virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-12-23 12:43:46 +02:00
Code Issues Releases Activity
b011bb5f8b
FFmpeg/libavcodec/aacsbr.c
Ganesh Ajjanagadde def3c83e1b lavc/aacsbr: sbr_dequant optimization 
This uses ff_exp2fi to get a speedup (~ 6x).

sample benchmark (Haswell, GNU/Linux):
old:
  19102 decicycles in sbr_dequant,    1023 runs,      1 skips
  19002 decicycles in sbr_dequant,    2045 runs,      3 skips
  17638 decicycles in sbr_dequant,    4093 runs,      3 skips
  15825 decicycles in sbr_dequant,    8189 runs,      3 skips
  16404 decicycles in sbr_dequant,   16379 runs,      5 skips

new:
   3063 decicycles in sbr_dequant,    1024 runs,      0 skips
   3049 decicycles in sbr_dequant,    2048 runs,      0 skips
   2968 decicycles in sbr_dequant,    4096 runs,      0 skips
   2818 decicycles in sbr_dequant,    8191 runs,      1 skips
   2853 decicycles in sbr_dequant,   16383 runs,      1 skips

Reviewed-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
2015-12-19 09:32:53 -08:00

370 lines
14 KiB
C
Raw Blame History

/*
* AAC Spectral Band Replication decoding functions
* Copyright (c) 2008-2009 Robert Swain ( rob opendot cl )
* Copyright (c) 2009-2010 Alex Converse <alex.converse@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
* AAC Spectral Band Replication decoding functions
* @author Robert Swain ( rob opendot cl )
*/
#define USE_FIXED 0
#include "aac.h"
#include "sbr.h"
#include "aacsbr.h"
#include "aacsbrdata.h"
#include "aacsbr_tablegen.h"
#include "fft.h"
#include "internal.h"
#include "aacps.h"
#include "sbrdsp.h"
#include "libavutil/internal.h"
#include "libavutil/libm.h"
#include "libavutil/avassert.h"
#include <stdint.h>
#include <float.h>
#include <math.h>
#if ARCH_MIPS
#include "mips/aacsbr_mips.h"
#endif /* ARCH_MIPS */
static VLC vlc_sbr[10];
static void aacsbr_func_ptr_init(AACSBRContext *c);
static void make_bands(int16_t* bands, int start, int stop, int num_bands)
{
int k, previous, present;
float base, prod;
base = powf((float)stop / start, 1.0f / num_bands);
prod = start;
previous = start;
for (k = 0; k < num_bands-1; k++) {
prod *= base;
present = lrintf(prod);
bands[k] = present - previous;
previous = present;
}
bands[num_bands-1] = stop - previous;
}
/// Dequantization and stereo decoding (14496-3 sp04 p203)
static void sbr_dequant(SpectralBandReplication *sbr, int id_aac)
{
int k, e;
int ch;
static const double exp2_tab[2] = {1, M_SQRT2};
if (id_aac == TYPE_CPE && sbr->bs_coupling) {
int pan_offset = sbr->data[0].bs_amp_res ? 12 : 24;
for (e = 1; e <= sbr->data[0].bs_num_env; e++) {
for (k = 0; k < sbr->n[sbr->data[0].bs_freq_res[e]]; k++) {
float temp1, temp2, fac;
if (sbr->data[0].bs_amp_res) {
temp1 = ff_exp2fi(sbr->data[0].env_facs_q[e][k] + 7);
temp2 = ff_exp2fi(pan_offset - sbr->data[1].env_facs_q[e][k]);
}
else {
temp1 = ff_exp2fi((sbr->data[0].env_facs_q[e][k]>>1) + 7) *
exp2_tab[sbr->data[0].env_facs_q[e][k] & 1];
temp2 = ff_exp2fi((pan_offset - sbr->data[1].env_facs_q[e][k])>>1) *
exp2_tab[(pan_offset - sbr->data[1].env_facs_q[e][k]) & 1];
}
if (temp1 > 1E20) {
av_log(NULL, AV_LOG_ERROR, "envelope scalefactor overflow in dequant\n");
temp1 = 1;
}
fac = temp1 / (1.0f + temp2);
sbr->data[0].env_facs[e][k] = fac;
sbr->data[1].env_facs[e][k] = fac * temp2;
}
}
for (e = 1; e <= sbr->data[0].bs_num_noise; e++) {
for (k = 0; k < sbr->n_q; k++) {
float temp1 = ff_exp2fi(NOISE_FLOOR_OFFSET - sbr->data[0].noise_facs_q[e][k] + 1);
float temp2 = ff_exp2fi(12 - sbr->data[1].noise_facs_q[e][k]);
float fac;
av_assert0(temp1 <= 1E20);
fac = temp1 / (1.0f + temp2);
sbr->data[0].noise_facs[e][k] = fac;
sbr->data[1].noise_facs[e][k] = fac * temp2;
}
}
} else { // SCE or one non-coupled CPE
for (ch = 0; ch < (id_aac == TYPE_CPE) + 1; ch++) {
for (e = 1; e <= sbr->data[ch].bs_num_env; e++)
for (k = 0; k < sbr->n[sbr->data[ch].bs_freq_res[e]]; k++){
if (sbr->data[ch].bs_amp_res)
sbr->data[ch].env_facs[e][k] = ff_exp2fi(sbr->data[ch].env_facs_q[e][k] + 6);
else
sbr->data[ch].env_facs[e][k] = ff_exp2fi((sbr->data[ch].env_facs_q[e][k]>>1) + 6)
* exp2_tab[sbr->data[ch].env_facs_q[e][k] & 1];
if (sbr->data[ch].env_facs[e][k] > 1E20) {
av_log(NULL, AV_LOG_ERROR, "envelope scalefactor overflow in dequant\n");
sbr->data[ch].env_facs[e][k] = 1;
}
}
for (e = 1; e <= sbr->data[ch].bs_num_noise; e++)
for (k = 0; k < sbr->n_q; k++)
sbr->data[ch].noise_facs[e][k] =
ff_exp2fi(NOISE_FLOOR_OFFSET - sbr->data[ch].noise_facs_q[e][k]);
}
}
}
/** High Frequency Generation (14496-3 sp04 p214+) and Inverse Filtering
* (14496-3 sp04 p214)
* Warning: This routine does not seem numerically stable.
*/
static void sbr_hf_inverse_filter(SBRDSPContext *dsp,
float (*alpha0)[2], float (*alpha1)[2],
const float X_low[32][40][2], int k0)
{
int k;
for (k = 0; k < k0; k++) {
LOCAL_ALIGNED_16(float, phi, [3], [2][2]);
float dk;
dsp->autocorrelate(X_low[k], phi);
dk = phi[2][1][0] * phi[1][0][0] -
(phi[1][1][0] * phi[1][1][0] + phi[1][1][1] * phi[1][1][1]) / 1.000001f;
if (!dk) {
alpha1[k][0] = 0;
alpha1[k][1] = 0;
} else {
float temp_real, temp_im;
temp_real = phi[0][0][0] * phi[1][1][0] -
phi[0][0][1] * phi[1][1][1] -
phi[0][1][0] * phi[1][0][0];
temp_im = phi[0][0][0] * phi[1][1][1] +
phi[0][0][1] * phi[1][1][0] -
phi[0][1][1] * phi[1][0][0];
alpha1[k][0] = temp_real / dk;
alpha1[k][1] = temp_im / dk;
}
if (!phi[1][0][0]) {
alpha0[k][0] = 0;
alpha0[k][1] = 0;
} else {
float temp_real, temp_im;
temp_real = phi[0][0][0] + alpha1[k][0] * phi[1][1][0] +
alpha1[k][1] * phi[1][1][1];
temp_im = phi[0][0][1] + alpha1[k][1] * phi[1][1][0] -
alpha1[k][0] * phi[1][1][1];
alpha0[k][0] = -temp_real / phi[1][0][0];
alpha0[k][1] = -temp_im / phi[1][0][0];
}
if (alpha1[k][0] * alpha1[k][0] + alpha1[k][1] * alpha1[k][1] >= 16.0f ||
alpha0[k][0] * alpha0[k][0] + alpha0[k][1] * alpha0[k][1] >= 16.0f) {
alpha1[k][0] = 0;
alpha1[k][1] = 0;
alpha0[k][0] = 0;
alpha0[k][1] = 0;
}
}
}
/// Chirp Factors (14496-3 sp04 p214)
static void sbr_chirp(SpectralBandReplication *sbr, SBRData *ch_data)
{
int i;
float new_bw;
static const float bw_tab[] = { 0.0f, 0.75f, 0.9f, 0.98f };
for (i = 0; i < sbr->n_q; i++) {
if (ch_data->bs_invf_mode[0][i] + ch_data->bs_invf_mode[1][i] == 1) {
new_bw = 0.6f;
} else
new_bw = bw_tab[ch_data->bs_invf_mode[0][i]];
if (new_bw < ch_data->bw_array[i]) {
new_bw = 0.75f * new_bw + 0.25f * ch_data->bw_array[i];
} else
new_bw = 0.90625f * new_bw + 0.09375f * ch_data->bw_array[i];
ch_data->bw_array[i] = new_bw < 0.015625f ? 0.0f : new_bw;
}
}
/**
* Calculation of levels of additional HF signal components (14496-3 sp04 p219)
* and Calculation of gain (14496-3 sp04 p219)
*/
static void sbr_gain_calc(AACContext *ac, SpectralBandReplication *sbr,
SBRData *ch_data, const int e_a[2])
{
int e, k, m;
// max gain limits : -3dB, 0dB, 3dB, inf dB (limiter off)
static const float limgain[4] = { 0.70795, 1.0, 1.41254, 10000000000 };
for (e = 0; e < ch_data->bs_num_env; e++) {
int delta = !((e == e_a[1]) || (e == e_a[0]));
for (k = 0; k < sbr->n_lim; k++) {
float gain_boost, gain_max;
float sum[2] = { 0.0f, 0.0f };
for (m = sbr->f_tablelim[k] - sbr->kx[1]; m < sbr->f_tablelim[k + 1] - sbr->kx[1]; m++) {
const float temp = sbr->e_origmapped[e][m] / (1.0f + sbr->q_mapped[e][m]);
sbr->q_m[e][m] = sqrtf(temp * sbr->q_mapped[e][m]);
sbr->s_m[e][m] = sqrtf(temp * ch_data->s_indexmapped[e + 1][m]);
if (!sbr->s_mapped[e][m]) {
sbr->gain[e][m] = sqrtf(sbr->e_origmapped[e][m] /
((1.0f + sbr->e_curr[e][m]) *
(1.0f + sbr->q_mapped[e][m] * delta)));
} else {
sbr->gain[e][m] = sqrtf(sbr->e_origmapped[e][m] * sbr->q_mapped[e][m] /
((1.0f + sbr->e_curr[e][m]) *
(1.0f + sbr->q_mapped[e][m])));
}
}
for (m = sbr->f_tablelim[k] - sbr->kx[1]; m < sbr->f_tablelim[k + 1] - sbr->kx[1]; m++) {
sum[0] += sbr->e_origmapped[e][m];
sum[1] += sbr->e_curr[e][m];
}
gain_max = limgain[sbr->bs_limiter_gains] * sqrtf((FLT_EPSILON + sum[0]) / (FLT_EPSILON + sum[1]));
gain_max = FFMIN(100000.f, gain_max);
for (m = sbr->f_tablelim[k] - sbr->kx[1]; m < sbr->f_tablelim[k + 1] - sbr->kx[1]; m++) {
float q_m_max = sbr->q_m[e][m] * gain_max / sbr->gain[e][m];
sbr->q_m[e][m] = FFMIN(sbr->q_m[e][m], q_m_max);
sbr->gain[e][m] = FFMIN(sbr->gain[e][m], gain_max);
}
sum[0] = sum[1] = 0.0f;
for (m = sbr->f_tablelim[k] - sbr->kx[1]; m < sbr->f_tablelim[k + 1] - sbr->kx[1]; m++) {
sum[0] += sbr->e_origmapped[e][m];
sum[1] += sbr->e_curr[e][m] * sbr->gain[e][m] * sbr->gain[e][m]
+ sbr->s_m[e][m] * sbr->s_m[e][m]
+ (delta && !sbr->s_m[e][m]) * sbr->q_m[e][m] * sbr->q_m[e][m];
}
gain_boost = sqrtf((FLT_EPSILON + sum[0]) / (FLT_EPSILON + sum[1]));
gain_boost = FFMIN(1.584893192f, gain_boost);
for (m = sbr->f_tablelim[k] - sbr->kx[1]; m < sbr->f_tablelim[k + 1] - sbr->kx[1]; m++) {
sbr->gain[e][m] *= gain_boost;
sbr->q_m[e][m] *= gain_boost;
sbr->s_m[e][m] *= gain_boost;
}
}
}
}
/// Assembling HF Signals (14496-3 sp04 p220)
static void sbr_hf_assemble(float Y1[38][64][2],
const float X_high[64][40][2],
SpectralBandReplication *sbr, SBRData *ch_data,
const int e_a[2])
{
int e, i, j, m;
const int h_SL = 4 * !sbr->bs_smoothing_mode;
const int kx = sbr->kx[1];
const int m_max = sbr->m[1];
static const float h_smooth[5] = {
0.33333333333333,
0.30150283239582,
0.21816949906249,
0.11516383427084,
0.03183050093751,
};
float (*g_temp)[48] = ch_data->g_temp, (*q_temp)[48] = ch_data->q_temp;
int indexnoise = ch_data->f_indexnoise;
int indexsine = ch_data->f_indexsine;
if (sbr->reset) {
for (i = 0; i < h_SL; i++) {
memcpy(g_temp[i + 2*ch_data->t_env[0]], sbr->gain[0], m_max * sizeof(sbr->gain[0][0]));
memcpy(q_temp[i + 2*ch_data->t_env[0]], sbr->q_m[0], m_max * sizeof(sbr->q_m[0][0]));
}
} else if (h_SL) {
for (i = 0; i < 4; i++) {
memcpy(g_temp[i + 2 * ch_data->t_env[0]],
g_temp[i + 2 * ch_data->t_env_num_env_old],
sizeof(g_temp[0]));
memcpy(q_temp[i + 2 * ch_data->t_env[0]],
q_temp[i + 2 * ch_data->t_env_num_env_old],
sizeof(q_temp[0]));
}
}
for (e = 0; e < ch_data->bs_num_env; e++) {
for (i = 2 * ch_data->t_env[e]; i < 2 * ch_data->t_env[e + 1]; i++) {
memcpy(g_temp[h_SL + i], sbr->gain[e], m_max * sizeof(sbr->gain[0][0]));
memcpy(q_temp[h_SL + i], sbr->q_m[e], m_max * sizeof(sbr->q_m[0][0]));
}
}
for (e = 0; e < ch_data->bs_num_env; e++) {
for (i = 2 * ch_data->t_env[e]; i < 2 * ch_data->t_env[e + 1]; i++) {
LOCAL_ALIGNED_16(float, g_filt_tab, [48]);
LOCAL_ALIGNED_16(float, q_filt_tab, [48]);
float *g_filt, *q_filt;
if (h_SL && e != e_a[0] && e != e_a[1]) {
g_filt = g_filt_tab;
q_filt = q_filt_tab;
for (m = 0; m < m_max; m++) {
const int idx1 = i + h_SL;
g_filt[m] = 0.0f;
q_filt[m] = 0.0f;
for (j = 0; j <= h_SL; j++) {
g_filt[m] += g_temp[idx1 - j][m] * h_smooth[j];
q_filt[m] += q_temp[idx1 - j][m] * h_smooth[j];
}
}
} else {
g_filt = g_temp[i + h_SL];
q_filt = q_temp[i];
}
sbr->dsp.hf_g_filt(Y1[i] + kx, X_high + kx, g_filt, m_max,
i + ENVELOPE_ADJUSTMENT_OFFSET);
if (e != e_a[0] && e != e_a[1]) {
sbr->dsp.hf_apply_noise[indexsine](Y1[i] + kx, sbr->s_m[e],
q_filt, indexnoise,
kx, m_max);
} else {
int idx = indexsine&1;
int A = (1-((indexsine+(kx & 1))&2));
int B = (A^(-idx)) + idx;
float *out = &Y1[i][kx][idx];
float *in = sbr->s_m[e];
for (m = 0; m+1 < m_max; m+=2) {
out[2*m ] += in[m ] * A;
out[2*m+2] += in[m+1] * B;
}
if(m_max&1)
out[2*m ] += in[m ] * A;
}
indexnoise = (indexnoise + m_max) & 0x1ff;
indexsine = (indexsine + 1) & 3;
}
}
ch_data->f_indexnoise = indexnoise;
ch_data->f_indexsine = indexsine;
}
#include "aacsbr_template.c"
Reference in New Issue View Git Blame Copy Permalink