/*************************** Fraunhofer IIS FDK Tools ********************** (C) Copyright Fraunhofer IIS (2011) All Rights Reserved Please be advised that this software and/or program delivery is Confidential Information of Fraunhofer and subject to and covered by the Fraunhofer IIS Software Evaluation Agreement between Google Inc. and Fraunhofer effective and in full force since March 1, 2012. You may use this software and/or program only under the terms and conditions described in the above mentioned Fraunhofer IIS Software Evaluation Agreement. Any other and/or further use requires a separate agreement. $Id$ Author(s): Markus Lohwasser Description: FDK Tools Hybrid Filterbank This software and/or program is protected by copyright law and international treaties. Any reproduction or distribution of this software and/or program, or any portion of it, may result in severe civil and criminal penalties, and will be prosecuted to the maximum extent possible under law. ******************************************************************************/ #include "FDK_hybrid.h" #include "fft.h" /*--------------- defines -----------------------------*/ #define FFT_IDX_R(a) (2*a) #define FFT_IDX_I(a) (2*a+1) #define HYB_COEF8_0 ( 0.00746082949812f ) #define HYB_COEF8_1 ( 0.02270420949825f ) #define HYB_COEF8_2 ( 0.04546865930473f ) #define HYB_COEF8_3 ( 0.07266113929591f ) #define HYB_COEF8_4 ( 0.09885108575264f ) #define HYB_COEF8_5 ( 0.11793710567217f ) #define HYB_COEF8_6 ( 0.12500000000000f ) #define HYB_COEF8_7 ( HYB_COEF8_5 ) #define HYB_COEF8_8 ( HYB_COEF8_4 ) #define HYB_COEF8_9 ( HYB_COEF8_3 ) #define HYB_COEF8_10 ( HYB_COEF8_2 ) #define HYB_COEF8_11 ( HYB_COEF8_1 ) #define HYB_COEF8_12 ( HYB_COEF8_0 ) /*--------------- structure definitions ---------------*/ #if defined(ARCH_PREFER_MULT_32x16) #define FIXP_HTB FIXP_SGL /* SGL data type. */ #define FIXP_HTP FIXP_SPK /* Packed SGL data type. */ #define HTC(a) (FX_DBL2FXCONST_SGL(a)) /* Cast to SGL */ #define FL2FXCONST_HTB FL2FXCONST_SGL #else #define FIXP_HTB FIXP_DBL /* SGL data type. */ #define FIXP_HTP FIXP_DPK /* Packed DBL data type. */ #define HTC(a) ((FIXP_DBL)(LONG)(a)) /* Cast to DBL */ #define FL2FXCONST_HTB FL2FXCONST_DBL #endif #define HTCP(real,imag) { { HTC(real), HTC(imag) } } /* How to arrange the packed values. */ struct FDK_HYBRID_SETUP { UCHAR nrQmfBands; /*!< Number of QMF bands to be converted to hybrid. */ UCHAR nHybBands[3]; /*!< Number of Hybrid bands generated by nrQmfBands. */ SCHAR kHybrid[3]; /*!< Filter configuration of each QMF band. */ UCHAR protoLen; /*!< Prototype filter length. */ UCHAR filterDelay; /*!< Delay caused by hybrid filter. */ const INT *pReadIdxTable; /*!< Helper table to access input data ringbuffer. */ }; /*--------------- constants ---------------------------*/ static const INT ringbuffIdxTab[2*13] = { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 }; static const FDK_HYBRID_SETUP setup_3_16 = { 3, { 8, 4, 4}, { 8, 4, 4}, 13, (13-1)/2, ringbuffIdxTab}; static const FDK_HYBRID_SETUP setup_3_12 = { 3, { 8, 2, 2}, { 8, 2, 2}, 13, (13-1)/2, ringbuffIdxTab}; static const FDK_HYBRID_SETUP setup_3_10 = { 3, { 6, 2, 2}, { -8, -2, 2}, 13, (13-1)/2, ringbuffIdxTab}; static const FIXP_HTP HybFilterCoef8[] = { HTCP(0x10000000, 0x00000000), HTCP(0x0df26407, 0xfa391882), HTCP(0xff532109, 0x00acdef7), HTCP(0x08f26d36, 0xf70d92ca), HTCP(0xfee34b5f, 0x02af570f), HTCP(0x038f276e, 0xf7684793), HTCP(0x00000000, 0x05d1eac2), HTCP(0x00000000, 0x05d1eac2), HTCP(0x038f276e, 0x0897b86d), HTCP(0xfee34b5f, 0xfd50a8f1), HTCP(0x08f26d36, 0x08f26d36), HTCP(0xff532109, 0xff532109), HTCP(0x0df26407, 0x05c6e77e) }; static const FIXP_HTB HybFilterCoef2[13] = { FL2FXCONST_HTB( 0.00000000000000f), FL2FXCONST_HTB( 0.01899487526049f), FL2FXCONST_HTB( 0.00000000000000f), FL2FXCONST_HTB(-0.07293139167538f), FL2FXCONST_HTB( 0.00000000000000f), FL2FXCONST_HTB( 0.30596630545168f), FL2FXCONST_HTB( 0.50000000000000f), FL2FXCONST_HTB( 0.30596630545168f), FL2FXCONST_HTB( 0.00000000000000f), FL2FXCONST_HTB(-0.07293139167538f), FL2FXCONST_HTB( 0.00000000000000f), FL2FXCONST_HTB( 0.01899487526049f), FL2FXCONST_HTB( 0.00000000000000f) }; static const FIXP_HTB HybFilterCoef4[13] = { FL2FXCONST_HTB(-0.00305151927305f), FL2FXCONST_HTB(-0.00794862316203f), FL2FXCONST_HTB( 0.0f), FL2FXCONST_HTB( 0.04318924038756f), FL2FXCONST_HTB( 0.12542448210445f), FL2FXCONST_HTB( 0.21227807049160f), FL2FXCONST_HTB( 0.25f), FL2FXCONST_HTB( 0.21227807049160f), FL2FXCONST_HTB( 0.12542448210445f), FL2FXCONST_HTB( 0.04318924038756f), FL2FXCONST_HTB( 0.0f), FL2FXCONST_HTB(-0.00794862316203f), FL2FXCONST_HTB(-0.00305151927305f) }; /*--------------- function declarations ---------------*/ static INT kChannelFiltering( const FIXP_DBL *const pQmfReal, const FIXP_DBL *const pQmfImag, const INT *const pReadIdx, FIXP_DBL *const mHybridReal, FIXP_DBL *const mHybridImag, const SCHAR hybridConfig ); /*--------------- function definitions ----------------*/ INT FDKhybridAnalysisOpen( HANDLE_FDK_ANA_HYB_FILTER hAnalysisHybFilter, FIXP_DBL *const pLFmemory, const UINT LFmemorySize, FIXP_DBL *const pHFmemory, const UINT HFmemorySize ) { INT err = 0; /* Save pointer to extern memory. */ hAnalysisHybFilter->pLFmemory = pLFmemory; hAnalysisHybFilter->LFmemorySize = LFmemorySize; hAnalysisHybFilter->pHFmemory = pHFmemory; hAnalysisHybFilter->HFmemorySize = HFmemorySize; return err; } INT FDKhybridAnalysisInit( HANDLE_FDK_ANA_HYB_FILTER hAnalysisHybFilter, const FDK_HYBRID_MODE mode, const INT qmfBands, const INT cplxBands, const INT initStatesFlag ) { int k; INT err = 0; FIXP_DBL *pMem = NULL; HANDLE_FDK_HYBRID_SETUP setup = NULL; switch (mode) { case THREE_TO_TEN: setup = (HANDLE_FDK_HYBRID_SETUP)&setup_3_10; break; case THREE_TO_TWELVE: setup = (HANDLE_FDK_HYBRID_SETUP)&setup_3_12; break; case THREE_TO_SIXTEEN: setup = (HANDLE_FDK_HYBRID_SETUP)&setup_3_16; break; default: err = -1; goto bail; } /* Initialize handle. */ hAnalysisHybFilter->pSetup = setup; hAnalysisHybFilter->bufferLFpos = setup->protoLen-1; hAnalysisHybFilter->bufferHFpos = 0; hAnalysisHybFilter->nrBands = qmfBands; hAnalysisHybFilter->cplxBands = cplxBands; hAnalysisHybFilter->hfMode = 0; /* Check available memory. */ if ( ((2*setup->nrQmfBands*setup->protoLen*sizeof(FIXP_DBL)) > hAnalysisHybFilter->LFmemorySize) || ((setup->filterDelay*((qmfBands-setup->nrQmfBands)+(cplxBands-setup->nrQmfBands))*sizeof(FIXP_DBL)) > hAnalysisHybFilter->HFmemorySize) ) { err = -2; goto bail; } /* Distribut LF memory. */ pMem = hAnalysisHybFilter->pLFmemory; for (k=0; knrQmfBands; k++) { hAnalysisHybFilter->bufferLFReal[k] = pMem; pMem += setup->protoLen; hAnalysisHybFilter->bufferLFImag[k] = pMem; pMem += setup->protoLen; } /* Distribut HF memory. */ pMem = hAnalysisHybFilter->pHFmemory; for (k=0; kfilterDelay; k++) { hAnalysisHybFilter->bufferHFReal[k] = pMem; pMem += (qmfBands-setup->nrQmfBands); hAnalysisHybFilter->bufferHFImag[k] = pMem; pMem += (cplxBands-setup->nrQmfBands); } if (initStatesFlag) { /* Clear LF buffer */ for (k=0; knrQmfBands; k++) { FDKmemclear(hAnalysisHybFilter->bufferLFReal[k], setup->protoLen*sizeof(FIXP_DBL)); FDKmemclear(hAnalysisHybFilter->bufferLFImag[k], setup->protoLen*sizeof(FIXP_DBL)); } if (qmfBands > setup->nrQmfBands) { /* Clear HF buffer */ for (k=0; kfilterDelay; k++) { FDKmemclear(hAnalysisHybFilter->bufferHFReal[k], (qmfBands-setup->nrQmfBands)*sizeof(FIXP_DBL)); FDKmemclear(hAnalysisHybFilter->bufferHFImag[k], (cplxBands-setup->nrQmfBands)*sizeof(FIXP_DBL)); } } } bail: return err; } INT FDKhybridAnalysisScaleStates( HANDLE_FDK_ANA_HYB_FILTER hAnalysisHybFilter, const INT scalingValue ) { INT err = 0; if (hAnalysisHybFilter==NULL) { err = 1; /* invalid handle */ } else { int k; HANDLE_FDK_HYBRID_SETUP setup = hAnalysisHybFilter->pSetup; /* Scale LF buffer */ for (k=0; knrQmfBands; k++) { scaleValues(hAnalysisHybFilter->bufferLFReal[k], setup->protoLen, scalingValue); scaleValues(hAnalysisHybFilter->bufferLFImag[k], setup->protoLen, scalingValue); } if (hAnalysisHybFilter->nrBands > setup->nrQmfBands) { /* Scale HF buffer */ for (k=0; kfilterDelay; k++) { scaleValues(hAnalysisHybFilter->bufferHFReal[k], (hAnalysisHybFilter->nrBands-setup->nrQmfBands), scalingValue); scaleValues(hAnalysisHybFilter->bufferHFImag[k], (hAnalysisHybFilter->cplxBands-setup->nrQmfBands), scalingValue); } } } return err; } INT FDKhybridAnalysisApply( HANDLE_FDK_ANA_HYB_FILTER hAnalysisHybFilter, const FIXP_DBL *const pQmfReal, const FIXP_DBL *const pQmfImag, FIXP_DBL *const pHybridReal, FIXP_DBL *const pHybridImag) { int k, hybOffset = 0; INT err = 0; const int nrQmfBandsLF = hAnalysisHybFilter->pSetup->nrQmfBands; /* number of QMF bands to be converted to hybrid */ const int writIndex = hAnalysisHybFilter->bufferLFpos; int readIndex = hAnalysisHybFilter->bufferLFpos; if (++readIndex>=hAnalysisHybFilter->pSetup->protoLen) readIndex = 0; const INT* pBufferLFreadIdx = &hAnalysisHybFilter->pSetup->pReadIdxTable[readIndex]; /* * LF buffer. */ for (k=0; kbufferLFReal[k][writIndex] = pQmfReal[k]; hAnalysisHybFilter->bufferLFImag[k][writIndex] = pQmfImag[k]; /* Perform hybrid filtering. */ kChannelFiltering( hAnalysisHybFilter->bufferLFReal[k], hAnalysisHybFilter->bufferLFImag[k], pBufferLFreadIdx, pHybridReal+hybOffset, pHybridImag+hybOffset, hAnalysisHybFilter->pSetup->kHybrid[k]); hybOffset += hAnalysisHybFilter->pSetup->nHybBands[k]; } hAnalysisHybFilter->bufferLFpos = readIndex; /* Index where to write next input sample. */ if (hAnalysisHybFilter->nrBands > nrQmfBandsLF) { /* * HF buffer. */ if (hAnalysisHybFilter->hfMode!=0) { /* HF delay compensation was applied outside. */ FDKmemcpy(pHybridReal+hybOffset, &pQmfReal[nrQmfBandsLF], (hAnalysisHybFilter->nrBands-nrQmfBandsLF)*sizeof(FIXP_DBL)); FDKmemcpy(pHybridImag+hybOffset, &pQmfImag[nrQmfBandsLF], (hAnalysisHybFilter->cplxBands-nrQmfBandsLF)*sizeof(FIXP_DBL)); } else { /* HF delay compensation, filterlength/2. */ FDKmemcpy(pHybridReal+hybOffset, hAnalysisHybFilter->bufferHFReal[hAnalysisHybFilter->bufferHFpos], (hAnalysisHybFilter->nrBands-nrQmfBandsLF)*sizeof(FIXP_DBL)); FDKmemcpy(pHybridImag+hybOffset, hAnalysisHybFilter->bufferHFImag[hAnalysisHybFilter->bufferHFpos], (hAnalysisHybFilter->cplxBands-nrQmfBandsLF)*sizeof(FIXP_DBL)); FDKmemcpy(hAnalysisHybFilter->bufferHFReal[hAnalysisHybFilter->bufferHFpos], &pQmfReal[nrQmfBandsLF], (hAnalysisHybFilter->nrBands-nrQmfBandsLF)*sizeof(FIXP_DBL)); FDKmemcpy(hAnalysisHybFilter->bufferHFImag[hAnalysisHybFilter->bufferHFpos], &pQmfImag[nrQmfBandsLF], (hAnalysisHybFilter->cplxBands-nrQmfBandsLF)*sizeof(FIXP_DBL)); if (++hAnalysisHybFilter->bufferHFpos>=hAnalysisHybFilter->pSetup->filterDelay) hAnalysisHybFilter->bufferHFpos = 0; } } /* process HF part*/ return err; } INT FDKhybridAnalysisClose( HANDLE_FDK_ANA_HYB_FILTER hAnalysisHybFilter ) { INT err = 0; if (hAnalysisHybFilter != NULL) { hAnalysisHybFilter->pLFmemory = NULL; hAnalysisHybFilter->pHFmemory = NULL; hAnalysisHybFilter->LFmemorySize = 0; hAnalysisHybFilter->HFmemorySize = 0; } return err; } INT FDKhybridSynthesisInit( HANDLE_FDK_SYN_HYB_FILTER hSynthesisHybFilter, const FDK_HYBRID_MODE mode, const INT qmfBands, const INT cplxBands ) { INT err = 0; HANDLE_FDK_HYBRID_SETUP setup = NULL; switch (mode) { case THREE_TO_TEN: setup = (HANDLE_FDK_HYBRID_SETUP)&setup_3_10; break; case THREE_TO_TWELVE: setup = (HANDLE_FDK_HYBRID_SETUP)&setup_3_12; break; case THREE_TO_SIXTEEN: setup = (HANDLE_FDK_HYBRID_SETUP)&setup_3_16; break; default: err = -1; goto bail; } hSynthesisHybFilter->pSetup = setup; hSynthesisHybFilter->nrBands = qmfBands; hSynthesisHybFilter->cplxBands = cplxBands; bail: return err; } INT FDKhybridSynthesisApply( HANDLE_FDK_SYN_HYB_FILTER hSynthesisHybFilter, const FIXP_DBL *const pHybridReal, const FIXP_DBL *const pHybridImag, FIXP_DBL *const pQmfReal, FIXP_DBL *const pQmfImag ) { int k, n, hybOffset=0; INT err = 0; const INT nrQmfBandsLF = hSynthesisHybFilter->pSetup->nrQmfBands; /* * LF buffer. */ for (k=0; kpSetup->nHybBands[k]; FIXP_DBL accu1 = FL2FXCONST_DBL(0.f); FIXP_DBL accu2 = FL2FXCONST_DBL(0.f); /* Perform hybrid filtering. */ for (n=0; nnrBands > nrQmfBandsLF) { /* * HF buffer. */ FDKmemcpy(&pQmfReal[nrQmfBandsLF], &pHybridReal[hybOffset], (hSynthesisHybFilter->nrBands-nrQmfBandsLF)*sizeof(FIXP_DBL)); FDKmemcpy(&pQmfImag[nrQmfBandsLF], &pHybridImag[hybOffset], (hSynthesisHybFilter->cplxBands-nrQmfBandsLF)*sizeof(FIXP_DBL)); } return err; } /*****************************************************************************/ /* **** FILTERBANK **** */ /* 2 channel filter Filter Coefs: 0.0, 0.01899487526049, 0.0, -0.07293139167538, 0.0, 0.30596630545168, 0.5, 0.30596630545168, 0.0, -0.07293139167538, 0.0, 0.01899487526049, 0.0 Filter design: h[q,n] = g[n] * cos(2pi/2 * q * (n-6) ); n = 0..12, q = 0,1; -> h[0,n] = g[n] * 1; -> h[1,n] = g[n] * pow(-1,n); */ static void dualChannelFiltering( const FIXP_DBL *const pQmfReal, const FIXP_DBL *const pQmfImag, const INT *const pReadIdx, FIXP_DBL *const mHybridReal, FIXP_DBL *const mHybridImag, const INT invert ) { const FIXP_HTB *p = HybFilterCoef2; FIXP_DBL r1, r6; FIXP_DBL i1, i6; /* symmetric filter coefficients */ r1 = fMultDiv2(p[1], pQmfReal[pReadIdx[1]]) + fMultDiv2(p[1], pQmfReal[pReadIdx[11]]) ; i1 = fMultDiv2(p[1], pQmfImag[pReadIdx[1]]) + fMultDiv2(p[1], pQmfImag[pReadIdx[11]]) ; r1 += fMultDiv2(p[3], pQmfReal[pReadIdx[3]]) + fMultDiv2(p[3], pQmfReal[pReadIdx[ 9]]) ; i1 += fMultDiv2(p[3], pQmfImag[pReadIdx[3]]) + fMultDiv2(p[3], pQmfImag[pReadIdx[ 9]]) ; r1 += fMultDiv2(p[5], pQmfReal[pReadIdx[5]]) + fMultDiv2(p[5], pQmfReal[pReadIdx[ 7]]) ; i1 += fMultDiv2(p[5], pQmfImag[pReadIdx[5]]) + fMultDiv2(p[5], pQmfImag[pReadIdx[ 7]]) ; r6 = fMultDiv2(p[6], pQmfReal[pReadIdx[6]]) ; i6 = fMultDiv2(p[6], pQmfImag[pReadIdx[6]]) ; if (invert) { mHybridReal[1] = (r1 + r6) << 1; mHybridImag[1] = (i1 + i6) << 1; mHybridReal[0] = (r6 - r1) << 1; mHybridImag[0] = (i6 - i1) << 1; } else { mHybridReal[0] = (r1 + r6) << 1; mHybridImag[0] = (i1 + i6) << 1; mHybridReal[1] = (r6 - r1) << 1; mHybridImag[1] = (i6 - i1) << 1; } } static void fourChannelFiltering( const FIXP_DBL *const pQmfReal, const FIXP_DBL *const pQmfImag, const INT *const pReadIdx, FIXP_DBL *const mHybridReal, FIXP_DBL *const mHybridImag, const INT invert ) { const FIXP_HTB *p = HybFilterCoef4; FIXP_DBL fft[8]; static const FIXP_DBL cr[13] = { FL2FXCONST_DBL( 0.f), FL2FXCONST_DBL(-0.70710678118655f), FL2FXCONST_DBL( -1.f), FL2FXCONST_DBL(-0.70710678118655f), FL2FXCONST_DBL( 0.f), FL2FXCONST_DBL( 0.70710678118655f), FL2FXCONST_DBL( 1.f), FL2FXCONST_DBL( 0.70710678118655f), FL2FXCONST_DBL( 0.f), FL2FXCONST_DBL(-0.70710678118655f), FL2FXCONST_DBL( -1.f), FL2FXCONST_DBL(-0.70710678118655f), FL2FXCONST_DBL( 0.f) }; static const FIXP_DBL ci[13] = { FL2FXCONST_DBL( -1.f), FL2FXCONST_DBL(-0.70710678118655f), FL2FXCONST_DBL( 0.f), FL2FXCONST_DBL( 0.70710678118655f), FL2FXCONST_DBL( 1.f), FL2FXCONST_DBL( 0.70710678118655f), FL2FXCONST_DBL( 0.f), FL2FXCONST_DBL(-0.70710678118655f), FL2FXCONST_DBL( -1.f), FL2FXCONST_DBL(-0.70710678118655f), FL2FXCONST_DBL( 0.f), FL2FXCONST_DBL( 0.70710678118655f), FL2FXCONST_DBL( 1.f) }; /* FIR filter. */ /* pre twiddeling with pre-twiddling coefficients c[n] */ /* multiplication with filter coefficients p[n] */ /* hint: (a + ib)*(c + id) = (a*c - b*d) + i(a*d + b*c) */ /* write to fft coefficient n' */ fft[FFT_IDX_R(0)] = ( fMult(p[10], ( fMultSub(fMultDiv2(cr[ 2], pQmfReal[pReadIdx[ 2]]), ci[ 2], pQmfImag[pReadIdx[ 2]]))) + fMult(p[ 6], ( fMultSub(fMultDiv2(cr[ 6], pQmfReal[pReadIdx[ 6]]), ci[ 6], pQmfImag[pReadIdx[ 6]]))) + fMult(p[ 2], ( fMultSub(fMultDiv2(cr[10], pQmfReal[pReadIdx[10]]), ci[10], pQmfImag[pReadIdx[10]]))) ); fft[FFT_IDX_I(0)] = ( fMult(p[10], ( fMultAdd(fMultDiv2(ci[ 2], pQmfReal[pReadIdx[ 2]]), cr[ 2], pQmfImag[pReadIdx[ 2]]))) + fMult(p[ 6], ( fMultAdd(fMultDiv2(ci[ 6], pQmfReal[pReadIdx[ 6]]), cr[ 6], pQmfImag[pReadIdx[ 6]]))) + fMult(p[ 2], ( fMultAdd(fMultDiv2(ci[10], pQmfReal[pReadIdx[10]]), cr[10], pQmfImag[pReadIdx[10]]))) ); /* twiddle dee dum */ fft[FFT_IDX_R(1)] = ( fMult(p[ 9], ( fMultSub(fMultDiv2(cr[ 3], pQmfReal[pReadIdx[ 3]]), ci[ 3], pQmfImag[pReadIdx[ 3]]))) + fMult(p[ 5], ( fMultSub(fMultDiv2(cr[ 7], pQmfReal[pReadIdx[ 7]]), ci[ 7], pQmfImag[pReadIdx[ 7]]))) + fMult(p[ 1], ( fMultSub(fMultDiv2(cr[11], pQmfReal[pReadIdx[11]]), ci[11], pQmfImag[pReadIdx[11]]))) ); fft[FFT_IDX_I(1)] = ( fMult(p[ 9], ( fMultAdd(fMultDiv2(ci[ 3], pQmfReal[pReadIdx[ 3]]), cr[ 3], pQmfImag[pReadIdx[ 3]]))) + fMult(p[ 5], ( fMultAdd(fMultDiv2(ci[ 7], pQmfReal[pReadIdx[ 7]]), cr[ 7], pQmfImag[pReadIdx[ 7]]))) + fMult(p[ 1], ( fMultAdd(fMultDiv2(ci[11], pQmfReal[pReadIdx[11]]), cr[11], pQmfImag[pReadIdx[11]]))) ); /* twiddle dee dee */ fft[FFT_IDX_R(2)] = ( fMult(p[12], ( fMultSub(fMultDiv2(cr[ 0], pQmfReal[pReadIdx[ 0]]), ci[ 0], pQmfImag[pReadIdx[ 0]]))) + fMult(p[ 8], ( fMultSub(fMultDiv2(cr[ 4], pQmfReal[pReadIdx[ 4]]), ci[ 4], pQmfImag[pReadIdx[ 4]]))) + fMult(p[ 4], ( fMultSub(fMultDiv2(cr[ 8], pQmfReal[pReadIdx[ 8]]), ci[ 8], pQmfImag[pReadIdx[ 8]]))) + fMult(p[ 0], ( fMultSub(fMultDiv2(cr[12], pQmfReal[pReadIdx[12]]), ci[12], pQmfImag[pReadIdx[12]]))) ); fft[FFT_IDX_I(2)] = ( fMult(p[12], ( fMultAdd(fMultDiv2(ci[ 0], pQmfReal[pReadIdx[ 0]]), cr[ 0], pQmfImag[pReadIdx[ 0]]))) + fMult(p[ 8], ( fMultAdd(fMultDiv2(ci[ 4], pQmfReal[pReadIdx[ 4]]), cr[ 4], pQmfImag[pReadIdx[ 4]]))) + fMult(p[ 4], ( fMultAdd(fMultDiv2(ci[ 8], pQmfReal[pReadIdx[ 8]]), cr[ 8], pQmfImag[pReadIdx[ 8]]))) + fMult(p[ 0], ( fMultAdd(fMultDiv2(ci[12], pQmfReal[pReadIdx[12]]), cr[12], pQmfImag[pReadIdx[12]]))) ); fft[FFT_IDX_R(3)] = ( fMult(p[11], ( fMultSub(fMultDiv2(cr[ 1], pQmfReal[pReadIdx[ 1]]), ci[ 1], pQmfImag[pReadIdx[ 1]]))) + fMult(p[ 7], ( fMultSub(fMultDiv2(cr[ 5], pQmfReal[pReadIdx[ 5]]), ci[ 5], pQmfImag[pReadIdx[ 5]]))) + fMult(p[ 3], ( fMultSub(fMultDiv2(cr[ 9], pQmfReal[pReadIdx[ 9]]), ci[ 9], pQmfImag[pReadIdx[ 9]]))) ); fft[FFT_IDX_I(3)] = ( fMult(p[11], ( fMultAdd(fMultDiv2(ci[ 1], pQmfReal[pReadIdx[ 1]]), cr[ 1], pQmfImag[pReadIdx[ 1]]))) + fMult(p[ 7], ( fMultAdd(fMultDiv2(ci[ 5], pQmfReal[pReadIdx[ 5]]), cr[ 5], pQmfImag[pReadIdx[ 5]]))) + fMult(p[ 3], ( fMultAdd(fMultDiv2(ci[ 9], pQmfReal[pReadIdx[ 9]]), cr[ 9], pQmfImag[pReadIdx[ 9]]))) ); /* fft modulation */ /* here: fast manual fft modulation for a fft of length M=4 */ /* fft_4{x[n]} = x[0]*exp(-i*2*pi/4*m*0) + x[1]*exp(-i*2*pi/4*m*1) + x[2]*exp(-i*2*pi/4*m*2) + x[3]*exp(-i*2*pi/4*m*3) */ /* fft bin m=0: X[0, n] = x[0] + x[1] + x[2] + x[3] */ mHybridReal[0] = fft[FFT_IDX_R(0)] + fft[FFT_IDX_R(1)] + fft[FFT_IDX_R(2)] + fft[FFT_IDX_R(3)]; mHybridImag[0] = fft[FFT_IDX_I(0)] + fft[FFT_IDX_I(1)] + fft[FFT_IDX_I(2)] + fft[FFT_IDX_I(3)]; /* fft bin m=1: X[1, n] = x[0] - i*x[1] - x[2] + i*x[3] */ mHybridReal[1] = fft[FFT_IDX_R(0)] + fft[FFT_IDX_I(1)] - fft[FFT_IDX_R(2)] - fft[FFT_IDX_I(3)]; mHybridImag[1] = fft[FFT_IDX_I(0)] - fft[FFT_IDX_R(1)] - fft[FFT_IDX_I(2)] + fft[FFT_IDX_R(3)]; /* fft bin m=2: X[2, n] = x[0] - x[1] + x[2] - x[3] */ mHybridReal[2] = fft[FFT_IDX_R(0)] - fft[FFT_IDX_R(1)] + fft[FFT_IDX_R(2)] - fft[FFT_IDX_R(3)]; mHybridImag[2] = fft[FFT_IDX_I(0)] - fft[FFT_IDX_I(1)] + fft[FFT_IDX_I(2)] - fft[FFT_IDX_I(3)]; /* fft bin m=3: X[3, n] = x[0] + j*x[1] - x[2] - j*x[3] */ mHybridReal[3] = fft[FFT_IDX_R(0)] - fft[FFT_IDX_I(1)] - fft[FFT_IDX_R(2)] + fft[FFT_IDX_I(3)]; mHybridImag[3] = fft[FFT_IDX_I(0)] + fft[FFT_IDX_R(1)] - fft[FFT_IDX_I(2)] - fft[FFT_IDX_R(3)]; } /* 8 channel filter Implementation using a FFT of length 8 prototype filter coefficients: 0.00746082949812 0.02270420949825 0.04546865930473 0.07266113929591 0.09885108575264 0.11793710567217 0.125 0.11793710567217 0.09885108575264 0.07266113929591 0.04546865930473 0.02270420949825 0.00746082949812 Filter design: N = 13; Q = 8; h[q,n] = g[n] * exp(j * 2 * pi / Q * (q + .5) * (n - 6)); n = 0..(N-1), q = 0..(Q-1); Time Signal: x[t]; Filter Bank Output y[q,t] = conv(x[t],h[q,t]) = conv(h[q,t],x[t]) = sum(x[k] * h[q, t - k] ) = sum(h[q, k] * x[t - k] ); k = 0..(N-1); y[q,t] = x[t - 12]*h[q, 12] + x[t - 11]*h[q, 11] + x[t - 10]*h[q, 10] + x[t - 9]*h[q, 9] + x[t - 8]*h[q, 8] + x[t - 7]*h[q, 7] + x[t - 6]*h[q, 6] + x[t - 5]*h[q, 5] + x[t - 4]*h[q, 4] + x[t - 3]*h[q, 3] + x[t - 2]*h[q, 2] + x[t - 1]*h[q, 1] + x[t - 0]*h[q, 0]; h'[q, n] = h[q,(N-1)-n] = g[n] * exp(j * 2 * pi / Q * (q + .5) * (6 - n)); n = 0..(N-1), q = 0..(Q-1); y[q,t] = x[t - 12]*h'[q, 0] + x[t - 11]*h'[q, 1] + x[t - 10]*h'[q, 2] + x[t - 9]*h'[q, 3] + x[t - 8]*h'[q, 4] + x[t - 7]*h'[q, 5] + x[t - 6]*h'[q, 6] + x[t - 5]*h'[q, 7] + x[t - 4]*h'[q, 8] + x[t - 3]*h'[q, 9] + x[t - 2]*h'[q, 10] + x[t - 1]*h'[q, 11] + x[t - 0]*h'[q, 12]; Try to split off FFT Modulation Term: FFT(x[t], q) = sum(x[t+k]*exp(-j*2*pi/N *q * k)) c m Step 1: h'[q,n] = g[n] * ( exp(j * 2 * pi / 8 * .5 * (6 - n)) ) * ( exp (j * 2 * pi / 8 * q * (6 - n)) ); h'[q,n] = g[n] *c[n] * m[q,n]; (see above) c[n] = exp( j * 2 * pi / 8 * .5 * (6 - n) ); m[q,n] = exp( j * 2 * pi / 8 * q * (6 - n) ); y[q,t] = x[t - 0]*g[0]*c[0]*m[q,0] + x[t - 1]*g[1]*c[ 1]*m[q, 1] + ... ... + x[t - 12]*g[2]*c[12]*m[q,12]; | n m *exp(-j*2*pi) | n' fft ------------------------------------------------------------------------------------------------------------------------- 0 exp( j * 2 * pi / 8 * q * 6) -> exp(-j * 2 * pi / 8 * q * 2) | 2 exp(-j * 2 * pi / 8 * q * 0) 1 exp( j * 2 * pi / 8 * q * 5) -> exp(-j * 2 * pi / 8 * q * 3) | 3 exp(-j * 2 * pi / 8 * q * 1) 2 exp( j * 2 * pi / 8 * q * 4) -> exp(-j * 2 * pi / 8 * q * 4) | 4 exp(-j * 2 * pi / 8 * q * 2) 3 exp( j * 2 * pi / 8 * q * 3) -> exp(-j * 2 * pi / 8 * q * 5) | 5 exp(-j * 2 * pi / 8 * q * 3) 4 exp( j * 2 * pi / 8 * q * 2) -> exp(-j * 2 * pi / 8 * q * 6) | 6 exp(-j * 2 * pi / 8 * q * 4) 5 exp( j * 2 * pi / 8 * q * 1) -> exp(-j * 2 * pi / 8 * q * 7) | 7 exp(-j * 2 * pi / 8 * q * 5) 6 exp( j * 2 * pi / 8 * q * 0) | 0 exp(-j * 2 * pi / 8 * q * 6) 7 exp(-j * 2 * pi / 8 * q * 1) | 1 exp(-j * 2 * pi / 8 * q * 7) 8 exp(-j * 2 * pi / 8 * q * 2) | 2 9 exp(-j * 2 * pi / 8 * q * 3) | 3 10 exp(-j * 2 * pi / 8 * q * 4) | 4 11 exp(-j * 2 * pi / 8 * q * 5) | 5 12 exp(-j * 2 * pi / 8 * q * 6) | 6 now use fft modulation coefficients m[6] = = fft[0] m[7] = = fft[1] m[8] = m[ 0] = fft[2] m[9] = m[ 1] = fft[3] m[10] = m[ 2] = fft[4] m[11] = m[ 3] = fft[5] m[12] = m[ 4] = fft[6] m[ 5] = fft[7] y[q,t] = ( x[t- 6]*g[ 6]*c[ 6] ) * fft[q,0] + ( x[t- 7]*g[ 7]*c[ 7] ) * fft[q,1] + ( x[t- 0]*g[ 0]*c[ 0] + x[t- 8]*g[ 8]*c[ 8] ) * fft[q,2] + ( x[t- 1]*g[ 1]*c[ 1] + x[t- 9]*g[ 9]*c[ 9] ) * fft[q,3] + ( x[t- 2]*g[ 2]*c[ 2] + x[t-10]*g[10]*c[10] ) * fft[q,4] + ( x[t- 3]*g[ 3]*c[ 3] + x[t-11]*g[11]*c[11] ) * fft[q,5] + ( x[t- 4]*g[ 4]*c[ 4] + x[t-12]*g[12]*c[12] ) * fft[q,6] + ( x[t- 5]*g[ 5]*c[ 5] ) * fft[q,7]; pre twiddle factors c[n] = exp(j * 2 * pi / 8 * .5 * (6 - n)); n c] | n c[n] | n c[n] --------------------------------------------------------------------------------------------------- 0 exp( j * 6 * pi / 8) | 1 exp( j * 5 * pi / 8) | 2 exp( j * 4 * pi / 8) 3 exp( j * 3 * pi / 8) | 4 exp( j * 2 * pi / 8) | 5 exp( j * 1 * pi / 8) 6 exp( j * 0 * pi / 8) | 7 exp(-j * 1 * pi / 8) | 8 exp(-j * 2 * pi / 8) 9 exp(-j * 3 * pi / 8) | 10 exp(-j * 4 * pi / 8) | 11 exp(-j * 5 * pi / 8) 12 exp(-j * 6 * pi / 8) | | */ static void eightChannelFiltering( const FIXP_DBL *const pQmfReal, const FIXP_DBL *const pQmfImag, const INT *const pReadIdx, FIXP_DBL *const mHybridReal, FIXP_DBL *const mHybridImag, const INT invert ) { const FIXP_HTP *p = HybFilterCoef8; INT k, sc; FIXP_DBL mfft[16+ALIGNMENT_DEFAULT]; FIXP_DBL *pfft = (FIXP_DBL*)ALIGN_PTR(mfft); FIXP_DBL accu1, accu2, accu3, accu4; /* pre twiddeling */ pfft[FFT_IDX_R(0)] = fMultDiv2(p[0].v.re, pQmfReal[pReadIdx[6]]); pfft[FFT_IDX_I(0)] = fMultDiv2(p[0].v.re, pQmfImag[pReadIdx[6]]); cplxMultDiv2(&accu1, &accu2, pQmfReal[pReadIdx[7]], pQmfImag[pReadIdx[7]], p[1]); pfft[FFT_IDX_R(1)] = accu1; pfft[FFT_IDX_I(1)] = accu2; cplxMultDiv2(&accu1, &accu2, pQmfReal[pReadIdx[0]], pQmfImag[pReadIdx[0]], p[2]); cplxMultDiv2(&accu3, &accu4, pQmfReal[pReadIdx[8]], pQmfImag[pReadIdx[8]], p[3]); pfft[FFT_IDX_R(2)] = accu1 + accu3; pfft[FFT_IDX_I(2)] = accu2 + accu4; cplxMultDiv2(&accu1, &accu2, pQmfReal[pReadIdx[1]], pQmfImag[pReadIdx[1]], p[4]); cplxMultDiv2(&accu3, &accu4, pQmfReal[pReadIdx[9]], pQmfImag[pReadIdx[9]], p[5]); pfft[FFT_IDX_R(3)] = accu1 + accu3; pfft[FFT_IDX_I(3)] = accu2 + accu4; pfft[FFT_IDX_R(4)] = fMultDiv2(pQmfImag[pReadIdx[10]], p[7].v.im) - fMultDiv2(pQmfImag[pReadIdx[ 2]], p[6].v.im); pfft[FFT_IDX_I(4)] = fMultDiv2(pQmfReal[pReadIdx[ 2]], p[6].v.im) - fMultDiv2(pQmfReal[pReadIdx[10]], p[7].v.im); cplxMultDiv2(&accu1, &accu2, pQmfReal[pReadIdx[ 3]], pQmfImag[pReadIdx[ 3]], p[8]); cplxMultDiv2(&accu3, &accu4, pQmfReal[pReadIdx[11]], pQmfImag[pReadIdx[11]], p[9]); pfft[FFT_IDX_R(5)] = accu1 + accu3; pfft[FFT_IDX_I(5)] = accu2 + accu4; cplxMultDiv2(&accu1, &accu2, pQmfReal[pReadIdx[ 4]], pQmfImag[pReadIdx[ 4]], p[10]); cplxMultDiv2(&accu3, &accu4, pQmfReal[pReadIdx[12]], pQmfImag[pReadIdx[12]], p[11]); pfft[FFT_IDX_R(6)] = accu1 + accu3; pfft[FFT_IDX_I(6)] = accu2 + accu4; cplxMultDiv2(&accu1, &accu2, pQmfReal[pReadIdx[ 5]], pQmfImag[pReadIdx[ 5]], p[12]); pfft[FFT_IDX_R(7)] = accu1; pfft[FFT_IDX_I(7)] = accu2; /* fft modulation */ fft_8 (pfft); sc = 1 + 2; if (invert) { mHybridReal[0] = pfft[FFT_IDX_R(7)] << sc; mHybridImag[0] = pfft[FFT_IDX_I(7)] << sc; mHybridReal[1] = pfft[FFT_IDX_R(0)] << sc; mHybridImag[1] = pfft[FFT_IDX_I(0)] << sc; mHybridReal[2] = pfft[FFT_IDX_R(6)] << sc; mHybridImag[2] = pfft[FFT_IDX_I(6)] << sc; mHybridReal[3] = pfft[FFT_IDX_R(1)] << sc; mHybridImag[3] = pfft[FFT_IDX_I(1)] << sc; mHybridReal[4] = pfft[FFT_IDX_R(2)] << sc; mHybridReal[4] += pfft[FFT_IDX_R(5)] << sc; mHybridImag[4] = pfft[FFT_IDX_I(2)] << sc; mHybridImag[4] += pfft[FFT_IDX_I(5)] << sc; mHybridReal[5] = pfft[FFT_IDX_R(3)] << sc; mHybridReal[5] += pfft[FFT_IDX_R(4)] << sc; mHybridImag[5] = pfft[FFT_IDX_I(3)] << sc; mHybridImag[5] += pfft[FFT_IDX_I(4)] << sc; } else { for(k=0; k<8;k++ ) { mHybridReal[k] = pfft[FFT_IDX_R(k)] << sc; mHybridImag[k] = pfft[FFT_IDX_I(k)] << sc; } } } static INT kChannelFiltering( const FIXP_DBL *const pQmfReal, const FIXP_DBL *const pQmfImag, const INT *const pReadIdx, FIXP_DBL *const mHybridReal, FIXP_DBL *const mHybridImag, const SCHAR hybridConfig ) { INT err = 0; switch (hybridConfig) { case 2: case -2: dualChannelFiltering(pQmfReal, pQmfImag, pReadIdx, mHybridReal, mHybridImag, (hybridConfig<0) ? 1 : 0 ); break; case 4: case -4: fourChannelFiltering(pQmfReal, pQmfImag, pReadIdx, mHybridReal, mHybridImag, (hybridConfig<0) ? 1 : 0 ); break; case 8: case -8: eightChannelFiltering(pQmfReal, pQmfImag, pReadIdx, mHybridReal, mHybridImag, (hybridConfig<0) ? 1 : 0 ); break; default: err = -1; } return err; }