FFmpeg
cafenc.c
Go to the documentation of this file.
1 /*
2  * Core Audio Format muxer
3  * Copyright (c) 2011 Carl Eugen Hoyos
4  *
5  * This file is part of FFmpeg.
6  *
7  * FFmpeg 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  * FFmpeg 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 FFmpeg; if not, write to the Free Software
19  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20  */
21 
22 #include "avformat.h"
23 #include "caf.h"
24 #include "isom.h"
25 #include "avio_internal.h"
26 #include "mux.h"
27 #include "libavutil/intfloat.h"
28 #include "libavutil/dict.h"
29 #include "libavutil/mem.h"
30 
31 #define FRAME_SIZE_OFFSET 40
32 
33 typedef struct {
37  int packets;
38 } CAFContext;
39 
40 static uint32_t codec_flags(enum AVCodecID codec_id) {
41  switch (codec_id) {
44  return 1; //< kCAFLinearPCMFormatFlagIsFloat
48  return 2; //< kCAFLinearPCMFormatFlagIsLittleEndian
51  return 3; //< kCAFLinearPCMFormatFlagIsFloat | kCAFLinearPCMFormatFlagIsLittleEndian
52  default:
53  return 0;
54  }
55 }
56 
57 static uint32_t samples_per_packet(const AVCodecParameters *par) {
58  enum AVCodecID codec_id = par->codec_id;
59  int channels = par->ch_layout.nb_channels, block_align = par->block_align;
60  int frame_size = par->frame_size, sample_rate = par->sample_rate;
61 
62  switch (codec_id) {
63  case AV_CODEC_ID_PCM_S8:
76  return 1;
77  case AV_CODEC_ID_MACE3:
78  case AV_CODEC_ID_MACE6:
79  return 6;
81  return 64;
82  case AV_CODEC_ID_AMR_NB:
83  case AV_CODEC_ID_GSM:
84  case AV_CODEC_ID_ILBC:
85  case AV_CODEC_ID_QCELP:
86  return 160;
87  case AV_CODEC_ID_GSM_MS:
88  return 320;
89  case AV_CODEC_ID_MP1:
90  return 384;
91  case AV_CODEC_ID_OPUS:
92  return frame_size * 48000 / sample_rate;
93  case AV_CODEC_ID_MP2:
94  case AV_CODEC_ID_MP3:
95  return 1152;
96  case AV_CODEC_ID_AC3:
97  return 1536;
98  case AV_CODEC_ID_QDM2:
99  case AV_CODEC_ID_QDMC:
100  return 2048 * channels;
101  case AV_CODEC_ID_ALAC:
102  return 4096;
104  return (block_align - 4 * channels) * 8 / (4 * channels) + 1;
106  return (block_align - 7 * channels) * 2 / channels + 2;
107  default:
108  return 0;
109  }
110 }
111 
113 {
114  AVIOContext *pb = s->pb;
115  AVCodecParameters *par = s->streams[0]->codecpar;
116  CAFContext *caf = s->priv_data;
117  const AVDictionaryEntry *t = NULL;
118  unsigned int codec_tag = ff_codec_get_tag(ff_codec_caf_tags, par->codec_id);
119  int64_t chunk_size = 0;
120  int frame_size = par->frame_size, sample_rate = par->sample_rate;
121 
122  switch (par->codec_id) {
123  case AV_CODEC_ID_AAC:
124  av_log(s, AV_LOG_ERROR, "muxing codec currently unsupported\n");
125  return AVERROR_PATCHWELCOME;
126  }
127 
128  if (par->codec_id == AV_CODEC_ID_OPUS && par->ch_layout.nb_channels > 2) {
129  av_log(s, AV_LOG_ERROR, "Only mono and stereo are supported for Opus\n");
130  return AVERROR_INVALIDDATA;
131  }
132 
133  if (!codec_tag) {
134  av_log(s, AV_LOG_ERROR, "unsupported codec\n");
135  return AVERROR_INVALIDDATA;
136  }
137 
138  if (!par->block_align && !(pb->seekable & AVIO_SEEKABLE_NORMAL)) {
139  av_log(s, AV_LOG_ERROR, "Muxing variable packet size not supported on non seekable output\n");
140  return AVERROR_INVALIDDATA;
141  }
142 
143  if (par->codec_id != AV_CODEC_ID_MP3 || frame_size != 576)
145 
146  if (par->codec_id == AV_CODEC_ID_OPUS)
147  sample_rate = 48000;
148 
149  ffio_wfourcc(pb, "caff"); //< mFileType
150  avio_wb16(pb, 1); //< mFileVersion
151  avio_wb16(pb, 0); //< mFileFlags
152 
153  ffio_wfourcc(pb, "desc"); //< Audio Description chunk
154  avio_wb64(pb, 32); //< mChunkSize
155  avio_wb64(pb, av_double2int(sample_rate)); //< mSampleRate
156  avio_wl32(pb, codec_tag); //< mFormatID
157  avio_wb32(pb, codec_flags(par->codec_id)); //< mFormatFlags
158  avio_wb32(pb, par->block_align); //< mBytesPerPacket
159  avio_wb32(pb, frame_size); //< mFramesPerPacket
160  avio_wb32(pb, par->ch_layout.nb_channels); //< mChannelsPerFrame
161  avio_wb32(pb, av_get_bits_per_sample(par->codec_id)); //< mBitsPerChannel
162 
164  ffio_wfourcc(pb, "chan");
165  avio_wb64(pb, 12);
166  ff_mov_write_chan(pb, par->ch_layout.u.mask);
167  }
168 
169  if (par->codec_id == AV_CODEC_ID_ALAC) {
170  ffio_wfourcc(pb, "kuki");
171  avio_wb64(pb, 12 + par->extradata_size);
172  avio_write(pb, "\0\0\0\14frmaalac", 12);
173  avio_write(pb, par->extradata, par->extradata_size);
174  } else if (par->codec_id == AV_CODEC_ID_AMR_NB) {
175  ffio_wfourcc(pb, "kuki");
176  avio_wb64(pb, 29);
177  avio_write(pb, "\0\0\0\14frmasamr", 12);
178  avio_wb32(pb, 0x11); /* size */
179  avio_write(pb, "samrFFMP", 8);
180  avio_w8(pb, 0); /* decoder version */
181 
182  avio_wb16(pb, 0x81FF); /* Mode set (all modes for AMR_NB) */
183  avio_w8(pb, 0x00); /* Mode change period (no restriction) */
184  avio_w8(pb, 0x01); /* Frames per sample */
185  } else if (par->codec_id == AV_CODEC_ID_QDM2 || par->codec_id == AV_CODEC_ID_QDMC) {
186  ffio_wfourcc(pb, "kuki");
187  avio_wb64(pb, par->extradata_size);
188  avio_write(pb, par->extradata, par->extradata_size);
189  }
190 
192  if (av_dict_count(s->metadata)) {
193  ffio_wfourcc(pb, "info"); //< Information chunk
194  while ((t = av_dict_iterate(s->metadata, t))) {
195  chunk_size += strlen(t->key) + strlen(t->value) + 2;
196  }
197  avio_wb64(pb, chunk_size + 4);
198  avio_wb32(pb, av_dict_count(s->metadata));
199  t = NULL;
200  while ((t = av_dict_iterate(s->metadata, t))) {
201  avio_put_str(pb, t->key);
202  avio_put_str(pb, t->value);
203  }
204  }
205 
206  ffio_wfourcc(pb, "data"); //< Audio Data chunk
207  caf->data = avio_tell(pb);
208  avio_wb64(pb, -1); //< mChunkSize
209  avio_wb32(pb, 0); //< mEditCount
210 
211  return 0;
212 }
213 
215 {
216  CAFContext *caf = s->priv_data;
217  AVStream *const st = s->streams[0];
218 
219  if (!st->codecpar->block_align) {
220  uint8_t *pkt_sizes;
221  int i, alloc_size = caf->size_entries_used + 5U;
222  if (alloc_size < 0)
223  return AVERROR(ERANGE);
224 
225  pkt_sizes = av_fast_realloc(st->priv_data,
226  &caf->size_buffer_size,
227  alloc_size);
228  if (!pkt_sizes)
229  return AVERROR(ENOMEM);
230  st->priv_data = pkt_sizes;
231  for (i = 4; i > 0; i--) {
232  unsigned top = pkt->size >> i * 7;
233  if (top)
234  pkt_sizes[caf->size_entries_used++] = 128 | top;
235  }
236  pkt_sizes[caf->size_entries_used++] = pkt->size & 127;
237  caf->packets++;
238  }
239  avio_write(s->pb, pkt->data, pkt->size);
240  return 0;
241 }
242 
244 {
245  CAFContext *caf = s->priv_data;
246  AVIOContext *pb = s->pb;
247  AVStream *st = s->streams[0];
248  AVCodecParameters *par = st->codecpar;
249 
250  if (pb->seekable & AVIO_SEEKABLE_NORMAL) {
251  int64_t file_size = avio_tell(pb);
252 
253  avio_seek(pb, caf->data, SEEK_SET);
254  avio_wb64(pb, file_size - caf->data - 8);
255  if (!par->block_align) {
256  int packet_size = samples_per_packet(par);
257  if (!packet_size) {
258  packet_size = st->duration / (caf->packets - 1);
259  avio_seek(pb, FRAME_SIZE_OFFSET, SEEK_SET);
260  avio_wb32(pb, packet_size);
261  }
262  avio_seek(pb, file_size, SEEK_SET);
263  ffio_wfourcc(pb, "pakt");
264  avio_wb64(pb, caf->size_entries_used + 24U);
265  avio_wb64(pb, caf->packets); ///< mNumberPackets
266  avio_wb64(pb, caf->packets * packet_size); ///< mNumberValidFrames
267  avio_wb32(pb, 0); ///< mPrimingFrames
268  avio_wb32(pb, 0); ///< mRemainderFrames
269  avio_write(pb, st->priv_data, caf->size_entries_used);
270  }
271  }
272  return 0;
273 }
274 
276  .p.name = "caf",
277  .p.long_name = NULL_IF_CONFIG_SMALL("Apple CAF (Core Audio Format)"),
278  .p.mime_type = "audio/x-caf",
279  .p.extensions = "caf",
280  .priv_data_size = sizeof(CAFContext),
281  .p.audio_codec = AV_CODEC_ID_PCM_S16BE,
282  .p.video_codec = AV_CODEC_ID_NONE,
283  .p.subtitle_codec = AV_CODEC_ID_NONE,
284  .flags_internal = FF_OFMT_FLAG_MAX_ONE_OF_EACH,
285  .write_header = caf_write_header,
286  .write_packet = caf_write_packet,
287  .write_trailer = caf_write_trailer,
288  .p.codec_tag = ff_caf_codec_tags_list,
289 };
CAFContext::data
int64_t data
Definition: cafenc.c:34
AV_CODEC_ID_PCM_S16LE
@ AV_CODEC_ID_PCM_S16LE
Definition: codec_id.h:334
AV_CODEC_ID_MACE6
@ AV_CODEC_ID_MACE6
Definition: codec_id.h:456
AV_CODEC_ID_PCM_F32BE
@ AV_CODEC_ID_PCM_F32BE
Definition: codec_id.h:354
AVCodecParameters::extradata
uint8_t * extradata
Extra binary data needed for initializing the decoder, codec-dependent.
Definition: codec_par.h:69
AV_CODEC_ID_ADPCM_MS
@ AV_CODEC_ID_ADPCM_MS
Definition: codec_id.h:379
AV_CODEC_ID_ADPCM_IMA_QT
@ AV_CODEC_ID_ADPCM_IMA_QT
Definition: codec_id.h:373
AV_CODEC_ID_AC3
@ AV_CODEC_ID_AC3
Definition: codec_id.h:449
AVOutputFormat::name
const char * name
Definition: avformat.h:510
AVERROR
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
ffio_wfourcc
static av_always_inline void ffio_wfourcc(AVIOContext *pb, const uint8_t *s)
Definition: avio_internal.h:124
caf.h
AVCodecParameters
This struct describes the properties of an encoded stream.
Definition: codec_par.h:47
caf_write_packet
static int caf_write_packet(AVFormatContext *s, AVPacket *pkt)
Definition: cafenc.c:214
AVStream::priv_data
void * priv_data
Definition: avformat.h:773
av_dict_count
int av_dict_count(const AVDictionary *m)
Get number of entries in dictionary.
Definition: dict.c:39
int64_t
long long int64_t
Definition: coverity.c:34
AVPacket::data
uint8_t * data
Definition: packet.h:539
AV_CODEC_ID_ALAC
@ AV_CODEC_ID_ALAC
Definition: codec_id.h:462
AV_CODEC_ID_AMR_NB
@ AV_CODEC_ID_AMR_NB
Definition: codec_id.h:427
AVChannelLayout::order
enum AVChannelOrder order
Channel order used in this layout.
Definition: channel_layout.h:322
CAFContext::size_entries_used
int size_entries_used
Definition: cafenc.c:36
AVChannelLayout::mask
uint64_t mask
This member must be used for AV_CHANNEL_ORDER_NATIVE, and may be used for AV_CHANNEL_ORDER_AMBISONIC ...
Definition: channel_layout.h:349
AVChannelLayout::nb_channels
int nb_channels
Number of channels in this layout.
Definition: channel_layout.h:327
intfloat.h
FFOutputFormat::p
AVOutputFormat p
The public AVOutputFormat.
Definition: mux.h:65
AV_CODEC_ID_PCM_S16BE
@ AV_CODEC_ID_PCM_S16BE
Definition: codec_id.h:335
avio_tell
static av_always_inline int64_t avio_tell(AVIOContext *s)
ftell() equivalent for AVIOContext.
Definition: avio.h:494
AV_CODEC_ID_MP3
@ AV_CODEC_ID_MP3
preferred ID for decoding MPEG audio layer 1, 2 or 3
Definition: codec_id.h:447
AVStream::duration
int64_t duration
Decoding: duration of the stream, in stream time base.
Definition: avformat.h:807
av_get_bits_per_sample
int av_get_bits_per_sample(enum AVCodecID codec_id)
Return codec bits per sample.
Definition: utils.c:550
AV_CODEC_ID_PCM_S8
@ AV_CODEC_ID_PCM_S8
Definition: codec_id.h:338
pkt
AVPacket * pkt
Definition: movenc.c:60
AV_LOG_ERROR
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
Definition: log.h:209
AV_CODEC_ID_MACE3
@ AV_CODEC_ID_MACE3
Definition: codec_id.h:455
AVCodecParameters::frame_size
int frame_size
Audio only.
Definition: codec_par.h:195
av_fast_realloc
void * av_fast_realloc(void *ptr, unsigned int *size, size_t min_size)
Reallocate the given buffer if it is not large enough, otherwise do nothing.
Definition: mem.c:497
s
#define s(width, name)
Definition: cbs_vp9.c:198
AVDictionaryEntry::key
char * key
Definition: dict.h:90
frame_size
int frame_size
Definition: mxfenc.c:2429
AVChannelLayout::u
union AVChannelLayout::@428 u
Details about which channels are present in this layout.
AV_CODEC_ID_MP2
@ AV_CODEC_ID_MP2
Definition: codec_id.h:446
channels
channels
Definition: aptx.h:31
AV_CODEC_ID_PCM_MULAW
@ AV_CODEC_ID_PCM_MULAW
Definition: codec_id.h:340
codec_id
enum AVCodecID codec_id
Definition: vaapi_decode.c:410
caf_write_header
static int caf_write_header(AVFormatContext *s)
Definition: cafenc.c:112
AVFormatContext
Format I/O context.
Definition: avformat.h:1300
AV_CODEC_ID_PCM_ALAW
@ AV_CODEC_ID_PCM_ALAW
Definition: codec_id.h:341
AVStream::codecpar
AVCodecParameters * codecpar
Codec parameters associated with this stream.
Definition: avformat.h:771
NULL
#define NULL
Definition: coverity.c:32
AVERROR_PATCHWELCOME
#define AVERROR_PATCHWELCOME
Not yet implemented in FFmpeg, patches welcome.
Definition: error.h:64
isom.h
FFOutputFormat
Definition: mux.h:61
avio_w8
void avio_w8(AVIOContext *s, int b)
Definition: aviobuf.c:179
AV_CODEC_ID_QDM2
@ AV_CODEC_ID_QDM2
Definition: codec_id.h:465
AVCodecParameters::ch_layout
AVChannelLayout ch_layout
Audio only.
Definition: codec_par.h:180
AVCodecParameters::sample_rate
int sample_rate
Audio only.
Definition: codec_par.h:184
AV_CODEC_ID_GSM
@ AV_CODEC_ID_GSM
as in Berlin toast format
Definition: codec_id.h:464
AVCodecID
AVCodecID
Identify the syntax and semantics of the bitstream.
Definition: codec_id.h:49
AVCodecParameters::extradata_size
int extradata_size
Size of the extradata content in bytes.
Definition: codec_par.h:73
AV_CODEC_ID_AAC
@ AV_CODEC_ID_AAC
Definition: codec_id.h:448
AV_CODEC_ID_QCELP
@ AV_CODEC_ID_QCELP
Definition: codec_id.h:470
AVIOContext
Bytestream IO Context.
Definition: avio.h:160
AV_CODEC_ID_PCM_S24LE
@ AV_CODEC_ID_PCM_S24LE
Definition: codec_id.h:346
AVPacket::size
int size
Definition: packet.h:540
NULL_IF_CONFIG_SMALL
#define NULL_IF_CONFIG_SMALL(x)
Return NULL if CONFIG_SMALL is true, otherwise the argument without modification.
Definition: internal.h:94
AVIOContext::seekable
int seekable
A combination of AVIO_SEEKABLE_ flags or 0 when the stream is not seekable.
Definition: avio.h:261
ff_standardize_creation_time
int ff_standardize_creation_time(AVFormatContext *s)
Standardize creation_time metadata in AVFormatContext to an ISO-8601 timestamp string.
Definition: mux_utils.c:155
FRAME_SIZE_OFFSET
#define FRAME_SIZE_OFFSET
Definition: cafenc.c:31
AV_CODEC_ID_QDMC
@ AV_CODEC_ID_QDMC
Definition: codec_id.h:496
AV_CODEC_ID_OPUS
@ AV_CODEC_ID_OPUS
Definition: codec_id.h:506
avio_write
void avio_write(AVIOContext *s, const unsigned char *buf, int size)
Definition: aviobuf.c:201
avio_wb32
void avio_wb32(AVIOContext *s, unsigned int val)
Definition: aviobuf.c:365
avio_wl32
void avio_wl32(AVIOContext *s, unsigned int val)
Definition: aviobuf.c:357
AV_CHANNEL_ORDER_NATIVE
@ AV_CHANNEL_ORDER_NATIVE
The native channel order, i.e.
Definition: channel_layout.h:125
av_double2int
static av_always_inline uint64_t av_double2int(double f)
Reinterpret a double as a 64-bit integer.
Definition: intfloat.h:70
AV_CODEC_ID_NONE
@ AV_CODEC_ID_NONE
Definition: codec_id.h:50
i
#define i(width, name, range_min, range_max)
Definition: cbs_h2645.c:256
avio_internal.h
AVCodecParameters::block_align
int block_align
Audio only.
Definition: codec_par.h:191
FF_OFMT_FLAG_MAX_ONE_OF_EACH
#define FF_OFMT_FLAG_MAX_ONE_OF_EACH
If this flag is set, it indicates that for each codec type whose corresponding default codec (i....
Definition: mux.h:50
AV_CODEC_ID_PCM_F64BE
@ AV_CODEC_ID_PCM_F64BE
Definition: codec_id.h:356
AV_CODEC_ID_PCM_S32BE
@ AV_CODEC_ID_PCM_S32BE
Definition: codec_id.h:343
ff_codec_caf_tags
const AVCodecTag ff_codec_caf_tags[]
Known codec tags for CAF.
Definition: caf.c:34
caf_write_trailer
static int caf_write_trailer(AVFormatContext *s)
Definition: cafenc.c:243
AV_CODEC_ID_GSM_MS
@ AV_CODEC_ID_GSM_MS
Definition: codec_id.h:476
AVStream
Stream structure.
Definition: avformat.h:748
avio_seek
int64_t avio_seek(AVIOContext *s, int64_t offset, int whence)
fseek() equivalent for AVIOContext.
Definition: aviobuf.c:231
avformat.h
dict.h
U
#define U(x)
Definition: vpx_arith.h:37
ff_caf_codec_tags_list
const AVCodecTag *const ff_caf_codec_tags_list[]
Definition: caf.c:82
AVIO_SEEKABLE_NORMAL
#define AVIO_SEEKABLE_NORMAL
Seeking works like for a local file.
Definition: avio.h:41
ff_codec_get_tag
unsigned int ff_codec_get_tag(const AVCodecTag *tags, enum AVCodecID id)
Definition: utils.c:123
CAFContext::packets
int packets
Definition: cafenc.c:37
avio_wb64
void avio_wb64(AVIOContext *s, uint64_t val)
Definition: aviobuf.c:431
AV_CODEC_ID_PCM_S32LE
@ AV_CODEC_ID_PCM_S32LE
Definition: codec_id.h:342
mem.h
AV_CODEC_ID_PCM_F64LE
@ AV_CODEC_ID_PCM_F64LE
Definition: codec_id.h:357
AVDictionaryEntry
Definition: dict.h:89
CAFContext
Definition: cafenc.c:33
AVCodecParameters::codec_id
enum AVCodecID codec_id
Specific type of the encoded data (the codec used).
Definition: codec_par.h:55
AVPacket
This structure stores compressed data.
Definition: packet.h:516
AV_CODEC_ID_ADPCM_IMA_WAV
@ AV_CODEC_ID_ADPCM_IMA_WAV
Definition: codec_id.h:374
AV_CODEC_ID_ILBC
@ AV_CODEC_ID_ILBC
Definition: codec_id.h:505
CAFContext::size_buffer_size
int size_buffer_size
Definition: cafenc.c:35
avio_wb16
void avio_wb16(AVIOContext *s, unsigned int val)
Definition: aviobuf.c:443
AV_CODEC_ID_PCM_F32LE
@ AV_CODEC_ID_PCM_F32LE
Definition: codec_id.h:355
av_log
#define av_log(a,...)
Definition: tableprint_vlc.h:27
codec_flags
static uint32_t codec_flags(enum AVCodecID codec_id)
Definition: cafenc.c:40
AVERROR_INVALIDDATA
#define AVERROR_INVALIDDATA
Invalid data found when processing input.
Definition: error.h:61
ff_caf_muxer
const FFOutputFormat ff_caf_muxer
Definition: cafenc.c:275
AVDictionaryEntry::value
char * value
Definition: dict.h:91
ff_mov_write_chan
void ff_mov_write_chan(AVIOContext *pb, int64_t channel_layout)
Definition: isom.c:416
avio_put_str
int avio_put_str(AVIOContext *s, const char *str)
Write a NULL-terminated string.
Definition: aviobuf.c:373
AV_CODEC_ID_MP1
@ AV_CODEC_ID_MP1
Definition: codec_id.h:488
samples_per_packet
static uint32_t samples_per_packet(const AVCodecParameters *par)
Definition: cafenc.c:57
AV_CODEC_ID_PCM_S24BE
@ AV_CODEC_ID_PCM_S24BE
Definition: codec_id.h:347
av_dict_iterate
const AVDictionaryEntry * av_dict_iterate(const AVDictionary *m, const AVDictionaryEntry *prev)
Iterate over a dictionary.
Definition: dict.c:44
mux.h