dv.c
Go to the documentation of this file.
1 /*
2  * General DV muxer/demuxer
3  * Copyright (c) 2003 Roman Shaposhnik
4  *
5  * Many thanks to Dan Dennedy <dan@dennedy.org> for providing wealth
6  * of DV technical info.
7  *
8  * Raw DV format
9  * Copyright (c) 2002 Fabrice Bellard
10  *
11  * 50 Mbps (DVCPRO50) and 100 Mbps (DVCPRO HD) support
12  * Copyright (c) 2006 Daniel Maas <dmaas@maasdigital.com>
13  * Funded by BBC Research & Development
14  *
15  * This file is part of Libav.
16  *
17  * Libav is free software; you can redistribute it and/or
18  * modify it under the terms of the GNU Lesser General Public
19  * License as published by the Free Software Foundation; either
20  * version 2.1 of the License, or (at your option) any later version.
21  *
22  * Libav is distributed in the hope that it will be useful,
23  * but WITHOUT ANY WARRANTY; without even the implied warranty of
24  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
25  * Lesser General Public License for more details.
26  *
27  * You should have received a copy of the GNU Lesser General Public
28  * License along with Libav; if not, write to the Free Software
29  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
30  */
31 #include <time.h>
32 #include "avformat.h"
33 #include "internal.h"
34 #include "libavcodec/dvdata.h"
35 #include "libavutil/intreadwrite.h"
36 #include "libavutil/mathematics.h"
37 #include "dv.h"
38 
40  const DVprofile* sys; /* Current DV profile. E.g.: 525/60, 625/50 */
45  uint8_t audio_buf[4][8192];
46  int ach;
47  int frames;
48  uint64_t abytes;
49 };
50 
51 static inline uint16_t dv_audio_12to16(uint16_t sample)
52 {
53  uint16_t shift, result;
54 
55  sample = (sample < 0x800) ? sample : sample | 0xf000;
56  shift = (sample & 0xf00) >> 8;
57 
58  if (shift < 0x2 || shift > 0xd) {
59  result = sample;
60  } else if (shift < 0x8) {
61  shift--;
62  result = (sample - (256 * shift)) << shift;
63  } else {
64  shift = 0xe - shift;
65  result = ((sample + ((256 * shift) + 1)) << shift) - 1;
66  }
67 
68  return result;
69 }
70 
71 /*
72  * This is the dumbest implementation of all -- it simply looks at
73  * a fixed offset and if pack isn't there -- fails. We might want
74  * to have a fallback mechanism for complete search of missing packs.
75  */
76 static const uint8_t* dv_extract_pack(uint8_t* frame, enum dv_pack_type t)
77 {
78  int offs;
79 
80  switch (t) {
81  case dv_audio_source:
82  offs = (80*6 + 80*16*3 + 3);
83  break;
84  case dv_audio_control:
85  offs = (80*6 + 80*16*4 + 3);
86  break;
87  case dv_video_control:
88  offs = (80*5 + 48 + 5);
89  break;
90  default:
91  return NULL;
92  }
93 
94  return frame[offs] == t ? &frame[offs] : NULL;
95 }
96 
97 /*
98  * There's a couple of assumptions being made here:
99  * 1. By default we silence erroneous (0x8000/16bit 0x800/12bit) audio samples.
100  * We can pass them upwards when libavcodec will be ready to deal with them.
101  * 2. We don't do software emphasis.
102  * 3. Audio is always returned as 16bit linear samples: 12bit nonlinear samples
103  * are converted into 16bit linear ones.
104  */
105 static int dv_extract_audio(uint8_t *frame, uint8_t **ppcm,
106  const DVprofile *sys)
107 {
108  int size, chan, i, j, d, of, smpls, freq, quant, half_ch;
109  uint16_t lc, rc;
110  const uint8_t* as_pack;
111  uint8_t *pcm, ipcm;
112 
113  as_pack = dv_extract_pack(frame, dv_audio_source);
114  if (!as_pack) /* No audio ? */
115  return 0;
116 
117  smpls = as_pack[1] & 0x3f; /* samples in this frame - min. samples */
118  freq = (as_pack[4] >> 3) & 0x07; /* 0 - 48kHz, 1 - 44,1kHz, 2 - 32kHz */
119  quant = as_pack[4] & 0x07; /* 0 - 16bit linear, 1 - 12bit nonlinear */
120 
121  if (quant > 1)
122  return -1; /* unsupported quantization */
123 
124  if (freq >= FF_ARRAY_ELEMS(dv_audio_frequency))
125  return AVERROR_INVALIDDATA;
126 
127  size = (sys->audio_min_samples[freq] + smpls) * 4; /* 2ch, 2bytes */
128  half_ch = sys->difseg_size / 2;
129 
130  /* We work with 720p frames split in half, thus even frames have
131  * channels 0,1 and odd 2,3. */
132  ipcm = (sys->height == 720 && !(frame[1] & 0x0C)) ? 2 : 0;
133 
134  /* for each DIF channel */
135  for (chan = 0; chan < sys->n_difchan; chan++) {
136  /* next stereo channel (50Mbps and 100Mbps only) */
137  pcm = ppcm[ipcm++];
138  if (!pcm)
139  break;
140 
141  /* for each DIF segment */
142  for (i = 0; i < sys->difseg_size; i++) {
143  frame += 6 * 80; /* skip DIF segment header */
144  if (quant == 1 && i == half_ch) {
145  /* next stereo channel (12bit mode only) */
146  pcm = ppcm[ipcm++];
147  if (!pcm)
148  break;
149  }
150 
151  /* for each AV sequence */
152  for (j = 0; j < 9; j++) {
153  for (d = 8; d < 80; d += 2) {
154  if (quant == 0) { /* 16bit quantization */
155  of = sys->audio_shuffle[i][j] + (d - 8) / 2 * sys->audio_stride;
156  if (of*2 >= size)
157  continue;
158 
159  pcm[of*2] = frame[d+1]; // FIXME: maybe we have to admit
160  pcm[of*2+1] = frame[d]; // that DV is a big-endian PCM
161  if (pcm[of*2+1] == 0x80 && pcm[of*2] == 0x00)
162  pcm[of*2+1] = 0;
163  } else { /* 12bit quantization */
164  lc = ((uint16_t)frame[d] << 4) |
165  ((uint16_t)frame[d+2] >> 4);
166  rc = ((uint16_t)frame[d+1] << 4) |
167  ((uint16_t)frame[d+2] & 0x0f);
168  lc = (lc == 0x800 ? 0 : dv_audio_12to16(lc));
169  rc = (rc == 0x800 ? 0 : dv_audio_12to16(rc));
170 
171  of = sys->audio_shuffle[i%half_ch][j] + (d - 8) / 3 * sys->audio_stride;
172  if (of*2 >= size)
173  continue;
174 
175  pcm[of*2] = lc & 0xff; // FIXME: maybe we have to admit
176  pcm[of*2+1] = lc >> 8; // that DV is a big-endian PCM
177  of = sys->audio_shuffle[i%half_ch+half_ch][j] +
178  (d - 8) / 3 * sys->audio_stride;
179  pcm[of*2] = rc & 0xff; // FIXME: maybe we have to admit
180  pcm[of*2+1] = rc >> 8; // that DV is a big-endian PCM
181  ++d;
182  }
183  }
184 
185  frame += 16 * 80; /* 15 Video DIFs + 1 Audio DIF */
186  }
187  }
188  }
189 
190  return size;
191 }
192 
193 static int dv_extract_audio_info(DVDemuxContext* c, uint8_t* frame)
194 {
195  const uint8_t* as_pack;
196  int freq, stype, smpls, quant, i, ach;
197 
198  as_pack = dv_extract_pack(frame, dv_audio_source);
199  if (!as_pack || !c->sys) { /* No audio ? */
200  c->ach = 0;
201  return 0;
202  }
203 
204  smpls = as_pack[1] & 0x3f; /* samples in this frame - min. samples */
205  freq = (as_pack[4] >> 3) & 0x07; /* 0 - 48kHz, 1 - 44,1kHz, 2 - 32kHz */
206  stype = (as_pack[3] & 0x1f); /* 0 - 2CH, 2 - 4CH, 3 - 8CH */
207  quant = as_pack[4] & 0x07; /* 0 - 16bit linear, 1 - 12bit nonlinear */
208 
209  if (freq >= FF_ARRAY_ELEMS(dv_audio_frequency)) {
211  "Unrecognized audio sample rate index (%d)\n", freq);
212  return 0;
213  }
214 
215  if (stype > 3) {
216  av_log(c->fctx, AV_LOG_ERROR, "stype %d is invalid\n", stype);
217  c->ach = 0;
218  return 0;
219  }
220 
221  /* note: ach counts PAIRS of channels (i.e. stereo channels) */
222  ach = ((int[4]){ 1, 0, 2, 4})[stype];
223  if (ach == 1 && quant && freq == 2)
224  ach = 2;
225 
226  /* Dynamic handling of the audio streams in DV */
227  for (i = 0; i < ach; i++) {
228  if (!c->ast[i]) {
229  c->ast[i] = avformat_new_stream(c->fctx, NULL);
230  if (!c->ast[i])
231  break;
232  avpriv_set_pts_info(c->ast[i], 64, 1, 30000);
235 
236  av_init_packet(&c->audio_pkt[i]);
237  c->audio_pkt[i].size = 0;
238  c->audio_pkt[i].data = c->audio_buf[i];
239  c->audio_pkt[i].stream_index = c->ast[i]->index;
241  }
242  c->ast[i]->codec->sample_rate = dv_audio_frequency[freq];
243  c->ast[i]->codec->channels = 2;
244  c->ast[i]->codec->bit_rate = 2 * dv_audio_frequency[freq] * 16;
245  c->ast[i]->start_time = 0;
246  }
247  c->ach = i;
248 
249  return (c->sys->audio_min_samples[freq] + smpls) * 4; /* 2ch, 2bytes */;
250 }
251 
252 static int dv_extract_video_info(DVDemuxContext *c, uint8_t* frame)
253 {
254  const uint8_t* vsc_pack;
255  AVCodecContext* avctx;
256  int apt, is16_9;
257  int size = 0;
258 
259  if (c->sys) {
260  avctx = c->vst->codec;
261 
263  c->sys->time_base.den);
264  avctx->time_base= c->sys->time_base;
265  if (!avctx->width){
266  avctx->width = c->sys->width;
267  avctx->height = c->sys->height;
268  }
269  avctx->pix_fmt = c->sys->pix_fmt;
270 
271  /* finding out SAR is a little bit messy */
272  vsc_pack = dv_extract_pack(frame, dv_video_control);
273  apt = frame[4] & 0x07;
274  is16_9 = (vsc_pack && ((vsc_pack[2] & 0x07) == 0x02 ||
275  (!apt && (vsc_pack[2] & 0x07) == 0x07)));
276  c->vst->sample_aspect_ratio = c->sys->sar[is16_9];
277  avctx->bit_rate = av_rescale_q(c->sys->frame_size, (AVRational){8,1},
278  c->sys->time_base);
279  size = c->sys->frame_size;
280  }
281  return size;
282 }
283 
284 /*
285  * The following 3 functions constitute our interface to the world
286  */
287 
289 {
290  DVDemuxContext *c;
291 
292  c = av_mallocz(sizeof(DVDemuxContext));
293  if (!c)
294  return NULL;
295 
296  c->vst = avformat_new_stream(s, NULL);
297  if (!c->vst) {
298  av_free(c);
299  return NULL;
300  }
301 
302  c->sys = NULL;
303  c->fctx = s;
304  memset(c->ast, 0, sizeof(c->ast));
305  c->ach = 0;
306  c->frames = 0;
307  c->abytes = 0;
308 
311  c->vst->codec->bit_rate = 25000000;
312  c->vst->start_time = 0;
313 
314  return c;
315 }
316 
318 {
319  int size = -1;
320  int i;
321 
322  for (i = 0; i < c->ach; i++) {
323  if (c->ast[i] && c->audio_pkt[i].size) {
324  *pkt = c->audio_pkt[i];
325  c->audio_pkt[i].size = 0;
326  size = pkt->size;
327  break;
328  }
329  }
330 
331  return size;
332 }
333 
335  uint8_t* buf, int buf_size)
336 {
337  int size, i;
338  uint8_t *ppcm[5] = { 0 };
339 
340  if (buf_size < DV_PROFILE_BYTES ||
341  !(c->sys = avpriv_dv_frame_profile(c->sys, buf, buf_size)) ||
342  buf_size < c->sys->frame_size) {
343  return -1; /* Broken frame, or not enough data */
344  }
345 
346  /* Queueing audio packet */
347  /* FIXME: in case of no audio/bad audio we have to do something */
348  size = dv_extract_audio_info(c, buf);
349  for (i = 0; i < c->ach; i++) {
350  c->audio_pkt[i].size = size;
351  c->audio_pkt[i].pts = c->abytes * 30000*8 / c->ast[i]->codec->bit_rate;
352  ppcm[i] = c->audio_buf[i];
353  }
354  if (c->ach)
355  dv_extract_audio(buf, ppcm, c->sys);
356 
357  /* We work with 720p frames split in half, thus even frames have
358  * channels 0,1 and odd 2,3. */
359  if (c->sys->height == 720) {
360  if (buf[1] & 0x0C) {
361  c->audio_pkt[2].size = c->audio_pkt[3].size = 0;
362  } else {
363  c->audio_pkt[0].size = c->audio_pkt[1].size = 0;
364  c->abytes += size;
365  }
366  } else {
367  c->abytes += size;
368  }
369 
370  /* Now it's time to return video packet */
371  size = dv_extract_video_info(c, buf);
372  av_init_packet(pkt);
373  pkt->data = buf;
374  pkt->size = size;
375  pkt->flags |= AV_PKT_FLAG_KEY;
376  pkt->stream_index = c->vst->id;
377  pkt->pts = c->frames;
378 
379  c->frames++;
380 
381  return size;
382 }
383 
385  int64_t timestamp, int flags)
386 {
387  // FIXME: sys may be wrong if last dv_read_packet() failed (buffer is junk)
388  const DVprofile* sys = avpriv_dv_codec_profile(c->vst->codec);
389  int64_t offset;
390  int64_t size = avio_size(s->pb) - s->data_offset;
391  int64_t max_offset = ((size-1) / sys->frame_size) * sys->frame_size;
392 
393  offset = sys->frame_size * timestamp;
394 
395  if (size >= 0 && offset > max_offset) offset = max_offset;
396  else if (offset < 0) offset = 0;
397 
398  return offset + s->data_offset;
399 }
400 
401 void dv_offset_reset(DVDemuxContext *c, int64_t frame_offset)
402 {
403  c->frames= frame_offset;
404  if (c->ach)
405  c->abytes= av_rescale_q(c->frames, c->sys->time_base,
406  (AVRational){8, c->ast[0]->codec->bit_rate});
407  c->audio_pkt[0].size = c->audio_pkt[1].size = 0;
408  c->audio_pkt[2].size = c->audio_pkt[3].size = 0;
409 }
410 
411 /************************************************************
412  * Implementation of the easiest DV storage of all -- raw DV.
413  ************************************************************/
414 
415 typedef struct RawDVContext {
418 } RawDVContext;
419 
421  AVFormatParameters *ap)
422 {
423  unsigned state, marker_pos = 0;
424  RawDVContext *c = s->priv_data;
425 
427  if (!c->dv_demux)
428  return -1;
429 
430  state = avio_rb32(s->pb);
431  while ((state & 0xffffff7f) != 0x1f07003f) {
432  if (s->pb->eof_reached) {
433  av_log(s, AV_LOG_ERROR, "Cannot find DV header.\n");
434  return -1;
435  }
436  if (state == 0x003f0700 || state == 0xff3f0700)
437  marker_pos = avio_tell(s->pb);
438  if (state == 0xff3f0701 && avio_tell(s->pb) - marker_pos == 80) {
439  avio_seek(s->pb, -163, SEEK_CUR);
440  state = avio_rb32(s->pb);
441  break;
442  }
443  state = (state << 8) | avio_r8(s->pb);
444  }
445  AV_WB32(c->buf, state);
446 
447  if (avio_read(s->pb, c->buf + 4, DV_PROFILE_BYTES - 4) <= 0 ||
448  avio_seek(s->pb, -DV_PROFILE_BYTES, SEEK_CUR) < 0)
449  return AVERROR(EIO);
450 
452  if (!c->dv_demux->sys) {
453  av_log(s, AV_LOG_ERROR, "Can't determine profile of DV input stream.\n");
454  return -1;
455  }
456 
458  c->dv_demux->sys->time_base);
459 
460  return 0;
461 }
462 
463 
465 {
466  int size;
467  RawDVContext *c = s->priv_data;
468 
469  size = avpriv_dv_get_packet(c->dv_demux, pkt);
470 
471  if (size < 0) {
472  if (!c->dv_demux->sys)
473  return AVERROR(EIO);
474  size = c->dv_demux->sys->frame_size;
475  if (avio_read(s->pb, c->buf, size) <= 0)
476  return AVERROR(EIO);
477 
478  size = avpriv_dv_produce_packet(c->dv_demux, pkt, c->buf, size);
479  }
480 
481  return size;
482 }
483 
484 static int dv_read_seek(AVFormatContext *s, int stream_index,
485  int64_t timestamp, int flags)
486 {
487  RawDVContext *r = s->priv_data;
488  DVDemuxContext *c = r->dv_demux;
489  int64_t offset = dv_frame_offset(s, c, timestamp, flags);
490 
491  dv_offset_reset(c, offset / c->sys->frame_size);
492 
493  offset = avio_seek(s->pb, offset, SEEK_SET);
494  return (offset < 0) ? offset : 0;
495 }
496 
498 {
499  RawDVContext *c = s->priv_data;
500  av_free(c->dv_demux);
501  return 0;
502 }
503 
504 static int dv_probe(AVProbeData *p)
505 {
506  unsigned state, marker_pos = 0;
507  int i;
508  int matches = 0;
509  int secondary_matches = 0;
510 
511  if (p->buf_size < 5)
512  return 0;
513 
514  state = AV_RB32(p->buf);
515  for (i = 4; i < p->buf_size; i++) {
516  if ((state & 0xffffff7f) == 0x1f07003f)
517  matches++;
518  // any section header, also with seq/chan num != 0,
519  // should appear around every 12000 bytes, at least 10 per frame
520  if ((state & 0xff07ff7f) == 0x1f07003f)
521  secondary_matches++;
522  if (state == 0x003f0700 || state == 0xff3f0700)
523  marker_pos = i;
524  if (state == 0xff3f0701 && i - marker_pos == 80)
525  matches++;
526  state = (state << 8) | p->buf[i];
527  }
528 
529  if (matches && p->buf_size / matches < 1024*1024) {
530  if (matches > 4 || (secondary_matches >= 10 && p->buf_size / secondary_matches < 24000))
531  return AVPROBE_SCORE_MAX*3/4; // not max to avoid dv in mov to match
532  return AVPROBE_SCORE_MAX/4;
533  }
534  return 0;
535 }
536 
537 #if CONFIG_DV_DEMUXER
538 AVInputFormat ff_dv_demuxer = {
539  .name = "dv",
540  .long_name = NULL_IF_CONFIG_SMALL("DV video format"),
541  .priv_data_size = sizeof(RawDVContext),
542  .read_probe = dv_probe,
547  .extensions = "dv,dif",
548 };
549 #endif