dolphin/Externals/soundtouch/TDStretch.cpp

///////////////////////////////////////////////////////////////////////////////
/// 
/// Sampled sound tempo changer/time stretch algorithm. Changes the sound tempo 
/// while maintaining the original pitch by using a time domain WSOLA-like 
/// method with several performance-increasing tweaks.
///
/// Notes : MMX optimized functions reside in a separate, platform-specific 
/// file, e.g. 'mmx_win.cpp' or 'mmx_gcc.cpp'.
///
/// This source file contains OpenMP optimizations that allow speeding up the
/// corss-correlation algorithm by executing it in several threads / CPU cores 
/// in parallel. See the following article link for more detailed discussion 
/// about SoundTouch OpenMP optimizations:
/// http://www.softwarecoven.com/parallel-computing-in-embedded-mobile-devices
///
/// Author        : Copyright (c) Olli Parviainen
/// Author e-mail : oparviai 'at' iki.fi
/// SoundTouch WWW: http://www.surina.net/soundtouch
///
////////////////////////////////////////////////////////////////////////////////
//
// License :
//
//  SoundTouch audio processing library
//  Copyright (c) Olli Parviainen
//
//  This library 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.
//
//  This library 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 this library; if not, write to the Free Software
//  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
//
////////////////////////////////////////////////////////////////////////////////

#include <string.h>
#include <limits.h>
#include <assert.h>
#include <math.h>
#include <float.h>

#include "STTypes.h"
#include "cpu_detect.h"
#include "TDStretch.h"

using namespace soundtouch;

#define max(x, y) (((x) > (y)) ? (x) : (y))

/*****************************************************************************
 *
 * Implementation of the class 'TDStretch'
 *
 *****************************************************************************/


TDStretch::TDStretch() : FIFOProcessor(&outputBuffer)
{
    bQuickSeek = false;
    channels = 2;

    pMidBuffer = nullptr;
    pMidBufferUnaligned = nullptr;
    overlapLength = 0;

    bAutoSeqSetting = true;
    bAutoSeekSetting = true;

    tempo = 1.0f;
    setParameters(44100, DEFAULT_SEQUENCE_MS, DEFAULT_SEEKWINDOW_MS, DEFAULT_OVERLAP_MS);
    setTempo(1.0f);

    clear();
}



TDStretch::~TDStretch()
{
    delete[] pMidBufferUnaligned;
}



// Sets routine control parameters. These control are certain time constants
// defining how the sound is stretched to the desired duration.
//
// 'sampleRate' = sample rate of the sound
// 'sequenceMS' = one processing sequence length in milliseconds (default = 82 ms)
// 'seekwindowMS' = seeking window length for scanning the best overlapping 
//      position (default = 28 ms)
// 'overlapMS' = overlapping length (default = 12 ms)

void TDStretch::setParameters(int aSampleRate, int aSequenceMS, 
                              int aSeekWindowMS, int aOverlapMS)
{
    // accept only positive parameter values - if zero or negative, use old values instead
    if (aSampleRate > 0)
    {
        if (aSampleRate > 192000) ST_THROW_RT_ERROR("Error: Excessive samplerate");
        this->sampleRate = aSampleRate;
    }

    if (aOverlapMS > 0) this->overlapMs = aOverlapMS;

    if (aSequenceMS > 0)
    {
        this->sequenceMs = aSequenceMS;
        bAutoSeqSetting = false;
    } 
    else if (aSequenceMS == 0)
    {
        // if zero, use automatic setting
        bAutoSeqSetting = true;
    }

    if (aSeekWindowMS > 0) 
    {
        this->seekWindowMs = aSeekWindowMS;
        bAutoSeekSetting = false;
    } 
    else if (aSeekWindowMS == 0) 
    {
        // if zero, use automatic setting
        bAutoSeekSetting = true;
    }

    calcSeqParameters();

    calculateOverlapLength(overlapMs);

    // set tempo to recalculate 'sampleReq'
    setTempo(tempo);
}



/// Get routine control parameters, see setParameters() function.
/// Any of the parameters to this function can be nullptr, in such case corresponding parameter
/// value isn't returned.
void TDStretch::getParameters(int *pSampleRate, int *pSequenceMs, int *pSeekWindowMs, int *pOverlapMs) const
{
    if (pSampleRate)
    {
        *pSampleRate = sampleRate;
    }

    if (pSequenceMs)
    {
        *pSequenceMs = (bAutoSeqSetting) ? (USE_AUTO_SEQUENCE_LEN) : sequenceMs;
    }

    if (pSeekWindowMs)
    {
        *pSeekWindowMs = (bAutoSeekSetting) ? (USE_AUTO_SEEKWINDOW_LEN) : seekWindowMs;
    }

    if (pOverlapMs)
    {
        *pOverlapMs = overlapMs;
    }
}


// Overlaps samples in 'midBuffer' with the samples in 'pInput'
void TDStretch::overlapMono(SAMPLETYPE *pOutput, const SAMPLETYPE *pInput) const
{
    int i;
    SAMPLETYPE m1, m2;

    m1 = (SAMPLETYPE)0;
    m2 = (SAMPLETYPE)overlapLength;

    for (i = 0; i < overlapLength ; i ++)
    {
        pOutput[i] = (pInput[i] * m1 + pMidBuffer[i] * m2 ) / overlapLength;
        m1 += 1;
        m2 -= 1;
    }
}



void TDStretch::clearMidBuffer()
{
    memset(pMidBuffer, 0, channels * sizeof(SAMPLETYPE) * overlapLength);
}


void TDStretch::clearInput()
{
    inputBuffer.clear();
    clearMidBuffer();
    isBeginning = true;
    maxnorm = 0;
    maxnormf = 1e8;
    skipFract = 0;
}


// Clears the sample buffers
void TDStretch::clear()
{
    outputBuffer.clear();
    clearInput();
}



// Enables/disables the quick position seeking algorithm. Zero to disable, nonzero
// to enable
void TDStretch::enableQuickSeek(bool enable)
{
    bQuickSeek = enable;
}


// Returns nonzero if the quick seeking algorithm is enabled.
bool TDStretch::isQuickSeekEnabled() const
{
    return bQuickSeek;
}


// Seeks for the optimal overlap-mixing position.
int TDStretch::seekBestOverlapPosition(const SAMPLETYPE *refPos)
{
    if (bQuickSeek) 
    {
        return seekBestOverlapPositionQuick(refPos);
    }
    else 
    {
        return seekBestOverlapPositionFull(refPos);
    }
}


// Overlaps samples in 'midBuffer' with the samples in 'pInputBuffer' at position
// of 'ovlPos'.
inline void TDStretch::overlap(SAMPLETYPE *pOutput, const SAMPLETYPE *pInput, uint ovlPos) const
{
#ifndef USE_MULTICH_ALWAYS
    if (channels == 1)
    {
        // mono sound.
        overlapMono(pOutput, pInput + ovlPos);
    }
    else if (channels == 2)
    {
        // stereo sound
        overlapStereo(pOutput, pInput + 2 * ovlPos);
    } 
    else 
#endif // USE_MULTICH_ALWAYS
    {
        assert(channels > 0);
        overlapMulti(pOutput, pInput + channels * ovlPos);
    }
}


// Seeks for the optimal overlap-mixing position. The 'stereo' version of the
// routine
//
// The best position is determined as the position where the two overlapped
// sample sequences are 'most alike', in terms of the highest cross-correlation
// value over the overlapping period
int TDStretch::seekBestOverlapPositionFull(const SAMPLETYPE *refPos) 
{
    int bestOffs;
    double bestCorr;
    int i;
    double norm;

    bestCorr = -FLT_MAX;
    bestOffs = 0;

    // Scans for the best correlation value by testing each possible position
    // over the permitted range.
    bestCorr = calcCrossCorr(refPos, pMidBuffer, norm);
    bestCorr = (bestCorr + 0.1) * 0.75;

    #pragma omp parallel for
    for (i = 1; i < seekLength; i ++)
    {
        double corr;
        // Calculates correlation value for the mixing position corresponding to 'i'
#if defined(_OPENMP) || defined(ST_SIMD_AVOID_UNALIGNED)
        // in parallel OpenMP mode, can't use norm accumulator version as parallel executor won't
        // iterate the loop in sequential order
        // in SIMD mode, avoid accumulator version to allow avoiding unaligned positions
        corr = calcCrossCorr(refPos + channels * i, pMidBuffer, norm);
#else
        // In non-parallel version call "calcCrossCorrAccumulate" that is otherwise same
        // as "calcCrossCorr", but saves time by reusing & updating previously stored 
        // "norm" value
        corr = calcCrossCorrAccumulate(refPos + channels * i, pMidBuffer, norm);
#endif
        // heuristic rule to slightly favour values close to mid of the range
        double tmp = (double)(2 * i - seekLength) / (double)seekLength;
        corr = ((corr + 0.1) * (1.0 - 0.25 * tmp * tmp));

        // Checks for the highest correlation value
        if (corr > bestCorr) 
        {
            // For optimal performance, enter critical section only in case that best value found.
            // in such case repeat 'if' condition as it's possible that parallel execution may have
            // updated the bestCorr value in the mean time
            #pragma omp critical
            if (corr > bestCorr)
            {
                bestCorr = corr;
                bestOffs = i;
            }
        }
    }

#ifdef SOUNDTOUCH_INTEGER_SAMPLES
    adaptNormalizer();
#endif

    // clear cross correlation routine state if necessary (is so e.g. in MMX routines).
    clearCrossCorrState();

    return bestOffs;
}


// Quick seek algorithm for improved runtime-performance: First roughly scans through the 
// correlation area, and then scan surroundings of two best preliminary correlation candidates
// with improved precision
//
// Based on testing:
// - This algorithm gives on average 99% as good match as the full algorithm
// - this quick seek algorithm finds the best match on ~90% of cases
// - on those 10% of cases when this algorithm doesn't find best match, 
//   it still finds on average ~90% match vs. the best possible match
int TDStretch::seekBestOverlapPositionQuick(const SAMPLETYPE *refPos)
{
#define _MIN(a, b)   (((a) < (b)) ? (a) : (b))
#define SCANSTEP    16
#define SCANWIND    8

    int bestOffs;
    int i;
    int bestOffs2;
    float bestCorr, corr;
    float bestCorr2;
    double norm;

    // note: 'float' types used in this function in case that the platform would need to use software-fp

    bestCorr =
    bestCorr2 = -FLT_MAX;
    bestOffs = 
    bestOffs2 = SCANWIND;

    // Scans for the best correlation value by testing each possible position
    // over the permitted range. Look for two best matches on the first pass to
    // increase possibility of ideal match.
    //
    // Begin from "SCANSTEP" instead of SCANWIND to make the calculation
    // catch the 'middlepoint' of seekLength vector as that's the a-priori 
    // expected best match position
    //
    // Roughly:
    // - 15% of cases find best result directly on the first round,
    // - 75% cases find better match on 2nd round around the best match from 1st round
    // - 10% cases find better match on 2nd round around the 2nd-best-match from 1st round
    for (i = SCANSTEP; i < seekLength - SCANWIND - 1; i += SCANSTEP)
    {
        // Calculates correlation value for the mixing position corresponding
        // to 'i'
        corr = (float)calcCrossCorr(refPos + channels*i, pMidBuffer, norm);
        // heuristic rule to slightly favour values close to mid of the seek range
        float tmp = (float)(2 * i - seekLength - 1) / (float)seekLength;
        corr = ((corr + 0.1f) * (1.0f - 0.25f * tmp * tmp));

        // Checks for the highest correlation value
        if (corr > bestCorr)
        {
            // found new best match. keep the previous best as 2nd best match
            bestCorr2 = bestCorr;
            bestOffs2 = bestOffs;
            bestCorr = corr;
            bestOffs = i;
        }
        else if (corr > bestCorr2)
        {
            // not new best, but still new 2nd best match
            bestCorr2 = corr;
            bestOffs2 = i;
        }
    }

    // Scans surroundings of the found best match with small stepping
    int end = _MIN(bestOffs + SCANWIND + 1, seekLength);
    for (i = bestOffs - SCANWIND; i < end; i++)
    {
        if (i == bestOffs) continue;    // this offset already calculated, thus skip

        // Calculates correlation value for the mixing position corresponding
        // to 'i'
        corr = (float)calcCrossCorr(refPos + channels*i, pMidBuffer, norm);
        // heuristic rule to slightly favour values close to mid of the range
        float tmp = (float)(2 * i - seekLength - 1) / (float)seekLength;
        corr = ((corr + 0.1f) * (1.0f - 0.25f * tmp * tmp));

        // Checks for the highest correlation value
        if (corr > bestCorr)
        {
            bestCorr = corr;
            bestOffs = i;
        }
    }

    // Scans surroundings of the 2nd best match with small stepping
    end = _MIN(bestOffs2 + SCANWIND + 1, seekLength);
    for (i = bestOffs2 - SCANWIND; i < end; i++)
    {
        if (i == bestOffs2) continue;    // this offset already calculated, thus skip

        // Calculates correlation value for the mixing position corresponding
        // to 'i'
        corr = (float)calcCrossCorr(refPos + channels*i, pMidBuffer, norm);
        // heuristic rule to slightly favour values close to mid of the range
        float tmp = (float)(2 * i - seekLength - 1) / (float)seekLength;
        corr = ((corr + 0.1f) * (1.0f - 0.25f * tmp * tmp));

        // Checks for the highest correlation value
        if (corr > bestCorr)
        {
            bestCorr = corr;
            bestOffs = i;
        }
    }

    // clear cross correlation routine state if necessary (is so e.g. in MMX routines).
    clearCrossCorrState();

#ifdef SOUNDTOUCH_INTEGER_SAMPLES
    adaptNormalizer();
#endif

    return bestOffs;
}




/// For integer algorithm: adapt normalization factor divider with music so that 
/// it'll not be pessimistically restrictive that can degrade quality on quieter sections
/// yet won't cause integer overflows either
void TDStretch::adaptNormalizer()
{
    // Do not adapt normalizer over too silent sequences to avoid averaging filter depleting to
    // too low values during pauses in music
    if ((maxnorm > 1000) || (maxnormf > 40000000))
    { 
        //norm averaging filter
        maxnormf = 0.9f * maxnormf + 0.1f * (float)maxnorm;

        if ((maxnorm > 800000000) && (overlapDividerBitsNorm < 16))
        {
            // large values, so increase divider
            overlapDividerBitsNorm++;
            if (maxnorm > 1600000000) overlapDividerBitsNorm++; // extra large value => extra increase
        }
        else if ((maxnormf < 1000000) && (overlapDividerBitsNorm > 0))
        {
            // extra small values, decrease divider
            overlapDividerBitsNorm--;
        }
    }

    maxnorm = 0;
}


/// clear cross correlation routine state if necessary 
void TDStretch::clearCrossCorrState()
{
    // default implementation is empty.
}


/// Calculates processing sequence length according to tempo setting
void TDStretch::calcSeqParameters()
{
    // Adjust tempo param according to tempo, so that variating processing sequence length is used
    // at various tempo settings, between the given low...top limits
    #define AUTOSEQ_TEMPO_LOW   0.5     // auto setting low tempo range (-50%)
    #define AUTOSEQ_TEMPO_TOP   2.0     // auto setting top tempo range (+100%)

    // sequence-ms setting values at above low & top tempo
    #define AUTOSEQ_AT_MIN      90.0
    #define AUTOSEQ_AT_MAX      40.0
    #define AUTOSEQ_K           ((AUTOSEQ_AT_MAX - AUTOSEQ_AT_MIN) / (AUTOSEQ_TEMPO_TOP - AUTOSEQ_TEMPO_LOW))
    #define AUTOSEQ_C           (AUTOSEQ_AT_MIN - (AUTOSEQ_K) * (AUTOSEQ_TEMPO_LOW))

    // seek-window-ms setting values at above low & top tempoq
    #define AUTOSEEK_AT_MIN     20.0
    #define AUTOSEEK_AT_MAX     15.0
    #define AUTOSEEK_K          ((AUTOSEEK_AT_MAX - AUTOSEEK_AT_MIN) / (AUTOSEQ_TEMPO_TOP - AUTOSEQ_TEMPO_LOW))
    #define AUTOSEEK_C          (AUTOSEEK_AT_MIN - (AUTOSEEK_K) * (AUTOSEQ_TEMPO_LOW))

    #define CHECK_LIMITS(x, mi, ma) (((x) < (mi)) ? (mi) : (((x) > (ma)) ? (ma) : (x)))

    double seq, seek;
    
    if (bAutoSeqSetting)
    {
        seq = AUTOSEQ_C + AUTOSEQ_K * tempo;
        seq = CHECK_LIMITS(seq, AUTOSEQ_AT_MAX, AUTOSEQ_AT_MIN);
        sequenceMs = (int)(seq + 0.5);
    }

    if (bAutoSeekSetting)
    {
        seek = AUTOSEEK_C + AUTOSEEK_K * tempo;
        seek = CHECK_LIMITS(seek, AUTOSEEK_AT_MAX, AUTOSEEK_AT_MIN);
        seekWindowMs = (int)(seek + 0.5);
    }

    // Update seek window lengths
    seekWindowLength = (sampleRate * sequenceMs) / 1000;
    if (seekWindowLength < 2 * overlapLength) 
    {
        seekWindowLength = 2 * overlapLength;
    }
    seekLength = (sampleRate * seekWindowMs) / 1000;
}



// Sets new target tempo. Normal tempo = 'SCALE', smaller values represent slower 
// tempo, larger faster tempo.
void TDStretch::setTempo(double newTempo)
{
    int intskip;

    tempo = newTempo;

    // Calculate new sequence duration
    calcSeqParameters();

    // Calculate ideal skip length (according to tempo value) 
    nominalSkip = tempo * (seekWindowLength - overlapLength);
    intskip = (int)(nominalSkip + 0.5);

    // Calculate how many samples are needed in the 'inputBuffer' to 
    // process another batch of samples
    //sampleReq = max(intskip + overlapLength, seekWindowLength) + seekLength / 2;
    sampleReq = max(intskip + overlapLength, seekWindowLength) + seekLength;
}



// Sets the number of channels, 1 = mono, 2 = stereo
void TDStretch::setChannels(int numChannels)
{
    if (!verifyNumberOfChannels(numChannels) ||
        (channels == numChannels)) return;

    channels = numChannels;
    inputBuffer.setChannels(channels);
    outputBuffer.setChannels(channels);

    // re-init overlap/buffer
    overlapLength=0;
    setParameters(sampleRate);
}


// nominal tempo, no need for processing, just pass the samples through
// to outputBuffer
/*
void TDStretch::processNominalTempo()
{
    assert(tempo == 1.0f);

    if (bMidBufferDirty) 
    {
        // If there are samples in pMidBuffer waiting for overlapping,
        // do a single sliding overlapping with them in order to prevent a 
        // clicking distortion in the output sound
        if (inputBuffer.numSamples() < overlapLength) 
        {
            // wait until we've got overlapLength input samples
            return;
        }
        // Mix the samples in the beginning of 'inputBuffer' with the 
        // samples in 'midBuffer' using sliding overlapping 
        overlap(outputBuffer.ptrEnd(overlapLength), inputBuffer.ptrBegin(), 0);
        outputBuffer.putSamples(overlapLength);
        inputBuffer.receiveSamples(overlapLength);
        clearMidBuffer();
        // now we've caught the nominal sample flow and may switch to
        // bypass mode
    }

    // Simply bypass samples from input to output
    outputBuffer.moveSamples(inputBuffer);
}
*/


// Processes as many processing frames of the samples 'inputBuffer', store
// the result into 'outputBuffer'
void TDStretch::processSamples()
{
    int ovlSkip;
    int offset = 0;
    int temp;

    /* Removed this small optimization - can introduce a click to sound when tempo setting
       crosses the nominal value
    if (tempo == 1.0f) 
    {
        // tempo not changed from the original, so bypass the processing
        processNominalTempo();
        return;
    }
    */

    // Process samples as long as there are enough samples in 'inputBuffer'
    // to form a processing frame.
    while ((int)inputBuffer.numSamples() >= sampleReq) 
    {
        if (isBeginning == false)
        {
            // apart from the very beginning of the track, 
            // scan for the best overlapping position & do overlap-add
            offset = seekBestOverlapPosition(inputBuffer.ptrBegin());

            // Mix the samples in the 'inputBuffer' at position of 'offset' with the 
            // samples in 'midBuffer' using sliding overlapping
            // ... first partially overlap with the end of the previous sequence
            // (that's in 'midBuffer')
            overlap(outputBuffer.ptrEnd((uint)overlapLength), inputBuffer.ptrBegin(), (uint)offset);
            outputBuffer.putSamples((uint)overlapLength);
            offset += overlapLength;
        }
        else
        {
            // Adjust processing offset at beginning of track by not perform initial overlapping
            // and compensating that in the 'input buffer skip' calculation
            isBeginning = false;
            int skip = (int)(tempo * overlapLength + 0.5 * seekLength + 0.5);

            #ifdef ST_SIMD_AVOID_UNALIGNED
            // in SIMD mode, round the skip amount to value corresponding to aligned memory address
            if (channels == 1)
            {
                skip &= -4;
            }
            else if (channels == 2)
            {
                skip &= -2;
            }
            #endif
            skipFract -= skip;
            if (skipFract <= -nominalSkip)
            {
                skipFract = -nominalSkip;
            }
        }

        // ... then copy sequence samples from 'inputBuffer' to output:

        // crosscheck that we don't have buffer overflow...
        if ((int)inputBuffer.numSamples() < (offset + seekWindowLength - overlapLength))
        {
            continue;    // just in case, shouldn't really happen
        }

        // length of sequence
        temp = (seekWindowLength - 2 * overlapLength);
        outputBuffer.putSamples(inputBuffer.ptrBegin() + channels * offset, (uint)temp);

        // Copies the end of the current sequence from 'inputBuffer' to 
        // 'midBuffer' for being mixed with the beginning of the next 
        // processing sequence and so on
        assert((offset + temp + overlapLength) <= (int)inputBuffer.numSamples());
        memcpy(pMidBuffer, inputBuffer.ptrBegin() + channels * (offset + temp), 
            channels * sizeof(SAMPLETYPE) * overlapLength);

        // Remove the processed samples from the input buffer. Update
        // the difference between integer & nominal skip step to 'skipFract'
        // in order to prevent the error from accumulating over time.
        skipFract += nominalSkip;   // real skip size
        ovlSkip = (int)skipFract;   // rounded to integer skip
        skipFract -= ovlSkip;       // maintain the fraction part, i.e. real vs. integer skip
        inputBuffer.receiveSamples((uint)ovlSkip);
    }
}


// Adds 'numsamples' pcs of samples from the 'samples' memory position into
// the input of the object.
void TDStretch::putSamples(const SAMPLETYPE *samples, uint nSamples)
{
    // Add the samples into the input buffer
    inputBuffer.putSamples(samples, nSamples);
    // Process the samples in input buffer
    processSamples();
}



/// Set new overlap length parameter & reallocate RefMidBuffer if necessary.
void TDStretch::acceptNewOverlapLength(int newOverlapLength)
{
    int prevOvl;

    assert(newOverlapLength >= 0);
    prevOvl = overlapLength;
    overlapLength = newOverlapLength;

    if (overlapLength > prevOvl)
    {
        delete[] pMidBufferUnaligned;

        pMidBufferUnaligned = new SAMPLETYPE[overlapLength * channels + 16 / sizeof(SAMPLETYPE)];
        // ensure that 'pMidBuffer' is aligned to 16 byte boundary for efficiency
        pMidBuffer = (SAMPLETYPE *)SOUNDTOUCH_ALIGN_POINTER_16(pMidBufferUnaligned);

        clearMidBuffer();
    }
}


// Operator 'new' is overloaded so that it automatically creates a suitable instance 
// depending on if we've a MMX/SSE/etc-capable CPU available or not.
void * TDStretch::operator new(size_t)
{
    // Notice! don't use "new TDStretch" directly, use "newInstance" to create a new instance instead!
    ST_THROW_RT_ERROR("Error in TDStretch::new: Don't use 'new TDStretch' directly, use 'newInstance' member instead!");
    return newInstance();
}


TDStretch * TDStretch::newInstance()
{
    uint uExtensions;

    uExtensions = detectCPUextensions();

    // Check if MMX/SSE instruction set extensions supported by CPU

#ifdef SOUNDTOUCH_ALLOW_MMX
    // MMX routines available only with integer sample types
    if (uExtensions & SUPPORT_MMX)
    {
        return ::new TDStretchMMX;
    }
    else
#endif // SOUNDTOUCH_ALLOW_MMX


#ifdef SOUNDTOUCH_ALLOW_SSE
    if (uExtensions & SUPPORT_SSE)
    {
        // SSE support
        return ::new TDStretchSSE;
    }
    else
#endif // SOUNDTOUCH_ALLOW_SSE

    {
        // ISA optimizations not supported, use plain C version
        return ::new TDStretch;
    }
}


//////////////////////////////////////////////////////////////////////////////
//
// Integer arithmetic specific algorithm implementations.
//
//////////////////////////////////////////////////////////////////////////////

#ifdef SOUNDTOUCH_INTEGER_SAMPLES

// Overlaps samples in 'midBuffer' with the samples in 'input'. The 'Stereo' 
// version of the routine.
void TDStretch::overlapStereo(short *poutput, const short *input) const
{
    int i;
    short temp;
    int cnt2;

    for (i = 0; i < overlapLength ; i ++)
    {
        temp = (short)(overlapLength - i);
        cnt2 = 2 * i;
        poutput[cnt2] = (input[cnt2] * i + pMidBuffer[cnt2] * temp )  / overlapLength;
        poutput[cnt2 + 1] = (input[cnt2 + 1] * i + pMidBuffer[cnt2 + 1] * temp ) / overlapLength;
    }
}


// Overlaps samples in 'midBuffer' with the samples in 'input'. The 'Multi'
// version of the routine.
void TDStretch::overlapMulti(short *poutput, const short *input) const
{
    short m1;
    int i = 0;

    for (m1 = 0; m1 < overlapLength; m1 ++)
    {
        short m2 = (short)(overlapLength - m1);
        for (int c = 0; c < channels; c ++)
        {
            poutput[i] = (input[i] * m1 + pMidBuffer[i] * m2)  / overlapLength;
            i++;
        }
    }
}

// Calculates the x having the closest 2^x value for the given value
static int _getClosest2Power(double value)
{
    return (int)(log(value) / log(2.0) + 0.5);
}


/// Calculates overlap period length in samples.
/// Integer version rounds overlap length to closest power of 2
/// for a divide scaling operation.
void TDStretch::calculateOverlapLength(int aoverlapMs)
{
    int newOvl;

    assert(aoverlapMs >= 0);

    // calculate overlap length so that it's power of 2 - thus it's easy to do
    // integer division by right-shifting. Term "-1" at end is to account for 
    // the extra most significatnt bit left unused in result by signed multiplication 
    overlapDividerBitsPure = _getClosest2Power((sampleRate * aoverlapMs) / 1000.0) - 1;
    if (overlapDividerBitsPure > 9) overlapDividerBitsPure = 9;
    if (overlapDividerBitsPure < 3) overlapDividerBitsPure = 3;
    newOvl = (int)pow(2.0, (int)overlapDividerBitsPure + 1);    // +1 => account for -1 above

    acceptNewOverlapLength(newOvl);

    overlapDividerBitsNorm = overlapDividerBitsPure;

    // calculate sloping divider so that crosscorrelation operation won't 
    // overflow 32-bit register. Max. sum of the crosscorrelation sum without 
    // divider would be 2^30*(N^3-N)/3, where N = overlap length
    slopingDivider = (newOvl * newOvl - 1) / 3;
}


double TDStretch::calcCrossCorr(const short *mixingPos, const short *compare, double &norm)
{
    long corr;
    unsigned long lnorm;
    int i;

    #ifdef ST_SIMD_AVOID_UNALIGNED
        // in SIMD mode skip 'mixingPos' positions that aren't aligned to 16-byte boundary
        if (((ulongptr)mixingPos) & 15) return -1e50;
    #endif

    // hint compiler autovectorization that loop length is divisible by 8
    int ilength = (channels * overlapLength) & -8;

    corr = lnorm = 0;
    // Same routine for stereo and mono
    for (i = 0; i < ilength; i += 2)
    {
        corr += (mixingPos[i] * compare[i] + 
                 mixingPos[i + 1] * compare[i + 1]) >> overlapDividerBitsNorm;
        lnorm += (mixingPos[i] * mixingPos[i] + 
                  mixingPos[i + 1] * mixingPos[i + 1]) >> overlapDividerBitsNorm;
        // do intermediate scalings to avoid integer overflow
    }

    if (lnorm > maxnorm)
    {
        // modify 'maxnorm' inside critical section to avoid multi-access conflict if in OpenMP mode
        #pragma omp critical
        if (lnorm > maxnorm)
        {
            maxnorm = lnorm;
        }
    }
    // Normalize result by dividing by sqrt(norm) - this step is easiest 
    // done using floating point operation
    norm = (double)lnorm;
    return (double)corr / sqrt((norm < 1e-9) ? 1.0 : norm);
}


/// Update cross-correlation by accumulating "norm" coefficient by previously calculated value
double TDStretch::calcCrossCorrAccumulate(const short *mixingPos, const short *compare, double &norm)
{
    long corr;
    long lnorm;
    int i;

    // hint compiler autovectorization that loop length is divisible by 8
    int ilength = (channels * overlapLength) & -8;

    // cancel first normalizer tap from previous round
    lnorm = 0;
    for (i = 1; i <= channels; i ++)
    {
        lnorm -= (mixingPos[-i] * mixingPos[-i]) >> overlapDividerBitsNorm;
    }

    corr = 0;
    // Same routine for stereo and mono.
    for (i = 0; i < ilength; i += 2) 
    {
        corr += (mixingPos[i] * compare[i] + 
                 mixingPos[i + 1] * compare[i + 1]) >> overlapDividerBitsNorm;
    }

    // update normalizer with last samples of this round
    for (int j = 0; j < channels; j ++)
    {
        i --;
        lnorm += (mixingPos[i] * mixingPos[i]) >> overlapDividerBitsNorm;
    }

    norm += (double)lnorm;
    if (norm > maxnorm)
    {
        maxnorm = (unsigned long)norm;
    }

    // Normalize result by dividing by sqrt(norm) - this step is easiest 
    // done using floating point operation
    return (double)corr / sqrt((norm < 1e-9) ? 1.0 : norm);
}

#endif // SOUNDTOUCH_INTEGER_SAMPLES

//////////////////////////////////////////////////////////////////////////////
//
// Floating point arithmetic specific algorithm implementations.
//

#ifdef SOUNDTOUCH_FLOAT_SAMPLES

// Overlaps samples in 'midBuffer' with the samples in 'pInput'
void TDStretch::overlapStereo(float *pOutput, const float *pInput) const
{
    int i;
    float fScale;
    float f1;
    float f2;

    fScale = 1.0f / (float)overlapLength;

    f1 = 0;
    f2 = 1.0f;

    for (i = 0; i < 2 * (int)overlapLength ; i += 2) 
    {
        pOutput[i + 0] = pInput[i + 0] * f1 + pMidBuffer[i + 0] * f2;
        pOutput[i + 1] = pInput[i + 1] * f1 + pMidBuffer[i + 1] * f2;

        f1 += fScale;
        f2 -= fScale;
    }
}


// Overlaps samples in 'midBuffer' with the samples in 'input'. 
void TDStretch::overlapMulti(float *pOutput, const float *pInput) const
{
    int i;
    float fScale;
    float f1;
    float f2;

    fScale = 1.0f / (float)overlapLength;

    f1 = 0;
    f2 = 1.0f;

    i=0;
    for (int i2 = 0; i2 < overlapLength; i2 ++)
    {
        // note: Could optimize this slightly by taking into account that always channels > 2
        for (int c = 0; c < channels; c ++)
        {
            pOutput[i] = pInput[i] * f1 + pMidBuffer[i] * f2;
            i++;
        }
        f1 += fScale;
        f2 -= fScale;
    }
}


/// Calculates overlapInMsec period length in samples.
void TDStretch::calculateOverlapLength(int overlapInMsec)
{
    int newOvl;

    assert(overlapInMsec >= 0);
    newOvl = (sampleRate * overlapInMsec) / 1000;
    if (newOvl < 16) newOvl = 16;

    // must be divisible by 8
    newOvl -= newOvl % 8;

    acceptNewOverlapLength(newOvl);
}


/// Calculate cross-correlation
double TDStretch::calcCrossCorr(const float *mixingPos, const float *compare, double &anorm)
{
    float corr;
    float norm;
    int i;

    #ifdef ST_SIMD_AVOID_UNALIGNED
        // in SIMD mode skip 'mixingPos' positions that aren't aligned to 16-byte boundary
        if (((ulongptr)mixingPos) & 15) return -1e50;
    #endif

    // hint compiler autovectorization that loop length is divisible by 8
    int ilength = (channels * overlapLength) & -8;

    corr = norm = 0;
    // Same routine for stereo and mono
    for (i = 0; i < ilength; i ++)
    {
        corr += mixingPos[i] * compare[i];
        norm += mixingPos[i] * mixingPos[i];
    }

    anorm = norm;
    return corr / sqrt((norm < 1e-9 ? 1.0 : norm));
}


/// Update cross-correlation by accumulating "norm" coefficient by previously calculated value
double TDStretch::calcCrossCorrAccumulate(const float *mixingPos, const float *compare, double &norm)
{
    float corr;
    int i;

    corr = 0;

    // cancel first normalizer tap from previous round
    for (i = 1; i <= channels; i ++)
    {
        norm -= mixingPos[-i] * mixingPos[-i];
    }

    // hint compiler autovectorization that loop length is divisible by 8
    int ilength = (channels * overlapLength) & -8;

    // Same routine for stereo and mono
    for (i = 0; i < ilength; i ++)
    {
        corr += mixingPos[i] * compare[i];
    }

    // update normalizer with last samples of this round
    for (int j = 0; j < channels; j ++)
    {
        i --;
        norm += mixingPos[i] * mixingPos[i];
    }

    return corr / sqrt((norm < 1e-9 ? 1.0 : norm));
}


#endif // SOUNDTOUCH_FLOAT_SAMPLES