shithub: aacenc

ref: f462a48362fda61eba348d50b22892ca2be6b563
dir: /libfaac/psych.c/

View raw version
/*
 * FAAC - Freeware Advanced Audio Coder
 * Copyright (C) 2001 Menno Bakker
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program 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 General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 *
 * $Id: psych.c,v 1.11 2001/06/08 18:01:09 menno Exp $
 */

#include <math.h>
#ifdef _DEBUG
#include <stdio.h>
#endif

#include "psych.h"
#include "coder.h"
#include "fft.h"
#include "util.h"

#define NS_INTERP(x,y,r) (pow((x),(r))*pow((y),1-(r)))
#define SQRT2 1.41421356237309504880

void PsyInit(GlobalPsyInfo *gpsyInfo, PsyInfo *psyInfo, unsigned int numChannels,
             unsigned int sampleRate, unsigned int sampleRateIdx)
{
    unsigned int channel;
    int i, j, b, bb, high, low, size;
    double tmpx,tmpy,tmp,x;
    double bval[MAX_NPART], SNR;

    gpsyInfo->ath = (double*)AllocMemory(NPART_LONG*sizeof(double));
    gpsyInfo->athS = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
    gpsyInfo->mld = (double*)AllocMemory(NPART_LONG*sizeof(double));
    gpsyInfo->mldS = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
    gpsyInfo->window = (double*)AllocMemory(2*BLOCK_LEN_LONG*sizeof(double));
    gpsyInfo->windowS = (double*)AllocMemory(2*BLOCK_LEN_SHORT*sizeof(double));

    for(i = 0; i < BLOCK_LEN_LONG*2; i++)
        gpsyInfo->window[i] = 0.42-0.5*cos(2*M_PI*(i+.5)/(BLOCK_LEN_LONG*2))+
            0.08*cos(4*M_PI*(i+.5)/(BLOCK_LEN_LONG*2));
    for(i = 0; i < BLOCK_LEN_SHORT*2; i++)
        gpsyInfo->windowS[i] = 0.5 * (1-cos(2.0*M_PI*(i+0.5)/(BLOCK_LEN_SHORT*2)));
    gpsyInfo->sampleRate = (double)sampleRate;

    size = BLOCK_LEN_LONG;
    for (channel = 0; channel < numChannels; channel++) {
        psyInfo[channel].size = size;

        psyInfo[channel].lastPe = 0.0;
        psyInfo[channel].lastEnr = 0.0;
        psyInfo[channel].threeInARow = 0;
        psyInfo[channel].tonality = (double*)AllocMemory(NPART_LONG*sizeof(double));
        psyInfo[channel].nb = (double*)AllocMemory(NPART_LONG*sizeof(double));
        psyInfo[channel].maskThr = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
        psyInfo[channel].maskEn = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
        psyInfo[channel].maskThrNext = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
        psyInfo[channel].maskEnNext = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
        psyInfo[channel].maskThrMS = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
        psyInfo[channel].maskEnMS = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
        psyInfo[channel].maskThrNextMS = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
        psyInfo[channel].maskEnNextMS = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
        psyInfo[channel].prevSamples = (double*)AllocMemory(size*sizeof(double));
        SetMemory(psyInfo[channel].prevSamples, 0, size*sizeof(double));

        psyInfo[channel].lastNb = (double*)AllocMemory(NPART_LONG*sizeof(double));
        psyInfo[channel].lastNbMS = (double*)AllocMemory(NPART_LONG*sizeof(double));
        for (j = 0; j < NPART_LONG; j++) {
            psyInfo[channel].lastNb[j] = 2.;
            psyInfo[channel].lastNbMS[j] = 2.;
        }

        psyInfo[channel].energy = (double*)AllocMemory(size*sizeof(double));
        psyInfo[channel].energyMS = (double*)AllocMemory(size*sizeof(double));
        psyInfo[channel].transBuff = (double*)AllocMemory(2*size*sizeof(double));
    }

    gpsyInfo->psyPart = &psyPartTableLong[sampleRateIdx];
    gpsyInfo->psyPartS = &psyPartTableShort[sampleRateIdx];

    size = BLOCK_LEN_SHORT;
    for (channel = 0; channel < numChannels; channel++) {
        psyInfo[channel].sizeS = size;

        psyInfo[channel].prevSamplesS = (double*)AllocMemory(size*sizeof(double));
        SetMemory(psyInfo[channel].prevSamplesS, 0, size*sizeof(double));

        for (j = 0; j < 8; j++) {
            psyInfo[channel].nbS[j] = (double*)AllocMemory(NPART_SHORT*sizeof(double));
            psyInfo[channel].maskThrS[j] = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
            psyInfo[channel].maskEnS[j] = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
            psyInfo[channel].maskThrNextS[j] = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
            psyInfo[channel].maskEnNextS[j] = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
            psyInfo[channel].maskThrSMS[j] = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
            psyInfo[channel].maskEnSMS[j] = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
            psyInfo[channel].maskThrNextSMS[j] = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));
            psyInfo[channel].maskEnNextSMS[j] = (double*)AllocMemory(MAX_SCFAC_BANDS*sizeof(double));

            psyInfo[channel].energyS[j] = (double*)AllocMemory(size*sizeof(double));
            psyInfo[channel].energySMS[j] = (double*)AllocMemory(size*sizeof(double));
            psyInfo[channel].transBuffS[j] = (double*)AllocMemory(2*size*sizeof(double));
        }
    }

    size = BLOCK_LEN_LONG;
    high = 0;
    for(b = 0; b < gpsyInfo->psyPart->len; b++) {
        low = high;
        high += gpsyInfo->psyPart->width[b];

        bval[b] = 0.5 * (freq2bark(gpsyInfo->sampleRate*low/(2*size)) +
            freq2bark(gpsyInfo->sampleRate*(high-1)/(2*size)));
    }

    for(b = 0; b < gpsyInfo->psyPart->len; b++) {
        for(bb = 0; bb < gpsyInfo->psyPart->len; bb++) {
            if (bval[b] >= bval[bb]) tmpx = (bval[b] - bval[bb])*3.0;
            else tmpx = (bval[b] - bval[bb])*1.5;

            if(tmpx >= 0.5 && tmpx <= 2.5)
            {
                tmp = tmpx - 0.5;
                x = 8.0 * (tmp*tmp - 2.0 * tmp);
            } else
                x = 0.0;

            tmpx += 0.474;
            tmpy = 15.811389 + 7.5*tmpx - 17.5*sqrt(1.0+tmpx*tmpx);

            if (tmpy < -100.0) gpsyInfo->spreading[b][bb] = 0.0;
            else gpsyInfo->spreading[b][bb] = exp((x + tmpy)*0.2302585093);
        }
    }
    for(b = 0; b < gpsyInfo->psyPart->len; b++) {
        for(bb = 0; bb < gpsyInfo->psyPart->len; bb++) {
            if (gpsyInfo->spreading[b][bb] != 0.0)
                break;
        }
        gpsyInfo->sprInd[b][0] = bb;
        for(bb = gpsyInfo->psyPart->len-1; bb > 0; bb--) {
            if (gpsyInfo->spreading[b][bb] != 0.0)
                break;
        }
        gpsyInfo->sprInd[b][1] = bb;
    }

    for( b = 0; b < gpsyInfo->psyPart->len; b++){
        tmp = 0.0;
        for( bb = gpsyInfo->sprInd[b][0]; bb < gpsyInfo->sprInd[b][1]; bb++)
            tmp += gpsyInfo->spreading[b][bb];
        for( bb = gpsyInfo->sprInd[b][0]; bb < gpsyInfo->sprInd[b][1]; bb++)
            gpsyInfo->spreading[b][bb] /= tmp;
    }

    j = 0;
    for( b = 0; b < gpsyInfo->psyPart->len; b++){
        gpsyInfo->ath[b] = 1.e37;

        for (bb = 0; bb < gpsyInfo->psyPart->width[b]; bb++, j++) {
            double freq = gpsyInfo->sampleRate*j/(1000.0*2*size);
            double level;
            level = ATHformula(freq*1000.0) - 20.0;
            level = pow(10., 0.1*level);
            level *= gpsyInfo->psyPart->width[b];
            if (level < gpsyInfo->ath[b])
                gpsyInfo->ath[b] = level;
        }
    }

    low = 0;
    for (b = 0; b < gpsyInfo->psyPart->len; b++) {
        tmp = freq2bark(gpsyInfo->sampleRate*low/(2*size));
        tmp = (min(tmp, 15.5)/15.5);

        gpsyInfo->mld[b] = pow(10.0, 1.25*(1-cos(M_PI*tmp))-2.5);
        low += gpsyInfo->psyPart->width[b];
    }


    size = BLOCK_LEN_SHORT;
    high = 0;
    for(b = 0; b < gpsyInfo->psyPartS->len; b++) {
        low = high;
        high += gpsyInfo->psyPartS->width[b];

        bval[b] = 0.5 * (freq2bark(gpsyInfo->sampleRate*low/(2*size)) +
            freq2bark(gpsyInfo->sampleRate*(high-1)/(2*size)));
    }

    for(b = 0; b < gpsyInfo->psyPartS->len; b++) {
        for(bb = 0; bb < gpsyInfo->psyPartS->len; bb++) {
            if (bval[b] >= bval[bb]) tmpx = (bval[b] - bval[bb])*3.0;
            else tmpx = (bval[b] - bval[bb])*1.5;

            if(tmpx >= 0.5 && tmpx <= 2.5)
            {
                tmp = tmpx - 0.5;
                x = 8.0 * (tmp*tmp - 2.0 * tmp);
            } else
                x = 0.0;

            tmpx += 0.474;
            tmpy = 15.811389 + 7.5*tmpx - 17.5*sqrt(1.0+tmpx*tmpx);

            if (tmpy < -100.0) gpsyInfo->spreadingS[b][bb] = 0.0;
            else gpsyInfo->spreadingS[b][bb] = exp((x + tmpy)*0.2302585093);
        }
    }
    for(b = 0; b < gpsyInfo->psyPartS->len; b++) {
        for(bb = 0; bb < gpsyInfo->psyPartS->len; bb++) {
            if (gpsyInfo->spreadingS[b][bb] != 0.0)
                break;
        }
        gpsyInfo->sprIndS[b][0] = bb;
        for(bb = gpsyInfo->psyPartS->len-1; bb > 0; bb--) {
            if (gpsyInfo->spreadingS[b][bb] != 0.0)
                break;
        }
        gpsyInfo->sprIndS[b][1] = bb;
    }

    j = 0;
    for( b = 0; b < gpsyInfo->psyPartS->len; b++){
        gpsyInfo->athS[b] = 1.e37;

        for (bb = 0; bb < gpsyInfo->psyPartS->width[b]; bb++, j++) {
            double freq = gpsyInfo->sampleRate*j/(1000.0*2*size);
            double level;
            level = ATHformula(freq*1000.0) - 20.0;
            level = pow(10., 0.1*level);
            level *= gpsyInfo->psyPartS->width[b];
            if (level < gpsyInfo->athS[b])
                gpsyInfo->athS[b] = level;
        }
    }

    for( b = 0; b < gpsyInfo->psyPartS->len; b++){
        tmp = 0.0;
        for( bb = gpsyInfo->sprIndS[b][0]; bb < gpsyInfo->sprIndS[b][1]; bb++)
            tmp += gpsyInfo->spreadingS[b][bb];

        /* SNR formula */
        if (bval[b] < 13) SNR = -8.25;
        else SNR = -4.5 * (bval[b]-13)/(24.0-13.0) +
            -8.25*(bval[b]-24)/(13.0-24.0);
        SNR = pow(10.0, SNR/10.0);

        for( bb = gpsyInfo->sprIndS[b][0]; bb < gpsyInfo->sprIndS[b][1]; bb++)
            gpsyInfo->spreadingS[b][bb] *= SNR / tmp;
    }

    low = 0;
    for (b = 0; b < gpsyInfo->psyPartS->len; b++) {
        tmp = freq2bark(gpsyInfo->sampleRate*low/(2*size));
        tmp = (min(tmp, 15.5)/15.5);

        gpsyInfo->mldS[b] = pow(10.0, 1.25*(1-cos(M_PI*tmp))-2.5);
        low += gpsyInfo->psyPartS->width[b];
    }
}

void PsyEnd(GlobalPsyInfo *gpsyInfo, PsyInfo *psyInfo, unsigned int numChannels)
{
    unsigned int channel;
    int j;

    if (gpsyInfo->ath) FreeMemory(gpsyInfo->ath);
    if (gpsyInfo->athS) FreeMemory(gpsyInfo->athS);
    if (gpsyInfo->mld) FreeMemory(gpsyInfo->mld);
    if (gpsyInfo->mldS) FreeMemory(gpsyInfo->mldS);
    if (gpsyInfo->window) FreeMemory(gpsyInfo->window);
    if (gpsyInfo->windowS) FreeMemory(gpsyInfo->windowS);

    for (channel = 0; channel < numChannels; channel++) {
        if (psyInfo[channel].nb) FreeMemory(psyInfo[channel].nb);
        if (psyInfo[channel].tonality) FreeMemory(psyInfo[channel].tonality);
        if (psyInfo[channel].prevSamples) FreeMemory(psyInfo[channel].prevSamples);
        if (psyInfo[channel].maskThr) FreeMemory(psyInfo[channel].maskThr);
        if (psyInfo[channel].maskEn) FreeMemory(psyInfo[channel].maskEn);
        if (psyInfo[channel].maskThrNext) FreeMemory(psyInfo[channel].maskThrNext);
        if (psyInfo[channel].maskEnNext) FreeMemory(psyInfo[channel].maskEnNext);
        if (psyInfo[channel].maskThrMS) FreeMemory(psyInfo[channel].maskThrMS);
        if (psyInfo[channel].maskEnMS) FreeMemory(psyInfo[channel].maskEnMS);
        if (psyInfo[channel].maskThrNextMS) FreeMemory(psyInfo[channel].maskThrNextMS);
        if (psyInfo[channel].maskEnNextMS) FreeMemory(psyInfo[channel].maskEnNextMS);

        if (psyInfo[channel].lastNb) FreeMemory(psyInfo[channel].lastNb);
        if (psyInfo[channel].lastNbMS) FreeMemory(psyInfo[channel].lastNbMS);

        if (psyInfo[channel].energy) FreeMemory(psyInfo[channel].energy);
        if (psyInfo[channel].energyMS) FreeMemory(psyInfo[channel].energyMS);
        if (psyInfo[channel].transBuff) FreeMemory(psyInfo[channel].transBuff);
    }

    for (channel = 0; channel < numChannels; channel++) {
        if(psyInfo[channel].prevSamplesS) FreeMemory(psyInfo[channel].prevSamplesS);
        for (j = 0; j < 8; j++) {
            if (psyInfo[channel].nbS[j]) FreeMemory(psyInfo[channel].nbS[j]);
            if (psyInfo[channel].maskThrS[j]) FreeMemory(psyInfo[channel].maskThrS[j]);
            if (psyInfo[channel].maskEnS[j]) FreeMemory(psyInfo[channel].maskEnS[j]);
            if (psyInfo[channel].maskThrNextS[j]) FreeMemory(psyInfo[channel].maskThrNextS[j]);
            if (psyInfo[channel].maskEnNextS[j]) FreeMemory(psyInfo[channel].maskEnNextS[j]);
            if (psyInfo[channel].maskThrSMS[j]) FreeMemory(psyInfo[channel].maskThrSMS[j]);
            if (psyInfo[channel].maskEnSMS[j]) FreeMemory(psyInfo[channel].maskEnSMS[j]);
            if (psyInfo[channel].maskThrNextSMS[j]) FreeMemory(psyInfo[channel].maskThrNextSMS[j]);
            if (psyInfo[channel].maskEnNextSMS[j]) FreeMemory(psyInfo[channel].maskEnNextSMS[j]);

            if (psyInfo[channel].energyS[j]) FreeMemory(psyInfo[channel].energyS[j]);
            if (psyInfo[channel].energySMS[j]) FreeMemory(psyInfo[channel].energySMS[j]);
            if (psyInfo[channel].transBuffS[j]) FreeMemory(psyInfo[channel].transBuffS[j]);
        }
    }
}

/* Do psychoacoustical analysis */
void PsyCalculate(ChannelInfo *channelInfo, GlobalPsyInfo *gpsyInfo, PsyInfo *psyInfo,
                  int *cb_width_long, int num_cb_long, int *cb_width_short,
                  int num_cb_short, unsigned int numChannels)
{
    unsigned int channel;

    for (channel = 0; channel < numChannels; channel++) {
        if (channelInfo[channel].present) {

            if (channelInfo[channel].cpe &&
                channelInfo[channel].ch_is_left) { /* CPE */

                int leftChan = channel;
                int rightChan = channelInfo[channel].paired_ch;

                PsyBufferUpdateMS(gpsyInfo, &psyInfo[leftChan], &psyInfo[rightChan]);

                /* Calculate the threshold */
                PsyThreshold(gpsyInfo, &psyInfo[leftChan], cb_width_long, num_cb_long,
                    cb_width_short, num_cb_short);
                PsyThreshold(gpsyInfo, &psyInfo[rightChan], cb_width_long, num_cb_long,
                    cb_width_short, num_cb_short);

                /* And for MS */
                PsyThresholdMS(&channelInfo[leftChan], gpsyInfo, &psyInfo[leftChan],
                    &psyInfo[rightChan], cb_width_long, num_cb_long, cb_width_short,
                    num_cb_short);

            } else if (!channelInfo[channel].cpe &&
                channelInfo[channel].lfe) { /* LFE */

                /* NOT FINISHED */

            } else if (!channelInfo[channel].cpe) { /* SCE */

                /* Calculate the threshold */
                PsyThreshold(gpsyInfo, &psyInfo[channel], cb_width_long, num_cb_long,
                    cb_width_short, num_cb_short);
            }
        }
    }
}

static void Hann(GlobalPsyInfo *gpsyInfo, double *inSamples, int size)
{
    int i;

    /* Applying Hann window */
    if (size == BLOCK_LEN_LONG*2) {
        for(i = 0; i < size; i++)
            inSamples[i] *= gpsyInfo->window[i];
    } else {
        for(i = 0; i < size; i++)
            inSamples[i] *= gpsyInfo->windowS[i];
    }
}

void PsyBufferUpdate(GlobalPsyInfo *gpsyInfo, PsyInfo *psyInfo, double *newSamples)
{
    int i, j;
    double a, b;
    double temp[2048];

    memcpy(psyInfo->transBuff, psyInfo->prevSamples, psyInfo->size*sizeof(double));
    memcpy(psyInfo->transBuff + psyInfo->size, newSamples, psyInfo->size*sizeof(double));


    Hann(gpsyInfo, psyInfo->transBuff, 2*psyInfo->size);
    rsfft(psyInfo->transBuff, 11);

    /* Calculate magnitude of new data */
    for (i = 0; i < psyInfo->size; i++) {
        a = psyInfo->transBuff[i];
        b = psyInfo->transBuff[i+psyInfo->size];
        psyInfo->energy[i] = 0.5 * (a*a + b*b);
    }

    memcpy(temp, psyInfo->prevSamples, psyInfo->size*sizeof(double));
    memcpy(temp + psyInfo->size, newSamples, psyInfo->size*sizeof(double));

    for (j = 0; j < 8; j++) {

        memcpy(psyInfo->transBuffS[j], temp+(j*128)+(1024-128), 2*psyInfo->sizeS*sizeof(double));

        Hann(gpsyInfo, psyInfo->transBuffS[j], 2*psyInfo->sizeS);
        rsfft(psyInfo->transBuffS[j], 8);

        /* Calculate magnitude of new data */
        for(i = 0; i < psyInfo->sizeS; i++){
            a = psyInfo->transBuffS[j][i];
            b = psyInfo->transBuffS[j][i+psyInfo->sizeS];
            psyInfo->energyS[j][i] = 0.5 * (a*a + b*b);
        }
    }

    memcpy(psyInfo->prevSamples, newSamples, psyInfo->size*sizeof(double));
}

void PsyBufferUpdateMS(GlobalPsyInfo *gpsyInfo, PsyInfo *psyInfoL, PsyInfo *psyInfoR)
{
    int i, j;
    double a, b;
    double dataL[2048], dataR[2048];

    for (i = 0; i < psyInfoL->size*2; i++) {
        a = psyInfoL->transBuff[i];
        b = psyInfoR->transBuff[i];
        dataL[i] = (a+b)*SQRT2*0.5;
        dataR[i] = (a-b)*SQRT2*0.5;
    }

    /* Calculate magnitude of new data */
    for (i = 0; i < psyInfoL->size; i++) {
        a = dataL[i];
        b = dataL[i+psyInfoL->size];
        psyInfoL->energyMS[i] = 0.5 * (a*a + b*b);

        a = dataR[i];
        b = dataR[i+psyInfoL->size];
        psyInfoR->energyMS[i] = 0.5 * (a*a + b*b);
    }

    for (j = 0; j < 8; j++) {

        for (i = 0; i < psyInfoL->sizeS*2; i++) {
            a = psyInfoL->transBuffS[j][i];
            b = psyInfoR->transBuffS[j][i];
            dataL[i] = (a+b)*SQRT2*0.5;
            dataR[i] = (a-b)*SQRT2*0.5;
        }

        /* Calculate magnitude of new data */
        for (i = 0; i < psyInfoL->sizeS; i++) {
            a = dataL[i];
            b = dataL[i+psyInfoL->sizeS];
            psyInfoL->energySMS[j][i] = 0.5 * (a*a + b*b);

            a = dataR[i];
            b = dataR[i+psyInfoL->sizeS];
            psyInfoR->energySMS[j][i] = 0.5 * (a*a + b*b);
        }
    }
}

/* addition of simultaneous masking */
__inline double mask_add(double m1, double m2, int k, int b, double *ath)
{
    static const double table1[] = {
        3.3246 *3.3246 ,3.23837*3.23837,3.15437*3.15437,3.00412*3.00412,2.86103*2.86103,2.65407*2.65407,2.46209*2.46209,2.284  *2.284  ,
        2.11879*2.11879,1.96552*1.96552,1.82335*1.82335,1.69146*1.69146,1.56911*1.56911,1.46658*1.46658,1.37074*1.37074,1.31036*1.31036,
        1.25264*1.25264,1.20648*1.20648,1.16203*1.16203,1.12765*1.12765,1.09428*1.09428,1.0659 *1.0659 ,1.03826*1.03826,1.01895*1.01895,
        1
    };

    static const double table2[] = {
        1.33352*1.33352,1.35879*1.35879,1.38454*1.38454,1.39497*1.39497,1.40548*1.40548,1.3537 *1.3537 ,1.30382*1.30382,1.22321*1.22321,
        1.14758*1.14758
    };

    static const double table3[] = {
        2.35364*2.35364,2.29259*2.29259,2.23313*2.23313,2.12675*2.12675,2.02545*2.02545,1.87894*1.87894,1.74303*1.74303,1.61695*1.61695,
        1.49999*1.49999,1.39148*1.39148,1.29083*1.29083,1.19746*1.19746,1.11084*1.11084,1.03826*1.03826
    };


    int i;
    double m;

    if (m1 == 0) return m2;

    if (b < 0) b = -b;

    i = (int)(10*log10(m2 / m1)/10*16);
    m = 10*log10((m1+m2)/ath[k]);

    if (i < 0) i = -i;

    if (b <= 3) { /* approximately, 1 bark = 3 partitions */
        if (i > 8) return m1+m2;
        return (m1+m2)*table2[i];
    }

    if (m<15) {
        if (m > 0) {
            double f=1.0,r;
            if (i > 24) return m1+m2;
            if (i > 13) f = 1; else f = table3[i];
            r = (m-0)/15;
            return (m1+m2)*(table1[i]*r+f*(1-r));
        }
        if (i > 13) return m1+m2;
        return (m1+m2)*table3[i];
    }

    if (i > 24) return m1+m2;
    return (m1+m2)*table1[i];
}

static void PsyThreshold(GlobalPsyInfo *gpsyInfo, PsyInfo *psyInfo, int *cb_width_long,
                         int num_cb_long, int *cb_width_short, int num_cb_short)
{
    int b, bb, w, low, high, j;
    double tmp, ecb;

    double e[MAX_NPART];
    double c[MAX_NPART];
    double maxi[MAX_NPART];
    double avg[MAX_NPART];
    double eb;

    double nb_tmp[1024], epart, npart;

    double tot, mx, estot[8];
    double pe = 0.0;

    /* Energy in each partition and weighted unpredictability */
    high = 0;
    for (b = 0; b < gpsyInfo->psyPart->len; b++)
    {
        double m, a;
        low = high;
        high += gpsyInfo->psyPart->width[b];

        eb = psyInfo->energy[low];
        m = a = eb;

        for (w = low+1; w < high; w++)
        {
            double el = psyInfo->energy[w];
            eb += el;
            a += el;
            m = m < el ? el : m;
        }
        e[b] = eb;
        maxi[b] = m;
        avg[b] = a / gpsyInfo->psyPart->width[b];
    }

    for (b = 0; b < gpsyInfo->psyPart->len; b++)
    {
        static double tab[20] = {
            1,0.79433,0.63096,0.63096,0.63096,0.63096,0.63096,0.25119,0.11749,0.11749,
            0.11749,0.11749,0.11749,0.11749,0.11749,0.11749,0.11749,0.11749,0.11749,0.11749
        };
        int c1,c2,t;
        double m, a, tonality;

        c1 = c2 = 0;
        m = a = 0;
        for(w = b-1; w <= b+1; w++)
        {
            if (w >= 0 && w < gpsyInfo->psyPart->len) {
                c1++;
                c2 += gpsyInfo->psyPart->width[w];
                a += avg[w];
                m = m < maxi[w] ? maxi[w] : m;
            }
        }

        a /= c1;
        tonality = (a == 0) ? 0 : (m / a - 1)/(c2-1);

        t = (int)(20*tonality);
        if (t > 19) t = 19;
        psyInfo->tonality[b] = tab[t];
        c[b] = e[b] * tab[t];
    }

    /* Convolve the partitioned energy and unpredictability
       with the spreading function */
    for (b = 0; b < gpsyInfo->psyPart->len; b++)
    {
        ecb = 0;
        for (bb = gpsyInfo->sprInd[b][0]; bb < gpsyInfo->sprInd[b][1]; bb++)
        {
            ecb = mask_add(ecb, gpsyInfo->spreading[b][bb] * c[bb], bb, bb-b, gpsyInfo->ath);
        }
        ecb *= 0.158489319246111;

        /* Actual energy threshold */
        psyInfo->nb[b] = NS_INTERP(min(ecb, 2*psyInfo->lastNb[b]), ecb, 1/*pcfact*/);
/*
        psyInfo->nb[b] = max(psyInfo->nb[b], gpsyInfo->ath[b]);
*/
        psyInfo->lastNb[b] = ecb;

        /* Perceptual entropy */
        tmp = gpsyInfo->psyPart->width[b]
            * log((psyInfo->nb[b] + 0.0000000001)
            / (e[b] + 0.0000000001));
        tmp = min(0,tmp);

        pe -= tmp;
    }

    high = 0;
    for (b = 0; b < gpsyInfo->psyPart->len; b++)
    {
        low = high;
        high += gpsyInfo->psyPart->width[b];

        for (w = low; w < high; w++)
        {
            nb_tmp[w] = psyInfo->nb[b] / gpsyInfo->psyPart->width[b];
        }
    }

    high = 0;
    for (b = 0; b < num_cb_long; b++)
    {
        low = high;
        high += cb_width_long[b];

        epart = psyInfo->energy[low];
        npart = nb_tmp[low];
        for (w = low+1; w < high; w++)
        {
            epart += psyInfo->energy[w];

            if (nb_tmp[w] < npart)
                npart = nb_tmp[w];
        }
        npart *= cb_width_long[b];

        psyInfo->maskThr[b] = psyInfo->maskThrNext[b];
        psyInfo->maskEn[b] = psyInfo->maskEnNext[b];
        tmp = npart / epart;
        psyInfo->maskThrNext[b] = npart;
        psyInfo->maskEnNext[b] = epart;
    }

    /* Short windows */
    for (j = 0; j < 8; j++)
    {
        /* Energy in each partition and weighted unpredictability */
        high = 0;
        for (b = 0; b < gpsyInfo->psyPartS->len; b++)
        {
            low = high;
            high += gpsyInfo->psyPartS->width[b];

            eb = psyInfo->energyS[j][low];

            for (w = low+1; w < high; w++)
            {
                double el = psyInfo->energyS[j][w];
                eb += el;
            }
            e[b] = eb;
        }

        estot[j] = 0.0;

        /* Convolve the partitioned energy and unpredictability
        with the spreading function */
        for (b = 0; b < gpsyInfo->psyPartS->len; b++)
        {
            ecb = 0;
            for (bb = gpsyInfo->sprIndS[b][0]; bb <= gpsyInfo->sprIndS[b][1]; bb++)
            {
                ecb += gpsyInfo->spreadingS[b][bb] * e[bb];
            }

            /* Actual energy threshold */
            psyInfo->nbS[j][b] = max(1e-6, ecb);
/*
            psyInfo->nbS[j][b] = max(psyInfo->nbS[j][b], gpsyInfo->athS[b]);
*/

            estot[j] += e[b];
        }

        if (estot[j] != 0.0)
            estot[j] /= gpsyInfo->psyPartS->len;

        high = 0;
        for (b = 0; b < gpsyInfo->psyPartS->len; b++)
        {
            low = high;
            high += gpsyInfo->psyPartS->width[b];

            for (w = low; w < high; w++)
            {
                nb_tmp[w] = psyInfo->nbS[j][b] / gpsyInfo->psyPartS->width[b];
            }
        }

        high = 0;
        for (b = 0; b < num_cb_short; b++)
        {
            low = high;
            high += cb_width_short[b];

            epart = psyInfo->energyS[j][low];
            npart = nb_tmp[low];
            for (w = low+1; w < high; w++)
            {
                epart += psyInfo->energyS[j][w];

                if (nb_tmp[w] < npart)
                    npart = nb_tmp[w];
            }
            npart *= cb_width_short[b];

            psyInfo->maskThrS[j][b] = psyInfo->maskThrNextS[j][b];
            psyInfo->maskEnS[j][b] = psyInfo->maskEnNextS[j][b];
            psyInfo->maskThrNextS[j][b] = npart;
            psyInfo->maskEnNextS[j][b] = epart;
        }
    }

    tot = mx = estot[0];
    for (j = 1; j < 8; j++) {
        tot += estot[j];
        mx = max(mx, estot[j]);
    }

#ifdef _DEBUG
    printf("%4f %2.2f ", pe, mx/tot);
#endif

    tot = max(tot, 1.e-12);
    if (((mx/tot) > 0.35) && (pe > 1800.0) || ((mx/tot) > 0.5) || (pe > 3000.0)) {
        psyInfo->block_type = ONLY_SHORT_WINDOW;
        psyInfo->threeInARow++;
    } else if ((psyInfo->lastEnr > 0.5) || (psyInfo->lastPe > 3000.0)) {
        psyInfo->block_type = ONLY_SHORT_WINDOW;
        psyInfo->threeInARow++;
    } else if (psyInfo->threeInARow >= 3) {
        psyInfo->block_type = ONLY_SHORT_WINDOW;
        psyInfo->threeInARow = 0;
    } else {
        psyInfo->block_type = ONLY_LONG_WINDOW;
    }

    psyInfo->lastEnr = mx/tot;
    psyInfo->pe = psyInfo->lastPe;
    psyInfo->lastPe = pe;
}

static void PsyThresholdMS(ChannelInfo *channelInfoL, GlobalPsyInfo *gpsyInfo,
                           PsyInfo *psyInfoL, PsyInfo *psyInfoR,
                           int *cb_width_long, int num_cb_long, int *cb_width_short,
                           int num_cb_short)
{
    int b, bb, w, low, high, j;
    double ecb, tmp1, tmp2;

    double nb_tmpM[1024];
    double nb_tmpS[1024];
    double epartM, epartS, npartM, npartS;

    double nbM[MAX_NPART];
    double nbS[MAX_NPART];
    double eM[MAX_NPART];
    double eS[MAX_NPART];
    double cM[MAX_NPART];
    double cS[MAX_NPART];

    double mld;

#ifdef _DEBUG
    int ms_used = 0;
    int ms_usedS = 0;
#endif

    /* Energy in each partition and weighted unpredictability */
    high = 0;
    for (b = 0; b < gpsyInfo->psyPart->len; b++)
    {
        double mid, side, ebM, ebS;
        low = high;
        high += gpsyInfo->psyPart->width[b];

        mid  = psyInfoL->energyMS[low];
        side = psyInfoR->energyMS[low];

        ebM = mid;
        ebS = side;

        for (w = low+1; w < high; w++)
        {
            mid  = psyInfoL->energyMS[w];
            side = psyInfoR->energyMS[w];

            ebM += mid;
            ebS += side;
        }
        eM[b] = ebM;
        eS[b] = ebS;
        cM[b] = ebM * min(psyInfoL->tonality[b], psyInfoR->tonality[b]);
        cS[b] = ebS * min(psyInfoL->tonality[b], psyInfoR->tonality[b]);
    }

    /* Convolve the partitioned energy and unpredictability
       with the spreading function */
    for (b = 0; b < gpsyInfo->psyPart->len; b++)
    {
        /* Mid channel */

        ecb = 0;
        for (bb = gpsyInfo->sprInd[b][0]; bb <= gpsyInfo->sprInd[b][1]; bb++)
        {
            ecb = mask_add(ecb, gpsyInfo->spreading[bb][b] * cM[bb], bb, bb-b, gpsyInfo->ath);
        }
        ecb *= 0.158489319246111;

        /* Actual energy threshold */
        nbM[b] = NS_INTERP(min(ecb, 2*psyInfoL->lastNbMS[b]), ecb, 1/*pcfact*/);
/*
        nbM[b] = max(nbM[b], gpsyInfo->ath[b]);
*/
        psyInfoL->lastNbMS[b] = ecb;


        /* Side channel */

        ecb = 0;
        for (bb = gpsyInfo->sprInd[b][0]; bb <= gpsyInfo->sprInd[b][1]; bb++)
        {
            ecb = mask_add(ecb, gpsyInfo->spreading[bb][b] * cS[bb], bb, bb-b, gpsyInfo->ath);
        }
        ecb *= 0.158489319246111;

        /* Actual energy threshold */
        nbS[b] = NS_INTERP(min(ecb, 2*psyInfoR->lastNbMS[b]), ecb, 1/*pcfact*/);
/*
        nbS[b] = max(nbS[b], gpsyInfo->ath[b]);
*/
        psyInfoR->lastNbMS[b] = ecb;

        if (psyInfoL->nb[b] <= 1.58*psyInfoR->nb[b]
            && psyInfoR->nb[b] <= 1.58*psyInfoL->nb[b]) {

            mld = gpsyInfo->mld[b]*eM[b];
            tmp1 = max(nbM[b], min(nbS[b],mld));

            mld = gpsyInfo->mld[b]*eS[b];
            tmp2 = max(nbS[b], min(nbM[b],mld));

            nbM[b] = tmp1;
            nbS[b] = tmp2;
        }
    }

    high = 0;
    for (b = 0; b < gpsyInfo->psyPart->len; b++)
    {
        low = high;
        high += gpsyInfo->psyPart->width[b];

        for (w = low; w < high; w++)
        {
            nb_tmpM[w] = nbM[b] / gpsyInfo->psyPart->width[b];
            nb_tmpS[w] = nbS[b] / gpsyInfo->psyPart->width[b];
        }
    }

    high = 0;
    for (b = 0; b < num_cb_long; b++)
    {
        low = high;
        high += cb_width_long[b];

        epartM = psyInfoL->energyMS[low];
        npartM = nb_tmpM[low];
        epartS = psyInfoR->energyMS[low];
        npartS = nb_tmpS[low];

        for (w = low+1; w < high; w++)
        {
            epartM += psyInfoL->energyMS[w];
            epartS += psyInfoR->energyMS[w];

            if (nb_tmpM[w] < npartM)
                npartM = nb_tmpM[w];
            if (nb_tmpS[w] < npartS)
                npartS = nb_tmpS[w];
        }
        npartM *= cb_width_long[b];
        npartS *= cb_width_long[b];

        psyInfoL->maskThrMS[b] = psyInfoL->maskThrNextMS[b];
        psyInfoR->maskThrMS[b] = psyInfoR->maskThrNextMS[b];
        psyInfoL->maskEnMS[b] = psyInfoL->maskEnNextMS[b];
        psyInfoR->maskEnMS[b] = psyInfoR->maskEnNextMS[b];
        psyInfoL->maskThrNextMS[b] = npartM;
        psyInfoR->maskThrNextMS[b] = npartS;
        psyInfoL->maskEnNextMS[b] = epartM;
        psyInfoR->maskEnNextMS[b] = epartS;

        {
            double thmL = psyInfoL->maskThr[b];
            double thmR = psyInfoR->maskThr[b];
            double thmM = psyInfoL->maskThrMS[b];
            double thmS = psyInfoR->maskThrMS[b];
            double msfix = 3.5;

            if (thmL*msfix < (thmM+thmS)/2) {
                double f = thmL*msfix / ((thmM+thmS)/2);
                thmM *= f;
                thmS *= f;
            }
            if (thmR*msfix < (thmM+thmS)/2) {
                double f = thmR*msfix / ((thmM+thmS)/2);
                thmM *= f;
                thmS *= f;
            }

            psyInfoL->maskThrMS[b] = min(thmM,psyInfoL->maskThrMS[b]);
            psyInfoR->maskThrMS[b] = min(thmS,psyInfoR->maskThrMS[b]);
            if (psyInfoL->maskThr[b] * psyInfoR->maskThr[b] < psyInfoL->maskThrMS[b] * psyInfoR->maskThrMS[b])
                channelInfoL->msInfo.ms_used[b] = 0;
            else
                channelInfoL->msInfo.ms_used[b] = 1;
        }
    }


#ifdef _DEBUG
    printf("MSL:%3d ", ms_used);
#endif

    /* Short windows */
    for (j = 0; j < 8; j++)
    {
        /* Energy in each partition and weighted unpredictability */
        high = 0;
        for (b = 0; b < gpsyInfo->psyPartS->len; b++)
        {
            double ebM, ebS;
            low = high;
            high += gpsyInfo->psyPartS->width[b];

            ebM = psyInfoL->energySMS[j][low];
            ebS = psyInfoR->energySMS[j][low];

            for (w = low+1; w < high; w++)
            {
                ebM += psyInfoL->energySMS[j][w];
                ebS += psyInfoR->energySMS[j][w];
            }
            eM[b] = ebM;
            eS[b] = ebS;
        }

        /* Convolve the partitioned energy and unpredictability
        with the spreading function */
        for (b = 0; b < gpsyInfo->psyPartS->len; b++)
        {
            /* Mid channel */

            /* Get power ratio */
            ecb = 0;
            for (bb = gpsyInfo->sprIndS[b][0]; bb <= gpsyInfo->sprIndS[b][1]; bb++)
            {
                ecb += gpsyInfo->spreadingS[b][bb] * eM[bb];
            }

            /* Actual energy threshold */
            nbM[b] = max(1e-6, ecb);
/*
            nbM[b] = max(nbM[b], gpsyInfo->athS[b]);
*/

            /* Side channel */

            /* Get power ratio */
            ecb = 0;
            for (bb = gpsyInfo->sprIndS[b][0]; bb <= gpsyInfo->sprIndS[b][1]; bb++)
            {
                ecb += gpsyInfo->spreadingS[b][bb] * eS[bb];
            }

            /* Actual energy threshold */
            nbS[b] = max(1e-6, ecb);
/*
            nbS[b] = max(nbS[b], gpsyInfo->athS[b]);
*/

            if (psyInfoL->nbS[j][b] <= 1.58*psyInfoR->nbS[j][b]
                && psyInfoR->nbS[j][b] <= 1.58*psyInfoL->nbS[j][b]) {

                mld = gpsyInfo->mldS[b]*eM[b];
                tmp1 = max(nbM[b], min(nbS[b],mld));

                mld = gpsyInfo->mldS[b]*eS[b];
                tmp2 = max(nbS[b], min(nbM[b],mld));

                nbM[b] = tmp1;
                nbS[b] = tmp2;
            }
        }

        high = 0;
        for (b = 0; b < gpsyInfo->psyPartS->len; b++)
        {
            low = high;
            high += gpsyInfo->psyPartS->width[b];

            for (w = low; w < high; w++)
            {
                nb_tmpM[w] = nbM[b] / gpsyInfo->psyPartS->width[b];
                nb_tmpS[w] = nbS[b] / gpsyInfo->psyPartS->width[b];
            }
        }

        high = 0;
        for (b = 0; b < num_cb_short; b++)
        {
            low = high;
            high += cb_width_short[b];

            epartM = psyInfoL->energySMS[j][low];
            epartS = psyInfoR->energySMS[j][low];
            npartM = nb_tmpM[low];
            npartS = nb_tmpS[low];

            for (w = low+1; w < high; w++)
            {
                epartM += psyInfoL->energySMS[j][w];
                epartS += psyInfoR->energySMS[j][w];

                if (nb_tmpM[w] < npartM)
                    npartM = nb_tmpM[w];
                if (nb_tmpS[w] < npartS)
                    npartS = nb_tmpS[w];
            }
            npartM *= cb_width_short[b];
            npartS *= cb_width_short[b];

            psyInfoL->maskThrSMS[j][b] = psyInfoL->maskThrNextSMS[j][b];
            psyInfoR->maskThrSMS[j][b] = psyInfoR->maskThrNextSMS[j][b];
            psyInfoL->maskEnSMS[j][b] = psyInfoL->maskEnNextSMS[j][b];
            psyInfoR->maskEnSMS[j][b] = psyInfoR->maskEnNextSMS[j][b];
            psyInfoL->maskThrNextSMS[j][b] = npartM;
            psyInfoR->maskThrNextSMS[j][b] = npartS;
            psyInfoL->maskEnNextSMS[j][b] = epartM;
            psyInfoR->maskEnNextSMS[j][b] = epartS;

            {
                double thmL = psyInfoL->maskThrS[j][b];
                double thmR = psyInfoR->maskThrS[j][b];
                double thmM = psyInfoL->maskThrSMS[j][b];
                double thmS = psyInfoR->maskThrSMS[j][b];
                double msfix = 3.5;

                if (thmL*msfix < (thmM+thmS)/2) {
                    double f = thmL*msfix / ((thmM+thmS)/2);
                    thmM *= f;
                    thmS *= f;
                }
                if (thmR*msfix < (thmM+thmS)/2) {
                    double f = thmR*msfix / ((thmM+thmS)/2);
                    thmM *= f;
                    thmS *= f;
                }

                psyInfoL->maskThrSMS[j][b] = min(thmM,psyInfoL->maskThrSMS[j][b]);
                psyInfoR->maskThrSMS[j][b] = min(thmS,psyInfoR->maskThrSMS[j][b]);
                if (psyInfoL->maskThrS[j][b] * psyInfoR->maskThrS[j][b] <
                    psyInfoL->maskThrSMS[j][b] * psyInfoR->maskThrSMS[j][b])
                    channelInfoL->msInfo.ms_usedS[j][b] = 0;
                else
                    channelInfoL->msInfo.ms_usedS[j][b] = 1;
            }
        }
    }

#ifdef _DEBUG
    printf("MSS:%3d ", ms_usedS);
#endif
}

void BlockSwitch(CoderInfo *coderInfo, PsyInfo *psyInfo, unsigned int numChannels)
{
    unsigned int channel;
    int desire = ONLY_LONG_WINDOW;

    /* Use the same block type for all channels
       If there is 1 channel that wants a short block,
       use a short block on all channels.
    */
    for (channel = 0; channel < numChannels; channel++)
    {
        if (psyInfo[channel].block_type == ONLY_SHORT_WINDOW)
            desire = ONLY_SHORT_WINDOW;
    }

    for (channel = 0; channel < numChannels; channel++)
    {
        if ((coderInfo[channel].block_type == ONLY_SHORT_WINDOW) ||
            (coderInfo[channel].block_type == LONG_SHORT_WINDOW) ) {
            if ((coderInfo[channel].desired_block_type==ONLY_LONG_WINDOW) &&
                (desire == ONLY_LONG_WINDOW) ) {
                coderInfo[channel].block_type = SHORT_LONG_WINDOW;
            } else {
                coderInfo[channel].block_type = ONLY_SHORT_WINDOW;
            }
        } else if (desire == ONLY_SHORT_WINDOW) {
            coderInfo[channel].block_type = LONG_SHORT_WINDOW;
        } else {
            coderInfo[channel].block_type = ONLY_LONG_WINDOW;
        }
        coderInfo[channel].desired_block_type = desire;
    }

#ifdef _DEBUG
    printf("%s ", (coderInfo[0].block_type == ONLY_SHORT_WINDOW) ? "SHORT" : "LONG ");
#endif
}

static double freq2bark(double freq)
{
    double bark;

    if(freq > 200.0)
        bark = 26.81 / (1 + (1960 / freq)) - 0.53;
    else
        bark = freq / 102.9;

    return (bark);
}

static double ATHformula(double f)
{
    double ath;
    f /= 1000;  /* convert to khz */
    f  = max(0.01, f);
    f  = min(18.0,f);

    /* from Painter & Spanias, 1997 */
    /* modified by Gabriel Bouvigne to better fit to the reality */
    ath =    3.640 * pow(f,-0.8)
        - 6.800 * exp(-0.6*pow(f-3.4,2.0))
        + 6.000 * exp(-0.15*pow(f-8.7,2.0))
        + 0.6* 0.001 * pow(f,4.0);
    return ath;
}

static PsyPartTable psyPartTableLong[12+1] =
{
  { 96000, 71,
     { /* width */
      1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,
      3,3,3,3,3,4,4,4,5,5,5,6,6,7,7,8,8,9,10,10,11,12,13,14,15,16,
      18,19,21,24,26,30,34,39,45,53,64,78,98,127,113
     }
  },
  { 88200, 72,
     { /* width */
      1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,
      3,3,3,3,3,4,4,4,4,5,5,5,6,6,7,7,8,8,9,10,10,11,12,13,14,15,
      16,18,19,21,23,26,29,32,37,42,49,58,69,85,106,137,35
     }
  },
  { 64000, 67,
     { /* width */
      2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,3,3,3,3,3,3,
      4,4,4,4,5,5,5,6,6,7,7,8,8,9,10,10,11,12,13,14,15,16,17,
      18,20,21,23,25,28,30,34,37,42,47,54,63,73,87,105,57
     }
  },
  { 48000, 69,
     { /* width */
      2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3,
      3, 3, 3, 4, 4, 4, 4, 4, 5, 5, 5, 6, 6, 7, 7, 8, 8, 9, 10, 10, 11, 12,
      13, 14, 15, 16, 17, 18, 20, 21, 23, 24, 26, 28, 31, 34, 37, 40, 45, 50,
      56, 63, 72, 84, 86
     }
  },
  { 44100, 70,
     { /* width */
      2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3,
      3, 3, 3, 3, 4, 4, 4, 4, 4, 5, 5, 5, 6, 6, 7, 7, 8, 8, 9, 9, 10, 11,
      12, 13, 14, 15, 16, 17, 18, 20, 21, 23, 24, 26, 28, 30, 33, 36, 39,
      43, 47, 53, 59, 67, 76, 88, 27
     }
  },
  { 32000, 66,
     { /* width */
       3,3,3,3,3,3,3,3,3,3,3,
       3,3,3,3,3,3,3,3,4,4,4,
       4,4,4,4,5,5,5,5,6,6,6,
       7,7,8,8,9,10,10,11,12,13,14,
       15,16,17,19,20,22,23,25,27,29,31,
       33,35,38,41,45,48,53,58,64,71,62
     }
  },
  { 24000, 66,
     { /* width */
       3,3,3,3,3,3,3,3,3,3,3,
       4,4,4,4,4,4,4,4,4,4,4,
       5,5,5,5,5,6,6,6,6,7,7,
       7,8,8,9,9,10,11,12,12,13,14,
       15,17,18,19,21,22,24,26,28,30,32,
       34,37,39,42,45,49,53,57,62,67,34
     }
  },
  { 22050, 63,
     { /* width */
      4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5,
      6, 6, 6, 6, 7, 7, 7, 8, 8, 9, 9, 10, 10, 11, 12, 13, 14, 15, 16, 17,
      19, 20, 22, 23, 25, 27, 29, 31, 33, 36, 38, 41, 44, 47, 51, 55, 59,
      64, 61
     }
  },
  { 16000, 60,
     { /* width */
       5,5,5,5,5,5,5,5,5,5,
       5,5,5,5,5,6,6,6,6,6,
       6,6,7,7,7,7,8,8,8,9,
       9,10,10,11,11,12,13,14,15,16,
       17,18,19,21,22,24,26,28,30,33,
       35,38,41,44,47,50,54,58,62,58
     }
  },
  { 12000, 57,
     { /* width */
       6,6,6,6,6,6,6,6,6,6,6,7,7,7,7,7,7,7,
       8,8,8,8,8,9,9,9,10,10,11,11,12,12,13,13,
       14,15,16,17,18,19,20,22,23,25,27,29,31,
       34,36,39,42,45,49,53,57,61,58
    }
  },
  { 11025, 56,
     { /* width */
       7,7,7,7,7,7,7,7,7,7,7,7,7,7,8,8,8,8,8,
       9,9,9,9,10,10,10,11,11,12,12,13,13,14,15,16,17,18,19,20,
       21,23,24,26,28,30,33,35,38,41,44,48,51,55,59,64,9
     }
  },
  { 8000, 52,
     { /* width */
      9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 10, 10, 10, 10, 10, 11, 11, 11, 11,
      12, 12, 12, 13, 13, 14, 14, 15, 15, 16, 17, 18, 18, 19, 20, 21, 23, 24,
      26, 27, 29, 31, 33, 36, 38, 41, 44, 48, 52, 56, 60, 14
     }
  },
  { -1 }
};

static PsyPartTable psyPartTableShort[12+1] =
{
  { 96000, 36,
     { /* width */
      1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,3,3,3,4,4,5,5,
      6,7,9,11,14,18,7
     }
   },
  { 88200, 37,
    { /* width */
      1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,3,3,3,4,4,
      5,5,6,7,8,10,12,16,1
     }
  },
  { 64000, 39,
     { /* width */
      1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,3,3,3,3,4,4,4,
      5,5,6,7,8,9,11,13,10
     }
  },
  { 48000, 42,
    { /* width */
      1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2,
      2, 3, 3, 3, 3, 3, 4, 4, 4, 5, 5, 6, 6, 7, 8, 9, 10, 12, 1
     }
  },
  { 44100, 42,
    { /* width */
      1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2,
      2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 6, 6, 7, 8, 9, 10, 12
     }
  },
  { 32000, 44,
     { /* width */
       1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
       2,2,2,2,2,2,2,3,3,3,3,4,4,4,4,5,5,5,6,6,7,8,8,9,8
     }
  },
  { 24000, 46,
     { /* width */
       1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
       2,2,2,2,2,2,2,3,3,3,3,3,4,4,4,5,5,5,6,6,7,7,8,8,9,1
     }
  },
  { 22050, 46,
     { /* width */
      1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2,
      2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 6, 6, 7, 7, 8, 8, 7
     }
  },
  { 16000, 47,
     { /* width */
       1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
       2,2,2,2,2,2,2,2,3,3,3,3,3,4,4,4,5,5,5,6,6,7,7,8,8,7
     }
  },
  { 12000, 48,
     { /* width */
       1,1,1,1,1,1,1,1,1,1,1,1,
       1,1,1,1,1,1,1,2,2,2,2,2,
       2,2,2,2,2,2,3,3,3,3,3,4,
       4,4,5,5,5,6,6,7,7,8,8,3
     }
  },
  { 11025, 47,
     { /* width */
       1,1,1,1,1,1,1,1,1,1,
       1,1,1,1,1,1,1,1,2,2,
       2,2,2,2,2,2,2,2,2,3,
       3,3,3,3,4,4,4,4,5,5,
       5,6,6,7,7,8,8
     }
  },
  { 8000, 40,
    { /* width */
     2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3,
     3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 6, 6, 7, 7, 8, 3
    }
  },
  { -1 }
};