FFmpeg
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
libfdk-aacdec.c
Go to the documentation of this file.
1 /*
2  * AAC decoder wrapper
3  * Copyright (c) 2012 Martin Storsjo
4  *
5  * This file is part of FFmpeg.
6  *
7  * Permission to use, copy, modify, and/or distribute this software for any
8  * purpose with or without fee is hereby granted, provided that the above
9  * copyright notice and this permission notice appear in all copies.
10  *
11  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
12  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
13  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
14  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
15  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
16  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
17  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
18  */
19 
20 #include <fdk-aac/aacdecoder_lib.h>
21 
23 #include "libavutil/common.h"
24 #include "libavutil/opt.h"
25 #include "avcodec.h"
26 #include "internal.h"
27 
28 /* The version macro is introduced the same time as the setting enum was
29  * changed, so this check should suffice. */
30 #ifndef AACDECODER_LIB_VL0
31 #define AAC_PCM_MAX_OUTPUT_CHANNELS AAC_PCM_OUTPUT_CHANNELS
32 #endif
33 
39 };
40 
41 typedef struct FDKAACDecContext {
42  const AVClass *class;
43  HANDLE_AACDECODER handle;
48  int drc_level;
49  int drc_boost;
50  int drc_heavy;
51  int drc_cut;
54 
55 
56 #define DMX_ANC_BUFFSIZE 128
57 #define DECODER_MAX_CHANNELS 8
58 #define DECODER_BUFFSIZE 2048 * sizeof(INT_PCM)
59 
60 #define OFFSET(x) offsetof(FDKAACDecContext, x)
61 #define AD AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_DECODING_PARAM
62 static const AVOption fdk_aac_dec_options[] = {
63  { "conceal", "Error concealment method", OFFSET(conceal_method), AV_OPT_TYPE_INT, { .i64 = CONCEAL_METHOD_NOISE_SUBSTITUTION }, CONCEAL_METHOD_SPECTRAL_MUTING, CONCEAL_METHOD_NB - 1, AD, "conceal" },
64  { "spectral", "Spectral muting", 0, AV_OPT_TYPE_CONST, { .i64 = CONCEAL_METHOD_SPECTRAL_MUTING }, INT_MIN, INT_MAX, AD, "conceal" },
65  { "noise", "Noise Substitution", 0, AV_OPT_TYPE_CONST, { .i64 = CONCEAL_METHOD_NOISE_SUBSTITUTION }, INT_MIN, INT_MAX, AD, "conceal" },
66  { "energy", "Energy Interpolation", 0, AV_OPT_TYPE_CONST, { .i64 = CONCEAL_METHOD_ENERGY_INTERPOLATION }, INT_MIN, INT_MAX, AD, "conceal" },
67  { "drc_boost", "Dynamic Range Control: boost, where [0] is none and [127] is max boost",
68  OFFSET(drc_boost), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 127, AD, NULL },
69  { "drc_cut", "Dynamic Range Control: attenuation factor, where [0] is none and [127] is max compression",
70  OFFSET(drc_cut), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 127, AD, NULL },
71  { "drc_level", "Dynamic Range Control: reference level, quantized to 0.25dB steps where [0] is 0dB and [127] is -31.75dB",
72  OFFSET(drc_level), AV_OPT_TYPE_INT, { .i64 = -1}, -1, 127, AD, NULL },
73  { "drc_heavy", "Dynamic Range Control: heavy compression, where [1] is on (RF mode) and [0] is off",
74  OFFSET(drc_heavy), AV_OPT_TYPE_INT, { .i64 = -1}, -1, 1, AD, NULL },
75 #ifdef AACDECODER_LIB_VL0
76  { "level_limit", "Signal level limiting", OFFSET(level_limit), AV_OPT_TYPE_INT, { .i64 = 0 }, -1, 1, AD },
77 #endif
78  { NULL }
79 };
80 
81 static const AVClass fdk_aac_dec_class = {
82  .class_name = "libfdk-aac decoder",
83  .item_name = av_default_item_name,
84  .option = fdk_aac_dec_options,
85  .version = LIBAVUTIL_VERSION_INT,
86 };
87 
88 static int get_stream_info(AVCodecContext *avctx)
89 {
90  FDKAACDecContext *s = avctx->priv_data;
91  CStreamInfo *info = aacDecoder_GetStreamInfo(s->handle);
92  int channel_counts[0x24] = { 0 };
93  int i, ch_error = 0;
94  uint64_t ch_layout = 0;
95 
96  if (!info) {
97  av_log(avctx, AV_LOG_ERROR, "Unable to get stream info\n");
98  return AVERROR_UNKNOWN;
99  }
100 
101  if (info->sampleRate <= 0) {
102  av_log(avctx, AV_LOG_ERROR, "Stream info not initialized\n");
103  return AVERROR_UNKNOWN;
104  }
105  avctx->sample_rate = info->sampleRate;
106  avctx->frame_size = info->frameSize;
107 
108  for (i = 0; i < info->numChannels; i++) {
109  AUDIO_CHANNEL_TYPE ctype = info->pChannelType[i];
110  if (ctype <= ACT_NONE || ctype >= FF_ARRAY_ELEMS(channel_counts)) {
111  av_log(avctx, AV_LOG_WARNING, "unknown channel type\n");
112  break;
113  }
114  channel_counts[ctype]++;
115  }
116  av_log(avctx, AV_LOG_DEBUG,
117  "%d channels - front:%d side:%d back:%d lfe:%d top:%d\n",
118  info->numChannels,
119  channel_counts[ACT_FRONT], channel_counts[ACT_SIDE],
120  channel_counts[ACT_BACK], channel_counts[ACT_LFE],
121  channel_counts[ACT_FRONT_TOP] + channel_counts[ACT_SIDE_TOP] +
122  channel_counts[ACT_BACK_TOP] + channel_counts[ACT_TOP]);
123 
124  switch (channel_counts[ACT_FRONT]) {
125  case 4:
128  break;
129  case 3:
131  break;
132  case 2:
133  ch_layout |= AV_CH_LAYOUT_STEREO;
134  break;
135  case 1:
136  ch_layout |= AV_CH_FRONT_CENTER;
137  break;
138  default:
139  av_log(avctx, AV_LOG_WARNING,
140  "unsupported number of front channels: %d\n",
141  channel_counts[ACT_FRONT]);
142  ch_error = 1;
143  break;
144  }
145  if (channel_counts[ACT_SIDE] > 0) {
146  if (channel_counts[ACT_SIDE] == 2) {
147  ch_layout |= AV_CH_SIDE_LEFT | AV_CH_SIDE_RIGHT;
148  } else {
149  av_log(avctx, AV_LOG_WARNING,
150  "unsupported number of side channels: %d\n",
151  channel_counts[ACT_SIDE]);
152  ch_error = 1;
153  }
154  }
155  if (channel_counts[ACT_BACK] > 0) {
156  switch (channel_counts[ACT_BACK]) {
157  case 3:
159  break;
160  case 2:
161  ch_layout |= AV_CH_BACK_LEFT | AV_CH_BACK_RIGHT;
162  break;
163  case 1:
164  ch_layout |= AV_CH_BACK_CENTER;
165  break;
166  default:
167  av_log(avctx, AV_LOG_WARNING,
168  "unsupported number of back channels: %d\n",
169  channel_counts[ACT_BACK]);
170  ch_error = 1;
171  break;
172  }
173  }
174  if (channel_counts[ACT_LFE] > 0) {
175  if (channel_counts[ACT_LFE] == 1) {
176  ch_layout |= AV_CH_LOW_FREQUENCY;
177  } else {
178  av_log(avctx, AV_LOG_WARNING,
179  "unsupported number of LFE channels: %d\n",
180  channel_counts[ACT_LFE]);
181  ch_error = 1;
182  }
183  }
184  if (!ch_error &&
185  av_get_channel_layout_nb_channels(ch_layout) != info->numChannels) {
186  av_log(avctx, AV_LOG_WARNING, "unsupported channel configuration\n");
187  ch_error = 1;
188  }
189  if (ch_error)
190  avctx->channel_layout = 0;
191  else
192  avctx->channel_layout = ch_layout;
193 
194  avctx->channels = info->numChannels;
195 
196  return 0;
197 }
198 
200 {
201  FDKAACDecContext *s = avctx->priv_data;
202 
203  if (s->handle)
204  aacDecoder_Close(s->handle);
206  av_freep(&s->anc_buffer);
207 
208  return 0;
209 }
210 
212 {
213  FDKAACDecContext *s = avctx->priv_data;
214  AAC_DECODER_ERROR err;
215 
216  s->handle = aacDecoder_Open(avctx->extradata_size ? TT_MP4_RAW : TT_MP4_ADTS, 1);
217  if (!s->handle) {
218  av_log(avctx, AV_LOG_ERROR, "Error opening decoder\n");
219  return AVERROR_UNKNOWN;
220  }
221 
222  if (avctx->extradata_size) {
223  if ((err = aacDecoder_ConfigRaw(s->handle, &avctx->extradata,
224  &avctx->extradata_size)) != AAC_DEC_OK) {
225  av_log(avctx, AV_LOG_ERROR, "Unable to set extradata\n");
226  return AVERROR_INVALIDDATA;
227  }
228  }
229 
230  if ((err = aacDecoder_SetParam(s->handle, AAC_CONCEAL_METHOD,
231  s->conceal_method)) != AAC_DEC_OK) {
232  av_log(avctx, AV_LOG_ERROR, "Unable to set error concealment method\n");
233  return AVERROR_UNKNOWN;
234  }
235 
236  if (avctx->request_channel_layout > 0 &&
238  int downmix_channels = -1;
239 
240  switch (avctx->request_channel_layout) {
241  case AV_CH_LAYOUT_STEREO:
243  downmix_channels = 2;
244  break;
245  case AV_CH_LAYOUT_MONO:
246  downmix_channels = 1;
247  break;
248  default:
249  av_log(avctx, AV_LOG_WARNING, "Invalid request_channel_layout\n");
250  break;
251  }
252 
253  if (downmix_channels != -1) {
254  if (aacDecoder_SetParam(s->handle, AAC_PCM_MAX_OUTPUT_CHANNELS,
255  downmix_channels) != AAC_DEC_OK) {
256  av_log(avctx, AV_LOG_WARNING, "Unable to set output channels in the decoder\n");
257  } else {
259  if (!s->anc_buffer) {
260  av_log(avctx, AV_LOG_ERROR, "Unable to allocate ancillary buffer for the decoder\n");
261  return AVERROR(ENOMEM);
262  }
263  if (aacDecoder_AncDataInit(s->handle, s->anc_buffer, DMX_ANC_BUFFSIZE)) {
264  av_log(avctx, AV_LOG_ERROR, "Unable to register downmix ancillary buffer in the decoder\n");
265  return AVERROR_UNKNOWN;
266  }
267  }
268  }
269  }
270 
271  if (s->drc_boost != -1) {
272  if (aacDecoder_SetParam(s->handle, AAC_DRC_BOOST_FACTOR, s->drc_boost) != AAC_DEC_OK) {
273  av_log(avctx, AV_LOG_ERROR, "Unable to set DRC boost factor in the decoder\n");
274  return AVERROR_UNKNOWN;
275  }
276  }
277 
278  if (s->drc_cut != -1) {
279  if (aacDecoder_SetParam(s->handle, AAC_DRC_ATTENUATION_FACTOR, s->drc_cut) != AAC_DEC_OK) {
280  av_log(avctx, AV_LOG_ERROR, "Unable to set DRC attenuation factor in the decoder\n");
281  return AVERROR_UNKNOWN;
282  }
283  }
284 
285  if (s->drc_level != -1) {
286  if (aacDecoder_SetParam(s->handle, AAC_DRC_REFERENCE_LEVEL, s->drc_level) != AAC_DEC_OK) {
287  av_log(avctx, AV_LOG_ERROR, "Unable to set DRC reference level in the decoder\n");
288  return AVERROR_UNKNOWN;
289  }
290  }
291 
292  if (s->drc_heavy != -1) {
293  if (aacDecoder_SetParam(s->handle, AAC_DRC_HEAVY_COMPRESSION, s->drc_heavy) != AAC_DEC_OK) {
294  av_log(avctx, AV_LOG_ERROR, "Unable to set DRC heavy compression in the decoder\n");
295  return AVERROR_UNKNOWN;
296  }
297  }
298 
299 #ifdef AACDECODER_LIB_VL0
300  if (aacDecoder_SetParam(s->handle, AAC_PCM_LIMITER_ENABLE, s->level_limit) != AAC_DEC_OK) {
301  av_log(avctx, AV_LOG_ERROR, "Unable to set in signal level limiting in the decoder\n");
302  return AVERROR_UNKNOWN;
303  }
304 #endif
305 
306  avctx->sample_fmt = AV_SAMPLE_FMT_S16;
307 
310  if (!s->decoder_buffer)
311  return AVERROR(ENOMEM);
312 
313  return 0;
314 }
315 
316 static int fdk_aac_decode_frame(AVCodecContext *avctx, void *data,
317  int *got_frame_ptr, AVPacket *avpkt)
318 {
319  FDKAACDecContext *s = avctx->priv_data;
320  AVFrame *frame = data;
321  int ret;
322  AAC_DECODER_ERROR err;
323  UINT valid = avpkt->size;
324 
325  err = aacDecoder_Fill(s->handle, &avpkt->data, &avpkt->size, &valid);
326  if (err != AAC_DEC_OK) {
327  av_log(avctx, AV_LOG_ERROR, "aacDecoder_Fill() failed: %x\n", err);
328  return AVERROR_INVALIDDATA;
329  }
330 
331  err = aacDecoder_DecodeFrame(s->handle, (INT_PCM *) s->decoder_buffer, s->decoder_buffer_size / sizeof(INT_PCM), 0);
332  if (err == AAC_DEC_NOT_ENOUGH_BITS) {
333  ret = avpkt->size - valid;
334  goto end;
335  }
336  if (err != AAC_DEC_OK) {
337  av_log(avctx, AV_LOG_ERROR,
338  "aacDecoder_DecodeFrame() failed: %x\n", err);
339  ret = AVERROR_UNKNOWN;
340  goto end;
341  }
342 
343  if ((ret = get_stream_info(avctx)) < 0)
344  goto end;
345  frame->nb_samples = avctx->frame_size;
346 
347  if ((ret = ff_get_buffer(avctx, frame, 0)) < 0)
348  goto end;
349 
350  memcpy(frame->extended_data[0], s->decoder_buffer,
351  avctx->channels * avctx->frame_size *
353 
354  *got_frame_ptr = 1;
355  ret = avpkt->size - valid;
356 
357 end:
358  return ret;
359 }
360 
362 {
363  FDKAACDecContext *s = avctx->priv_data;
364  AAC_DECODER_ERROR err;
365 
366  if (!s->handle)
367  return;
368 
369  if ((err = aacDecoder_SetParam(s->handle,
370  AAC_TPDEC_CLEAR_BUFFER, 1)) != AAC_DEC_OK)
371  av_log(avctx, AV_LOG_WARNING, "failed to clear buffer when flushing\n");
372 }
373 
375  .name = "libfdk_aac",
376  .long_name = NULL_IF_CONFIG_SMALL("Fraunhofer FDK AAC"),
377  .type = AVMEDIA_TYPE_AUDIO,
378  .id = AV_CODEC_ID_AAC,
379  .priv_data_size = sizeof(FDKAACDecContext),
382  .close = fdk_aac_decode_close,
385  .priv_class = &fdk_aac_dec_class,
386  .caps_internal = FF_CODEC_CAP_INIT_THREADSAFE |
388 };
#define FF_CODEC_CAP_INIT_CLEANUP
The codec allows calling the close function for deallocation even if the init function returned a fai...
Definition: internal.h:48
#define NULL
Definition: coverity.c:32
const char * s
Definition: avisynth_c.h:768
#define AVERROR_INVALIDDATA
Invalid data found when processing input.
Definition: error.h:59
#define DECODER_BUFFSIZE
Definition: libfdk-aacdec.c:58
This structure describes decoded (raw) audio or video data.
Definition: frame.h:201
AVOption.
Definition: opt.h:246
ptrdiff_t const GLvoid * data
Definition: opengl_enc.c:101
static void flush(AVCodecContext *avctx)
#define AV_LOG_WARNING
Something somehow does not look correct.
Definition: log.h:182
#define LIBAVUTIL_VERSION_INT
Definition: version.h:85
static av_cold int init(AVCodecContext *avctx)
Definition: avrndec.c:35
int size
Definition: avcodec.h:1401
const char * av_default_item_name(void *ptr)
Return the context name.
Definition: log.c:193
#define AV_CODEC_CAP_CHANNEL_CONF
Codec should fill in channel configuration and samplerate instead of container.
Definition: avcodec.h:1008
#define AV_CH_LAYOUT_STEREO
#define DMX_ANC_BUFFSIZE
Definition: libfdk-aacdec.c:56
AVCodec.
Definition: avcodec.h:3351
static void decode(AVCodecContext *dec_ctx, AVPacket *pkt, AVFrame *frame, FILE *outfile)
Definition: decode_audio.c:42
int av_get_channel_layout_nb_channels(uint64_t channel_layout)
Return the number of channels in the channel layout.
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:72
#define FF_CODEC_CAP_INIT_THREADSAFE
The codec does not modify any global variables in the init function, allowing to call the init functi...
Definition: internal.h:40
enum AVSampleFormat sample_fmt
audio sample format
Definition: avcodec.h:2151
uint8_t
#define OFFSET(x)
Definition: libfdk-aacdec.c:60
#define av_cold
Definition: attributes.h:82
#define av_malloc(s)
#define AD
Definition: libfdk-aacdec.c:61
AVOptions.
static av_cold int end(AVCodecContext *avctx)
Definition: avrndec.c:90
uint8_t * extradata
some codecs need / can use extradata like Huffman tables.
Definition: avcodec.h:1588
#define AV_CH_LOW_FREQUENCY
static AVFrame * frame
uint8_t * data
Definition: avcodec.h:1400
static int get_stream_info(AVCodecContext *avctx)
Definition: libfdk-aacdec.c:88
#define AV_CH_BACK_LEFT
#define av_log(a,...)
HANDLE_AACDECODER handle
Definition: libfdk-aacdec.c:43
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
Definition: log.h:176
static av_cold void fdk_aac_decode_flush(AVCodecContext *avctx)
#define AVERROR(e)
Definition: error.h:43
#define NULL_IF_CONFIG_SMALL(x)
Return NULL if CONFIG_SMALL is true, otherwise the argument without modification. ...
Definition: internal.h:186
AVCodec ff_libfdk_aac_decoder
#define AV_LOG_DEBUG
Stuff which is only useful for libav* developers.
Definition: log.h:197
const char * name
Name of the codec implementation.
Definition: avcodec.h:3358
uint64_t channel_layout
Audio channel layout.
Definition: avcodec.h:2194
#define AAC_PCM_MAX_OUTPUT_CHANNELS
Definition: libfdk-aacdec.c:31
uint8_t * decoder_buffer
Definition: libfdk-aacdec.c:44
ConcealMethod
Definition: libfdk-aacdec.c:34
audio channel layout utility functions
#define AV_CH_LAYOUT_STEREO_DOWNMIX
static const AVOption fdk_aac_dec_options[]
Definition: libfdk-aacdec.c:62
#define AV_CH_FRONT_LEFT_OF_CENTER
#define AV_CH_FRONT_CENTER
static const uint8_t channel_counts[7]
Definition: dca_lbr.c:109
#define FF_ARRAY_ELEMS(a)
#define AV_CH_FRONT_RIGHT_OF_CENTER
int frame_size
Number of samples per channel in an audio frame.
Definition: avcodec.h:2163
Libavcodec external API header.
int sample_rate
samples per second
Definition: avcodec.h:2143
static int fdk_aac_decode_frame(AVCodecContext *avctx, void *data, int *got_frame_ptr, AVPacket *avpkt)
static const AVClass fdk_aac_dec_class
Definition: libfdk-aacdec.c:81
#define AV_CH_LAYOUT_NATIVE
Channel mask value used for AVCodecContext.request_channel_layout to indicate that the user requests ...
main external API structure.
Definition: avcodec.h:1488
int ff_get_buffer(AVCodecContext *avctx, AVFrame *frame, int flags)
Get a buffer for a frame.
Definition: decode.c:1886
int extradata_size
Definition: avcodec.h:1589
Describe the class of an AVClass context structure.
Definition: log.h:67
uint8_t * anc_buffer
Definition: libfdk-aacdec.c:46
#define AV_CH_BACK_CENTER
#define AV_CH_SIDE_RIGHT
int av_get_bytes_per_sample(enum AVSampleFormat sample_fmt)
Return number of bytes per sample.
Definition: samplefmt.c:106
common internal api header.
static av_cold int fdk_aac_decode_close(AVCodecContext *avctx)
common internal and external API header
signed 16 bits
Definition: samplefmt.h:61
static av_cold int fdk_aac_decode_init(AVCodecContext *avctx)
#define AVERROR_UNKNOWN
Unknown error, typically from an external library.
Definition: error.h:71
void * priv_data
Definition: avcodec.h:1515
int channels
number of audio channels
Definition: avcodec.h:2144
#define DECODER_MAX_CHANNELS
Definition: libfdk-aacdec.c:57
#define av_freep(p)
unsigned int UINT
#define AV_CH_SIDE_LEFT
uint8_t ** extended_data
pointers to the data planes/channels.
Definition: frame.h:248
#define AV_CH_LAYOUT_MONO
uint64_t request_channel_layout
Request decoder to use this channel layout if it can (0 for default)
Definition: avcodec.h:2201
This structure stores compressed data.
Definition: avcodec.h:1377
int nb_samples
number of audio samples (per channel) described by this frame
Definition: frame.h:267
#define AV_CODEC_CAP_DR1
Codec uses get_buffer() for allocating buffers and supports custom allocators.
Definition: avcodec.h:956
#define AV_CH_BACK_RIGHT