ffmpeg/libavcodec/ac3enc.h

288 lines
12 KiB
C
Raw Permalink Normal View History

/*
* AC-3 encoder & E-AC-3 encoder common header
* Copyright (c) 2000 Fabrice Bellard
* Copyright (c) 2006-2010 Justin Ruggles <justin.ruggles@gmail.com>
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
* @file
* AC-3 encoder & E-AC-3 encoder common header
*/
#ifndef AVCODEC_AC3ENC_H
#define AVCODEC_AC3ENC_H
#include <stdint.h>
#include "libavutil/mem_internal.h"
#include "libavutil/opt.h"
2022-10-26 13:12:58 +02:00
#include "libavutil/tx.h"
#include "ac3.h"
#include "ac3defs.h"
#include "ac3dsp.h"
#include "avcodec.h"
#include "codec_internal.h"
#include "mathops.h"
#include "me_cmp.h"
#include "put_bits.h"
#include "audiodsp.h"
#ifndef AC3ENC_FLOAT
#define AC3ENC_FLOAT 0
#endif
#if AC3ENC_FLOAT
ac3enc_fixed: convert to 32-bit sample format The AC3 encoder used to be a separate library called "Aften", which got merged into libavcodec (literally, SVN commits and all). The merge preserved as much features from the library as possible. The code had two versions - a fixed point version and a floating point version. FFmpeg had floating point DSP code used by other codecs, the AC3 decoder including, so the floating-point DSP was simply replaced with FFmpeg's own functions. However, FFmpeg had no fixed-point audio code at that point. So the encoder brought along its own fixed-point DSP functions, including a fixed-point MDCT. The fixed-point MDCT itself is trivially just a float MDCT with a different type and each multiply being a fixed-point multiply. So over time, it got refactored, and the FFT used for all other codecs was templated. Due to design decisions at the time, the fixed-point version of the encoder operates at 16-bits of precision. Although convenient, this, even at the time, was inadequate and inefficient. The encoder is noisy, does not produce output comparable to the float encoder, and even rings at higher frequencies due to the badly approximated winow function. Enter MIPS (owned by Imagination Technologies at the time). They wanted quick fixed-point decoding on their FPUless cores. So they contributed patches to template the AC3 decoder so it had both a fixed-point and a floating-point version. They also did the same for the AAC decoder. They however, used 32-bit samples. Not 16-bits. And we did not have 32-bit fixed-point DSP functions, including an MDCT. But instead of templating our MDCT to output 3 versions (float, 32-bit fixed and 16-bit fixed), they simply copy-pasted their own MDCT into ours, and completely ifdeffed our own MDCT code out if a 32-bit fixed point MDCT was selected. This is also the status quo nowadays - 2 separate MDCTs, one which produces floating point and 16-bit fixed point versions, and one sort-of integrated which produces 32-bit MDCT. MIPS weren't all that interested in encoding, so they left the encoder as-is, and they didn't care much about the ifdeffery, mess or quality - it's not their problem. So the MDCT/FFT code has always been a thorn in anyone looking to clean up code's eye. Backstory over. Internally AC3 operates on 25-bit fixed-point coefficients. So for the floating point version, the encoder simply runs the float MDCT, and converts the resulting coefficients to 25-bit fixed-point, as AC3 is inherently a fixed-point codec. For the fixed-point version, the input is 16-bit samples, so to maximize precision the frame samples are analyzed and the highest set bit is detected via ac3_max_msb_abs_int16(), and the coefficients are then scaled up via ac3_lshift_int16(), so the input for the FFT is always at least 14 bits, computed in normalize_samples(). After FFT, the coefficients are scaled up to 25 bits. This patch simply changes the encoder to accept 32-bit samples, reusing the already well-optimized 32-bit MDCT code, allowing us to clean up and drop a large part of a very messy code of ours, as well as prepare for the future lavu/tx conversion. The coefficients are simply scaled down to 25 bits during windowing, skipping 2 separate scalings, as the hacks to extend precision are simply no longer necessary. There's no point in running the MDCT always at 32 bits when you're going to drop 6 bits off anyway, the headroom is plenty, and the MDCT rounds properly. This also makes the encoder even slightly more accurate over the float version, as there's no coefficient conversion step necessary. SIZE SAVINGS: ARM32: HARDCODED TABLES: BASE - 10709590 DROP DSP - 10702872 - diff: -6.56KiB DROP MDCT - 10667932 - diff: -34.12KiB - both: -40.68KiB DROP FFT - 10336652 - diff: -323.52KiB - all: -364.20KiB SOFTCODED TABLES: BASE - 9685096 DROP DSP - 9678378 - diff: -6.56KiB DROP MDCT - 9643466 - diff: -34.09KiB - both: -40.65KiB DROP FFT - 9573918 - diff: -67.92KiB - all: -108.57KiB ARM64: HARDCODED TABLES: BASE - 14641112 DROP DSP - 14633806 - diff: -7.13KiB DROP MDCT - 14604812 - diff: -28.31KiB - both: -35.45KiB DROP FFT - 14286826 - diff: -310.53KiB - all: -345.98KiB SOFTCODED TABLES: BASE - 13636238 DROP DSP - 13628932 - diff: -7.13KiB DROP MDCT - 13599866 - diff: -28.38KiB - both: -35.52KiB DROP FFT - 13542080 - diff: -56.43KiB - all: -91.95KiB x86: HARDCODED TABLES: BASE - 12367336 DROP DSP - 12354698 - diff: -12.34KiB DROP MDCT - 12331024 - diff: -23.12KiB - both: -35.46KiB DROP FFT - 12029788 - diff: -294.18KiB - all: -329.64KiB SOFTCODED TABLES: BASE - 11358094 DROP DSP - 11345456 - diff: -12.34KiB DROP MDCT - 11321742 - diff: -23.16KiB - both: -35.50KiB DROP FFT - 11276946 - diff: -43.75KiB - all: -79.25KiB PERFORMANCE (10min random s32le): ARM32 - before - 39.9x - 0m15.046s ARM32 - after - 28.2x - 0m21.525s Speed: -30% ARM64 - before - 36.1x - 0m16.637s ARM64 - after - 36.0x - 0m16.727s Speed: -0.5% x86 - before - 184x - 0m3.277s x86 - after - 190x - 0m3.187s Speed: +3%
2021-01-09 01:51:52 +01:00
#include "libavutil/float_dsp.h"
#define MAC_COEF(d,a,b) ((d)+=(a)*(b))
#define COEF_MIN (-16777215.0/16777216.0)
#define COEF_MAX ( 16777215.0/16777216.0)
#define NEW_CPL_COORD_THRESHOLD 0.03
typedef float SampleType;
typedef float CoefType;
typedef float CoefSumType;
#else
ac3enc_fixed: convert to 32-bit sample format The AC3 encoder used to be a separate library called "Aften", which got merged into libavcodec (literally, SVN commits and all). The merge preserved as much features from the library as possible. The code had two versions - a fixed point version and a floating point version. FFmpeg had floating point DSP code used by other codecs, the AC3 decoder including, so the floating-point DSP was simply replaced with FFmpeg's own functions. However, FFmpeg had no fixed-point audio code at that point. So the encoder brought along its own fixed-point DSP functions, including a fixed-point MDCT. The fixed-point MDCT itself is trivially just a float MDCT with a different type and each multiply being a fixed-point multiply. So over time, it got refactored, and the FFT used for all other codecs was templated. Due to design decisions at the time, the fixed-point version of the encoder operates at 16-bits of precision. Although convenient, this, even at the time, was inadequate and inefficient. The encoder is noisy, does not produce output comparable to the float encoder, and even rings at higher frequencies due to the badly approximated winow function. Enter MIPS (owned by Imagination Technologies at the time). They wanted quick fixed-point decoding on their FPUless cores. So they contributed patches to template the AC3 decoder so it had both a fixed-point and a floating-point version. They also did the same for the AAC decoder. They however, used 32-bit samples. Not 16-bits. And we did not have 32-bit fixed-point DSP functions, including an MDCT. But instead of templating our MDCT to output 3 versions (float, 32-bit fixed and 16-bit fixed), they simply copy-pasted their own MDCT into ours, and completely ifdeffed our own MDCT code out if a 32-bit fixed point MDCT was selected. This is also the status quo nowadays - 2 separate MDCTs, one which produces floating point and 16-bit fixed point versions, and one sort-of integrated which produces 32-bit MDCT. MIPS weren't all that interested in encoding, so they left the encoder as-is, and they didn't care much about the ifdeffery, mess or quality - it's not their problem. So the MDCT/FFT code has always been a thorn in anyone looking to clean up code's eye. Backstory over. Internally AC3 operates on 25-bit fixed-point coefficients. So for the floating point version, the encoder simply runs the float MDCT, and converts the resulting coefficients to 25-bit fixed-point, as AC3 is inherently a fixed-point codec. For the fixed-point version, the input is 16-bit samples, so to maximize precision the frame samples are analyzed and the highest set bit is detected via ac3_max_msb_abs_int16(), and the coefficients are then scaled up via ac3_lshift_int16(), so the input for the FFT is always at least 14 bits, computed in normalize_samples(). After FFT, the coefficients are scaled up to 25 bits. This patch simply changes the encoder to accept 32-bit samples, reusing the already well-optimized 32-bit MDCT code, allowing us to clean up and drop a large part of a very messy code of ours, as well as prepare for the future lavu/tx conversion. The coefficients are simply scaled down to 25 bits during windowing, skipping 2 separate scalings, as the hacks to extend precision are simply no longer necessary. There's no point in running the MDCT always at 32 bits when you're going to drop 6 bits off anyway, the headroom is plenty, and the MDCT rounds properly. This also makes the encoder even slightly more accurate over the float version, as there's no coefficient conversion step necessary. SIZE SAVINGS: ARM32: HARDCODED TABLES: BASE - 10709590 DROP DSP - 10702872 - diff: -6.56KiB DROP MDCT - 10667932 - diff: -34.12KiB - both: -40.68KiB DROP FFT - 10336652 - diff: -323.52KiB - all: -364.20KiB SOFTCODED TABLES: BASE - 9685096 DROP DSP - 9678378 - diff: -6.56KiB DROP MDCT - 9643466 - diff: -34.09KiB - both: -40.65KiB DROP FFT - 9573918 - diff: -67.92KiB - all: -108.57KiB ARM64: HARDCODED TABLES: BASE - 14641112 DROP DSP - 14633806 - diff: -7.13KiB DROP MDCT - 14604812 - diff: -28.31KiB - both: -35.45KiB DROP FFT - 14286826 - diff: -310.53KiB - all: -345.98KiB SOFTCODED TABLES: BASE - 13636238 DROP DSP - 13628932 - diff: -7.13KiB DROP MDCT - 13599866 - diff: -28.38KiB - both: -35.52KiB DROP FFT - 13542080 - diff: -56.43KiB - all: -91.95KiB x86: HARDCODED TABLES: BASE - 12367336 DROP DSP - 12354698 - diff: -12.34KiB DROP MDCT - 12331024 - diff: -23.12KiB - both: -35.46KiB DROP FFT - 12029788 - diff: -294.18KiB - all: -329.64KiB SOFTCODED TABLES: BASE - 11358094 DROP DSP - 11345456 - diff: -12.34KiB DROP MDCT - 11321742 - diff: -23.16KiB - both: -35.50KiB DROP FFT - 11276946 - diff: -43.75KiB - all: -79.25KiB PERFORMANCE (10min random s32le): ARM32 - before - 39.9x - 0m15.046s ARM32 - after - 28.2x - 0m21.525s Speed: -30% ARM64 - before - 36.1x - 0m16.637s ARM64 - after - 36.0x - 0m16.727s Speed: -0.5% x86 - before - 184x - 0m3.277s x86 - after - 190x - 0m3.187s Speed: +3%
2021-01-09 01:51:52 +01:00
#include "libavutil/fixed_dsp.h"
#define MAC_COEF(d,a,b) MAC64(d,a,b)
#define COEF_MIN -16777215
#define COEF_MAX 16777215
#define NEW_CPL_COORD_THRESHOLD 503317
ac3enc_fixed: convert to 32-bit sample format The AC3 encoder used to be a separate library called "Aften", which got merged into libavcodec (literally, SVN commits and all). The merge preserved as much features from the library as possible. The code had two versions - a fixed point version and a floating point version. FFmpeg had floating point DSP code used by other codecs, the AC3 decoder including, so the floating-point DSP was simply replaced with FFmpeg's own functions. However, FFmpeg had no fixed-point audio code at that point. So the encoder brought along its own fixed-point DSP functions, including a fixed-point MDCT. The fixed-point MDCT itself is trivially just a float MDCT with a different type and each multiply being a fixed-point multiply. So over time, it got refactored, and the FFT used for all other codecs was templated. Due to design decisions at the time, the fixed-point version of the encoder operates at 16-bits of precision. Although convenient, this, even at the time, was inadequate and inefficient. The encoder is noisy, does not produce output comparable to the float encoder, and even rings at higher frequencies due to the badly approximated winow function. Enter MIPS (owned by Imagination Technologies at the time). They wanted quick fixed-point decoding on their FPUless cores. So they contributed patches to template the AC3 decoder so it had both a fixed-point and a floating-point version. They also did the same for the AAC decoder. They however, used 32-bit samples. Not 16-bits. And we did not have 32-bit fixed-point DSP functions, including an MDCT. But instead of templating our MDCT to output 3 versions (float, 32-bit fixed and 16-bit fixed), they simply copy-pasted their own MDCT into ours, and completely ifdeffed our own MDCT code out if a 32-bit fixed point MDCT was selected. This is also the status quo nowadays - 2 separate MDCTs, one which produces floating point and 16-bit fixed point versions, and one sort-of integrated which produces 32-bit MDCT. MIPS weren't all that interested in encoding, so they left the encoder as-is, and they didn't care much about the ifdeffery, mess or quality - it's not their problem. So the MDCT/FFT code has always been a thorn in anyone looking to clean up code's eye. Backstory over. Internally AC3 operates on 25-bit fixed-point coefficients. So for the floating point version, the encoder simply runs the float MDCT, and converts the resulting coefficients to 25-bit fixed-point, as AC3 is inherently a fixed-point codec. For the fixed-point version, the input is 16-bit samples, so to maximize precision the frame samples are analyzed and the highest set bit is detected via ac3_max_msb_abs_int16(), and the coefficients are then scaled up via ac3_lshift_int16(), so the input for the FFT is always at least 14 bits, computed in normalize_samples(). After FFT, the coefficients are scaled up to 25 bits. This patch simply changes the encoder to accept 32-bit samples, reusing the already well-optimized 32-bit MDCT code, allowing us to clean up and drop a large part of a very messy code of ours, as well as prepare for the future lavu/tx conversion. The coefficients are simply scaled down to 25 bits during windowing, skipping 2 separate scalings, as the hacks to extend precision are simply no longer necessary. There's no point in running the MDCT always at 32 bits when you're going to drop 6 bits off anyway, the headroom is plenty, and the MDCT rounds properly. This also makes the encoder even slightly more accurate over the float version, as there's no coefficient conversion step necessary. SIZE SAVINGS: ARM32: HARDCODED TABLES: BASE - 10709590 DROP DSP - 10702872 - diff: -6.56KiB DROP MDCT - 10667932 - diff: -34.12KiB - both: -40.68KiB DROP FFT - 10336652 - diff: -323.52KiB - all: -364.20KiB SOFTCODED TABLES: BASE - 9685096 DROP DSP - 9678378 - diff: -6.56KiB DROP MDCT - 9643466 - diff: -34.09KiB - both: -40.65KiB DROP FFT - 9573918 - diff: -67.92KiB - all: -108.57KiB ARM64: HARDCODED TABLES: BASE - 14641112 DROP DSP - 14633806 - diff: -7.13KiB DROP MDCT - 14604812 - diff: -28.31KiB - both: -35.45KiB DROP FFT - 14286826 - diff: -310.53KiB - all: -345.98KiB SOFTCODED TABLES: BASE - 13636238 DROP DSP - 13628932 - diff: -7.13KiB DROP MDCT - 13599866 - diff: -28.38KiB - both: -35.52KiB DROP FFT - 13542080 - diff: -56.43KiB - all: -91.95KiB x86: HARDCODED TABLES: BASE - 12367336 DROP DSP - 12354698 - diff: -12.34KiB DROP MDCT - 12331024 - diff: -23.12KiB - both: -35.46KiB DROP FFT - 12029788 - diff: -294.18KiB - all: -329.64KiB SOFTCODED TABLES: BASE - 11358094 DROP DSP - 11345456 - diff: -12.34KiB DROP MDCT - 11321742 - diff: -23.16KiB - both: -35.50KiB DROP FFT - 11276946 - diff: -43.75KiB - all: -79.25KiB PERFORMANCE (10min random s32le): ARM32 - before - 39.9x - 0m15.046s ARM32 - after - 28.2x - 0m21.525s Speed: -30% ARM64 - before - 36.1x - 0m16.637s ARM64 - after - 36.0x - 0m16.727s Speed: -0.5% x86 - before - 184x - 0m3.277s x86 - after - 190x - 0m3.187s Speed: +3%
2021-01-09 01:51:52 +01:00
typedef int32_t SampleType;
typedef int32_t CoefType;
typedef int64_t CoefSumType;
#endif
/* common option values */
#define AC3ENC_OPT_NONE -1
#define AC3ENC_OPT_AUTO -1
#define AC3ENC_OPT_OFF 0
#define AC3ENC_OPT_ON 1
#define AC3ENC_OPT_NOT_INDICATED 0
#define AC3ENC_OPT_MODE_ON 2
#define AC3ENC_OPT_MODE_OFF 1
#define AC3ENC_OPT_DSUREX_DPLIIZ 3
/* specific option values */
#define AC3ENC_OPT_LARGE_ROOM 1
#define AC3ENC_OPT_SMALL_ROOM 2
#define AC3ENC_OPT_DOWNMIX_LTRT 1
#define AC3ENC_OPT_DOWNMIX_LORO 2
#define AC3ENC_OPT_DOWNMIX_DPLII 3 // reserved value in A/52, but used by encoders to indicate DPL2
#define AC3ENC_OPT_ADCONV_STANDARD 0
#define AC3ENC_OPT_ADCONV_HDCD 1
/**
* Encoding Options used by AVOption.
*/
typedef struct AC3EncOptions {
/* AC-3 metadata options*/
int dialogue_level;
int bitstream_mode;
float center_mix_level;
float surround_mix_level;
int dolby_surround_mode;
int audio_production_info;
int mixing_level;
int room_type;
int copyright;
int original;
int extended_bsi_1;
int preferred_stereo_downmix;
float ltrt_center_mix_level;
float ltrt_surround_mix_level;
float loro_center_mix_level;
float loro_surround_mix_level;
int extended_bsi_2;
int dolby_surround_ex_mode;
int dolby_headphone_mode;
int ad_converter_type;
int eac3_mixing_metadata;
int eac3_info_metadata;
/* other encoding options */
int allow_per_frame_metadata;
int stereo_rematrixing;
int channel_coupling;
int cpl_start;
} AC3EncOptions;
/**
* Data for a single audio block.
*/
typedef struct AC3Block {
CoefType *mdct_coef[AC3_MAX_CHANNELS]; ///< MDCT coefficients
int32_t *fixed_coef[AC3_MAX_CHANNELS]; ///< fixed-point MDCT coefficients
uint8_t *exp[AC3_MAX_CHANNELS]; ///< original exponents
uint8_t *grouped_exp[AC3_MAX_CHANNELS]; ///< grouped exponents
int16_t *psd[AC3_MAX_CHANNELS]; ///< psd per frequency bin
int16_t *band_psd[AC3_MAX_CHANNELS]; ///< psd per critical band
int16_t *mask[AC3_MAX_CHANNELS]; ///< masking curve
uint16_t *qmant[AC3_MAX_CHANNELS]; ///< quantized mantissas
uint8_t *cpl_coord_exp[AC3_MAX_CHANNELS]; ///< coupling coord exponents (cplcoexp)
uint8_t *cpl_coord_mant[AC3_MAX_CHANNELS]; ///< coupling coord mantissas (cplcomant)
uint8_t new_rematrixing_strategy; ///< send new rematrixing flags in this block
int num_rematrixing_bands; ///< number of rematrixing bands
uint8_t rematrixing_flags[4]; ///< rematrixing flags
int new_cpl_strategy; ///< send new coupling strategy
int cpl_in_use; ///< coupling in use for this block (cplinu)
uint8_t channel_in_cpl[AC3_MAX_CHANNELS]; ///< channel in coupling (chincpl)
int num_cpl_channels; ///< number of channels in coupling
uint8_t new_cpl_coords[AC3_MAX_CHANNELS]; ///< send new coupling coordinates (cplcoe)
uint8_t cpl_master_exp[AC3_MAX_CHANNELS]; ///< coupling coord master exponents (mstrcplco)
int new_snr_offsets; ///< send new SNR offsets
int new_cpl_leak; ///< send new coupling leak info
int end_freq[AC3_MAX_CHANNELS]; ///< end frequency bin (endmant)
} AC3Block;
/**
* AC-3 encoder private context.
*/
typedef struct AC3EncodeContext {
AVClass *av_class; ///< AVClass used for AVOption
AC3EncOptions options; ///< encoding options
AVCodecContext *avctx; ///< parent AVCodecContext
PutBitContext pb; ///< bitstream writer context
AudioDSPContext adsp;
ac3enc_fixed: convert to 32-bit sample format The AC3 encoder used to be a separate library called "Aften", which got merged into libavcodec (literally, SVN commits and all). The merge preserved as much features from the library as possible. The code had two versions - a fixed point version and a floating point version. FFmpeg had floating point DSP code used by other codecs, the AC3 decoder including, so the floating-point DSP was simply replaced with FFmpeg's own functions. However, FFmpeg had no fixed-point audio code at that point. So the encoder brought along its own fixed-point DSP functions, including a fixed-point MDCT. The fixed-point MDCT itself is trivially just a float MDCT with a different type and each multiply being a fixed-point multiply. So over time, it got refactored, and the FFT used for all other codecs was templated. Due to design decisions at the time, the fixed-point version of the encoder operates at 16-bits of precision. Although convenient, this, even at the time, was inadequate and inefficient. The encoder is noisy, does not produce output comparable to the float encoder, and even rings at higher frequencies due to the badly approximated winow function. Enter MIPS (owned by Imagination Technologies at the time). They wanted quick fixed-point decoding on their FPUless cores. So they contributed patches to template the AC3 decoder so it had both a fixed-point and a floating-point version. They also did the same for the AAC decoder. They however, used 32-bit samples. Not 16-bits. And we did not have 32-bit fixed-point DSP functions, including an MDCT. But instead of templating our MDCT to output 3 versions (float, 32-bit fixed and 16-bit fixed), they simply copy-pasted their own MDCT into ours, and completely ifdeffed our own MDCT code out if a 32-bit fixed point MDCT was selected. This is also the status quo nowadays - 2 separate MDCTs, one which produces floating point and 16-bit fixed point versions, and one sort-of integrated which produces 32-bit MDCT. MIPS weren't all that interested in encoding, so they left the encoder as-is, and they didn't care much about the ifdeffery, mess or quality - it's not their problem. So the MDCT/FFT code has always been a thorn in anyone looking to clean up code's eye. Backstory over. Internally AC3 operates on 25-bit fixed-point coefficients. So for the floating point version, the encoder simply runs the float MDCT, and converts the resulting coefficients to 25-bit fixed-point, as AC3 is inherently a fixed-point codec. For the fixed-point version, the input is 16-bit samples, so to maximize precision the frame samples are analyzed and the highest set bit is detected via ac3_max_msb_abs_int16(), and the coefficients are then scaled up via ac3_lshift_int16(), so the input for the FFT is always at least 14 bits, computed in normalize_samples(). After FFT, the coefficients are scaled up to 25 bits. This patch simply changes the encoder to accept 32-bit samples, reusing the already well-optimized 32-bit MDCT code, allowing us to clean up and drop a large part of a very messy code of ours, as well as prepare for the future lavu/tx conversion. The coefficients are simply scaled down to 25 bits during windowing, skipping 2 separate scalings, as the hacks to extend precision are simply no longer necessary. There's no point in running the MDCT always at 32 bits when you're going to drop 6 bits off anyway, the headroom is plenty, and the MDCT rounds properly. This also makes the encoder even slightly more accurate over the float version, as there's no coefficient conversion step necessary. SIZE SAVINGS: ARM32: HARDCODED TABLES: BASE - 10709590 DROP DSP - 10702872 - diff: -6.56KiB DROP MDCT - 10667932 - diff: -34.12KiB - both: -40.68KiB DROP FFT - 10336652 - diff: -323.52KiB - all: -364.20KiB SOFTCODED TABLES: BASE - 9685096 DROP DSP - 9678378 - diff: -6.56KiB DROP MDCT - 9643466 - diff: -34.09KiB - both: -40.65KiB DROP FFT - 9573918 - diff: -67.92KiB - all: -108.57KiB ARM64: HARDCODED TABLES: BASE - 14641112 DROP DSP - 14633806 - diff: -7.13KiB DROP MDCT - 14604812 - diff: -28.31KiB - both: -35.45KiB DROP FFT - 14286826 - diff: -310.53KiB - all: -345.98KiB SOFTCODED TABLES: BASE - 13636238 DROP DSP - 13628932 - diff: -7.13KiB DROP MDCT - 13599866 - diff: -28.38KiB - both: -35.52KiB DROP FFT - 13542080 - diff: -56.43KiB - all: -91.95KiB x86: HARDCODED TABLES: BASE - 12367336 DROP DSP - 12354698 - diff: -12.34KiB DROP MDCT - 12331024 - diff: -23.12KiB - both: -35.46KiB DROP FFT - 12029788 - diff: -294.18KiB - all: -329.64KiB SOFTCODED TABLES: BASE - 11358094 DROP DSP - 11345456 - diff: -12.34KiB DROP MDCT - 11321742 - diff: -23.16KiB - both: -35.50KiB DROP FFT - 11276946 - diff: -43.75KiB - all: -79.25KiB PERFORMANCE (10min random s32le): ARM32 - before - 39.9x - 0m15.046s ARM32 - after - 28.2x - 0m21.525s Speed: -30% ARM64 - before - 36.1x - 0m16.637s ARM64 - after - 36.0x - 0m16.727s Speed: -0.5% x86 - before - 184x - 0m3.277s x86 - after - 190x - 0m3.187s Speed: +3%
2021-01-09 01:51:52 +01:00
#if AC3ENC_FLOAT
AVFloatDSPContext *fdsp;
ac3enc_fixed: convert to 32-bit sample format The AC3 encoder used to be a separate library called "Aften", which got merged into libavcodec (literally, SVN commits and all). The merge preserved as much features from the library as possible. The code had two versions - a fixed point version and a floating point version. FFmpeg had floating point DSP code used by other codecs, the AC3 decoder including, so the floating-point DSP was simply replaced with FFmpeg's own functions. However, FFmpeg had no fixed-point audio code at that point. So the encoder brought along its own fixed-point DSP functions, including a fixed-point MDCT. The fixed-point MDCT itself is trivially just a float MDCT with a different type and each multiply being a fixed-point multiply. So over time, it got refactored, and the FFT used for all other codecs was templated. Due to design decisions at the time, the fixed-point version of the encoder operates at 16-bits of precision. Although convenient, this, even at the time, was inadequate and inefficient. The encoder is noisy, does not produce output comparable to the float encoder, and even rings at higher frequencies due to the badly approximated winow function. Enter MIPS (owned by Imagination Technologies at the time). They wanted quick fixed-point decoding on their FPUless cores. So they contributed patches to template the AC3 decoder so it had both a fixed-point and a floating-point version. They also did the same for the AAC decoder. They however, used 32-bit samples. Not 16-bits. And we did not have 32-bit fixed-point DSP functions, including an MDCT. But instead of templating our MDCT to output 3 versions (float, 32-bit fixed and 16-bit fixed), they simply copy-pasted their own MDCT into ours, and completely ifdeffed our own MDCT code out if a 32-bit fixed point MDCT was selected. This is also the status quo nowadays - 2 separate MDCTs, one which produces floating point and 16-bit fixed point versions, and one sort-of integrated which produces 32-bit MDCT. MIPS weren't all that interested in encoding, so they left the encoder as-is, and they didn't care much about the ifdeffery, mess or quality - it's not their problem. So the MDCT/FFT code has always been a thorn in anyone looking to clean up code's eye. Backstory over. Internally AC3 operates on 25-bit fixed-point coefficients. So for the floating point version, the encoder simply runs the float MDCT, and converts the resulting coefficients to 25-bit fixed-point, as AC3 is inherently a fixed-point codec. For the fixed-point version, the input is 16-bit samples, so to maximize precision the frame samples are analyzed and the highest set bit is detected via ac3_max_msb_abs_int16(), and the coefficients are then scaled up via ac3_lshift_int16(), so the input for the FFT is always at least 14 bits, computed in normalize_samples(). After FFT, the coefficients are scaled up to 25 bits. This patch simply changes the encoder to accept 32-bit samples, reusing the already well-optimized 32-bit MDCT code, allowing us to clean up and drop a large part of a very messy code of ours, as well as prepare for the future lavu/tx conversion. The coefficients are simply scaled down to 25 bits during windowing, skipping 2 separate scalings, as the hacks to extend precision are simply no longer necessary. There's no point in running the MDCT always at 32 bits when you're going to drop 6 bits off anyway, the headroom is plenty, and the MDCT rounds properly. This also makes the encoder even slightly more accurate over the float version, as there's no coefficient conversion step necessary. SIZE SAVINGS: ARM32: HARDCODED TABLES: BASE - 10709590 DROP DSP - 10702872 - diff: -6.56KiB DROP MDCT - 10667932 - diff: -34.12KiB - both: -40.68KiB DROP FFT - 10336652 - diff: -323.52KiB - all: -364.20KiB SOFTCODED TABLES: BASE - 9685096 DROP DSP - 9678378 - diff: -6.56KiB DROP MDCT - 9643466 - diff: -34.09KiB - both: -40.65KiB DROP FFT - 9573918 - diff: -67.92KiB - all: -108.57KiB ARM64: HARDCODED TABLES: BASE - 14641112 DROP DSP - 14633806 - diff: -7.13KiB DROP MDCT - 14604812 - diff: -28.31KiB - both: -35.45KiB DROP FFT - 14286826 - diff: -310.53KiB - all: -345.98KiB SOFTCODED TABLES: BASE - 13636238 DROP DSP - 13628932 - diff: -7.13KiB DROP MDCT - 13599866 - diff: -28.38KiB - both: -35.52KiB DROP FFT - 13542080 - diff: -56.43KiB - all: -91.95KiB x86: HARDCODED TABLES: BASE - 12367336 DROP DSP - 12354698 - diff: -12.34KiB DROP MDCT - 12331024 - diff: -23.12KiB - both: -35.46KiB DROP FFT - 12029788 - diff: -294.18KiB - all: -329.64KiB SOFTCODED TABLES: BASE - 11358094 DROP DSP - 11345456 - diff: -12.34KiB DROP MDCT - 11321742 - diff: -23.16KiB - both: -35.50KiB DROP FFT - 11276946 - diff: -43.75KiB - all: -79.25KiB PERFORMANCE (10min random s32le): ARM32 - before - 39.9x - 0m15.046s ARM32 - after - 28.2x - 0m21.525s Speed: -30% ARM64 - before - 36.1x - 0m16.637s ARM64 - after - 36.0x - 0m16.727s Speed: -0.5% x86 - before - 184x - 0m3.277s x86 - after - 190x - 0m3.187s Speed: +3%
2021-01-09 01:51:52 +01:00
#else
AVFixedDSPContext *fdsp;
#endif
MECmpContext mecc;
AC3DSPContext ac3dsp; ///< AC-3 optimized functions
2022-10-26 13:12:58 +02:00
AVTXContext *tx; ///< FFT context for MDCT calculation
av_tx_fn tx_fn;
AC3Block blocks[AC3_MAX_BLOCKS]; ///< per-block info
int fixed_point; ///< indicates if fixed-point encoder is being used
int eac3; ///< indicates if this is E-AC-3 vs. AC-3
int bitstream_id; ///< bitstream id (bsid)
int bitstream_mode; ///< bitstream mode (bsmod)
int bit_rate; ///< target bit rate, in bits-per-second
int sample_rate; ///< sampling frequency, in Hz
int num_blks_code; ///< number of blocks code (numblkscod)
int num_blocks; ///< number of blocks per frame
int frame_size_min; ///< minimum frame size in case rounding is necessary
int frame_size; ///< current frame size in bytes
int frame_size_code; ///< frame size code (frmsizecod)
uint16_t crc_inv[2];
int64_t bits_written; ///< bit count (used to avg. bitrate)
int64_t samples_written; ///< sample count (used to avg. bitrate)
int fbw_channels; ///< number of full-bandwidth channels (nfchans)
int channels; ///< total number of channels (nchans)
int lfe_on; ///< indicates if there is an LFE channel (lfeon)
int lfe_channel; ///< channel index of the LFE channel
int has_center; ///< indicates if there is a center channel
int has_surround; ///< indicates if there are one or more surround channels
int channel_mode; ///< channel mode (acmod)
const uint8_t *channel_map; ///< channel map used to reorder channels
int center_mix_level; ///< center mix level code
int surround_mix_level; ///< surround mix level code
int ltrt_center_mix_level; ///< Lt/Rt center mix level code
int ltrt_surround_mix_level; ///< Lt/Rt surround mix level code
int loro_center_mix_level; ///< Lo/Ro center mix level code
int loro_surround_mix_level; ///< Lo/Ro surround mix level code
int cutoff; ///< user-specified cutoff frequency, in Hz
int bandwidth_code; ///< bandwidth code (0 to 60) (chbwcod)
int start_freq[AC3_MAX_CHANNELS]; ///< start frequency bin (strtmant)
int cpl_end_freq; ///< coupling channel end frequency bin
int cpl_on; ///< coupling turned on for this frame
int cpl_enabled; ///< coupling enabled for all frames
int num_cpl_subbands; ///< number of coupling subbands (ncplsubnd)
int num_cpl_bands; ///< number of coupling bands (ncplbnd)
uint8_t cpl_band_sizes[AC3_MAX_CPL_BANDS]; ///< number of coeffs in each coupling band
int rematrixing_enabled; ///< stereo rematrixing enabled
/* bitrate allocation control */
int slow_gain_code; ///< slow gain code (sgaincod)
int slow_decay_code; ///< slow decay code (sdcycod)
int fast_decay_code; ///< fast decay code (fdcycod)
int db_per_bit_code; ///< dB/bit code (dbpbcod)
int floor_code; ///< floor code (floorcod)
AC3BitAllocParameters bit_alloc; ///< bit allocation parameters
int coarse_snr_offset; ///< coarse SNR offsets (csnroffst)
int fast_gain_code[AC3_MAX_CHANNELS]; ///< fast gain codes (signal-to-mask ratio) (fgaincod)
int fine_snr_offset[AC3_MAX_CHANNELS]; ///< fine SNR offsets (fsnroffst)
int frame_bits_fixed; ///< number of non-coefficient bits for fixed parameters
int frame_bits; ///< all frame bits except exponents and mantissas
int exponent_bits; ///< number of bits used for exponents
uint8_t *planar_samples[AC3_MAX_CHANNELS - 1];
uint8_t *bap_buffer;
uint8_t *bap1_buffer;
CoefType *mdct_coef_buffer;
int32_t *fixed_coef_buffer;
uint8_t *exp_buffer;
uint8_t *grouped_exp_buffer;
int16_t *psd_buffer;
int16_t *band_psd_buffer;
int16_t *mask_buffer;
int16_t *qmant_buffer;
uint8_t *cpl_coord_buffer;
uint8_t exp_strategy[AC3_MAX_CHANNELS][AC3_MAX_BLOCKS]; ///< exponent strategies
uint8_t frame_exp_strategy[AC3_MAX_CHANNELS]; ///< frame exp strategy index
int use_frame_exp_strategy; ///< indicates use of frame exp strategy
uint8_t exp_ref_block[AC3_MAX_CHANNELS][AC3_MAX_BLOCKS]; ///< reference blocks for EXP_REUSE
uint8_t *ref_bap [AC3_MAX_CHANNELS][AC3_MAX_BLOCKS]; ///< bit allocation pointers (bap)
int ref_bap_set; ///< indicates if ref_bap pointers have been set
/** fixed vs. float function pointers */
void (*encode_frame)(struct AC3EncodeContext *s, uint8_t * const *samples);
/* AC-3 vs. E-AC-3 function pointers */
void (*output_frame_header)(struct AC3EncodeContext *s);
union {
DECLARE_ALIGNED(32, float, mdct_window_float)[AC3_BLOCK_SIZE];
DECLARE_ALIGNED(32, int32_t, mdct_window_fixed)[AC3_BLOCK_SIZE];
};
union {
DECLARE_ALIGNED(32, float, windowed_samples_float)[AC3_WINDOW_SIZE];
DECLARE_ALIGNED(32, int32_t, windowed_samples_fixed)[AC3_WINDOW_SIZE];
};
} AC3EncodeContext;
extern const AVChannelLayout ff_ac3_ch_layouts[19];
extern const AVOption ff_ac3_enc_options[];
extern const AVClass ff_ac3enc_class;
extern const FFCodecDefault ff_ac3_enc_defaults[];
int ff_ac3_encode_init(AVCodecContext *avctx);
int ff_ac3_float_encode_init(AVCodecContext *avctx);
int ff_ac3_encode_close(AVCodecContext *avctx);
void ff_ac3_compute_coupling_strategy(AC3EncodeContext *s);
int ff_ac3_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
const AVFrame *frame, int *got_packet_ptr);
#endif /* AVCODEC_AC3ENC_H */