FFmpeg
aac.h
Go to the documentation of this file.
1 /*
2  * AAC definitions and structures
3  * Copyright (c) 2005-2006 Oded Shimon ( ods15 ods15 dyndns org )
4  * Copyright (c) 2006-2007 Maxim Gavrilov ( maxim.gavrilov 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 /**
24  * @file
25  * AAC definitions and structures
26  * @author Oded Shimon ( ods15 ods15 dyndns org )
27  * @author Maxim Gavrilov ( maxim.gavrilov gmail com )
28  */
29 
30 #ifndef AVCODEC_AAC_H
31 #define AVCODEC_AAC_H
32 
33 
34 #include "aac_defines.h"
35 #include "libavutil/float_dsp.h"
36 #include "libavutil/fixed_dsp.h"
37 #include "avcodec.h"
38 #if !USE_FIXED
39 #include "mdct15.h"
40 #endif
41 #include "fft.h"
42 #include "mpeg4audio.h"
43 #include "sbr.h"
44 
45 #include <stdint.h>
46 
47 #define MAX_CHANNELS 64
48 #define MAX_ELEM_ID 16
49 
50 #define TNS_MAX_ORDER 20
51 #define MAX_LTP_LONG_SFB 40
52 
53 #define CLIP_AVOIDANCE_FACTOR 0.95f
54 
64 };
65 
71  EXT_SBR_DATA = 0xd,
73 };
74 
80 };
81 
82 enum BandType {
83  ZERO_BT = 0, ///< Scalefactors and spectral data are all zero.
84  FIRST_PAIR_BT = 5, ///< This and later band types encode two values (rather than four) with one code word.
85  ESC_BT = 11, ///< Spectral data are coded with an escape sequence.
86  RESERVED_BT = 12, ///< Band types following are encoded differently from others.
87  NOISE_BT = 13, ///< Spectral data are scaled white noise not coded in the bitstream.
88  INTENSITY_BT2 = 14, ///< Scalefactor data are intensity stereo positions (out of phase).
89  INTENSITY_BT = 15, ///< Scalefactor data are intensity stereo positions (in phase).
90 };
91 
92 #define IS_CODEBOOK_UNSIGNED(x) (((x) - 1) & 10)
93 
101 };
102 
103 /**
104  * The point during decoding at which channel coupling is applied.
105  */
110 };
111 
112 /**
113  * Output configuration status
114  */
115 enum OCStatus {
116  OC_NONE, ///< Output unconfigured
117  OC_TRIAL_PCE, ///< Output configuration under trial specified by an inband PCE
118  OC_TRIAL_FRAME, ///< Output configuration under trial specified by a frame header
119  OC_GLOBAL_HDR, ///< Output configuration set in a global header but not yet locked
120  OC_LOCKED, ///< Output configuration locked in place
121 };
122 
123 typedef struct OutputConfiguration {
127  int channels;
128  uint64_t channel_layout;
131 
132 /**
133  * Predictor State
134  */
135 typedef struct PredictorState {
145 
146 #define MAX_PREDICTORS 672
147 
148 #define SCALE_DIV_512 36 ///< scalefactor difference that corresponds to scale difference in 512 times
149 #define SCALE_ONE_POS 140 ///< scalefactor index that corresponds to scale=1.0
150 #define SCALE_MAX_POS 255 ///< scalefactor index maximum value
151 #define SCALE_MAX_DIFF 60 ///< maximum scalefactor difference allowed by standard
152 #define SCALE_DIFF_ZERO 60 ///< codebook index corresponding to zero scalefactor indices difference
153 
154 #define POW_SF2_ZERO 200 ///< ff_aac_pow2sf_tab index corresponding to pow(2, 0);
155 
156 #define NOISE_PRE 256 ///< preamble for NOISE_BT, put in bitstream with the first noise band
157 #define NOISE_PRE_BITS 9 ///< length of preamble
158 #define NOISE_OFFSET 90 ///< subtracted from global gain, used as offset for the preamble
159 
160 /**
161  * Long Term Prediction
162  */
163 typedef struct LongTermPrediction {
164  int8_t present;
165  int16_t lag;
166  int coef_idx;
170 
171 /**
172  * Individual Channel Stream
173  */
174 typedef struct IndividualChannelStream {
175  uint8_t max_sfb; ///< number of scalefactor bands per group
177  uint8_t use_kb_window[2]; ///< If set, use Kaiser-Bessel window, otherwise use a sine window.
181  const uint16_t *swb_offset; ///< table of offsets to the lowest spectral coefficient of a scalefactor band, sfb, for a particular window
182  const uint8_t *swb_sizes; ///< table of scalefactor band sizes for a particular window
183  int num_swb; ///< number of scalefactor window bands
189  int predictor_reset_count[31]; ///< used by encoder to count prediction resets
191  uint8_t window_clipping[8]; ///< set if a certain window is near clipping
192  float clip_avoidance_factor; ///< set if any window is near clipping to the necessary atennuation factor to avoid it
194 
195 /**
196  * Temporal Noise Shaping
197  */
198 typedef struct TemporalNoiseShaping {
199  int present;
200  int n_filt[8];
201  int length[8][4];
202  int direction[8][4];
203  int order[8][4];
207 
208 /**
209  * Dynamic Range Control - decoded from the bitstream but not processed further.
210  */
211 typedef struct DynamicRangeControl {
212  int pce_instance_tag; ///< Indicates with which program the DRC info is associated.
213  int dyn_rng_sgn[17]; ///< DRC sign information; 0 - positive, 1 - negative
214  int dyn_rng_ctl[17]; ///< DRC magnitude information
215  int exclude_mask[MAX_CHANNELS]; ///< Channels to be excluded from DRC processing.
216  int band_incr; ///< Number of DRC bands greater than 1 having DRC info.
217  int interpolation_scheme; ///< Indicates the interpolation scheme used in the SBR QMF domain.
218  int band_top[17]; ///< Indicates the top of the i-th DRC band in units of 4 spectral lines.
219  int prog_ref_level; /**< A reference level for the long-term program audio level for all
220  * channels combined.
221  */
223 
224 typedef struct Pulse {
226  int start;
227  int pos[4];
228  int amp[4];
229 } Pulse;
230 
231 /**
232  * coupling parameters
233  */
234 typedef struct ChannelCoupling {
235  enum CouplingPoint coupling_point; ///< The point during decoding at which coupling is applied.
236  int num_coupled; ///< number of target elements
237  enum RawDataBlockType type[8]; ///< Type of channel element to be coupled - SCE or CPE.
238  int id_select[8]; ///< element id
239  int ch_select[8]; /**< [0] shared list of gains; [1] list of gains for right channel;
240  * [2] list of gains for left channel; [3] lists of gains for both channels
241  */
242  INTFLOAT gain[16][120];
244 
245 /**
246  * Single Channel Element - used for both SCE and LFE elements.
247  */
248 typedef struct SingleChannelElement {
252  enum BandType band_type[128]; ///< band types
253  enum BandType band_alt[128]; ///< alternative band type (used by encoder)
254  int band_type_run_end[120]; ///< band type run end points
255  INTFLOAT sf[120]; ///< scalefactors
256  int sf_idx[128]; ///< scalefactor indices (used by encoder)
257  uint8_t zeroes[128]; ///< band is not coded (used by encoder)
258  uint8_t can_pns[128]; ///< band is allowed to PNS (informative)
259  float is_ener[128]; ///< Intensity stereo pos (used by encoder)
260  float pns_ener[128]; ///< Noise energy values (used by encoder)
261  DECLARE_ALIGNED(32, INTFLOAT, pcoeffs)[1024]; ///< coefficients for IMDCT, pristine
262  DECLARE_ALIGNED(32, INTFLOAT, coeffs)[1024]; ///< coefficients for IMDCT, maybe processed
263  DECLARE_ALIGNED(32, INTFLOAT, saved)[1536]; ///< overlap
264  DECLARE_ALIGNED(32, INTFLOAT, ret_buf)[2048]; ///< PCM output buffer
265  DECLARE_ALIGNED(16, INTFLOAT, ltp_state)[3072]; ///< time signal for LTP
266  DECLARE_ALIGNED(32, AAC_FLOAT, lcoeffs)[1024]; ///< MDCT of LTP coefficients (used by encoder)
267  DECLARE_ALIGNED(32, AAC_FLOAT, prcoeffs)[1024]; ///< Main prediction coefs (used by encoder)
269  INTFLOAT *ret; ///< PCM output
271 
272 /**
273  * channel element - generic struct for SCE/CPE/CCE/LFE
274  */
275 typedef struct ChannelElement {
276  int present;
277  // CPE specific
278  int common_window; ///< Set if channels share a common 'IndividualChannelStream' in bitstream.
279  int ms_mode; ///< Signals mid/side stereo flags coding mode (used by encoder)
280  uint8_t is_mode; ///< Set if any bands have been encoded using intensity stereo (used by encoder)
281  uint8_t ms_mask[128]; ///< Set if mid/side stereo is used for each scalefactor window band
282  uint8_t is_mask[128]; ///< Set if intensity stereo is used (used by encoder)
283  // shared
285  // CCE specific
289 
290 /**
291  * main AAC context
292  */
293 struct AACContext {
294  AVClass *class;
297 
298  int is_saved; ///< Set if elements have stored overlap from previous frame.
300 
301  /**
302  * @name Channel element related data
303  * @{
304  */
309  /** @} */
310 
311  /**
312  * @name temporary aligned temporary buffers
313  * (We do not want to have these on the stack.)
314  * @{
315  */
317  /** @} */
318 
319  /**
320  * @name Computed / set up during initialization
321  * @{
322  */
327 #if USE_FIXED
329 #else
334 #endif /* USE_FIXED */
336  /** @} */
337 
338  /**
339  * @name Members used for output
340  * @{
341  */
342  SingleChannelElement *output_element[MAX_CHANNELS]; ///< Points to each SingleChannelElement
343  /** @} */
344 
345 
346  /**
347  * @name Japanese DTV specific extension
348  * @{
349  */
350  int force_dmono_mode;///< 0->not dmono, 1->use first channel, 2->use second channel
351  int dmono_mode; ///< 0->not dmono, 1->use first channel, 2->use second channel
352  /** @} */
353 
355 
359  unsigned warned_71_wide;
361 
362  /* aacdec functions pointers */
365  void (*apply_tns)(INTFLOAT coef[1024], TemporalNoiseShaping *tns,
366  IndividualChannelStream *ics, int decode);
370  void (*vector_pow43)(int *coefs, int len);
371  void (*subband_scale)(int *dst, int *src, int scale, int offset, int len, void *log_context);
372 
373 };
374 
376 
377 #endif /* AVCODEC_AAC_H */
ChannelCoupling::type
enum RawDataBlockType type[8]
Type of channel element to be coupled - SCE or CPE.
Definition: aac.h:237
MAX_ELEM_ID
#define MAX_ELEM_ID
Definition: aac.h:48
AAC_CHANNEL_BACK
@ AAC_CHANNEL_BACK
Definition: aac.h:98
CouplingPoint
CouplingPoint
The point during decoding at which channel coupling is applied.
Definition: aac.h:106
SingleChannelElement::band_alt
enum BandType band_alt[128]
alternative band type (used by encoder)
Definition: aac.h:253
SingleChannelElement::can_pns
uint8_t can_pns[128]
band is allowed to PNS (informative)
Definition: aac.h:258
TemporalNoiseShaping::order
int order[8][4]
Definition: aac.h:203
SingleChannelElement::pulse
Pulse pulse
Definition: aac.h:251
TYPE_FIL
@ TYPE_FIL
Definition: aac.h:62
out
FILE * out
Definition: movenc.c:54
EXT_FILL
@ EXT_FILL
Definition: aac.h:67
TemporalNoiseShaping::direction
int direction[8][4]
Definition: aac.h:202
Pulse::num_pulse
int num_pulse
Definition: aac.h:225
SingleChannelElement::zeroes
uint8_t zeroes[128]
band is not coded (used by encoder)
Definition: aac.h:257
AACContext::subband_scale
void(* subband_scale)(int *dst, int *src, int scale, int offset, int len, void *log_context)
Definition: aac.h:371
PredictorState::var1
AAC_FLOAT var1
Definition: aac.h:139
TemporalNoiseShaping::coef_idx
int coef_idx[8][4][TNS_MAX_ORDER]
Definition: aac.h:204
AVFrame
This structure describes decoded (raw) audio or video data.
Definition: frame.h:300
PredictorState::var0
AAC_FLOAT var0
Definition: aac.h:138
TYPE_PCE
@ TYPE_PCE
Definition: aac.h:61
AACContext::tag_che_map
ChannelElement * tag_che_map[4][MAX_ELEM_ID]
Definition: aac.h:306
MAX_PREDICTORS
#define MAX_PREDICTORS
Definition: aac.h:146
mdct15.h
TemporalNoiseShaping::present
int present
Definition: aac.h:199
AVFixedDSPContext
Definition: fixed_dsp.h:56
LongTermPrediction::used
int8_t used[MAX_LTP_LONG_SFB]
Definition: aac.h:168
AACContext::random_state
int random_state
Definition: aac.h:335
SingleChannelElement::pcoeffs
INTFLOAT pcoeffs[1024]
coefficients for IMDCT, pristine
Definition: aac.h:261
SingleChannelElement::ret
INTFLOAT * ret
PCM output.
Definition: aac.h:269
EXT_DYNAMIC_RANGE
@ EXT_DYNAMIC_RANGE
Definition: aac.h:70
ChannelElement::present
int present
Definition: aac.h:276
sbr.h
MPEG4AudioConfig
Definition: mpeg4audio.h:33
DynamicRangeControl
Dynamic Range Control - decoded from the bitstream but not processed further.
Definition: aac.h:211
IndividualChannelStream::num_swb
int num_swb
number of scalefactor window bands
Definition: aac.h:183
AACContext::temp
INTFLOAT temp[128]
Definition: aac.h:354
ChannelCoupling::coupling_point
enum CouplingPoint coupling_point
The point during decoding at which coupling is applied.
Definition: aac.h:235
OC_LOCKED
@ OC_LOCKED
Output configuration locked in place.
Definition: aac.h:120
mpeg4audio.h
AACContext::apply_ltp
void(* apply_ltp)(AACContext *ac, SingleChannelElement *sce)
Definition: aac.h:364
SingleChannelElement::saved
INTFLOAT saved[1536]
overlap
Definition: aac.h:263
LongTermPrediction::coef
INTFLOAT coef
Definition: aac.h:167
SingleChannelElement::ret_buf
INTFLOAT ret_buf[2048]
PCM output buffer.
Definition: aac.h:264
ChannelElement::coup
ChannelCoupling coup
Definition: aac.h:286
ChannelCoupling::id_select
int id_select[8]
element id
Definition: aac.h:238
BEFORE_TNS
@ BEFORE_TNS
Definition: aac.h:107
TYPE_CPE
@ TYPE_CPE
Definition: aac.h:57
ChannelElement::ms_mode
int ms_mode
Signals mid/side stereo flags coding mode (used by encoder)
Definition: aac.h:279
Pulse::amp
int amp[4]
Definition: aac.h:228
Pulse::pos
int pos[4]
Definition: aac.h:227
OutputConfiguration::status
enum OCStatus status
Definition: aac.h:129
TemporalNoiseShaping::length
int length[8][4]
Definition: aac.h:201
Pulse::start
int start
Definition: aac.h:226
PredictorState::r0
AAC_FLOAT r0
Definition: aac.h:140
IndividualChannelStream::prediction_used
uint8_t prediction_used[41]
Definition: aac.h:190
MAX_LTP_LONG_SFB
#define MAX_LTP_LONG_SFB
Definition: aac.h:51
SingleChannelElement::ics
IndividualChannelStream ics
Definition: aac.h:249
AACContext::mdct
FFTContext mdct
Definition: aac.h:323
IndividualChannelStream::predictor_reset_count
int predictor_reset_count[31]
used by encoder to count prediction resets
Definition: aac.h:189
AACContext::warned_960_sbr
int warned_960_sbr
Definition: aac.h:358
AFTER_IMDCT
@ AFTER_IMDCT
Definition: aac.h:109
AACContext::is_saved
int is_saved
Set if elements have stored overlap from previous frame.
Definition: aac.h:298
ZERO_BT
@ ZERO_BT
Scalefactors and spectral data are all zero.
Definition: aac.h:83
DynamicRangeControl::exclude_mask
int exclude_mask[MAX_CHANNELS]
Channels to be excluded from DRC processing.
Definition: aac.h:215
IndividualChannelStream::clip_avoidance_factor
float clip_avoidance_factor
set if any window is near clipping to the necessary atennuation factor to avoid it
Definition: aac.h:192
AACContext::vector_pow43
void(* vector_pow43)(int *coefs, int len)
Definition: aac.h:370
AACContext::mdct_ld
FFTContext mdct_ld
Definition: aac.h:325
decode
static void decode(AVCodecContext *dec_ctx, AVPacket *pkt, AVFrame *frame, FILE *outfile)
Definition: decode_audio.c:71
NOISE_BT
@ NOISE_BT
Spectral data are scaled white noise not coded in the bitstream.
Definition: aac.h:87
OC_TRIAL_FRAME
@ OC_TRIAL_FRAME
Output configuration under trial specified by a frame header.
Definition: aac.h:118
SingleChannelElement::coeffs
INTFLOAT coeffs[1024]
coefficients for IMDCT, maybe processed
Definition: aac.h:262
ChannelCoupling::num_coupled
int num_coupled
number of target elements
Definition: aac.h:236
IndividualChannelStream::swb_sizes
const uint8_t * swb_sizes
table of scalefactor band sizes for a particular window
Definition: aac.h:182
EIGHT_SHORT_SEQUENCE
@ EIGHT_SHORT_SEQUENCE
Definition: aac.h:78
OC_NONE
@ OC_NONE
Output unconfigured.
Definition: aac.h:116
INTENSITY_BT2
@ INTENSITY_BT2
Scalefactor data are intensity stereo positions (out of phase).
Definition: aac.h:88
TYPE_DSE
@ TYPE_DSE
Definition: aac.h:60
IndividualChannelStream::group_len
uint8_t group_len[8]
Definition: aac.h:179
PredictorState
Predictor State.
Definition: aac.h:135
ChannelPosition
ChannelPosition
Definition: aac.h:94
AACContext::fdsp
AVFloatDSPContext * fdsp
Definition: aac.h:333
LongTermPrediction::present
int8_t present
Definition: aac.h:164
AACContext::force_dmono_mode
int force_dmono_mode
0->not dmono, 1->use first channel, 2->use second channel
Definition: aac.h:350
fixed_dsp.h
IndividualChannelStream
Individual Channel Stream.
Definition: aac.h:174
MDCT15Context
Definition: mdct15.h:28
TemporalNoiseShaping::coef
INTFLOAT coef[8][4][TNS_MAX_ORDER]
Definition: aac.h:205
IndividualChannelStream::swb_offset
const uint16_t * swb_offset
table of offsets to the lowest spectral coefficient of a scalefactor band, sfb, for a particular wind...
Definition: aac.h:181
AACContext::warned_num_aac_frames
int warned_num_aac_frames
Definition: aac.h:357
INTENSITY_BT
@ INTENSITY_BT
Scalefactor data are intensity stereo positions (in phase).
Definition: aac.h:89
AVClass
Describe the class of an AVClass context structure.
Definition: log.h:67
ChannelElement::is_mask
uint8_t is_mask[128]
Set if intensity stereo is used (used by encoder)
Definition: aac.h:282
SingleChannelElement::is_ener
float is_ener[128]
Intensity stereo pos (used by encoder)
Definition: aac.h:259
IndividualChannelStream::num_window_groups
int num_window_groups
Definition: aac.h:178
AACContext::mdct480
MDCT15Context * mdct480
Definition: aac.h:331
AAC_CHANNEL_SIDE
@ AAC_CHANNEL_SIDE
Definition: aac.h:97
AACContext::frame
AVFrame * frame
Definition: aac.h:296
src
#define src
Definition: vp8dsp.c:254
LongTermPrediction::coef_idx
int coef_idx
Definition: aac.h:166
IndividualChannelStream::predictor_present
int predictor_present
Definition: aac.h:186
DynamicRangeControl::band_top
int band_top[17]
Indicates the top of the i-th DRC band in units of 4 spectral lines.
Definition: aac.h:218
AACContext::che
ChannelElement * che[4][MAX_ELEM_ID]
Definition: aac.h:305
AAC_CHANNEL_FRONT
@ AAC_CHANNEL_FRONT
Definition: aac.h:96
SingleChannelElement::predictor_state
PredictorState predictor_state[MAX_PREDICTORS]
Definition: aac.h:268
PredictorState::k1
AAC_FLOAT k1
Definition: aac.h:142
TNS_MAX_ORDER
#define TNS_MAX_ORDER
Definition: aac.h:50
AAC_CHANNEL_OFF
@ AAC_CHANNEL_OFF
Definition: aac.h:95
PredictorState::r1
AAC_FLOAT r1
Definition: aac.h:141
c
Undefined Behavior In the C some operations are like signed integer dereferencing freed accessing outside allocated Undefined Behavior must not occur in a C it is not safe even if the output of undefined operations is unused The unsafety may seem nit picking but Optimizing compilers have in fact optimized code on the assumption that no undefined Behavior occurs Optimizing code based on wrong assumptions can and has in some cases lead to effects beyond the output of computations The signed integer overflow problem in speed critical code Code which is highly optimized and works with signed integers sometimes has the problem that often the output of the computation does not c
Definition: undefined.txt:32
SingleChannelElement::band_type_run_end
int band_type_run_end[120]
band type run end points
Definition: aac.h:254
SingleChannelElement::sf_idx
int sf_idx[128]
scalefactor indices (used by encoder)
Definition: aac.h:256
float_dsp.h
OutputConfiguration::layout_map_tags
int layout_map_tags
Definition: aac.h:126
OutputConfiguration::layout_map
uint8_t layout_map[MAX_ELEM_ID *4][3]
Definition: aac.h:125
AACContext::apply_tns
void(* apply_tns)(INTFLOAT coef[1024], TemporalNoiseShaping *tns, IndividualChannelStream *ics, int decode)
Definition: aac.h:365
SingleChannelElement::lcoeffs
AAC_FLOAT lcoeffs[1024]
MDCT of LTP coefficients (used by encoder)
Definition: aac.h:266
AACContext::tags_mapped
int tags_mapped
Definition: aac.h:307
AACContext::avctx
AVCodecContext * avctx
Definition: aac.h:295
ChannelElement::ch
SingleChannelElement ch[2]
Definition: aac.h:284
ONLY_LONG_SEQUENCE
@ ONLY_LONG_SEQUENCE
Definition: aac.h:76
TYPE_END
@ TYPE_END
Definition: aac.h:63
AACContext::warned_remapping_once
int warned_remapping_once
Definition: aac.h:308
TemporalNoiseShaping::n_filt
int n_filt[8]
Definition: aac.h:200
BandType
BandType
Definition: aac.h:82
AACContext::mdct120
MDCT15Context * mdct120
Definition: aac.h:330
ESC_BT
@ ESC_BT
Spectral data are coded with an escape sequence.
Definition: aac.h:85
DynamicRangeControl::prog_ref_level
int prog_ref_level
A reference level for the long-term program audio level for all channels combined.
Definition: aac.h:219
SpectralBandReplication
Spectral Band Replication.
Definition: sbr.h:139
OCStatus
OCStatus
Output configuration status.
Definition: aac.h:115
AVFloatDSPContext
Definition: float_dsp.h:24
OutputConfiguration
Definition: aac.h:123
ChannelElement::common_window
int common_window
Set if channels share a common 'IndividualChannelStream' in bitstream.
Definition: aac.h:278
SingleChannelElement::prcoeffs
AAC_FLOAT prcoeffs[1024]
Main prediction coefs (used by encoder)
Definition: aac.h:267
offset
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this just let it vf offset
Definition: writing_filters.txt:86
MAX_CHANNELS
#define MAX_CHANNELS
Definition: aac.h:47
ChannelElement::ms_mask
uint8_t ms_mask[128]
Set if mid/side stereo is used for each scalefactor window band.
Definition: aac.h:281
DynamicRangeControl::dyn_rng_ctl
int dyn_rng_ctl[17]
DRC magnitude information.
Definition: aac.h:214
EXT_FILL_DATA
@ EXT_FILL_DATA
Definition: aac.h:68
SingleChannelElement::pns_ener
float pns_ener[128]
Noise energy values (used by encoder)
Definition: aac.h:260
OC_GLOBAL_HDR
@ OC_GLOBAL_HDR
Output configuration set in a global header but not yet locked.
Definition: aac.h:119
DECLARE_ALIGNED
#define DECLARE_ALIGNED(n, t, v)
Definition: mem.h:112
in
uint8_t pi<< 24) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_U8, uint8_t,(*(const uint8_t *) pi - 0x80) *(1.0f/(1<< 7))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_U8, uint8_t,(*(const uint8_t *) pi - 0x80) *(1.0/(1<< 7))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S16, int16_t,(*(const int16_t *) pi >> 8)+0x80) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S16, int16_t, *(const int16_t *) pi *(1.0f/(1<< 15))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S16, int16_t, *(const int16_t *) pi *(1.0/(1<< 15))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S32, int32_t,(*(const int32_t *) pi >> 24)+0x80) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S32, int32_t, *(const int32_t *) pi *(1.0f/(1U<< 31))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S32, int32_t, *(const int32_t *) pi *(1.0/(1U<< 31))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_FLT, float, av_clip_uint8(lrintf(*(const float *) pi *(1<< 7))+0x80)) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_FLT, float, av_clip_int16(lrintf(*(const float *) pi *(1<< 15)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_FLT, float, av_clipl_int32(llrintf(*(const float *) pi *(1U<< 31)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_DBL, double, av_clip_uint8(lrint(*(const double *) pi *(1<< 7))+0x80)) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_DBL, double, av_clip_int16(lrint(*(const double *) pi *(1<< 15)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_DBL, double, av_clipl_int32(llrint(*(const double *) pi *(1U<< 31)))) #define SET_CONV_FUNC_GROUP(ofmt, ifmt) static void set_generic_function(AudioConvert *ac) { } void ff_audio_convert_free(AudioConvert **ac) { if(! *ac) return;ff_dither_free(&(*ac) ->dc);av_freep(ac);} AudioConvert *ff_audio_convert_alloc(AVAudioResampleContext *avr, enum AVSampleFormat out_fmt, enum AVSampleFormat in_fmt, int channels, int sample_rate, int apply_map) { AudioConvert *ac;int in_planar, out_planar;ac=av_mallocz(sizeof(*ac));if(!ac) return NULL;ac->avr=avr;ac->out_fmt=out_fmt;ac->in_fmt=in_fmt;ac->channels=channels;ac->apply_map=apply_map;if(avr->dither_method !=AV_RESAMPLE_DITHER_NONE &&av_get_packed_sample_fmt(out_fmt)==AV_SAMPLE_FMT_S16 &&av_get_bytes_per_sample(in_fmt) > 2) { ac->dc=ff_dither_alloc(avr, out_fmt, in_fmt, channels, sample_rate, apply_map);if(!ac->dc) { av_free(ac);return NULL;} return ac;} in_planar=ff_sample_fmt_is_planar(in_fmt, channels);out_planar=ff_sample_fmt_is_planar(out_fmt, channels);if(in_planar==out_planar) { ac->func_type=CONV_FUNC_TYPE_FLAT;ac->planes=in_planar ? ac->channels :1;} else if(in_planar) ac->func_type=CONV_FUNC_TYPE_INTERLEAVE;else ac->func_type=CONV_FUNC_TYPE_DEINTERLEAVE;set_generic_function(ac);if(ARCH_AARCH64) ff_audio_convert_init_aarch64(ac);if(ARCH_ARM) ff_audio_convert_init_arm(ac);if(ARCH_X86) ff_audio_convert_init_x86(ac);return ac;} int ff_audio_convert(AudioConvert *ac, AudioData *out, AudioData *in) { int use_generic=1;int len=in->nb_samples;int p;if(ac->dc) { av_log(ac->avr, AV_LOG_TRACE, "%d samples - audio_convert: %s to %s (dithered)\n", len, av_get_sample_fmt_name(ac->in_fmt), av_get_sample_fmt_name(ac->out_fmt));return ff_convert_dither(ac-> in
Definition: audio_convert.c:326
AACContext::mdct960
MDCT15Context * mdct960
Definition: aac.h:332
FFTContext
Definition: fft.h:88
AACContext::mdct_ltp
FFTContext mdct_ltp
Definition: aac.h:326
BETWEEN_TNS_AND_IMDCT
@ BETWEEN_TNS_AND_IMDCT
Definition: aac.h:108
RawDataBlockType
RawDataBlockType
Definition: aac.h:55
SingleChannelElement
Single Channel Element - used for both SCE and LFE elements.
Definition: aac.h:248
IndividualChannelStream::num_windows
int num_windows
Definition: aac.h:184
ChannelElement::sbr
SpectralBandReplication sbr
Definition: aac.h:287
LONG_STOP_SEQUENCE
@ LONG_STOP_SEQUENCE
Definition: aac.h:79
OutputConfiguration::channel_layout
uint64_t channel_layout
Definition: aac.h:128
ChannelElement
channel element - generic struct for SCE/CPE/CCE/LFE
Definition: aac.h:275
PredictorState::x_est
AAC_FLOAT x_est
Definition: aac.h:143
TYPE_LFE
@ TYPE_LFE
Definition: aac.h:59
uint8_t
uint8_t
Definition: audio_convert.c:194
LongTermPrediction::lag
int16_t lag
Definition: aac.h:165
TYPE_SCE
@ TYPE_SCE
Definition: aac.h:56
AACContext::oc
OutputConfiguration oc[2]
Definition: aac.h:356
len
int len
Definition: vorbis_enc_data.h:452
IndividualChannelStream::tns_max_bands
int tns_max_bands
Definition: aac.h:185
OC_TRIAL_PCE
@ OC_TRIAL_PCE
Output configuration under trial specified by an inband PCE.
Definition: aac.h:117
AACContext::che_drc
DynamicRangeControl che_drc
Definition: aac.h:299
avcodec.h
AAC_CHANNEL_LFE
@ AAC_CHANNEL_LFE
Definition: aac.h:99
DynamicRangeControl::pce_instance_tag
int pce_instance_tag
Indicates with which program the DRC info is associated.
Definition: aac.h:212
SingleChannelElement::sf
INTFLOAT sf[120]
scalefactors
Definition: aac.h:255
AACContext::output_element
SingleChannelElement * output_element[MAX_CHANNELS]
Points to each SingleChannelElement.
Definition: aac.h:342
RESERVED_BT
@ RESERVED_BT
Band types following are encoded differently from others.
Definition: aac.h:86
LONG_START_SEQUENCE
@ LONG_START_SEQUENCE
Definition: aac.h:77
AACContext::update_ltp
void(* update_ltp)(AACContext *ac, SingleChannelElement *sce)
Definition: aac.h:369
ChannelCoupling::ch_select
int ch_select[8]
[0] shared list of gains; [1] list of gains for right channel; [2] list of gains for left channel; [3...
Definition: aac.h:239
SingleChannelElement::tns
TemporalNoiseShaping tns
Definition: aac.h:250
AACContext::warned_71_wide
unsigned warned_71_wide
Definition: aac.h:359
EXT_SBR_DATA_CRC
@ EXT_SBR_DATA_CRC
Definition: aac.h:72
fft.h
AVCodecContext
main external API structure.
Definition: avcodec.h:526
LongTermPrediction
Long Term Prediction.
Definition: aac.h:163
EXT_SBR_DATA
@ EXT_SBR_DATA
Definition: aac.h:71
ff_aacdec_init_mips
void ff_aacdec_init_mips(AACContext *c)
Definition: aacdec_mips.c:433
IndividualChannelStream::window_sequence
enum WindowSequence window_sequence[2]
Definition: aac.h:176
FIRST_PAIR_BT
@ FIRST_PAIR_BT
This and later band types encode two values (rather than four) with one code word.
Definition: aac.h:84
TemporalNoiseShaping
Temporal Noise Shaping.
Definition: aac.h:198
ChannelCoupling::gain
INTFLOAT gain[16][120]
Definition: aac.h:242
ChannelElement::is_mode
uint8_t is_mode
Set if any bands have been encoded using intensity stereo (used by encoder)
Definition: aac.h:280
DynamicRangeControl::band_incr
int band_incr
Number of DRC bands greater than 1 having DRC info.
Definition: aac.h:216
AACContext::buf_mdct
INTFLOAT buf_mdct[1024]
Definition: aac.h:316
OutputConfiguration::m4ac
MPEG4AudioConfig m4ac
Definition: aac.h:124
TYPE_CCE
@ TYPE_CCE
Definition: aac.h:58
PredictorState::cor1
AAC_FLOAT cor1
Definition: aac.h:137
AACContext::windowing_and_mdct_ltp
void(* windowing_and_mdct_ltp)(AACContext *ac, INTFLOAT *out, INTFLOAT *in, IndividualChannelStream *ics)
Definition: aac.h:367
AAC_FLOAT
float AAC_FLOAT
Definition: aac_defines.h:90
DynamicRangeControl::dyn_rng_sgn
int dyn_rng_sgn[17]
DRC sign information; 0 - positive, 1 - negative.
Definition: aac.h:213
ExtensionPayloadID
ExtensionPayloadID
Definition: aac.h:66
IndividualChannelStream::window_clipping
uint8_t window_clipping[8]
set if a certain window is near clipping
Definition: aac.h:191
AACContext::mdct_small
FFTContext mdct_small
Definition: aac.h:324
ChannelCoupling
coupling parameters
Definition: aac.h:234
EXT_DATA_ELEMENT
@ EXT_DATA_ELEMENT
Definition: aac.h:69
AACContext
main AAC context
Definition: aac.h:293
aac_defines.h
IndividualChannelStream::max_sfb
uint8_t max_sfb
number of scalefactor bands per group
Definition: aac.h:175
Pulse
Definition: aac.h:224
AAC_CHANNEL_CC
@ AAC_CHANNEL_CC
Definition: aac.h:100
SingleChannelElement::ltp_state
INTFLOAT ltp_state[3072]
time signal for LTP
Definition: aac.h:265
DynamicRangeControl::interpolation_scheme
int interpolation_scheme
Indicates the interpolation scheme used in the SBR QMF domain.
Definition: aac.h:217
AACContext::warned_gain_control
int warned_gain_control
Definition: aac.h:360
WindowSequence
WindowSequence
Definition: aac.h:75
IndividualChannelStream::ltp
LongTermPrediction ltp
Definition: aac.h:180
AACContext::imdct_and_windowing
void(* imdct_and_windowing)(AACContext *ac, SingleChannelElement *sce)
Definition: aac.h:363
SingleChannelElement::band_type
enum BandType band_type[128]
band types
Definition: aac.h:252
IndividualChannelStream::use_kb_window
uint8_t use_kb_window[2]
If set, use Kaiser-Bessel window, otherwise use a sine window.
Definition: aac.h:177
INTFLOAT
float INTFLOAT
Definition: aac_defines.h:86
OutputConfiguration::channels
int channels
Definition: aac.h:127
AACContext::dmono_mode
int dmono_mode
0->not dmono, 1->use first channel, 2->use second channel
Definition: aac.h:351
IndividualChannelStream::predictor_reset_group
int predictor_reset_group
Definition: aac.h:188
IndividualChannelStream::predictor_initialized
int predictor_initialized
Definition: aac.h:187
PredictorState::cor0
AAC_FLOAT cor0
Definition: aac.h:136