
commit a2cc77cb2744a2cb0551b9bfdf06b97457b6d449 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Thu Nov 4 13:11:21 2010 -0400 Adding a switch to enable the post-filter (off by default) commit 8e860dc0dfbe57e59fcbd5352588c5edff020e27 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Thu Nov 4 11:57:12 2010 -0400 Allowing pitches up to 3000 Hz commit 837412d37bbca32bb34bfb5941e132ff4b0a568c Author: Jean-Marc Valin <jean-marc.valin@usherbrooke.ca> Date: Wed Nov 3 20:47:11 2010 -0400 Pitch estimation tuning to prevent some cases of pitch halving commit 34e20f24c85b40fffd1a15c5b632f2f78b26f081 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Wed Nov 3 16:31:51 2010 -0400 Resynthesis now purely a compile-time option with RESYNTH commit d83fb5a9cc2ec4b6cce938662997643da1c5ed0d Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Wed Nov 3 16:28:25 2010 -0400 Fixes a divide by zero in remove_doubling() commit bb91e05b7f8f91fd15a8a0daae3d8cb6bd8d81db Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Wed Nov 3 15:55:48 2010 -0400 Bring back resynthesis with RESYNTH macro commit 31fe6f6b4997af0a46b8c62f523fe2dfdb7f56ae Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Tue Nov 2 17:55:04 2010 -0400 Tuning the allocation tilt to give more bits to higher frequencies. Especially useful now that the post-filter can reduce low freq noise. commit 919ba48f0369a87885334756cdfac2a448ce52d0 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Mon Nov 1 17:27:19 2010 -0400 C89 fix commit ee0dbb1855a82ee8c132ddaffcab4d072bb3455e Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Mon Nov 1 11:45:10 2010 -0400 Complete fixed-point port of the pitch code (I think). commit 4c7b3fd12a8f7469607b5ac57c85301a5de9fa81 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Mon Nov 1 10:55:43 2010 -0400 More fixed-point pitch gain work commit 26f1412188900199b63e187fcb0bd04db53c898a Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Mon Nov 1 10:39:25 2010 -0400 Fixed-point version of the pitch gain calculation code commit 27c73d008e9f50d282c3ad08e2f05f7006013ae1 Author: Jean-Marc Valin <jean-marc.valin@usherbrooke.ca> Date: Sun Oct 31 16:50:26 2010 -0400 Some more fixed-point work in remove_doubling() commit 59354672cb3af794a0e46c0b2097d6441c75cdd1 Author: Jean-Marc Valin <jean-marc.valin@usherbrooke.ca> Date: Sun Oct 31 09:57:36 2010 -0400 Fixed a stupid fixed-point pf bug in the gain handling commit be9e7dabf6c8b32bc049da260b58ff6085dc1ac3 Author: Jean-Marc Valin <jean-marc.valin@usherbrooke.ca> Date: Sat Oct 30 01:52:09 2010 -0400 Fixed-point: fixed frac_div32() that was broken a few commits ago. commit 5b06270afc41a88915252cea14411be43650e704 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Fri Oct 29 17:45:44 2010 -0400 This fixes VBR when encoding the pitch period with raw bits commit 10e0488458ae558aa80d0b30cce70841ad081f73 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Fri Oct 29 16:50:31 2010 -0400 Pitch period is now encoder with equal probability for each octave (rather than each lag). Max pitch gain allowed is now 0.625. commit ca19396c1c1511c0e208b400efb51384fc7c200d Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Fri Oct 29 16:00:01 2010 -0400 More fixed-point post-filter work commit f3e42fde1b575bc587b2557b8b31a6085421a99c Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Fri Oct 29 14:39:23 2010 -0400 More fixed-point work for the prefilter/postfilter commit db945132d12b25ff25acc0701b91a1d8a81417d5 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Fri Oct 29 14:14:02 2010 -0400 Making the pitch estimation work in fixed-point Even if there's still lots of float operations left. commit acb3f96e04802ac4601295f83bef1f32593e261a Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Fri Oct 29 10:57:39 2010 -0400 Making the PLC code consistent with the prefilter/postfilter commit 8f64f5974ac846b8c35d0b692e0472f279206cf0 Author: Jean-Marc Valin <jean-marc.valin@usherbrooke.ca> Date: Thu Oct 28 00:33:53 2010 -0400 More tuning for remove_doubling() commit 0c08f2ee9dcc135dd222fef30f5ad93e95e0d364 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Wed Oct 27 17:48:02 2010 -0400 Doing an interpolation step to improve the accuracy of the pitch estimate Also increasing the gain slightly. commit 23d303e992f1fdc3d2668652603ae6311d3b91c5 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Wed Oct 27 16:56:42 2010 -0400 Implements a fixed 3-tap prefilter/postfilter to make the gain roll off with frequency commit 881c5928adc1af9eb75c4b68e9eba94ab1d65adc Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Wed Oct 27 14:47:30 2010 -0400 Partially whitening the down-sampled signal before the pitch search commit 4a8687deea8587007f14051cb966f6fd748893a1 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Wed Oct 27 14:27:47 2010 -0400 pitch_search() no longer computes the gain commit a7f85bb6b10d9c509caec521ca444efb3f27df05 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Wed Oct 27 14:00:53 2010 -0400 remove_doubling() now works on the down-sampled signal commit 06cb70e876873f79fed214ebbca35cb4c5057ec8 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Wed Oct 27 11:28:53 2010 -0400 Simplification to the pitch continuity code commit 5201927c284a424eb8f21f63d358844b3de8c285 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Wed Oct 27 11:04:02 2010 -0400 Some more pitch doubling prevention code commit 7ef63fbe1f78f79e1923bc42e06fbdf1ec28ffd3 Author: Jean-Marc Valin <jean-marc.valin@usherbrooke.ca> Date: Wed Oct 27 06:49:28 2010 -0400 Minor fix commit eb37eaab32e7df074a7ddf0ae4781e57f827c4ad Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Tue Oct 26 18:32:25 2010 -0400 Enforcing some pitch continuity commit 751ef6edf2ee7721252cedb264bdf9b3f6244a9d Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Tue Oct 26 17:29:47 2010 -0400 Code for preventing pitch doubling/halving commit c12647ecb55b645005efbeede91880db72936f8d Author: Jean-Marc Valin <jean-marc.valin@usherbrooke.ca> Date: Tue Oct 26 00:04:34 2010 -0400 Finally getting perfect reconstruction when pitch changes Post-filter now delays the filter coefs by the overlap so that the pre-filter and post-filter are synchronised. commit f854311d945bb375039a4a4a4fea782b648581f8 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Mon Oct 25 14:59:13 2010 -0400 Very simple/inefficient signalling of the prefilter period/gain commit b4e1215432e3d89a29c998639a6d8b07e28c5a2a Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Mon Oct 25 14:09:17 2010 -0400 using the actual pitch gain commit e7cd4f07bb073b6955a001e56c0bbf16156f4195 Author: Jean-Marc Valin <jean-marc.valin@octasic.com> Date: Mon Oct 25 12:16:11 2010 -0400 Adding some pitch prediction though side information still isn't coded commit 77a03aa27c9b6ed2fe80c27a1196b460ccb5079e Author: Jean-Marc Valin <jean-marc.valin@usherbrooke.ca> Date: Mon Oct 25 00:12:48 2010 -0400 prefilter implemented as well commit a3fd81b6ca213d4a9f8ddfa2883fd0e238d64d04 Author: Jean-Marc Valin <jean-marc.valin@usherbrooke.ca> Date: Sun Oct 24 01:14:10 2010 -0400 Implementing Raymond Chen's comb filter idea So far, only the post-filter is there.
334 lines
9.6 KiB
C
334 lines
9.6 KiB
C
/* Copyright (c) 2007-2008 CSIRO
|
|
Copyright (c) 2007-2009 Xiph.Org Foundation
|
|
Written by Jean-Marc Valin */
|
|
/*
|
|
Redistribution and use in source and binary forms, with or without
|
|
modification, are permitted provided that the following conditions
|
|
are met:
|
|
|
|
- Redistributions of source code must retain the above copyright
|
|
notice, this list of conditions and the following disclaimer.
|
|
|
|
- Redistributions in binary form must reproduce the above copyright
|
|
notice, this list of conditions and the following disclaimer in the
|
|
documentation and/or other materials provided with the distribution.
|
|
|
|
- Neither the name of the Xiph.org Foundation nor the names of its
|
|
contributors may be used to endorse or promote products derived from
|
|
this software without specific prior written permission.
|
|
|
|
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
|
|
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
|
|
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
|
|
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR
|
|
CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
|
|
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
|
|
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
|
|
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
|
|
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
|
|
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
|
|
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
*/
|
|
|
|
#ifdef HAVE_CONFIG_H
|
|
#include "config.h"
|
|
#endif
|
|
|
|
#include <math.h>
|
|
#include "modes.h"
|
|
#include "cwrs.h"
|
|
#include "arch.h"
|
|
#include "os_support.h"
|
|
|
|
#include "entcode.h"
|
|
#include "rate.h"
|
|
|
|
|
|
#ifndef STATIC_MODES
|
|
|
|
/*Determines if V(N,K) fits in a 32-bit unsigned integer.
|
|
N and K are themselves limited to 15 bits.*/
|
|
static int fits_in32(int _n, int _k)
|
|
{
|
|
static const celt_int16 maxN[15] = {
|
|
32767, 32767, 32767, 1476, 283, 109, 60, 40,
|
|
29, 24, 20, 18, 16, 14, 13};
|
|
static const celt_int16 maxK[15] = {
|
|
32767, 32767, 32767, 32767, 1172, 238, 95, 53,
|
|
36, 27, 22, 18, 16, 15, 13};
|
|
if (_n>=14)
|
|
{
|
|
if (_k>=14)
|
|
return 0;
|
|
else
|
|
return _n <= maxN[_k];
|
|
} else {
|
|
return _k <= maxK[_n];
|
|
}
|
|
}
|
|
|
|
void compute_pulse_cache(CELTMode *m, int LM)
|
|
{
|
|
int i;
|
|
int curr=0;
|
|
int nbEntries=0;
|
|
int entryN[100], entryK[100], entryI[100];
|
|
const celt_int16 *eBands = m->eBands;
|
|
PulseCache *cache = &m->cache;
|
|
celt_int16 *cindex;
|
|
unsigned char *bits;
|
|
|
|
cindex = celt_alloc(sizeof(cache->index[0])*m->nbEBands*(LM+2));
|
|
cache->index = cindex;
|
|
|
|
/* Scan for all unique band sizes */
|
|
for (i=0;i<=LM+1;i++)
|
|
{
|
|
int j;
|
|
for (j=0;j<m->nbEBands;j++)
|
|
{
|
|
int k;
|
|
int N = (eBands[j+1]-eBands[j])<<i>>1;
|
|
cindex[i*m->nbEBands+j] = -1;
|
|
/* Find other bands that have the same size */
|
|
for (k=0;k<=i;k++)
|
|
{
|
|
int n;
|
|
for (n=0;n<m->nbEBands && (k!=i || n<j);n++)
|
|
{
|
|
if (N == (eBands[n+1]-eBands[n])<<k>>1)
|
|
{
|
|
cindex[i*m->nbEBands+j] = cindex[k*m->nbEBands+n];
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
if (cache->index[i*m->nbEBands+j] == -1 && N!=0)
|
|
{
|
|
int K;
|
|
entryN[nbEntries] = N;
|
|
K = 0;
|
|
while (fits_in32(N,get_pulses(K+1)) && K<MAX_PSEUDO)
|
|
K++;
|
|
entryK[nbEntries] = K;
|
|
cindex[i*m->nbEBands+j] = curr;
|
|
entryI[nbEntries] = curr;
|
|
|
|
curr += K+1;
|
|
nbEntries++;
|
|
}
|
|
}
|
|
}
|
|
bits = celt_alloc(sizeof(unsigned char)*curr);
|
|
cache->bits = bits;
|
|
cache->size = curr;
|
|
/* Compute the cache for all unique sizes */
|
|
for (i=0;i<nbEntries;i++)
|
|
{
|
|
int j;
|
|
unsigned char *ptr = bits+entryI[i];
|
|
celt_int16 tmp[MAX_PULSES+1];
|
|
get_required_bits(tmp, entryN[i], get_pulses(entryK[i]), BITRES);
|
|
for (j=1;j<=entryK[i];j++)
|
|
ptr[j] = tmp[get_pulses(j)]-1;
|
|
ptr[0] = entryK[i];
|
|
}
|
|
}
|
|
|
|
#endif /* !STATIC_MODES */
|
|
|
|
|
|
#define ALLOC_STEPS 6
|
|
|
|
static inline int interp_bits2pulses(const CELTMode *m, int start, int end, int *bits1, int *bits2, int total, int *bits, int *ebits, int *fine_priority, int len, int _C, int LM)
|
|
{
|
|
int psum;
|
|
int lo, hi;
|
|
int i, j;
|
|
int logM;
|
|
const int C = CHANNELS(_C);
|
|
int codedBands=-1;
|
|
VARDECL(int, thresh);
|
|
SAVE_STACK;
|
|
|
|
ALLOC(thresh, len, int);
|
|
|
|
/* Threshold: don't allow any band to go below 3/8 bit/sample */
|
|
for (j=start;j<end;j++)
|
|
thresh[j] = 3*(C*(m->eBands[j+1]-m->eBands[j])<<LM<<BITRES)>>3;
|
|
logM = LM<<BITRES;
|
|
lo = 0;
|
|
hi = 1<<ALLOC_STEPS;
|
|
for (i=0;i<ALLOC_STEPS;i++)
|
|
{
|
|
int mid = (lo+hi)>>1;
|
|
psum = 0;
|
|
for (j=start;j<end;j++)
|
|
{
|
|
int tmp = bits1[j] + (mid*bits2[j]>>ALLOC_STEPS);
|
|
if (tmp >= thresh[j])
|
|
psum += tmp;
|
|
else if (tmp >= 1<<BITRES)
|
|
psum += 1<<BITRES;
|
|
}
|
|
if (psum > (total<<BITRES))
|
|
hi = mid;
|
|
else
|
|
lo = mid;
|
|
}
|
|
psum = 0;
|
|
/*printf ("interp bisection gave %d\n", lo);*/
|
|
for (j=start;j<end;j++)
|
|
{
|
|
int tmp = bits1[j] + (lo*bits2[j]>>ALLOC_STEPS);
|
|
if (tmp >= thresh[j])
|
|
{
|
|
bits[j] = tmp;
|
|
codedBands = j;
|
|
} else if (tmp >= 1<<BITRES)
|
|
bits[j] = 1<<BITRES;
|
|
else
|
|
bits[j] = 0;
|
|
psum += bits[j];
|
|
}
|
|
codedBands++;
|
|
/* Allocate the remaining bits */
|
|
if (codedBands) {
|
|
int left, perband;
|
|
left = (total<<BITRES)-psum;
|
|
perband = left/(codedBands-start);
|
|
for (j=start;j<codedBands;j++)
|
|
bits[j] += perband;
|
|
left = left-codedBands*perband;
|
|
for (j=start;j<start+left;j++)
|
|
bits[j]++;
|
|
}
|
|
for (j=start;j<end;j++)
|
|
{
|
|
int N0, N, den;
|
|
int offset;
|
|
int NClogN;
|
|
|
|
N0 = m->eBands[j+1]-m->eBands[j];
|
|
N=N0<<LM;
|
|
NClogN = N*C*(m->logN[j] + logM);
|
|
|
|
/* Compensate for the extra DoF in stereo */
|
|
den=(C*N+ ((C==2 && N>2) ? 1 : 0));
|
|
|
|
/* Offset for the number of fine bits by log2(N)/2 + FINE_OFFSET
|
|
compared to their "fair share" of total/N */
|
|
offset = (NClogN>>1)-N*C*FINE_OFFSET;
|
|
|
|
/* N=2 is the only point that doesn't match the curve */
|
|
if (N==2)
|
|
offset += N*C<<BITRES>>2;
|
|
|
|
/* Changing the offset for allocating the second and third fine energy bit */
|
|
if (bits[j] + offset < den*2<<BITRES)
|
|
offset += NClogN>>2;
|
|
else if (bits[j] + offset < den*3<<BITRES)
|
|
offset += NClogN>>3;
|
|
|
|
/* Divide with rounding */
|
|
ebits[j] = (bits[j] + offset + (den<<(BITRES-1))) / (den<<BITRES);
|
|
|
|
/* If we rounded down, make it a candidate for final fine energy pass */
|
|
fine_priority[j] = ebits[j]*(den<<BITRES) >= bits[j]+offset;
|
|
|
|
/* For N=1, all bits go to fine energy except for a single sign bit */
|
|
if (N==1)
|
|
ebits[j] = (bits[j]/C >> BITRES)-1;
|
|
/* Make sure the first bit is spent on fine energy */
|
|
if (ebits[j] < 1)
|
|
ebits[j] = 1;
|
|
|
|
/* Make sure not to bust */
|
|
if (C*ebits[j] > (bits[j]>>BITRES))
|
|
ebits[j] = bits[j]/C >> BITRES;
|
|
|
|
/* More than that is useless because that's about as far as PVQ can go */
|
|
if (ebits[j]>7)
|
|
ebits[j]=7;
|
|
|
|
/* The other bits are assigned to PVQ */
|
|
bits[j] -= C*ebits[j]<<BITRES;
|
|
if (bits[j] < 0)
|
|
bits[j] = 0;
|
|
}
|
|
RESTORE_STACK;
|
|
return codedBands;
|
|
}
|
|
|
|
int compute_allocation(const CELTMode *m, int start, int end, int *offsets, int alloc_trim,
|
|
int total, int *pulses, int *ebits, int *fine_priority, int _C, int LM)
|
|
{
|
|
int lo, hi, len, j;
|
|
const int C = CHANNELS(_C);
|
|
int codedBands;
|
|
VARDECL(int, bits1);
|
|
VARDECL(int, bits2);
|
|
VARDECL(int, thresh);
|
|
VARDECL(int, trim_offset);
|
|
SAVE_STACK;
|
|
|
|
len = m->nbEBands;
|
|
ALLOC(bits1, len, int);
|
|
ALLOC(bits2, len, int);
|
|
ALLOC(thresh, len, int);
|
|
ALLOC(trim_offset, len, int);
|
|
|
|
/* Below this threshold, we don't allocate any PVQ bits */
|
|
for (j=start;j<end;j++)
|
|
thresh[j] = 3*(C*(m->eBands[j+1]-m->eBands[j])<<LM<<BITRES)>>3;
|
|
/* Tilt of the allocation curve */
|
|
for (j=start;j<end;j++)
|
|
trim_offset[j] = C*(m->eBands[j+1]-m->eBands[j])*(2*alloc_trim-7)*(m->nbEBands-j-1)
|
|
<<(LM+BITRES)>>6;
|
|
|
|
lo = 0;
|
|
hi = m->nbAllocVectors - 1;
|
|
while (hi-lo != 1)
|
|
{
|
|
int psum = 0;
|
|
int mid = (lo+hi) >> 1;
|
|
for (j=start;j<end;j++)
|
|
{
|
|
int N = m->eBands[j+1]-m->eBands[j];
|
|
bits1[j] = C*N*m->allocVectors[mid*len+j]<<LM>>2;
|
|
if (bits1[j] > 0)
|
|
bits1[j] += trim_offset[j];
|
|
if (bits1[j] < 0)
|
|
bits1[j] = 0;
|
|
bits1[j] += offsets[j];
|
|
if (bits1[j] >= thresh[j])
|
|
psum += bits1[j];
|
|
else if (bits1[j] >= 1<<BITRES)
|
|
psum += 1<<BITRES;
|
|
|
|
/*printf ("%d ", bits[j]);*/
|
|
}
|
|
/*printf ("\n");*/
|
|
if (psum > (total<<BITRES))
|
|
hi = mid;
|
|
else
|
|
lo = mid;
|
|
/*printf ("lo = %d, hi = %d\n", lo, hi);*/
|
|
}
|
|
/*printf ("interp between %d and %d\n", lo, hi);*/
|
|
for (j=start;j<end;j++)
|
|
{
|
|
int N = m->eBands[j+1]-m->eBands[j];
|
|
bits1[j] = (C*N*m->allocVectors[lo*len+j]<<LM>>2);
|
|
bits2[j] = (C*N*m->allocVectors[hi*len+j]<<LM>>2) - bits1[j];
|
|
if (bits1[j] > 0)
|
|
bits1[j] += trim_offset[j];
|
|
if (bits1[j] < 0)
|
|
bits1[j] = 0;
|
|
bits1[j] += offsets[j];
|
|
}
|
|
codedBands = interp_bits2pulses(m, start, end, bits1, bits2, total, pulses, ebits, fine_priority, len, C, LM);
|
|
RESTORE_STACK;
|
|
return codedBands;
|
|
}
|
|
|