |
10 | * |
11 | * FFmpeg is free software; you can redistribute it and/or |
12 | * modify it under the terms of the GNU Lesser General Public |
13 | * License as published by the Free Software Foundation; either |
14 | * version 2.1 of the License, or (at your option) any later version. |
15 | * |
16 | * FFmpeg is distributed in the hope that it will be useful, |
17 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
18 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
19 | * Lesser General Public License for more details. |
20 | * |
21 | * You should have received a copy of the GNU Lesser General Public |
22 | * License along with FFmpeg; if not, write to the Free Software |
23 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
24 | */ |
25 | |
26 | #include <stdint.h> |
27 | |
28 | #include "libavutil/avassert.h" |
29 | #include "libavutil/dict.h" |
30 | #include "libavutil/intreadwrite.h" |
31 | #include "libavutil/log.h" |
32 | #include "libavutil/mathematics.h" |
33 | #include "libavutil/opt.h" |
34 | #include "avformat.h" |
35 | #include "avio.h" |
36 | #include "avio_internal.h" |
37 | #include "internal.h" |
38 | #include "metadata.h" |
39 | #include "pcm.h" |
40 | #include "riff.h" |
41 | #include "w64.h" |
42 | #include "spdif.h" |
43 | |
44 | typedef struct WAVDemuxContext { |
45 | const AVClass *class; |
46 | int64_t data_end; |
47 | int w64; |
48 | int64_t smv_data_ofs; |
49 | int smv_block_size; |
50 | int smv_frames_per_jpeg; |
51 | int smv_block; |
52 | int smv_last_stream; |
53 | int smv_eof; |
54 | int audio_eof; |
55 | int ignore_length; |
56 | int spdif; |
57 | int smv_cur_pt; |
58 | int smv_given_first; |
59 | int unaligned; // e.g. if an odd number of bytes ID3 tag was prepended |
60 | int rifx; // RIFX: integer byte order for parameters is big endian |
61 | } WAVDemuxContext; |
62 | |
63 | static void set_spdif(AVFormatContext *s, WAVDemuxContext *wav) |
64 | { |
65 | if (CONFIG_SPDIF_DEMUXER && s->streams[0]->codecpar->codec_tag == 1) { |
66 | enum AVCodecID codec; |
67 | int len = 1<<16; |
68 | int ret = ffio_ensure_seekback(s->pb, len); |
69 | |
70 | if (ret >= 0) { |
71 | uint8_t *buf = av_malloc(len); |
72 | if (!buf) { |
73 | ret = AVERROR(ENOMEM); |
74 | } else { |
75 | int64_t pos = avio_tell(s->pb); |
76 | len = ret = avio_read(s->pb, buf, len); |
77 | if (len >= 0) { |
78 | ret = ff_spdif_probe(buf, len, &codec); |
79 | if (ret > AVPROBE_SCORE_EXTENSION) { |
80 | s->streams[0]->codecpar->codec_id = codec; |
81 | wav->spdif = 1; |
82 | } |
83 | } |
84 | avio_seek(s->pb, pos, SEEK_SET); |
85 | av_free(buf); |
86 | } |
87 | } |
88 | |
89 | if (ret < 0) |
90 | av_log(s, AV_LOG_WARNING, "Cannot check for SPDIF\n"); |
91 | } |
92 | } |
93 | |
94 | #if CONFIG_WAV_DEMUXER |
95 | |
96 | static int64_t next_tag(AVIOContext *pb, uint32_t *tag, int big_endian) |
97 | { |
98 | *tag = avio_rl32(pb); |
99 | if (!big_endian) { |
100 | return avio_rl32(pb); |
101 | } else { |
102 | return avio_rb32(pb); |
103 | } |
104 | } |
105 | |
106 | /* RIFF chunks are always at even offsets relative to where they start. */ |
107 | static int64_t wav_seek_tag(WAVDemuxContext * wav, AVIOContext *s, int64_t offset, int whence) |
108 | { |
109 | offset += offset < INT64_MAX && offset + wav->unaligned & 1; |
110 | |
111 | return avio_seek(s, offset, whence); |
112 | } |
113 | |
114 | /* return the size of the found tag */ |
115 | static int64_t find_tag(WAVDemuxContext * wav, AVIOContext *pb, uint32_t tag1) |
116 | { |
117 | unsigned int tag; |
118 | int64_t size; |
119 | |
120 | for (;;) { |
121 | if (avio_feof(pb)) |
122 | return AVERROR_EOF; |
123 | size = next_tag(pb, &tag, wav->rifx); |
124 | if (tag == tag1) |
125 | break; |
126 | wav_seek_tag(wav, pb, size, SEEK_CUR); |
127 | } |
128 | return size; |
129 | } |
130 | |
131 | static int wav_probe(AVProbeData *p) |
132 | { |
133 | /* check file header */ |
134 | if (p->buf_size <= 32) |
135 | return 0; |
136 | if (!memcmp(p->buf + 8, "WAVE", 4)) { |
137 | if (!memcmp(p->buf, "RIFF", 4) || !memcmp(p->buf, "RIFX", 4)) |
138 | /* Since the ACT demuxer has a standard WAV header at the top of |
139 | * its own, the returned score is decreased to avoid a probe |
140 | * conflict between ACT and WAV. */ |
141 | return AVPROBE_SCORE_MAX - 1; |
142 | else if (!memcmp(p->buf, "RF64", 4) && |
143 | !memcmp(p->buf + 12, "ds64", 4)) |
144 | return AVPROBE_SCORE_MAX; |
145 | } |
146 | return 0; |
147 | } |
148 | |
149 | static void handle_stream_probing(AVStream *st) |
150 | { |
151 | if (st->codecpar->codec_id == AV_CODEC_ID_PCM_S16LE) { |
152 | st->request_probe = AVPROBE_SCORE_EXTENSION; |
153 | st->probe_packets = FFMIN(st->probe_packets, 32); |
154 | } |
155 | } |
156 | |
157 | static int wav_parse_fmt_tag(AVFormatContext *s, int64_t size, AVStream **st) |
158 | { |
159 | AVIOContext *pb = s->pb; |
160 | WAVDemuxContext *wav = s->priv_data; |
161 | int ret; |
162 | |
163 | /* parse fmt header */ |
164 | *st = avformat_new_stream(s, NULL); |
165 | if (!*st) |
166 | return AVERROR(ENOMEM); |
167 | |
168 | ret = ff_get_wav_header(s, pb, (*st)->codecpar, size, wav->rifx); |
169 | if (ret < 0) |
170 | return ret; |
171 | handle_stream_probing(*st); |
172 | |
173 | (*st)->need_parsing = AVSTREAM_PARSE_FULL_RAW; |
174 | |
175 | avpriv_set_pts_info(*st, 64, 1, (*st)->codecpar->sample_rate); |
176 | |
177 | return 0; |
178 | } |
179 | |
180 | static int wav_parse_xma2_tag(AVFormatContext *s, int64_t size, AVStream **st) |
181 | { |
182 | AVIOContext *pb = s->pb; |
183 | int num_streams, i, channels = 0; |
184 | |
185 | if (size < 44) |
186 | return AVERROR_INVALIDDATA; |
187 | |
188 | *st = avformat_new_stream(s, NULL); |
189 | if (!*st) |
190 | return AVERROR(ENOMEM); |
191 | |
192 | (*st)->codecpar->codec_type = AVMEDIA_TYPE_AUDIO; |
193 | (*st)->codecpar->codec_id = AV_CODEC_ID_XMA2; |
194 | (*st)->need_parsing = AVSTREAM_PARSE_FULL_RAW; |
195 | |
196 | avio_skip(pb, 1); |
197 | num_streams = avio_r8(pb); |
198 | if (size < 40 + num_streams * 4) |
199 | return AVERROR_INVALIDDATA; |
200 | avio_skip(pb, 10); |
201 | (*st)->codecpar->sample_rate = avio_rb32(pb); |
202 | avio_skip(pb, 12); |
203 | (*st)->duration = avio_rb32(pb); |
204 | avio_skip(pb, 8); |
205 | |
206 | for (i = 0; i < num_streams; i++) { |
207 | channels += avio_r8(pb); |
208 | avio_skip(pb, 3); |
209 | } |
210 | (*st)->codecpar->channels = channels; |
211 | |
212 | if ((*st)->codecpar->channels <= 0 || (*st)->codecpar->sample_rate <= 0) |
213 | return AVERROR_INVALIDDATA; |
214 | |
215 | avpriv_set_pts_info(*st, 64, 1, (*st)->codecpar->sample_rate); |
216 | if (ff_alloc_extradata((*st)->codecpar, 34)) |
217 | return AVERROR(ENOMEM); |
218 | memset((*st)->codecpar->extradata, 0, 34); |
219 | |
220 | return 0; |
221 | } |
222 | |
223 | static inline int wav_parse_bext_string(AVFormatContext *s, const char *key, |
224 | int length) |
225 | { |
226 | char temp[257]; |
227 | int ret; |
228 | |
229 | av_assert0(length <= sizeof(temp)); |
230 | if ((ret = avio_read(s->pb, temp, length)) < 0) |
231 | return ret; |
232 | |
233 | temp[length] = 0; |
234 | |
235 | if (strlen(temp)) |
236 | return av_dict_set(&s->metadata, key, temp, 0); |
237 | |
238 | return 0; |
239 | } |
240 | |
241 | static int wav_parse_bext_tag(AVFormatContext *s, int64_t size) |
242 | { |
243 | char temp[131], *coding_history; |
244 | int ret, x; |
245 | uint64_t time_reference; |
246 | int64_t umid_parts[8], umid_mask = 0; |
247 | |
248 | if ((ret = wav_parse_bext_string(s, "description", 256)) < 0 || |
249 | (ret = wav_parse_bext_string(s, "originator", 32)) < 0 || |
250 | (ret = wav_parse_bext_string(s, "originator_reference", 32)) < 0 || |
251 | (ret = wav_parse_bext_string(s, "origination_date", 10)) < 0 || |
252 | (ret = wav_parse_bext_string(s, "origination_time", 8)) < 0) |
253 | return ret; |
254 | |
255 | time_reference = avio_rl64(s->pb); |
256 | snprintf(temp, sizeof(temp), "%"PRIu64, time_reference); |
257 | if ((ret = av_dict_set(&s->metadata, "time_reference", temp, 0)) < 0) |
258 | return ret; |
259 | |
260 | /* check if version is >= 1, in which case an UMID may be present */ |
261 | if (avio_rl16(s->pb) >= 1) { |
262 | for (x = 0; x < 8; x++) |
263 | umid_mask |= umid_parts[x] = avio_rb64(s->pb); |
264 | |
265 | if (umid_mask) { |
266 | /* the string formatting below is per SMPTE 330M-2004 Annex C */ |
267 | if (umid_parts[4] == 0 && umid_parts[5] == 0 && |
268 | umid_parts[6] == 0 && umid_parts[7] == 0) { |
269 | /* basic UMID */ |
270 | snprintf(temp, sizeof(temp), |
271 | "0x%016"PRIX64"%016"PRIX64"%016"PRIX64"%016"PRIX64, |
272 | umid_parts[0], umid_parts[1], |
273 | umid_parts[2], umid_parts[3]); |
274 | } else { |
275 | /* extended UMID */ |
276 | snprintf(temp, sizeof(temp), |
277 | "0x%016"PRIX64"%016"PRIX64"%016"PRIX64"%016"PRIX64 |
278 | "%016"PRIX64"%016"PRIX64"%016"PRIX64"%016"PRIX64, |
279 | umid_parts[0], umid_parts[1], |
280 | umid_parts[2], umid_parts[3], |
281 | umid_parts[4], umid_parts[5], |
282 | umid_parts[6], umid_parts[7]); |
283 | } |
284 | |
285 | if ((ret = av_dict_set(&s->metadata, "umid", temp, 0)) < 0) |
286 | return ret; |
287 | } |
288 | |
289 | avio_skip(s->pb, 190); |
290 | } else |
291 | avio_skip(s->pb, 254); |
292 | |
293 | if (size > 602) { |
294 | /* CodingHistory present */ |
295 | size -= 602; |
296 | |
297 | if (!(coding_history = av_malloc(size + 1))) |
298 | return AVERROR(ENOMEM); |
299 | |
300 | if ((ret = avio_read(s->pb, coding_history, size)) < 0) |
301 | return ret; |
302 | |
303 | coding_history[size] = 0; |
304 | if ((ret = av_dict_set(&s->metadata, "coding_history", coding_history, |
305 | AV_DICT_DONT_STRDUP_VAL)) < 0) |
306 | return ret; |
307 | } |
308 | |
309 | return 0; |
310 | } |
311 | |
312 | static const AVMetadataConv wav_metadata_conv[] = { |
313 | { "description", "comment" }, |
314 | { "originator", "encoded_by" }, |
315 | { "origination_date", "date" }, |
316 | { "origination_time", "creation_time" }, |
317 | { 0 }, |
318 | }; |
319 | |
320 | /* wav input */ |
321 | static int wav_read_header(AVFormatContext *s) |
322 | { |
323 | int64_t size, av_uninit(data_size); |
324 | int64_t sample_count = 0; |
325 | int rf64 = 0; |
326 | uint32_t tag; |
327 | AVIOContext *pb = s->pb; |
328 | AVStream *st = NULL; |
329 | WAVDemuxContext *wav = s->priv_data; |
330 | int ret, got_fmt = 0, got_xma2 = 0; |
331 | int64_t next_tag_ofs, data_ofs = -1; |
332 | |
333 | wav->unaligned = avio_tell(s->pb) & 1; |
334 | |
335 | wav->smv_data_ofs = -1; |
336 | |
337 | /* read chunk ID */ |
338 | tag = avio_rl32(pb); |
339 | switch (tag) { |
340 | case MKTAG('R', 'I', 'F', 'F'): |
341 | break; |
342 | case MKTAG('R', 'I', 'F', 'X'): |
343 | wav->rifx = 1; |
344 | break; |
345 | case MKTAG('R', 'F', '6', '4'): |
346 | rf64 = 1; |
347 | break; |
348 | default: |
349 | av_log(s, AV_LOG_ERROR, "invalid start code %s in RIFF header\n", |
350 | av_fourcc2str(tag)); |
351 | return AVERROR_INVALIDDATA; |
352 | } |
353 | |
354 | /* read chunk size */ |
355 | avio_rl32(pb); |
356 | |
357 | /* read format */ |
358 | if (avio_rl32(pb) != MKTAG('W', 'A', 'V', 'E')) { |
359 | av_log(s, AV_LOG_ERROR, "invalid format in RIFF header\n"); |
360 | return AVERROR_INVALIDDATA; |
361 | } |
362 | |
363 | if (rf64) { |
364 | if (avio_rl32(pb) != MKTAG('d', 's', '6', '4')) |
365 | return AVERROR_INVALIDDATA; |
366 | size = avio_rl32(pb); |
367 | if (size < 24) |
368 | return AVERROR_INVALIDDATA; |
369 | avio_rl64(pb); /* RIFF size */ |
370 | |
371 | data_size = avio_rl64(pb); |
372 | sample_count = avio_rl64(pb); |
373 | |
374 | if (data_size < 0 || sample_count < 0) { |
375 | av_log(s, AV_LOG_ERROR, "negative data_size and/or sample_count in " |
376 | "ds64: data_size = %"PRId64", sample_count = %"PRId64"\n", |
377 | data_size, sample_count); |
378 | return AVERROR_INVALIDDATA; |
379 | } |
380 | avio_skip(pb, size - 24); /* skip rest of ds64 chunk */ |
381 | |
382 | } |
383 | |
384 | for (;;) { |
385 | AVStream *vst; |
386 | size = next_tag(pb, &tag, wav->rifx); |
387 | next_tag_ofs = avio_tell(pb) + size; |
388 | |
389 | if (avio_feof(pb)) |
390 | break; |
391 | |
392 | switch (tag) { |
393 | case MKTAG('f', 'm', 't', ' '): |
394 | /* only parse the first 'fmt ' tag found */ |
395 | if (!got_xma2 && !got_fmt && (ret = wav_parse_fmt_tag(s, size, &st)) < 0) { |
396 | return ret; |
397 | } else if (got_fmt) |
398 | av_log(s, AV_LOG_WARNING, "found more than one 'fmt ' tag\n"); |
399 | |
400 | got_fmt = 1; |
401 | break; |
402 | case MKTAG('X', 'M', 'A', '2'): |
403 | /* only parse the first 'XMA2' tag found */ |
404 | if (!got_fmt && !got_xma2 && (ret = wav_parse_xma2_tag(s, size, &st)) < 0) { |
405 | return ret; |
406 | } else if (got_xma2) |
407 | av_log(s, AV_LOG_WARNING, "found more than one 'XMA2' tag\n"); |
408 | |
409 | got_xma2 = 1; |
410 | break; |
411 | case MKTAG('d', 'a', 't', 'a'): |
412 | if (!(pb->seekable & AVIO_SEEKABLE_NORMAL) && !got_fmt && !got_xma2) { |
413 | av_log(s, AV_LOG_ERROR, |
414 | "found no 'fmt ' tag before the 'data' tag\n"); |
415 | return AVERROR_INVALIDDATA; |
416 | } |
417 | |
418 | if (rf64) { |
419 | next_tag_ofs = wav->data_end = avio_tell(pb) + data_size; |
420 | } else if (size != 0xFFFFFFFF) { |
421 | data_size = size; |
422 | next_tag_ofs = wav->data_end = size ? next_tag_ofs : INT64_MAX; |
423 | } else { |
424 | av_log(s, AV_LOG_WARNING, "Ignoring maximum wav data size, " |
425 | "file may be invalid\n"); |
426 | data_size = 0; |
427 | next_tag_ofs = wav->data_end = INT64_MAX; |
428 | } |
429 | |
430 | data_ofs = avio_tell(pb); |
431 | |
432 | /* don't look for footer metadata if we can't seek or if we don't |
433 | * know where the data tag ends |
434 | */ |
435 | if (!(pb->seekable & AVIO_SEEKABLE_NORMAL) || (!rf64 && !size)) |
436 | goto break_loop; |
437 | break; |
438 | case MKTAG('f', 'a', 'c', 't'): |
439 | if (!sample_count) |
440 | sample_count = (!wav->rifx ? avio_rl32(pb) : avio_rb32(pb)); |
441 | break; |
442 | case MKTAG('b', 'e', 'x', 't'): |
443 | if ((ret = wav_parse_bext_tag(s, size)) < 0) |
444 | return ret; |
445 | break; |
446 | case MKTAG('S','M','V','0'): |
447 | if (!got_fmt) { |
448 | av_log(s, AV_LOG_ERROR, "found no 'fmt ' tag before the 'SMV0' tag\n"); |
449 | return AVERROR_INVALIDDATA; |
450 | } |
451 | // SMV file, a wav file with video appended. |
452 | if (size != MKTAG('0','2','0','0')) { |
453 | av_log(s, AV_LOG_ERROR, "Unknown SMV version found\n"); |
454 | goto break_loop; |
455 | } |
456 | av_log(s, AV_LOG_DEBUG, "Found SMV data\n"); |
457 | wav->smv_given_first = 0; |
458 | vst = avformat_new_stream(s, NULL); |
459 | if (!vst) |
460 | return AVERROR(ENOMEM); |
461 | avio_r8(pb); |
462 | vst->id = 1; |
463 | vst->codecpar->codec_type = AVMEDIA_TYPE_VIDEO; |
464 | vst->codecpar->codec_id = AV_CODEC_ID_SMVJPEG; |
465 | vst->codecpar->width = avio_rl24(pb); |
466 | vst->codecpar->height = avio_rl24(pb); |
467 | if (ff_alloc_extradata(vst->codecpar, 4)) { |
468 | av_log(s, AV_LOG_ERROR, "Could not allocate extradata.\n"); |
469 | return AVERROR(ENOMEM); |
470 | } |
471 | size = avio_rl24(pb); |
472 | wav->smv_data_ofs = avio_tell(pb) + (size - 5) * 3; |
473 | avio_rl24(pb); |
474 | wav->smv_block_size = avio_rl24(pb); |
475 | avpriv_set_pts_info(vst, 32, 1, avio_rl24(pb)); |
476 | vst->duration = avio_rl24(pb); |
477 | avio_rl24(pb); |
478 | avio_rl24(pb); |
479 | wav->smv_frames_per_jpeg = avio_rl24(pb); |
480 | if (wav->smv_frames_per_jpeg > 65536) { |
481 | av_log(s, AV_LOG_ERROR, "too many frames per jpeg\n"); |
482 | return AVERROR_INVALIDDATA; |
483 | } |
484 | AV_WL32(vst->codecpar->extradata, wav->smv_frames_per_jpeg); |
485 | wav->smv_cur_pt = 0; |
486 | goto break_loop; |
487 | case MKTAG('L', 'I', 'S', 'T'): |
488 | if (size < 4) { |
489 | av_log(s, AV_LOG_ERROR, "too short LIST tag\n"); |
490 | return AVERROR_INVALIDDATA; |
491 | } |
492 | switch (avio_rl32(pb)) { |
493 | case MKTAG('I', 'N', 'F', 'O'): |
494 | ff_read_riff_info(s, size - 4); |
495 | } |
496 | break; |
497 | } |
498 | |
499 | /* seek to next tag unless we know that we'll run into EOF */ |
500 | if ((avio_size(pb) > 0 && next_tag_ofs >= avio_size(pb)) || |
501 | wav_seek_tag(wav, pb, next_tag_ofs, SEEK_SET) < 0) { |
502 | break; |
503 | } |
504 | } |
505 | |
506 | break_loop: |
507 | if (!got_fmt && !got_xma2) { |
508 | av_log(s, AV_LOG_ERROR, "no 'fmt ' or 'XMA2' tag found\n"); |
509 | return AVERROR_INVALIDDATA; |
510 | } |
511 | |
512 | if (data_ofs < 0) { |
513 | av_log(s, AV_LOG_ERROR, "no 'data' tag found\n"); |
514 | return AVERROR_INVALIDDATA; |
515 | } |
516 | |
517 | avio_seek(pb, data_ofs, SEEK_SET); |
518 | |
519 | if (data_size > (INT64_MAX>>3)) { |
520 | av_log(s, AV_LOG_WARNING, "Data size %"PRId64" is too large\n", data_size); |
521 | data_size = 0; |
522 | } |
523 | |
524 | if ( st->codecpar->bit_rate > 0 && data_size > 0 |
525 | && st->codecpar->sample_rate > 0 |
526 | && sample_count > 0 && st->codecpar->channels > 1 |
527 | && sample_count % st->codecpar->channels == 0) { |
528 | if (fabs(8.0 * data_size * st->codecpar->channels * st->codecpar->sample_rate / |
529 | sample_count /st->codecpar->bit_rate - 1.0) < 0.3) |
530 | sample_count /= st->codecpar->channels; |
531 | } |
532 | |
533 | if ( data_size > 0 && sample_count && st->codecpar->channels |
534 | && (data_size << 3) / sample_count / st->codecpar->channels > st->codecpar->bits_per_coded_sample + 1) { |
535 | av_log(s, AV_LOG_WARNING, "ignoring wrong sample_count %"PRId64"\n", sample_count); |
536 | sample_count = 0; |
537 | } |
538 | |
539 | /* G.729 hack (for Ticket4577) |
540 | * FIXME: Come up with cleaner, more general solution */ |
541 | if (st->codecpar->codec_id == AV_CODEC_ID_G729 && sample_count && (data_size << 3) > sample_count) { |
542 | av_log(s, AV_LOG_WARNING, "ignoring wrong sample_count %"PRId64"\n", sample_count); |
543 | sample_count = 0; |
544 | } |
545 | |
546 | if (!sample_count || av_get_exact_bits_per_sample(st->codecpar->codec_id) > 0) |
547 | if ( st->codecpar->channels |
548 | && data_size |
549 | && av_get_bits_per_sample(st->codecpar->codec_id) |
550 | && wav->data_end <= avio_size(pb)) |
551 | sample_count = (data_size << 3) |
552 | / |
553 | (st->codecpar->channels * (uint64_t)av_get_bits_per_sample(st->codecpar->codec_id)); |
554 | |
555 | if (sample_count) |
556 | st->duration = sample_count; |
557 | |
558 | if (st->codecpar->codec_id == AV_CODEC_ID_PCM_S32LE && |
559 | st->codecpar->block_align == st->codecpar->channels * 4 && |
560 | st->codecpar->bits_per_coded_sample == 32 && |
561 | st->codecpar->extradata_size == 2 && |
562 | AV_RL16(st->codecpar->extradata) == 1) { |
563 | st->codecpar->codec_id = AV_CODEC_ID_PCM_F16LE; |
564 | st->codecpar->bits_per_coded_sample = 16; |
565 | } else if (st->codecpar->codec_id == AV_CODEC_ID_PCM_S24LE && |
566 | st->codecpar->block_align == st->codecpar->channels * 4 && |
567 | st->codecpar->bits_per_coded_sample == 24) { |
568 | st->codecpar->codec_id = AV_CODEC_ID_PCM_F24LE; |
569 | } else if (st->codecpar->codec_id == AV_CODEC_ID_XMA1 || |
570 | st->codecpar->codec_id == AV_CODEC_ID_XMA2) { |
571 | st->codecpar->block_align = 2048; |
572 | } |
573 | |
574 | ff_metadata_conv_ctx(s, NULL, wav_metadata_conv); |
575 | ff_metadata_conv_ctx(s, NULL, ff_riff_info_conv); |
576 | |
577 | set_spdif(s, wav); |
578 | |
579 | return 0; |
580 | } |
581 | |
582 | /** |
583 | * Find chunk with w64 GUID by skipping over other chunks. |
584 | * @return the size of the found chunk |
585 | */ |
586 | static int64_t find_guid(AVIOContext *pb, const uint8_t guid1[16]) |
587 | { |
588 | uint8_t guid[16]; |
589 | int64_t size; |
590 | |
591 | while (!avio_feof(pb)) { |
592 | avio_read(pb, guid, 16); |
593 | size = avio_rl64(pb); |
594 | if (size <= 24) |
595 | return AVERROR_INVALIDDATA; |
596 | if (!memcmp(guid, guid1, 16)) |
597 | return size; |
598 | avio_skip(pb, FFALIGN(size, INT64_C(8)) - 24); |
599 | } |
600 | return AVERROR_EOF; |
601 | } |
602 | |
603 | #define MAX_SIZE 4096 |
604 | |
605 | static int wav_read_packet(AVFormatContext *s, AVPacket *pkt) |
606 | { |
607 | int ret, size; |
608 | int64_t left; |
609 | AVStream *st; |
610 | WAVDemuxContext *wav = s->priv_data; |
611 | |
612 | if (CONFIG_SPDIF_DEMUXER && wav->spdif == 1) |
613 | return ff_spdif_read_packet(s, pkt); |
614 | |
615 | if (wav->smv_data_ofs > 0) { |
616 | int64_t audio_dts, video_dts; |
617 | smv_retry: |
618 | audio_dts = (int32_t)s->streams[0]->cur_dts; |
619 | video_dts = (int32_t)s->streams[1]->cur_dts; |
620 | |
621 | if (audio_dts != AV_NOPTS_VALUE && video_dts != AV_NOPTS_VALUE) { |
622 | /*We always return a video frame first to get the pixel format first*/ |
623 | wav->smv_last_stream = wav->smv_given_first ? |
624 | av_compare_ts(video_dts, s->streams[1]->time_base, |
625 | audio_dts, s->streams[0]->time_base) > 0 : 0; |
626 | wav->smv_given_first = 1; |
627 | } |
628 | wav->smv_last_stream = !wav->smv_last_stream; |
629 | wav->smv_last_stream |= wav->audio_eof; |
630 | wav->smv_last_stream &= !wav->smv_eof; |
631 | if (wav->smv_last_stream) { |
632 | uint64_t old_pos = avio_tell(s->pb); |
633 | uint64_t new_pos = wav->smv_data_ofs + |
634 | wav->smv_block * wav->smv_block_size; |
635 | if (avio_seek(s->pb, new_pos, SEEK_SET) < 0) { |
636 | ret = AVERROR_EOF; |
637 | goto smv_out; |
638 | } |
639 | size = avio_rl24(s->pb); |
640 | ret = av_get_packet(s->pb, pkt, size); |
641 | if (ret < 0) |
642 | goto smv_out; |
643 | pkt->pos -= 3; |
644 | pkt->pts = wav->smv_block * wav->smv_frames_per_jpeg + wav->smv_cur_pt; |
645 | wav->smv_cur_pt++; |
646 | if (wav->smv_frames_per_jpeg > 0) |
647 | wav->smv_cur_pt %= wav->smv_frames_per_jpeg; |
648 | if (!wav->smv_cur_pt) |
649 | wav->smv_block++; |
650 | |
651 | pkt->stream_index = 1; |
652 | smv_out: |
653 | avio_seek(s->pb, old_pos, SEEK_SET); |
654 | if (ret == AVERROR_EOF) { |
655 | wav->smv_eof = 1; |
656 | goto smv_retry; |
657 | } |
658 | return ret; |
659 | } |
660 | } |
661 | |
662 | st = s->streams[0]; |
663 | |
664 | left = wav->data_end - avio_tell(s->pb); |
665 | if (wav->ignore_length) |
666 | left = INT_MAX; |
667 | if (left <= 0) { |
668 | if (CONFIG_W64_DEMUXER && wav->w64) |
669 | left = find_guid(s->pb, ff_w64_guid_data) - 24; |
670 | else |
671 | left = find_tag(wav, s->pb, MKTAG('d', 'a', 't', 'a')); |
672 | if (left < 0) { |
673 | wav->audio_eof = 1; |
674 | if (wav->smv_data_ofs > 0 && !wav->smv_eof) |
675 | goto smv_retry; |
676 | return AVERROR_EOF; |
677 | } |
678 | wav->data_end = avio_tell(s->pb) + left; |
679 | } |
680 | |
681 | size = MAX_SIZE; |
682 | if (st->codecpar->block_align > 1) { |
683 | if (size < st->codecpar->block_align) |
684 | size = st->codecpar->block_align; |
685 | size = (size / st->codecpar->block_align) * st->codecpar->block_align; |
686 | } |
687 | size = FFMIN(size, left); |
688 | ret = av_get_packet(s->pb, pkt, size); |
689 | if (ret < 0) |
690 | return ret; |
691 | pkt->stream_index = 0; |
692 | |
693 | return ret; |
694 | } |
695 | |
696 | static int wav_read_seek(AVFormatContext *s, |
697 | int stream_index, int64_t timestamp, int flags) |
698 | { |
699 | WAVDemuxContext *wav = s->priv_data; |
700 | AVStream *st; |
701 | wav->smv_eof = 0; |
702 | wav->audio_eof = 0; |
703 | if (wav->smv_data_ofs > 0) { |
704 | int64_t smv_timestamp = timestamp; |
705 | if (stream_index == 0) |
706 | smv_timestamp = av_rescale_q(timestamp, s->streams[0]->time_base, s->streams[1]->time_base); |
707 | else |
708 | timestamp = av_rescale_q(smv_timestamp, s->streams[1]->time_base, s->streams[0]->time_base); |
709 | if (wav->smv_frames_per_jpeg > 0) { |
710 | wav->smv_block = smv_timestamp / wav->smv_frames_per_jpeg; |
711 | wav->smv_cur_pt = smv_timestamp % wav->smv_frames_per_jpeg; |
712 | } |
713 | } |
714 | |
715 | st = s->streams[0]; |
716 | switch (st->codecpar->codec_id) { |
717 | case AV_CODEC_ID_MP2: |
718 | case AV_CODEC_ID_MP3: |
719 | case AV_CODEC_ID_AC3: |
720 | case AV_CODEC_ID_DTS: |
721 | case AV_CODEC_ID_XMA2: |
722 | /* use generic seeking with dynamically generated indexes */ |
723 | return -1; |
724 | default: |
725 | break; |
726 | } |
727 | return ff_pcm_read_seek(s, stream_index, timestamp, flags); |
728 | } |
729 | |
730 | #define OFFSET(x) offsetof(WAVDemuxContext, x) |
731 | #define DEC AV_OPT_FLAG_DECODING_PARAM |
732 | static const AVOption demux_options[] = { |
733 | { "ignore_length", "Ignore length", OFFSET(ignore_length), AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, DEC }, |
734 | { NULL }, |
735 | }; |
736 | |
737 | static const AVClass wav_demuxer_class = { |
738 | .class_name = "WAV demuxer", |
739 | .item_name = av_default_item_name, |
740 | .option = demux_options, |
741 | .version = LIBAVUTIL_VERSION_INT, |
742 | }; |
743 | AVInputFormat ff_wav_demuxer = { |
744 | .name = "wav", |
745 | .long_name = NULL_IF_CONFIG_SMALL("WAV / WAVE (Waveform Audio)"), |
746 | .priv_data_size = sizeof(WAVDemuxContext), |
747 | .read_probe = wav_probe, |
748 | .read_header = wav_read_header, |
749 | .read_packet = wav_read_packet, |
750 | .read_seek = wav_read_seek, |
751 | .flags = AVFMT_GENERIC_INDEX, |
752 | .codec_tag = (const AVCodecTag * const []) { ff_codec_wav_tags, 0 }, |
753 | .priv_class = &wav_demuxer_class, |
754 | }; |
755 | #endif /* CONFIG_WAV_DEMUXER */ |
756 | |
757 | #if CONFIG_W64_DEMUXER |
758 | static int w64_probe(AVProbeData *p) |
759 | { |
760 | if (p->buf_size <= 40) |
761 | return 0; |
762 | if (!memcmp(p->buf, ff_w64_guid_riff, 16) && |
763 | !memcmp(p->buf + 24, ff_w64_guid_wave, 16)) |
764 | return AVPROBE_SCORE_MAX; |
765 | else |
766 | return 0; |
767 | } |
768 | |
769 | static int w64_read_header(AVFormatContext *s) |
770 | { |
771 | int64_t size, data_ofs = 0; |
772 | AVIOContext *pb = s->pb; |
773 | WAVDemuxContext *wav = s->priv_data; |
774 | AVStream *st; |
775 | uint8_t guid[16]; |
776 | int ret; |
777 | |
778 | avio_read(pb, guid, 16); |
779 | if (memcmp(guid, ff_w64_guid_riff, 16)) |
780 | return AVERROR_INVALIDDATA; |
781 | |
782 | /* riff + wave + fmt + sizes */ |
783 | if (avio_rl64(pb) < 16 + 8 + 16 + 8 + 16 + 8) |
784 | return AVERROR_INVALIDDATA; |
785 | |
786 | avio_read(pb, guid, 16); |
787 | if (memcmp(guid, ff_w64_guid_wave, 16)) { |
788 | av_log(s, AV_LOG_ERROR, "could not find wave guid\n"); |
789 | return AVERROR_INVALIDDATA; |
790 | } |
791 | |
792 | wav->w64 = 1; |
793 | |
794 | st = avformat_new_stream(s, NULL); |
795 | if (!st) |
796 | return AVERROR(ENOMEM); |
797 | |
798 | while (!avio_feof(pb)) { |
799 | if (avio_read(pb, guid, 16) != 16) |
800 | break; |
801 | size = avio_rl64(pb); |
802 | if (size <= 24 || INT64_MAX - size < avio_tell(pb)) |
803 | return AVERROR_INVALIDDATA; |
804 | |
805 | if (!memcmp(guid, ff_w64_guid_fmt, 16)) { |
806 | /* subtract chunk header size - normal wav file doesn't count it */ |
807 | ret = ff_get_wav_header(s, pb, st->codecpar, size - 24, 0); |
808 | if (ret < 0) |
809 | return ret; |
810 | avio_skip(pb, FFALIGN(size, INT64_C(8)) - size); |
811 | |
812 | avpriv_set_pts_info(st, 64, 1, st->codecpar->sample_rate); |
813 | } else if (!memcmp(guid, ff_w64_guid_fact, 16)) { |
814 | int64_t samples; |
815 | |
816 | samples = avio_rl64(pb); |
817 | if (samples > 0) |
818 | st->duration = samples; |
819 | } else if (!memcmp(guid, ff_w64_guid_data, 16)) { |
820 | wav->data_end = avio_tell(pb) + size - 24; |
821 | |
822 | data_ofs = avio_tell(pb); |
823 | if (!(pb->seekable & AVIO_SEEKABLE_NORMAL)) |
824 | break; |
825 | |
826 | avio_skip(pb, size - 24); |
827 | } else if (!memcmp(guid, ff_w64_guid_summarylist, 16)) { |
828 | int64_t start, end, cur; |
829 | uint32_t count, chunk_size, i; |
830 | |
831 | start = avio_tell(pb); |
832 | end = start + FFALIGN(size, INT64_C(8)) - 24; |
833 | count = avio_rl32(pb); |
834 | |
835 | for (i = 0; i < count; i++) { |
836 | char chunk_key[5], *value; |
837 | |
838 | if (avio_feof(pb) || (cur = avio_tell(pb)) < 0 || cur > end - 8 /* = tag + size */) |
839 | break; |
840 | |
841 | chunk_key[4] = 0; |
842 | avio_read(pb, chunk_key, 4); |
843 | chunk_size = avio_rl32(pb); |
844 | |
845 | value = av_mallocz(chunk_size + 1); |
846 | if (!value) |
847 | return AVERROR(ENOMEM); |
848 | |
849 | ret = avio_get_str16le(pb, chunk_size, value, chunk_size); |
850 | avio_skip(pb, chunk_size - ret); |
851 | |
852 | av_dict_set(&s->metadata, chunk_key, value, AV_DICT_DONT_STRDUP_VAL); |
853 | } |
854 | |
855 | avio_skip(pb, end - avio_tell(pb)); |
856 | } else { |
857 | av_log(s, AV_LOG_DEBUG, "unknown guid: "FF_PRI_GUID"\n", FF_ARG_GUID(guid)); |
858 | avio_skip(pb, FFALIGN(size, INT64_C(8)) - 24); |
859 | } |
860 | } |
861 | |
862 | if (!data_ofs) |
863 | return AVERROR_EOF; |
864 | |
865 | ff_metadata_conv_ctx(s, NULL, wav_metadata_conv); |
866 | ff_metadata_conv_ctx(s, NULL, ff_riff_info_conv); |
867 | |
868 | handle_stream_probing(st); |
869 | st->need_parsing = AVSTREAM_PARSE_FULL_RAW; |
870 | |
871 | avio_seek(pb, data_ofs, SEEK_SET); |
872 | |
873 | set_spdif(s, wav); |
874 | |
875 | return 0; |
876 | } |
877 | |
878 | AVInputFormat ff_w64_demuxer = { |
879 | .name = "w64", |
880 | .long_name = NULL_IF_CONFIG_SMALL("Sony Wave64"), |
881 | .priv_data_size = sizeof(WAVDemuxContext), |
882 | .read_probe = w64_probe, |
883 | .read_header = w64_read_header, |
884 | .read_packet = wav_read_packet, |
885 | .read_seek = wav_read_seek, |
886 | .flags = AVFMT_GENERIC_INDEX, |
887 | .codec_tag = (const AVCodecTag * const []) { ff_codec_wav_tags, 0 }, |
888 | }; |
889 | #endif /* CONFIG_W64_DEMUXER */ |
890 |