FFmpeg
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
aacenc.h
Go to the documentation of this file.
1 /*
2  * AAC encoder
3  * Copyright (C) 2008 Konstantin Shishkov
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 #ifndef AVCODEC_AACENC_H
23 #define AVCODEC_AACENC_H
24 
25 #include "libavutil/float_dsp.h"
26 #include "avcodec.h"
27 #include "put_bits.h"
28 
29 #include "aac.h"
30 #include "audio_frame_queue.h"
31 #include "psymodel.h"
32 
33 #include "lpc.h"
34 
35 typedef enum AACCoder {
39 
41 }AACCoder;
42 
43 typedef struct AACEncOptions {
44  int coder;
45  int pns;
46  int tns;
47  int ltp;
48  int pred;
49  int mid_side;
52 
53 struct AACEncContext;
54 
55 typedef struct AACCoefficientsEncoder {
57  SingleChannelElement *sce, const float lambda);
59  int win, int group_len, const float lambda);
60  void (*quantize_and_encode_band)(struct AACEncContext *s, PutBitContext *pb, const float *in, float *out, int size,
61  int scale_idx, int cb, const float lambda, int rtz);
63  void (*encode_ltp_info)(struct AACEncContext *s, SingleChannelElement *sce, int common_window);
75  void (*search_for_ltp)(struct AACEncContext *s, SingleChannelElement *sce, int common_window);
80 
82 
84  float rd;
85  float energy;
86  int bits;
87  char cb;
88  char rtz;
89  uint16_t generation;
91 
92 /**
93  * AAC encoder context
94  */
95 typedef struct AACEncContext {
97  AACEncOptions options; ///< encoding options
99  FFTContext mdct1024; ///< long (1024 samples) frame transform context
100  FFTContext mdct128; ///< short (128 samples) frame transform context
102  float *planar_samples[8]; ///< saved preprocessed input
103 
104  int profile; ///< copied from avctx
105  LPCContext lpc; ///< used by TNS
106  int samplerate_index; ///< MPEG-4 samplerate index
107  int channels; ///< channel count
108  const uint8_t *chan_map; ///< channel configuration map
109 
110  ChannelElement *cpe; ///< channel elements
114  int cur_channel; ///< current channel for coder context
117  float lambda;
118  int last_frame_pb_count; ///< number of bits for the previous frame
119  float lambda_sum; ///< sum(lambda), for Qvg reporting
120  int lambda_count; ///< count(lambda), for Qvg reporting
121  enum RawDataBlockType cur_type; ///< channel group type cur_channel belongs to
122 
124  DECLARE_ALIGNED(16, int, qcoefs)[96]; ///< quantized coefficients
125  DECLARE_ALIGNED(32, float, scoefs)[1024]; ///< scaled coefficients
126 
128  AACQuantizeBandCostCacheEntry quantize_band_cost_cache[256][128]; ///< memoization area for quantize_band_cost
129 
130  void (*abs_pow34)(float *out, const float *in, const int size);
131  void (*quant_bands)(int *out, const float *in, const float *scaled,
132  int size, int is_signed, int maxval, const float Q34,
133  const float rounding);
134 
135  struct {
136  float *samples;
137  } buffer;
138 } AACEncContext;
139 
143 
144 
145 #endif /* AVCODEC_AACENC_H */
const char * s
Definition: avisynth_c.h:768
void(* search_for_pns)(struct AACEncContext *s, AVCodecContext *avctx, SingleChannelElement *sce)
Definition: aacenc.h:72
int coder
Definition: aacenc.h:44
enum RawDataBlockType cur_type
channel group type cur_channel belongs to
Definition: aacenc.h:121
Definition: lpc.h:52
AACQuantizeBandCostCacheEntry quantize_band_cost_cache[256][128]
memoization area for quantize_band_cost
Definition: aacenc.h:128
void(* encode_ltp_info)(struct AACEncContext *s, SingleChannelElement *sce, int common_window)
Definition: aacenc.h:63
AACCoefficientsEncoder * coder
Definition: aacenc.h:113
int lambda_count
count(lambda), for Qvg reporting
Definition: aacenc.h:120
float lambda
Definition: aacenc.h:117
void(* search_for_tns)(struct AACEncContext *s, SingleChannelElement *sce)
Definition: aacenc.h:74
static double cb(void *priv, double x, double y)
Definition: vf_geq.c:97
AACEncOptions options
encoding options
Definition: aacenc.h:97
AAC encoder context.
Definition: aacenc.h:95
uint8_t
void ff_quantize_band_cost_cache_init(struct AACEncContext *s)
Definition: aacenc.c:79
void(* quant_bands)(int *out, const float *in, const float *scaled, int size, int is_signed, int maxval, const float Q34, const float rounding)
Definition: aacenc.h:131
int intensity_stereo
Definition: aacenc.h:50
LPCContext lpc
used by TNS
Definition: aacenc.h:105
void ff_aac_coder_init_mips(AACEncContext *c)
int samplerate_index
MPEG-4 samplerate index.
Definition: aacenc.h:106
const uint8_t * chan_map
channel configuration map
Definition: aacenc.h:108
int qcoefs[96]
quantized coefficients
Definition: aacenc.h:124
AudioFrameQueue afq
Definition: aacenc.h:123
#define DECLARE_ALIGNED(n, t, v)
Declare a variable that is aligned in memory.
Definition: mem.h:101
context used by psychoacoustic model
Definition: psymodel.h:89
int profile
copied from avctx
Definition: aacenc.h:104
ptrdiff_t size
Definition: opengl_enc.c:101
float * planar_samples[8]
saved preprocessed input
Definition: aacenc.h:102
int last_frame
Definition: aacenc.h:115
AACCoder
Definition: aacenc.h:35
void(* set_special_band_scalefactors)(struct AACEncContext *s, SingleChannelElement *sce)
Definition: aacenc.h:71
void(* search_for_quantizers)(AVCodecContext *avctx, struct AACEncContext *s, SingleChannelElement *sce, const float lambda)
Definition: aacenc.h:56
struct AACEncContext::@43 buffer
void(* search_for_is)(struct AACEncContext *s, AVCodecContext *avctx, ChannelElement *cpe)
Definition: aacenc.h:77
void(* abs_pow34)(float *out, const float *in, const int size)
Definition: aacenc.h:130
Definition: fft.h:88
AVClass * av_class
Definition: aacenc.h:96
int cur_channel
current channel for coder context
Definition: aacenc.h:114
int last_frame_pb_count
number of bits for the previous frame
Definition: aacenc.h:118
void(* apply_tns_filt)(struct AACEncContext *s, SingleChannelElement *sce)
Definition: aacenc.h:68
typedef void(APIENTRY *FF_PFNGLACTIVETEXTUREPROC)(GLenum texture)
void(* quantize_and_encode_band)(struct AACEncContext *s, PutBitContext *pb, const float *in, float *out, int size, int scale_idx, int cb, const float lambda, int rtz)
Definition: aacenc.h:60
int channels
channel count
Definition: aacenc.h:107
void(* encode_main_pred)(struct AACEncContext *s, SingleChannelElement *sce)
Definition: aacenc.h:64
AAC definitions and structures.
void(* search_for_ltp)(struct AACEncContext *s, SingleChannelElement *sce, int common_window)
Definition: aacenc.h:75
FFTContext mdct128
short (128 samples) frame transform context
Definition: aacenc.h:100
PutBitContext pb
Definition: aacenc.h:98
void(* search_for_pred)(struct AACEncContext *s, SingleChannelElement *sce)
Definition: aacenc.h:78
AVFloatDSPContext * fdsp
Definition: aacenc.h:101
int mid_side
Definition: aacenc.h:49
void(* adjust_common_pred)(struct AACEncContext *s, ChannelElement *cpe)
Definition: aacenc.h:65
void ff_aac_dsp_init_x86(AACEncContext *s)
Libavcodec external API header.
void(* search_for_ms)(struct AACEncContext *s, ChannelElement *cpe)
Definition: aacenc.h:76
void(* apply_main_pred)(struct AACEncContext *s, SingleChannelElement *sce)
Definition: aacenc.h:67
main external API structure.
Definition: avcodec.h:1676
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(* encode_window_bands_info)(struct AACEncContext *s, SingleChannelElement *sce, int win, int group_len, const float lambda)
Definition: aacenc.h:58
void(* mark_pns)(struct AACEncContext *s, AVCodecContext *avctx, SingleChannelElement *sce)
Definition: aacenc.h:73
float lambda_sum
sum(lambda), for Qvg reporting
Definition: aacenc.h:119
Describe the class of an AVClass context structure.
Definition: log.h:67
uint16_t quantize_band_cost_cache_generation
Definition: aacenc.h:127
FFPsyContext psy
Definition: aacenc.h:111
struct FFPsyPreprocessContext * psypp
Definition: aacenc.h:112
void(* encode_tns_info)(struct AACEncContext *s, SingleChannelElement *sce)
Definition: aacenc.h:62
AACCoefficientsEncoder ff_aac_coders[]
Definition: aaccoder.c:897
float * samples
Definition: aacenc.h:136
Single Channel Element - used for both SCE and LFE elements.
Definition: aac.h:248
static double c[64]
ChannelElement * cpe
channel elements
Definition: aacenc.h:110
void(* ltp_insert_new_frame)(struct AACEncContext *s)
Definition: aacenc.h:70
channel element - generic struct for SCE/CPE/CCE/LFE
Definition: aac.h:275
FFTContext mdct1024
long (1024 samples) frame transform context
Definition: aacenc.h:99
int random_state
Definition: aacenc.h:116
FILE * out
Definition: movenc.c:54
void(* adjust_common_ltp)(struct AACEncContext *s, ChannelElement *cpe)
Definition: aacenc.h:66
float scoefs[1024]
scaled coefficients
Definition: aacenc.h:125
void(* update_ltp)(struct AACEncContext *s, SingleChannelElement *sce)
Definition: aacenc.h:69
int pred
Definition: aacenc.h:48
RawDataBlockType
Definition: aac.h:55
bitstream writer API