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
38 #if !USE_FIXED
40 #endif
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
73 };
74
80 };
81
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 */
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
121 };
122
131
132 /**
133 * Predictor State
134 */
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 */
170
171 /**
172 * Individual Channel Stream
173 */
181 const uint16_t *
swb_offset;
///< table of offsets to the lowest spectral coefficient of a scalefactor band, sfb, for a particular window
183 int num_swb;
///< number of scalefactor window bands
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 */
207
208 /**
209 * Dynamic Range Control - decoded from the bitstream but not processed further.
210 */
213 int dyn_rng_sgn[17];
///< DRC sign information; 0 - positive, 1 - negative
216 int band_incr;
///< Number of DRC bands greater than 1 having DRC info.
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
230
231 /**
232 * coupling parameters
233 */
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 */
244
245 /**
246 * Single Channel Element - used for both SCE and LFE elements.
247 */
256 int sf_idx[128];
///< scalefactor indices (used by encoder)
259 float is_ener[128];
///< Intensity stereo pos (used by encoder)
260 float pns_ener[128];
///< Noise energy values (used by encoder)
271
272 /**
273 * channel element - generic struct for SCE/CPE/CCE/LFE
274 */
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
283 // shared
285 // CCE specific
289
290 /**
291 * main AAC context
292 */
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 */
343 /** @} */
344
345
346 /**
347 * @name Japanese DTV specific extension
348 * @{
349 */
351 int dmono_mode;
///< 0->not dmono, 1->use first channel, 2->use second channel
352 /** @} */
353
355
359
361
362 /* aacdec functions pointers */
372
373 };
374
376
377 #endif /* AVCODEC_AAC_H */
int predictor_initialized
Band types following are encoded differently from others.
float pns_ener[128]
Noise energy values (used by encoder)
uint8_t use_kb_window[2]
If set, use Kaiser-Bessel window, otherwise use a sine window.
void(* subband_scale)(int *dst, int *src, int scale, int offset, int len)
This structure describes decoded (raw) audio or video data.
int coef_idx[8][4][TNS_MAX_ORDER]
uint8_t window_clipping[8]
set if a certain window is near clipping
ChannelElement * che[4][MAX_ELEM_ID]
INTFLOAT * ret
PCM output.
int common_window
Set if channels share a common 'IndividualChannelStream' in bitstream.
void(* update_ltp)(AACContext *ac, SingleChannelElement *sce)
void(* imdct_and_windowing)(AACContext *ac, SingleChannelElement *sce)
INTFLOAT sf[120]
scalefactors
uint8_t ms_mask[128]
Set if mid/side stereo is used for each scalefactor window band.
Dynamic Range Control - decoded from the bitstream but not processed further.
static void decode(AVCodecContext *dec_ctx, AVPacket *pkt, AVFrame *frame, FILE *outfile)
Spectral data are scaled white noise not coded in the bitstream.
AAC_FLOAT prcoeffs[1024]
Main prediction coefs (used by encoder)
int band_incr
Number of DRC bands greater than 1 having DRC info.
INTFLOAT pcoeffs[1024]
coefficients for IMDCT, pristine
int dmono_mode
0->not dmono, 1->use first channel, 2->use second channel
const uint16_t * swb_offset
table of offsets to the lowest spectral coefficient of a scalefactor band, sfb, for a particular wind...
uint8_t layout_map[MAX_ELEM_ID *4][3]
Output configuration under trial specified by an inband PCE.
SingleChannelElement ch[2]
CouplingPoint
The point during decoding at which channel coupling is applied.
int num_coupled
number of target elements
int exclude_mask[MAX_CHANNELS]
Channels to be excluded from DRC processing.
void(* vector_pow43)(int *coefs, int len)
SingleChannelElement * output_element[MAX_CHANNELS]
Points to each SingleChannelElement.
#define DECLARE_ALIGNED(n, t, v)
Declare a variable that is aligned in memory.
Scalefactor data are intensity stereo positions (in phase).
int id_select[8]
element id
ChannelElement * tag_che_map[4][MAX_ELEM_ID]
Output configuration set in a global header but not yet locked.
int dyn_rng_sgn[17]
DRC sign information; 0 - positive, 1 - negative.
PredictorState predictor_state[MAX_PREDICTORS]
SpectralBandReplication sbr
enum CouplingPoint coupling_point
The point during decoding at which coupling is applied.
float is_ener[128]
Intensity stereo pos (used by encoder)
Spectral Band Replication definitions and structures.
uint8_t max_sfb
number of scalefactor bands per group
static const uint8_t offset[127][2]
int num_swb
number of scalefactor window bands
int prog_ref_level
A reference level for the long-term program audio level for all channels combined.
Output configuration locked in place.
int warned_remapping_once
INTFLOAT ret_buf[2048]
PCM output buffer.
enum BandType band_alt[128]
alternative band type (used by encoder)
enum WindowSequence window_sequence[2]
INTFLOAT ltp_state[3072]
time signal for LTP
int predictor_reset_group
int dyn_rng_ctl[17]
DRC magnitude information.
uint8_t can_pns[128]
band is allowed to PNS (informative)
typedef void(APIENTRY *FF_PFNGLACTIVETEXTUREPROC)(GLenum texture)
void ff_aacdec_init_mips(AACContext *c)
int pce_instance_tag
Indicates with which program the DRC info is associated.
int interpolation_scheme
Indicates the interpolation scheme used in the SBR QMF domain.
int ch_select[8]
[0] shared list of gains; [1] list of gains for right channel; [2] list of gains for left channel; [3...
int force_dmono_mode
0->not dmono, 1->use first channel, 2->use second channel
int is_saved
Set if elements have stored overlap from previous frame.
int warned_num_aac_frames
Libavcodec external API header.
main external API structure.
IndividualChannelStream ics
uint8_t pi<< 24) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_U8, uint8_t,(*(constuint8_t *) pi-0x80)*(1.0f/(1<< 7))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_U8, uint8_t,(*(constuint8_t *) pi-0x80)*(1.0/(1<< 7))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S16, int16_t,(*(constint16_t *) pi >>8)+0x80) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S16, int16_t,*(constint16_t *) pi *(1.0f/(1<< 15))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S16, int16_t,*(constint16_t *) pi *(1.0/(1<< 15))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S32, int32_t,(*(constint32_t *) pi >>24)+0x80) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S32, int32_t,*(constint32_t *) pi *(1.0f/(1U<< 31))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S32, int32_t,*(constint32_t *) pi *(1.0/(1U<< 31))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_FLT, float, av_clip_uint8(lrintf(*(constfloat *) pi *(1<< 7))+0x80)) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_FLT, float, av_clip_int16(lrintf(*(constfloat *) pi *(1<< 15)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_FLT, float, av_clipl_int32(llrintf(*(constfloat *) pi *(1U<< 31)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_DBL, double, av_clip_uint8(lrint(*(constdouble *) pi *(1<< 7))+0x80)) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_DBL, double, av_clip_int16(lrint(*(constdouble *) pi *(1<< 15)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_DBL, double, av_clipl_int32(llrint(*(constdouble *) pi *(1U<< 31))))#defineSET_CONV_FUNC_GROUP(ofmt, ifmt) staticvoidset_generic_function(AudioConvert *ac){}voidff_audio_convert_free(AudioConvert **ac){if(!*ac) return;ff_dither_free(&(*ac) ->dc);av_freep(ac);}AudioConvert *ff_audio_convert_alloc(AVAudioResampleContext *avr, enumAVSampleFormatout_fmt, enumAVSampleFormatin_fmt, intchannels, intsample_rate, intapply_map){AudioConvert *ac;intin_planar, out_planar;ac=av_mallocz(sizeof(*ac));if(!ac) returnNULL;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);returnNULL;}returnac;}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;}elseif(in_planar) ac->func_type=CONV_FUNC_TYPE_INTERLEAVE;elseac->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);returnac;}intff_audio_convert(AudioConvert *ac, AudioData *out, AudioData *in){intuse_generic=1;intlen=in->nb_samples;intp;if(ac->dc){av_log(ac->avr, AV_LOG_TRACE,"%dsamples-audio_convert:%sto%s(dithered)\n", len, av_get_sample_fmt_name(ac->in_fmt), av_get_sample_fmt_name(ac->out_fmt));returnff_convert_dither(ac-> in
void(* apply_tns)(INTFLOAT coef[1024], TemporalNoiseShaping *tns, IndividualChannelStream *ics, int decode)
Describe the class of an AVClass context structure.
Spectral data are coded with an escape sequence.
const uint8_t * swb_sizes
table of scalefactor band sizes for a particular window
OCStatus
Output configuration status.
int ms_mode
Signals mid/side stereo flags coding mode (used by encoder)
Output configuration under trial specified by a frame header.
uint8_t zeroes[128]
band is not coded (used by encoder)
int band_type_run_end[120]
band type run end points
int sf_idx[128]
scalefactor indices (used by encoder)
uint8_t is_mode
Set if any bands have been encoded using intensity stereo (used by encoder)
int band_top[17]
Indicates the top of the i-th DRC band in units of 4 spectral lines.
INTFLOAT coeffs[1024]
coefficients for IMDCT, maybe processed
int predictor_reset_count[31]
used by encoder to count prediction resets
Scalefactor data are intensity stereo positions (out of phase).
INTFLOAT coef[8][4][TNS_MAX_ORDER]
DynamicRangeControl che_drc
OutputConfiguration oc[2]
void(* apply_ltp)(AACContext *ac, SingleChannelElement *sce)
uint8_t prediction_used[41]
INTFLOAT saved[1536]
overlap
Single Channel Element - used for both SCE and LFE elements.
void(* windowing_and_mdct_ltp)(AACContext *ac, INTFLOAT *out, INTFLOAT *in, IndividualChannelStream *ics)
Individual Channel Stream.
float clip_avoidance_factor
set if any window is near clipping to the necessary atennuation factor to avoid it ...
channel element - generic struct for SCE/CPE/CCE/LFE
Scalefactors and spectral data are all zero.
AAC_FLOAT lcoeffs[1024]
MDCT of LTP coefficients (used by encoder)
enum BandType band_type[128]
band types
int8_t used[MAX_LTP_LONG_SFB]
uint8_t is_mask[128]
Set if intensity stereo is used (used by encoder)
Spectral Band Replication.
This and later band types encode two values (rather than four) with one code word.