ffmpeg/libavcodec/aacenc_utils.h

267 lines
8.1 KiB
C
Raw Normal View History

/*
* AAC encoder utilities
* Copyright (C) 2015 Rostislav Pehlivanov
*
* 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 encoder utilities
* @author Rostislav Pehlivanov ( atomnuker gmail com )
*/
#ifndef AVCODEC_AACENC_UTILS_H
#define AVCODEC_AACENC_UTILS_H
#include "libavutil/ffmath.h"
#include "aac.h"
#include "aacenctab.h"
#include "aactab.h"
#define ROUND_STANDARD 0.4054f
#define ROUND_TO_ZERO 0.1054f
#define C_QUANT 0.4054f
static inline void abs_pow34_v(float *out, const float *in, const int size)
{
int i;
for (i = 0; i < size; i++) {
float a = fabsf(in[i]);
out[i] = sqrtf(a * sqrtf(a));
}
}
static inline float pos_pow34(float a)
{
return sqrtf(a * sqrtf(a));
}
/**
* Quantize one coefficient.
* @return absolute value of the quantized coefficient
* @see 3GPP TS26.403 5.6.2 "Scalefactor determination"
*/
static inline int quant(float coef, const float Q, const float rounding)
{
float a = coef * Q;
return sqrtf(a * sqrtf(a)) + rounding;
}
static inline void quantize_bands(int *out, const float *in, const float *scaled,
int size, float Q34, int is_signed, int maxval,
const float rounding)
{
int i;
for (i = 0; i < size; i++) {
float qc = scaled[i] * Q34;
int tmp = (int)FFMIN(qc + rounding, (float)maxval);
if (is_signed && in[i] < 0.0f) {
tmp = -tmp;
}
out[i] = tmp;
}
}
static inline float find_max_val(int group_len, int swb_size, const float *scaled)
{
float maxval = 0.0f;
int w2, i;
for (w2 = 0; w2 < group_len; w2++) {
for (i = 0; i < swb_size; i++) {
maxval = FFMAX(maxval, scaled[w2*128+i]);
}
}
return maxval;
}
static inline int find_min_book(float maxval, int sf)
{
float Q34 = ff_aac_pow34sf_tab[POW_SF2_ZERO - sf + SCALE_ONE_POS - SCALE_DIV_512];
int qmaxval, cb;
qmaxval = maxval * Q34 + C_QUANT;
if (qmaxval >= (FF_ARRAY_ELEMS(aac_maxval_cb)))
cb = 11;
else
cb = aac_maxval_cb[qmaxval];
return cb;
}
static inline float find_form_factor(int group_len, int swb_size, float thresh,
const float *scaled, float nzslope) {
AAC encoder: Extensive improvements This finalizes merging of the work in the patches in ticket #2686. Improvements to twoloop and RC logic are extensive. The non-exhaustive list of twoloop improvments includes: - Tweaks to distortion limits on the RD optimization phase of twoloop - Deeper search in twoloop - PNS information marking to let twoloop decide when to use it (turned out having the decision made separately wasn't working) - Tonal band detection and priorization - Better band energy conservation rules - Strict hole avoidance For rate control: - Use psymodel's bit allocation to allow proper use of the bit reservoir. Don't work against the bit reservoir by moving lambda in the opposite direction when psymodel decides to allocate more/less bits to a frame. - Retry the encode if the effective rate lies outside a reasonable margin of psymodel's allocation or the selected ABR. - Log average lambda at the end. Useful info for everyone, but especially for tuning of the various encoder constants that relate to lambda feedback. Psy: - Do not apply lowpass with a FIR filter, instead just let the coder zero bands above the cutoff. The FIR filter induces group delay, and while zeroing bands causes ripple, it's lost in the quantization noise. - Experimental VBR bit allocation code - Tweak automatic lowpass filter threshold to maximize audio bandwidth at all bitrates while still providing acceptable, stable quality. I/S: - Phase decision fixes. Unrelated to #2686, but the bugs only surfaced when the merge was finalized. Measure I/S band energy accounting for phase, and prevent I/S and M/S from being applied both. PNS: - Avoid marking short bands with PNS when they're part of a window group in which there's a large variation of energy from one window to the next. PNS can't preserve those and the effect is extremely noticeable. M/S: - Implement BMLD protection similar to the specified in ISO-IEC/13818:7-2003, Appendix C Section 6.1. Since M/S decision doesn't conform to section 6.1, a different method had to be implemented, but should provide equivalent protection. - Move the decision logic closer to the method specified in ISO-IEC/13818:7-2003, Appendix C Section 6.1. Specifically, make sure M/S needs less bits than dual stereo. - Don't apply M/S in bands that are using I/S Now, this of course needed adjustments in the compare targets and fuzz factors of the AAC encoder's fate tests, but if wondering why the targets go up (more distortion), consider the previous coder was using too many bits on LF content (far more than required by psy), and thus those signals will now be more distorted, not less. The extra distortion isn't audible though, I carried extensive ABX testing to make sure. A very similar patch was also extensively tested by Kamendo2 in the context of #2686.
2015-10-11 20:29:50 +00:00
const float iswb_size = 1.0f / swb_size;
const float iswb_sizem1 = 1.0f / (swb_size - 1);
const float ethresh = thresh;
float form = 0.0f, weight = 0.0f;
int w2, i;
for (w2 = 0; w2 < group_len; w2++) {
float e = 0.0f, e2 = 0.0f, var = 0.0f, maxval = 0.0f;
float nzl = 0;
for (i = 0; i < swb_size; i++) {
float s = fabsf(scaled[w2*128+i]);
maxval = FFMAX(maxval, s);
e += s;
e2 += s *= s;
/* We really don't want a hard non-zero-line count, since
* even below-threshold lines do add up towards band spectral power.
* So, fall steeply towards zero, but smoothly
*/
if (s >= ethresh) {
nzl += 1.0f;
} else {
if (nzslope == 2.f)
nzl += (s / ethresh) * (s / ethresh);
else
nzl += ff_fast_powf(s / ethresh, nzslope);
AAC encoder: Extensive improvements This finalizes merging of the work in the patches in ticket #2686. Improvements to twoloop and RC logic are extensive. The non-exhaustive list of twoloop improvments includes: - Tweaks to distortion limits on the RD optimization phase of twoloop - Deeper search in twoloop - PNS information marking to let twoloop decide when to use it (turned out having the decision made separately wasn't working) - Tonal band detection and priorization - Better band energy conservation rules - Strict hole avoidance For rate control: - Use psymodel's bit allocation to allow proper use of the bit reservoir. Don't work against the bit reservoir by moving lambda in the opposite direction when psymodel decides to allocate more/less bits to a frame. - Retry the encode if the effective rate lies outside a reasonable margin of psymodel's allocation or the selected ABR. - Log average lambda at the end. Useful info for everyone, but especially for tuning of the various encoder constants that relate to lambda feedback. Psy: - Do not apply lowpass with a FIR filter, instead just let the coder zero bands above the cutoff. The FIR filter induces group delay, and while zeroing bands causes ripple, it's lost in the quantization noise. - Experimental VBR bit allocation code - Tweak automatic lowpass filter threshold to maximize audio bandwidth at all bitrates while still providing acceptable, stable quality. I/S: - Phase decision fixes. Unrelated to #2686, but the bugs only surfaced when the merge was finalized. Measure I/S band energy accounting for phase, and prevent I/S and M/S from being applied both. PNS: - Avoid marking short bands with PNS when they're part of a window group in which there's a large variation of energy from one window to the next. PNS can't preserve those and the effect is extremely noticeable. M/S: - Implement BMLD protection similar to the specified in ISO-IEC/13818:7-2003, Appendix C Section 6.1. Since M/S decision doesn't conform to section 6.1, a different method had to be implemented, but should provide equivalent protection. - Move the decision logic closer to the method specified in ISO-IEC/13818:7-2003, Appendix C Section 6.1. Specifically, make sure M/S needs less bits than dual stereo. - Don't apply M/S in bands that are using I/S Now, this of course needed adjustments in the compare targets and fuzz factors of the AAC encoder's fate tests, but if wondering why the targets go up (more distortion), consider the previous coder was using too many bits on LF content (far more than required by psy), and thus those signals will now be more distorted, not less. The extra distortion isn't audible though, I carried extensive ABX testing to make sure. A very similar patch was also extensively tested by Kamendo2 in the context of #2686.
2015-10-11 20:29:50 +00:00
}
}
if (e2 > thresh) {
float frm;
e *= iswb_size;
/** compute variance */
for (i = 0; i < swb_size; i++) {
float d = fabsf(scaled[w2*128+i]) - e;
var += d*d;
}
var = sqrtf(var * iswb_sizem1);
e2 *= iswb_size;
frm = e / FFMIN(e+4*var,maxval);
form += e2 * sqrtf(frm) / FFMAX(0.5f,nzl);
weight += e2;
}
}
if (weight > 0) {
return form / weight;
} else {
return 1.0f;
}
}
/** Return the minimum scalefactor where the quantized coef does not clip. */
static inline uint8_t coef2minsf(float coef)
{
return av_clip_uint8(log2f(coef)*4 - 69 + SCALE_ONE_POS - SCALE_DIV_512);
}
/** Return the maximum scalefactor where the quantized coef is not zero. */
static inline uint8_t coef2maxsf(float coef)
{
return av_clip_uint8(log2f(coef)*4 + 6 + SCALE_ONE_POS - SCALE_DIV_512);
}
/*
* Returns the closest possible index to an array of float values, given a value.
*/
static inline int quant_array_idx(const float val, const float *arr, const int num)
{
int i, index = 0;
float quant_min_err = INFINITY;
for (i = 0; i < num; i++) {
float error = (val - arr[i])*(val - arr[i]);
if (error < quant_min_err) {
quant_min_err = error;
index = i;
}
}
return index;
}
AAC encoder: Extensive improvements This finalizes merging of the work in the patches in ticket #2686. Improvements to twoloop and RC logic are extensive. The non-exhaustive list of twoloop improvments includes: - Tweaks to distortion limits on the RD optimization phase of twoloop - Deeper search in twoloop - PNS information marking to let twoloop decide when to use it (turned out having the decision made separately wasn't working) - Tonal band detection and priorization - Better band energy conservation rules - Strict hole avoidance For rate control: - Use psymodel's bit allocation to allow proper use of the bit reservoir. Don't work against the bit reservoir by moving lambda in the opposite direction when psymodel decides to allocate more/less bits to a frame. - Retry the encode if the effective rate lies outside a reasonable margin of psymodel's allocation or the selected ABR. - Log average lambda at the end. Useful info for everyone, but especially for tuning of the various encoder constants that relate to lambda feedback. Psy: - Do not apply lowpass with a FIR filter, instead just let the coder zero bands above the cutoff. The FIR filter induces group delay, and while zeroing bands causes ripple, it's lost in the quantization noise. - Experimental VBR bit allocation code - Tweak automatic lowpass filter threshold to maximize audio bandwidth at all bitrates while still providing acceptable, stable quality. I/S: - Phase decision fixes. Unrelated to #2686, but the bugs only surfaced when the merge was finalized. Measure I/S band energy accounting for phase, and prevent I/S and M/S from being applied both. PNS: - Avoid marking short bands with PNS when they're part of a window group in which there's a large variation of energy from one window to the next. PNS can't preserve those and the effect is extremely noticeable. M/S: - Implement BMLD protection similar to the specified in ISO-IEC/13818:7-2003, Appendix C Section 6.1. Since M/S decision doesn't conform to section 6.1, a different method had to be implemented, but should provide equivalent protection. - Move the decision logic closer to the method specified in ISO-IEC/13818:7-2003, Appendix C Section 6.1. Specifically, make sure M/S needs less bits than dual stereo. - Don't apply M/S in bands that are using I/S Now, this of course needed adjustments in the compare targets and fuzz factors of the AAC encoder's fate tests, but if wondering why the targets go up (more distortion), consider the previous coder was using too many bits on LF content (far more than required by psy), and thus those signals will now be more distorted, not less. The extra distortion isn't audible though, I carried extensive ABX testing to make sure. A very similar patch was also extensively tested by Kamendo2 in the context of #2686.
2015-10-11 20:29:50 +00:00
/**
* approximates exp10f(-3.0f*(0.5f + 0.5f * cosf(FFMIN(b,15.5f) / 15.5f)))
*/
static av_always_inline float bval2bmax(float b)
{
return 0.001f + 0.0035f * (b*b*b) / (15.5f*15.5f*15.5f);
}
AAC encoder: improve SF range utilization This patch does 4 things, all of which interact and thus it woudln't be possible to commit them separately without causing either quality regressions or assertion failures. Fate comparison targets don't all reflect improvements in quality, yet listening tests show substantially improved quality and stability. 1. Increase SF range utilization. The spec requires SF delta values to be constrained within the range -60..60. The previous code was applying that range to the whole SF array and not only the deltas of consecutive values, because doing so requires smarter code: zeroing or otherwise skipping a band may invalidate lots of SF choices. This patch implements that logic to allow the coders to utilize the full dynamic range of scalefactors, increasing quality quite considerably, and fixing delta-SF-related assertion failures, since now the limitation is enforced rather than asserted. 2. PNS tweaks The previous modification makes big improvements in twoloop's efficiency, and every time that happens PNS logic needs to be tweaked accordingly to avoid it from stepping all over twoloop's decisions. This patch includes modifications of the sort. 3. Account for lowpass cutoff during PSY analysis The closer PSY's allocation is to final allocation the better the quality is, and given these modifications, twoloop is now very efficient at avoiding holes. Thus, to compute accurate thresholds, PSY needs to account for the lowpass applied implicitly during twoloop (by zeroing high bands). This patch makes twoloop set the cutoff in psymodel's context the first time it runs, and makes PSY account for it during threshold computation, making PE and threshold computations closer to the final allocation and thus achieving better subjective quality. 4. Tweaks to RC lambda tracking loop in relation to PNS Without this tweak some corner cases cause quality regressions. Basically, lambda needs to react faster to overall bitrate efficiency changes since now PNS can be quite successful in enforcing maximum bitrates, when PSY allocates too many bits to the lower bands, suppressing the signals RC logic uses to lower lambda in those cases and causing aggressive PNS. This tweak makes PNS much less aggressive, though it can still use some further tweaks. Also update MIPS specializations and adjust fuzz Also in lavc/mips/aacpsy_mips.h: remove trailing whitespace
2015-12-01 06:28:36 +00:00
/*
* Compute a nextband map to be used with SF delta constraint utilities.
* The nextband array should contain 128 elements, and positions that don't
* map to valid, nonzero bands of the form w*16+g (with w being the initial
* window of the window group, only) are left indetermined.
*/
static inline void ff_init_nextband_map(const SingleChannelElement *sce, uint8_t *nextband)
{
unsigned char prevband = 0;
int w, g;
/** Just a safe default */
for (g = 0; g < 128; g++)
nextband[g] = g;
/** Now really navigate the nonzero band chain */
for (w = 0; w < sce->ics.num_windows; w += sce->ics.group_len[w]) {
for (g = 0; g < sce->ics.num_swb; g++) {
if (!sce->zeroes[w*16+g] && sce->band_type[w*16+g] < RESERVED_BT)
prevband = nextband[prevband] = w*16+g;
}
}
nextband[prevband] = prevband; /* terminate */
}
/*
* Updates nextband to reflect a removed band (equivalent to
* calling ff_init_nextband_map after marking a band as zero)
*/
static inline void ff_nextband_remove(uint8_t *nextband, int prevband, int band)
{
nextband[prevband] = nextband[band];
}
/*
* Checks whether the specified band could be removed without inducing
* scalefactor delta that violates SF delta encoding constraints.
* prev_sf has to be the scalefactor of the previous nonzero, nonspecial
* band, in encoding order, or negative if there was no such band.
*/
static inline int ff_sfdelta_can_remove_band(const SingleChannelElement *sce,
const uint8_t *nextband, int prev_sf, int band)
{
return prev_sf >= 0
&& sce->sf_idx[nextband[band]] >= (prev_sf - SCALE_MAX_DIFF)
&& sce->sf_idx[nextband[band]] <= (prev_sf + SCALE_MAX_DIFF);
}
/*
* Checks whether the specified band's scalefactor could be replaced
* with another one without violating SF delta encoding constraints.
* prev_sf has to be the scalefactor of the previous nonzero, nonsepcial
* band, in encoding order, or negative if there was no such band.
*/
static inline int ff_sfdelta_can_replace(const SingleChannelElement *sce,
const uint8_t *nextband, int prev_sf, int new_sf, int band)
{
return new_sf >= (prev_sf - SCALE_MAX_DIFF)
&& new_sf <= (prev_sf + SCALE_MAX_DIFF)
&& sce->sf_idx[nextband[band]] >= (new_sf - SCALE_MAX_DIFF)
&& sce->sf_idx[nextband[band]] <= (new_sf + SCALE_MAX_DIFF);
}
#define ERROR_IF(cond, ...) \
if (cond) { \
av_log(avctx, AV_LOG_ERROR, __VA_ARGS__); \
return AVERROR(EINVAL); \
}
#define WARN_IF(cond, ...) \
if (cond) { \
av_log(avctx, AV_LOG_WARNING, __VA_ARGS__); \
}
#endif /* AVCODEC_AACENC_UTILS_H */