FFmpeg
latmenc.c
Go to the documentation of this file.
1 /*
2  * LATM/LOAS muxer
3  * Copyright (c) 2011 Kieran Kunhya <kieran@kunhya.com>
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 "libavcodec/get_bits.h"
23 #include "libavcodec/put_bits.h"
24 #include "libavcodec/codec_id.h"
25 #include "libavcodec/codec_par.h"
26 #include "libavcodec/packet.h"
27 #include "libavcodec/mpeg4audio.h"
28 #include "libavutil/opt.h"
29 #include "avformat.h"
30 #include "internal.h"
31 #include "mux.h"
32 #include "rawenc.h"
33 
34 #define MAX_EXTRADATA_SIZE 1024
35 
36 typedef struct LATMContext {
38  int off;
41  int counter;
42  int mod;
43  uint8_t buffer[0x1fff + MAX_EXTRADATA_SIZE + 1024];
44 } LATMContext;
45 
46 static const AVOption options[] = {
47  {"smc-interval", "StreamMuxConfig interval.",
48  offsetof(LATMContext, mod), AV_OPT_TYPE_INT, {.i64 = 0x0014}, 0x0001, 0xffff, AV_OPT_FLAG_ENCODING_PARAM},
49  {NULL},
50 };
51 
52 static const AVClass latm_muxer_class = {
53  .class_name = "LATM/LOAS muxer",
54  .item_name = av_default_item_name,
55  .option = options,
56  .version = LIBAVUTIL_VERSION_INT,
57 };
58 
59 static int latm_decode_extradata(AVFormatContext *s, uint8_t *buf, int size)
60 {
61  LATMContext *ctx = s->priv_data;
62  MPEG4AudioConfig m4ac;
63 
64  if (size > MAX_EXTRADATA_SIZE) {
65  av_log(s, AV_LOG_ERROR, "Extradata is larger than currently supported.\n");
66  return AVERROR_INVALIDDATA;
67  }
68  ctx->off = avpriv_mpeg4audio_get_config2(&m4ac, buf, size, 1, s);
69  if (ctx->off < 0)
70  return ctx->off;
71 
72  if (ctx->object_type == AOT_ALS && (ctx->off & 7)) {
73  // as long as avpriv_mpeg4audio_get_config works correctly this is impossible
74  av_log(s, AV_LOG_ERROR, "BUG: ALS offset is not byte-aligned\n");
75  return AVERROR_INVALIDDATA;
76  }
77  /* FIXME: are any formats not allowed in LATM? */
78 
79  if (m4ac.object_type > AOT_SBR && m4ac.object_type != AOT_ALS) {
80  av_log(s, AV_LOG_ERROR, "Muxing MPEG-4 AOT %d in LATM is not supported\n", m4ac.object_type);
81  return AVERROR_INVALIDDATA;
82  }
83  ctx->channel_conf = m4ac.chan_config;
84  ctx->object_type = m4ac.object_type;
85 
86  return 0;
87 }
88 
90 {
91  AVCodecParameters *par = s->streams[0]->codecpar;
92 
93  if (par->codec_id == AV_CODEC_ID_AAC_LATM)
94  return 0;
95  if (par->codec_id != AV_CODEC_ID_AAC && par->codec_id != AV_CODEC_ID_MP4ALS) {
96  av_log(s, AV_LOG_ERROR, "Only AAC, LATM and ALS are supported\n");
97  return AVERROR(EINVAL);
98  }
99 
100  if (par->extradata_size > 0 &&
102  return AVERROR_INVALIDDATA;
103 
104  return 0;
105 }
106 
107 static void copy_bits(PutBitContext *pb, const uint8_t *src, int length)
108 {
109  int words = length >> 4;
110  int bits = length & 15;
111  int i;
112  for (i = 0; i < words; i++)
113  put_bits(pb, 16, AV_RB16(src + 2 * i));
114  if (bits)
115  put_bits(pb, bits, AV_RB16(src + 2 * words) >> (16 - bits));
116 }
117 
119 {
120  LATMContext *ctx = s->priv_data;
121  AVCodecParameters *par = s->streams[0]->codecpar;
122  int header_size;
123 
124  /* AudioMuxElement */
125  put_bits(bs, 1, !!ctx->counter);
126 
127  if (!ctx->counter) {
128  /* StreamMuxConfig */
129  put_bits(bs, 1, 0); /* audioMuxVersion */
130  put_bits(bs, 1, 1); /* allStreamsSameTimeFraming */
131  put_bits(bs, 6, 0); /* numSubFrames */
132  put_bits(bs, 4, 0); /* numProgram */
133  put_bits(bs, 3, 0); /* numLayer */
134 
135  /* AudioSpecificConfig */
136  if (ctx->object_type == AOT_ALS) {
137  header_size = (par->extradata_size - (ctx->off >> 3)) * 8;
138  copy_bits(bs, &par->extradata[ctx->off >> 3], header_size);
139  } else {
140  // + 3 assumes not scalable and dependsOnCoreCoder == 0,
141  // see decode_ga_specific_config in libavcodec/aacdec.c
142  copy_bits(bs, par->extradata, ctx->off + 3);
143 
144  if (!ctx->channel_conf) {
145  GetBitContext gb;
146  int ret = init_get_bits8(&gb, par->extradata, par->extradata_size);
147  av_assert0(ret >= 0); // extradata size has been checked already, so this should not fail
148  skip_bits_long(&gb, ctx->off + 3);
149  ff_copy_pce_data(bs, &gb);
150  }
151  }
152 
153  put_bits(bs, 3, 0); /* frameLengthType */
154  put_bits(bs, 8, 0xff); /* latmBufferFullness */
155 
156  put_bits(bs, 1, 0); /* otherDataPresent */
157  put_bits(bs, 1, 0); /* crcCheckPresent */
158  }
159 
160  ctx->counter++;
161  ctx->counter %= ctx->mod;
162 }
163 
165 {
166  LATMContext *ctx = s->priv_data;
167  AVCodecParameters *par = s->streams[0]->codecpar;
168  AVIOContext *pb = s->pb;
169  PutBitContext bs;
170  int i, len;
171  uint8_t loas_header[] = "\x56\xe0\x00";
172 
173  if (par->codec_id == AV_CODEC_ID_AAC_LATM)
174  return ff_raw_write_packet(s, pkt);
175 
176  if (!par->extradata) {
177  if(pkt->size > 2 && pkt->data[0] == 0x56 && (pkt->data[1] >> 4) == 0xe &&
178  (AV_RB16(pkt->data + 1) & 0x1FFF) + 3 == pkt->size)
179  return ff_raw_write_packet(s, pkt);
180  else {
181  uint8_t *side_data;
182  size_t side_data_size;
183  int ret;
184 
186  &side_data_size);
187  if (side_data_size) {
188  if (latm_decode_extradata(s, side_data, side_data_size) < 0)
189  return AVERROR_INVALIDDATA;
190  ret = ff_alloc_extradata(par, side_data_size);
191  if (ret < 0)
192  return ret;
193  memcpy(par->extradata, side_data, side_data_size);
194  } else
195  return AVERROR_INVALIDDATA;
196  }
197  }
198 
199  if (pkt->size > 0x1fff)
200  goto too_large;
201 
202  init_put_bits(&bs, ctx->buffer, pkt->size+1024+MAX_EXTRADATA_SIZE);
203 
205 
206  /* PayloadLengthInfo() */
207  for (i = 0; i <= pkt->size-255; i+=255)
208  put_bits(&bs, 8, 255);
209 
210  put_bits(&bs, 8, pkt->size-i);
211 
212  /* The LATM payload is written unaligned */
213 
214  /* PayloadMux() */
215  if (pkt->size && (pkt->data[0] & 0xe1) == 0x81) {
216  // Convert byte-aligned DSE to non-aligned.
217  // Due to the input format encoding we know that
218  // it is naturally byte-aligned in the input stream,
219  // so there are no padding bits to account for.
220  // To avoid having to add padding bits and rearrange
221  // the whole stream we just remove the byte-align flag.
222  // This allows us to remux our FATE AAC samples into latm
223  // files that are still playable with minimal effort.
224  put_bits(&bs, 8, pkt->data[0] & 0xfe);
225  copy_bits(&bs, pkt->data + 1, 8*pkt->size - 8);
226  } else
227  copy_bits(&bs, pkt->data, 8*pkt->size);
228 
229  flush_put_bits(&bs);
230 
231  len = put_bytes_output(&bs);
232 
233  if (len > 0x1fff)
234  goto too_large;
235 
236  loas_header[1] |= (len >> 8) & 0x1f;
237  loas_header[2] |= len & 0xff;
238 
239  avio_write(pb, loas_header, 3);
240  avio_write(pb, ctx->buffer, len);
241 
242  return 0;
243 
244 too_large:
245  av_log(s, AV_LOG_ERROR, "LATM packet size larger than maximum size 0x1fff\n");
246  return AVERROR_INVALIDDATA;
247 }
248 
250  const AVPacket *pkt)
251 {
252  int ret = 1;
253 
254  if (st->codecpar->codec_id == AV_CODEC_ID_AAC) {
255  if (pkt->size > 2 && (AV_RB16(pkt->data) & 0xfff0) == 0xfff0)
256  ret = ff_stream_add_bitstream_filter(st, "aac_adtstoasc", NULL);
257  }
258 
259  return ret;
260 }
261 
263  .name = "latm",
264  .long_name = NULL_IF_CONFIG_SMALL("LOAS/LATM"),
265  .mime_type = "audio/MP4A-LATM",
266  .extensions = "latm,loas",
267  .priv_data_size = sizeof(LATMContext),
268  .audio_codec = AV_CODEC_ID_AAC,
269  .video_codec = AV_CODEC_ID_NONE,
272  .priv_class = &latm_muxer_class,
275 };
skip_bits_long
static void skip_bits_long(GetBitContext *s, int n)
Skips the specified number of bits.
Definition: get_bits.h:292
AVCodecParameters::extradata
uint8_t * extradata
Extra binary data needed for initializing the decoder, codec-dependent.
Definition: codec_par.h:75
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
opt.h
put_bytes_output
static int put_bytes_output(const PutBitContext *s)
Definition: put_bits.h:89
AVCodecParameters
This struct describes the properties of an encoded stream.
Definition: codec_par.h:53
latm_muxer_class
static const AVClass latm_muxer_class
Definition: latmenc.c:52
AVFMT_NOTIMESTAMPS
#define AVFMT_NOTIMESTAMPS
Format does not need / have any timestamps.
Definition: avformat.h:481
init_put_bits
static void init_put_bits(PutBitContext *s, uint8_t *buffer, int buffer_size)
Initialize the PutBitContext s.
Definition: put_bits.h:62
check_bitstream
static int check_bitstream(AVFormatContext *s, FFStream *sti, AVPacket *pkt)
Definition: mux.c:1069
put_bits
static void put_bits(Jpeg2000EncoderContext *s, int val, int n)
put n times val bit
Definition: j2kenc.c:221
AVPacket::data
uint8_t * data
Definition: packet.h:374
AVOption
AVOption.
Definition: opt.h:251
MPEG4AudioConfig
Definition: mpeg4audio.h:32
LATMContext::av_class
AVClass * av_class
Definition: latmenc.c:37
mpeg4audio.h
latm_check_bitstream
static int latm_check_bitstream(AVFormatContext *s, AVStream *st, const AVPacket *pkt)
Definition: latmenc.c:249
GetBitContext
Definition: get_bits.h:62
pkt
AVPacket * pkt
Definition: movenc.c:59
AV_LOG_ERROR
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
Definition: log.h:180
init_get_bits8
static int init_get_bits8(GetBitContext *s, const uint8_t *buffer, int byte_size)
Initialize GetBitContext.
Definition: get_bits.h:678
codec_id.h
avpriv_mpeg4audio_get_config2
int avpriv_mpeg4audio_get_config2(MPEG4AudioConfig *c, const uint8_t *buf, int size, int sync_extension, void *logctx)
Parse MPEG-4 systems extradata from a raw buffer to retrieve audio configuration.
Definition: mpeg4audio.c:165
s
#define s(width, name)
Definition: cbs_vp9.c:256
LATMContext::off
int off
Definition: latmenc.c:38
AV_OPT_FLAG_ENCODING_PARAM
#define AV_OPT_FLAG_ENCODING_PARAM
a generic parameter which can be set by the user for muxing or encoding
Definition: opt.h:281
bits
uint8_t bits
Definition: vp3data.h:141
av_assert0
#define av_assert0(cond)
assert() equivalent, that is always enabled.
Definition: avassert.h:37
ctx
AVFormatContext * ctx
Definition: movenc.c:48
ff_raw_write_packet
int ff_raw_write_packet(AVFormatContext *s, AVPacket *pkt)
Definition: rawenc.c:31
get_bits.h
LATMContext::object_type
int object_type
Definition: latmenc.c:40
PutBitContext
Definition: put_bits.h:50
AVFormatContext
Format I/O context.
Definition: avformat.h:1213
internal.h
AVStream::codecpar
AVCodecParameters * codecpar
Codec parameters associated with this stream.
Definition: avformat.h:1108
LIBAVUTIL_VERSION_INT
#define LIBAVUTIL_VERSION_INT
Definition: version.h:85
AVClass
Describe the class of an AVClass context structure.
Definition: log.h:66
NULL
#define NULL
Definition: coverity.c:32
av_default_item_name
const char * av_default_item_name(void *ptr)
Return the context name.
Definition: log.c:237
ff_copy_pce_data
static int ff_copy_pce_data(PutBitContext *pb, GetBitContext *gb)
Definition: mpeg4audio.h:132
AV_CODEC_ID_MP4ALS
@ AV_CODEC_ID_MP4ALS
Definition: codec_id.h:471
AVCodecParameters::extradata_size
int extradata_size
Size of the extradata content in bytes.
Definition: codec_par.h:79
AV_CODEC_ID_AAC
@ AV_CODEC_ID_AAC
Definition: codec_id.h:428
latm_decode_extradata
static int latm_decode_extradata(AVFormatContext *s, uint8_t *buf, int size)
Definition: latmenc.c:59
AVIOContext
Bytestream IO Context.
Definition: avio.h:162
AVPacket::size
int size
Definition: packet.h:375
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:117
latm_write_packet
static int latm_write_packet(AVFormatContext *s, AVPacket *pkt)
Definition: latmenc.c:164
size
int size
Definition: twinvq_data.h:10344
LATMContext::buffer
uint8_t buffer[0x1fff+MAX_EXTRADATA_SIZE+1024]
Definition: latmenc.c:43
avio_write
void avio_write(AVIOContext *s, const unsigned char *buf, int size)
Definition: aviobuf.c:232
LATMContext::counter
int counter
Definition: latmenc.c:41
latm_write_header
static int latm_write_header(AVFormatContext *s)
Definition: latmenc.c:89
ff_latm_muxer
const AVOutputFormat ff_latm_muxer
Definition: latmenc.c:262
rawenc.h
LATMContext
Definition: aacdec.c:266
LATMContext::mod
int mod
Definition: latmenc.c:42
AV_CODEC_ID_NONE
@ AV_CODEC_ID_NONE
Definition: codec_id.h:48
AVOutputFormat
Definition: avformat.h:509
i
#define i(width, name, range_min, range_max)
Definition: cbs_h2645.c:271
packet.h
av_packet_get_side_data
uint8_t * av_packet_get_side_data(const AVPacket *pkt, enum AVPacketSideDataType type, size_t *size)
Get side information from packet.
Definition: avpacket.c:251
MAX_EXTRADATA_SIZE
#define MAX_EXTRADATA_SIZE
Definition: latmenc.c:34
LATMContext::channel_conf
int channel_conf
Definition: latmenc.c:39
MPEG4AudioConfig::chan_config
int chan_config
Definition: mpeg4audio.h:36
len
int len
Definition: vorbis_enc_data.h:426
mod
static int mod(int a, int b)
Modulo operation with only positive remainders.
Definition: vf_v360.c:750
ret
ret
Definition: filter_design.txt:187
AVStream
Stream structure.
Definition: avformat.h:948
copy_bits
static void copy_bits(PutBitContext *pb, const uint8_t *src, int length)
Definition: latmenc.c:107
AVClass::class_name
const char * class_name
The name of the class; usually it is the same name as the context structure type to which the AVClass...
Definition: log.h:71
avformat.h
MPEG4AudioConfig::object_type
int object_type
Definition: mpeg4audio.h:33
ff_stream_add_bitstream_filter
int ff_stream_add_bitstream_filter(AVStream *st, const char *name, const char *args)
Add a bitstream filter to a stream.
Definition: mux.c:1307
AV_PKT_DATA_NEW_EXTRADATA
@ AV_PKT_DATA_NEW_EXTRADATA
The AV_PKT_DATA_NEW_EXTRADATA is used to notify the codec or the format that the extradata buffer was...
Definition: packet.h:56
AV_OPT_TYPE_INT
@ AV_OPT_TYPE_INT
Definition: opt.h:225
AOT_SBR
@ AOT_SBR
Y Spectral Band Replication.
Definition: mpeg4audio.h:80
latm_write_frame_header
static void latm_write_frame_header(AVFormatContext *s, PutBitContext *bs)
Definition: latmenc.c:118
flush_put_bits
static void flush_put_bits(PutBitContext *s)
Pad the end of the output stream with zeros.
Definition: put_bits.h:143
AOT_ALS
@ AOT_ALS
Y Audio LosslesS.
Definition: mpeg4audio.h:108
codec_par.h
AVCodecParameters::codec_id
enum AVCodecID codec_id
Specific type of the encoded data (the codec used).
Definition: codec_par.h:61
AVPacket
This structure stores compressed data.
Definition: packet.h:351
src
INIT_CLIP pixel * src
Definition: h264pred_template.c:418
flags
#define flags(name, subs,...)
Definition: cbs_av1.c:561
av_log
#define av_log(a,...)
Definition: tableprint_vlc.h:27
AV_CODEC_ID_AAC_LATM
@ AV_CODEC_ID_AAC_LATM
Definition: codec_id.h:475
AVERROR_INVALIDDATA
#define AVERROR_INVALIDDATA
Invalid data found when processing input.
Definition: error.h:61
write_packet
static int write_packet(AVFormatContext *s1, AVPacket *pkt)
Definition: v4l2enc.c:92
write_header
static void write_header(FFV1Context *f)
Definition: ffv1enc.c:346
put_bits.h
options
static const AVOption options[]
Definition: latmenc.c:46
AV_RB16
uint64_t_TMPL AV_WL64 unsigned int_TMPL AV_WL32 unsigned int_TMPL AV_WL24 unsigned int_TMPL AV_WL16 uint64_t_TMPL AV_WB64 unsigned int_TMPL AV_WB32 unsigned int_TMPL AV_WB24 unsigned int_TMPL AV_RB16
Definition: bytestream.h:98
ff_alloc_extradata
int ff_alloc_extradata(AVCodecParameters *par, int size)
Allocate extradata with additional AV_INPUT_BUFFER_PADDING_SIZE at end which is always set to 0.
Definition: utils.c:238
mux.h