FFmpeg
oggparsedaala.c
Go to the documentation of this file.
1 /*
2  * Ogg Daala parser
3  * Copyright (C) 2015 Rostislav Pehlivanov <atomnuker gmail com>
4  * Copyright (C) 2015 Vittorio Giovara <vittorio.giovara gmail com>
5  *
6  * This file is part of FFmpeg.
7  *
8  * FFmpeg is free software; you can redistribute it and/or
9  * modify it under the terms of the GNU Lesser General Public
10  * License as published by the Free Software Foundation; either
11  * version 2.1 of the License, or (at your option) any later version.
12  *
13  * FFmpeg is distributed in the hope that it will be useful,
14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16  * Lesser General Public License for more details.
17  *
18  * You should have received a copy of the GNU Lesser General Public
19  * License along with FFmpeg; if not, write to the Free Software
20  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21  */
22 
23 #include <stdlib.h>
24 #include "libavcodec/bytestream.h"
25 #include "avformat.h"
26 #include "internal.h"
27 #include "oggdec.h"
28 
31  int depth;
32  int planes;
33  int xdec[4];
34  int ydec[4];
35 };
36 
37 /* Currently supported formats only */
38 static const struct DaalaPixFmtMap list_fmts[] = {
39  { AV_PIX_FMT_YUV420P, 8, 3, {0, 1, 1, 0}, {0, 1, 1, 0} },
40  { AV_PIX_FMT_YUV444P, 8, 3, {0, 0, 0, 0}, {0, 0, 0, 0} }
41 };
42 
43 typedef struct DaalaInfoHeader {
44  int init_d;
45  int fpr;
46  int gpshift;
47  int gpmask;
55 
56 static inline int daala_match_pix_fmt(struct DaalaPixFmtMap *fmt)
57 {
58  int i, j;
59  for (i = 0; i < FF_ARRAY_ELEMS(list_fmts); i++) {
60  int match = 0;
61  if (fmt->depth != list_fmts[i].depth)
62  continue;
63  if (fmt->planes != list_fmts[i].planes)
64  continue;
65  for (j = 0; j < fmt->planes; j++) {
66  if (fmt->xdec[j] != list_fmts[i].xdec[j])
67  continue;
68  if (fmt->ydec[j] != list_fmts[i].ydec[j])
69  continue;
70  match++;
71  }
72  if (match == fmt->planes)
73  return list_fmts[i].ffmpeg_fmt;
74  }
75  return -1;
76 }
77 
78 static int daala_header(AVFormatContext *s, int idx)
79 {
80  int i, err;
81  uint8_t *cdp;
82  GetByteContext gb;
83  AVRational timebase;
84  struct ogg *ogg = s->priv_data;
85  struct ogg_stream *os = ogg->streams + idx;
86  AVStream *st = s->streams[idx];
87  int cds = st->codecpar->extradata_size + os->psize + 2;
88  DaalaInfoHeader *hdr = os->private;
89 
90  if (!(os->buf[os->pstart] & 0x80))
91  return 0;
92 
93  if (!hdr) {
94  hdr = av_mallocz(sizeof(*hdr));
95  if (!hdr)
96  return AVERROR(ENOMEM);
97  os->private = hdr;
98  }
99 
100  switch (os->buf[os->pstart]) {
101  case 0x80:
102  bytestream2_init(&gb, os->buf + os->pstart, os->psize);
104 
105  hdr->version_maj = bytestream2_get_byte(&gb);
106  hdr->version_min = bytestream2_get_byte(&gb);
107  hdr->version_sub = bytestream2_get_byte(&gb);
108 
109  st->codecpar->width = bytestream2_get_ne32(&gb);
111 
114 
115  timebase.num = bytestream2_get_ne32(&gb);
116  timebase.den = bytestream2_get_ne32(&gb);
117  if (timebase.num < 0 && timebase.den < 0) {
118  av_log(s, AV_LOG_WARNING, "Invalid timebase, assuming 30 FPS\n");
119  timebase.num = 1;
120  timebase.den = 30;
121  }
122  avpriv_set_pts_info(st, 64, timebase.den, timebase.num);
123 
125  hdr->gpshift = bytestream2_get_byte(&gb);
126  if (hdr->gpshift >= 32) {
127  av_log(s, AV_LOG_ERROR, "Too large gpshift %d (>= 32).\n",
128  hdr->gpshift);
129  hdr->gpshift = 0;
130  return AVERROR_INVALIDDATA;
131  }
132  hdr->gpmask = (1U << hdr->gpshift) - 1;
133 
134  hdr->format.depth = 8 + 2*(bytestream2_get_byte(&gb)-1);
135 
136  hdr->fpr = bytestream2_get_byte(&gb);
137 
138  hdr->format.planes = bytestream2_get_byte(&gb);
139  if (hdr->format.planes > 4) {
140  av_log(s, AV_LOG_ERROR,
141  "Invalid number of planes %d in daala pixel format map.\n",
142  hdr->format.planes);
143  return AVERROR_INVALIDDATA;
144  }
145  for (i = 0; i < hdr->format.planes; i++) {
146  hdr->format.xdec[i] = bytestream2_get_byte(&gb);
147  hdr->format.ydec[i] = bytestream2_get_byte(&gb);
148  }
149 
150  if ((st->codecpar->format = daala_match_pix_fmt(&hdr->format)) < 0)
151  av_log(s, AV_LOG_ERROR, "Unsupported pixel format - %i %i\n",
152  hdr->format.depth, hdr->format.planes);
153 
157 
158  hdr->init_d = 1;
159  break;
160  case 0x81:
161  if (!hdr->init_d)
162  return AVERROR_INVALIDDATA;
164  os->buf + os->pstart + ff_daala_codec.magicsize,
166  break;
167  case 0x82:
168  if (!hdr->init_d)
169  return AVERROR_INVALIDDATA;
170  break;
171  default:
172  av_log(s, AV_LOG_ERROR, "Unknown header type %X\n", os->buf[os->pstart]);
173  return AVERROR_INVALIDDATA;
174  break;
175  }
176 
177  if ((err = av_reallocp(&st->codecpar->extradata,
178  cds + AV_INPUT_BUFFER_PADDING_SIZE)) < 0) {
179  st->codecpar->extradata_size = 0;
180  return err;
181  }
182 
183  memset(st->codecpar->extradata + cds, 0, AV_INPUT_BUFFER_PADDING_SIZE);
184  cdp = st->codecpar->extradata + st->codecpar->extradata_size;
185  *cdp++ = os->psize >> 8;
186  *cdp++ = os->psize & 0xff;
187  memcpy(cdp, os->buf + os->pstart, os->psize);
188  st->codecpar->extradata_size = cds;
189 
190  return 1;
191 }
192 
193 static uint64_t daala_gptopts(AVFormatContext *ctx, int idx, uint64_t gp,
194  int64_t *dts)
195 {
196  uint64_t iframe, pframe;
197  struct ogg *ogg = ctx->priv_data;
198  struct ogg_stream *os = ogg->streams + idx;
199  DaalaInfoHeader *hdr = os->private;
200 
201  if (!hdr)
202  return AV_NOPTS_VALUE;
203 
204  iframe = gp >> hdr->gpshift;
205  pframe = gp & hdr->gpmask;
206 
207  if (!pframe)
208  os->pflags |= AV_PKT_FLAG_KEY;
209 
210  if (dts)
211  *dts = iframe + pframe;
212 
213  return iframe + pframe;
214 }
215 
216 static int daala_packet(AVFormatContext *s, int idx)
217 {
218  int seg, duration = 1;
219  struct ogg *ogg = s->priv_data;
220  struct ogg_stream *os = ogg->streams + idx;
221  int64_t pts;
222 
223  /*
224  * first packet handling: here we parse the duration of each packet in the
225  * first page and compare the total duration to the page granule to find the
226  * encoder delay and set the first timestamp
227  */
228 
229  if ((!os->lastpts || os->lastpts == AV_NOPTS_VALUE) && !(os->flags & OGG_FLAG_EOS)) {
230  for (seg = os->segp; seg < os->nsegs; seg++)
231  if (os->segments[seg] < 255)
232  duration++;
233 
234  pts = daala_gptopts(s, idx, os->granule, NULL);
235  if (pts != AV_NOPTS_VALUE)
236  pts -= duration;
237  os->lastpts = os->lastdts = pts;
238  if(s->streams[idx]->start_time == AV_NOPTS_VALUE) {
239  s->streams[idx]->start_time = os->lastpts;
240  if (s->streams[idx]->duration != AV_NOPTS_VALUE)
241  s->streams[idx]->duration -= s->streams[idx]->start_time;
242  }
243  }
244 
245  /* parse packet duration */
246  if (os->psize > 0)
247  os->pduration = 1;
248 
249  return 0;
250 }
251 
252 const struct ogg_codec ff_daala_codec = {
253  .name = "Daala",
254  .magic = "\200daala",
255  .magicsize = 6,
256  .header = daala_header,
257  .packet = daala_packet,
258  .gptopts = daala_gptopts,
259  .granule_is_start = 1,
260  .nb_header = 3,
261 };
#define NULL
Definition: coverity.c:32
#define AVERROR_INVALIDDATA
Invalid data found when processing input.
Definition: error.h:59
Copyright (C) 2005 Michael Ahlberg, Måns Rullgård.
Definition: oggdec.h:31
unsigned int pflags
Definition: oggdec.h:67
planar YUV 4:4:4, 24bpp, (1 Cr & Cb sample per 1x1 Y samples)
Definition: pixfmt.h:71
#define AV_LOG_WARNING
Something somehow does not look correct.
Definition: log.h:182
void avpriv_set_pts_info(AVStream *s, int pts_wrap_bits, unsigned int pts_num, unsigned int pts_den)
Set the time base and wrapping info for a given stream.
Definition: utils.c:4938
enum AVCodecID codec_id
Specific type of the encoded data (the codec used).
Definition: codec_par.h:60
AVRational sample_aspect_ratio
sample aspect ratio (0 if unknown)
Definition: avformat.h:938
int num
Numerator.
Definition: rational.h:59
static av_always_inline void bytestream2_init(GetByteContext *g, const uint8_t *buf, int buf_size)
Definition: bytestream.h:133
void * av_mallocz(size_t size)
Allocate a memory block with alignment suitable for all memory accesses (including vectors if availab...
Definition: mem.c:239
int flags
Definition: oggdec.h:76
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample format(the sample packing is implied by the sample format) and sample rate.The lists are not just lists
int64_t lastpts
Definition: oggdec.h:72
#define bytestream2_get_ne32
Definition: bytestream.h:117
Format I/O context.
Definition: avformat.h:1351
unsigned int psize
Definition: oggdec.h:66
uint8_t
int width
Video only.
Definition: codec_par.h:126
struct DaalaPixFmtMap format
Definition: oggparsedaala.c:53
enum AVStreamParseType need_parsing
Definition: avformat.h:1094
static uint64_t daala_gptopts(AVFormatContext *ctx, int idx, uint64_t gp, int64_t *dts)
AVStream ** streams
A list of all streams in the file.
Definition: avformat.h:1419
int64_t duration
Definition: movenc.c:63
int keyframe_granule_shift
Definition: oggparsedaala.c:52
#define av_log(a,...)
#define AV_PKT_FLAG_KEY
The packet contains a keyframe.
Definition: packet.h:388
#define U(x)
Definition: vp56_arith.h:37
#define i(width, name, range_min, range_max)
Definition: cbs_h2645.c:269
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
Definition: log.h:176
static av_always_inline void bytestream2_skip(GetByteContext *g, unsigned int size)
Definition: bytestream.h:164
int ff_vorbis_stream_comment(AVFormatContext *as, AVStream *st, const uint8_t *buf, int size)
#define OGG_FLAG_EOS
Definition: oggdec.h:113
enum AVMediaType codec_type
General type of the encoded data.
Definition: codec_par.h:56
uint8_t segments[255]
Definition: oggdec.h:80
int extradata_size
Size of the extradata content in bytes.
Definition: codec_par.h:78
Only parse headers, do not repack.
Definition: avformat.h:796
uint64_t granule
Definition: oggdec.h:70
unsigned int pstart
Definition: oggdec.h:65
AVFormatContext * ctx
Definition: movenc.c:48
struct ogg_stream * streams
Definition: oggdec.h:103
int segp
Definition: oggdec.h:79
#define s(width, name)
Definition: cbs_vp9.c:257
#define FF_ARRAY_ELEMS(a)
Stream structure.
Definition: avformat.h:876
int av_reallocp(void *ptr, size_t size)
Allocate, reallocate, or free a block of memory through a pointer to a pointer.
Definition: mem.c:163
unsigned int pduration
Definition: oggdec.h:68
int nsegs
Definition: oggdec.h:79
const int8_t * name
Definition: oggdec.h:34
enum AVPixelFormat ffmpeg_fmt
Definition: oggparsedaala.c:30
void * private
Definition: oggdec.h:91
Rational number (pair of numerator and denominator).
Definition: rational.h:58
int64_t lastdts
Definition: oggdec.h:73
static int64_t pts
static int daala_packet(AVFormatContext *s, int idx)
const struct ogg_codec ff_daala_codec
int64_t duration
Decoding: duration of the stream, in stream time base.
Definition: avformat.h:925
uint8_t * buf
Definition: oggdec.h:62
Main libavformat public API header.
planar YUV 4:2:0, 12bpp, (1 Cr & Cb sample per 2x2 Y samples)
Definition: pixfmt.h:66
int64_t start_time
Decoding: pts of the first frame of the stream in presentation order, in stream time base...
Definition: avformat.h:915
static int daala_match_pix_fmt(struct DaalaPixFmtMap *fmt)
Definition: oggparsedaala.c:56
int den
Denominator.
Definition: rational.h:60
#define AV_INPUT_BUFFER_PADDING_SIZE
Required number of additionally allocated bytes at the end of the input bitstream for decoding...
Definition: avcodec.h:214
Definition: oggdec.h:102
void * priv_data
Format private data.
Definition: avformat.h:1379
static int daala_header(AVFormatContext *s, int idx)
Definition: oggparsedaala.c:78
uint8_t * extradata
Extra binary data needed for initializing the decoder, codec-dependent.
Definition: codec_par.h:74
uint8_t magicsize
Definition: oggdec.h:33
#define gp
Definition: regdef.h:62
AVCodecParameters * codecpar
Codec parameters associated with this stream.
Definition: avformat.h:1023
static const struct DaalaPixFmtMap list_fmts[]
Definition: oggparsedaala.c:38
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a all references to both lists are replaced with a reference to the intersection And when a single format is eventually chosen for a link amongst the remaining all references to the list are updated That means that if a filter requires that its input and output have the same format amongst a supported all it has to do is use a reference to the same list of formats query_formats can leave some formats unset and return AVERROR(EAGAIN) to cause the negotiation mechanism toagain later.That can be used by filters with complex requirements to use the format negotiated on one link to set the formats supported on another.Frame references ownership and permissions
AVPixelFormat
Pixel format.
Definition: pixfmt.h:64
#define AV_NOPTS_VALUE
Undefined timestamp value.
Definition: avutil.h:248