diff options
Diffstat (limited to 'fdk-aac/libSACenc/src/sacenc_vectorfunctions.h')
-rw-r--r-- | fdk-aac/libSACenc/src/sacenc_vectorfunctions.h | 488 |
1 files changed, 488 insertions, 0 deletions
diff --git a/fdk-aac/libSACenc/src/sacenc_vectorfunctions.h b/fdk-aac/libSACenc/src/sacenc_vectorfunctions.h new file mode 100644 index 0000000..e9c4abd --- /dev/null +++ b/fdk-aac/libSACenc/src/sacenc_vectorfunctions.h @@ -0,0 +1,488 @@ +/* ----------------------------------------------------------------------------- +Software License for The Fraunhofer FDK AAC Codec Library for Android + +© Copyright 1995 - 2018 Fraunhofer-Gesellschaft zur Förderung der angewandten +Forschung e.V. All rights reserved. + + 1. INTRODUCTION +The Fraunhofer FDK AAC Codec Library for Android ("FDK AAC Codec") is software +that implements the MPEG Advanced Audio Coding ("AAC") encoding and decoding +scheme for digital audio. This FDK AAC Codec software is intended to be used on +a wide variety of Android devices. + +AAC's HE-AAC and HE-AAC v2 versions are regarded as today's most efficient +general perceptual audio codecs. AAC-ELD is considered the best-performing +full-bandwidth communications codec by independent studies and is widely +deployed. AAC has been standardized by ISO and IEC as part of the MPEG +specifications. + +Patent licenses for necessary patent claims for the FDK AAC Codec (including +those of Fraunhofer) may be obtained through Via Licensing +(www.vialicensing.com) or through the respective patent owners individually for +the purpose of encoding or decoding bit streams in products that are compliant +with the ISO/IEC MPEG audio standards. Please note that most manufacturers of +Android devices already license these patent claims through Via Licensing or +directly from the patent owners, and therefore FDK AAC Codec software may +already be covered under those patent licenses when it is used for those +licensed purposes only. + +Commercially-licensed AAC software libraries, including floating-point versions +with enhanced sound quality, are also available from Fraunhofer. Users are +encouraged to check the Fraunhofer website for additional applications +information and documentation. + +2. COPYRIGHT LICENSE + +Redistribution and use in source and binary forms, with or without modification, +are permitted without payment of copyright license fees provided that you +satisfy the following conditions: + +You must retain the complete text of this software license in redistributions of +the FDK AAC Codec or your modifications thereto in source code form. + +You must retain the complete text of this software license in the documentation +and/or other materials provided with redistributions of the FDK AAC Codec or +your modifications thereto in binary form. You must make available free of +charge copies of the complete source code of the FDK AAC Codec and your +modifications thereto to recipients of copies in binary form. + +The name of Fraunhofer may not be used to endorse or promote products derived +from this library without prior written permission. + +You may not charge copyright license fees for anyone to use, copy or distribute +the FDK AAC Codec software or your modifications thereto. + +Your modified versions of the FDK AAC Codec must carry prominent notices stating +that you changed the software and the date of any change. For modified versions +of the FDK AAC Codec, the term "Fraunhofer FDK AAC Codec Library for Android" +must be replaced by the term "Third-Party Modified Version of the Fraunhofer FDK +AAC Codec Library for Android." + +3. NO PATENT LICENSE + +NO EXPRESS OR IMPLIED LICENSES TO ANY PATENT CLAIMS, including without +limitation the patents of Fraunhofer, ARE GRANTED BY THIS SOFTWARE LICENSE. +Fraunhofer provides no warranty of patent non-infringement with respect to this +software. + +You may use this FDK AAC Codec software or modifications thereto only for +purposes that are authorized by appropriate patent licenses. + +4. DISCLAIMER + +This FDK AAC Codec software is provided by Fraunhofer on behalf of the copyright +holders and contributors "AS IS" and WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES, +including but not limited to the implied warranties of merchantability and +fitness for a particular purpose. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR +CONTRIBUTORS BE LIABLE for any direct, indirect, incidental, special, exemplary, +or consequential damages, including but not limited to procurement of substitute +goods or services; loss of use, data, or profits, or business interruption, +however caused and on any theory of liability, whether in contract, strict +liability, or tort (including negligence), arising in any way out of the use of +this software, even if advised of the possibility of such damage. + +5. CONTACT INFORMATION + +Fraunhofer Institute for Integrated Circuits IIS +Attention: Audio and Multimedia Departments - FDK AAC LL +Am Wolfsmantel 33 +91058 Erlangen, Germany + +www.iis.fraunhofer.de/amm +amm-info@iis.fraunhofer.de +----------------------------------------------------------------------------- */ + +/*********************** MPEG surround encoder library ************************* + + Author(s): Josef Hoepfl + + Description: Encoder Library Interface + vector functions + +*******************************************************************************/ + +/***************************************************************************** +\file +This file contains vector functions +******************************************************************************/ + +#ifndef SACENC_VECTORFUNCTIONS_H +#define SACENC_VECTORFUNCTIONS_H + +/* Includes ******************************************************************/ +#include "common_fix.h" + +/* Defines *******************************************************************/ +#define SUM_UP_STATIC_SCALE 0 +#define SUM_UP_DYNAMIC_SCALE 1 + +/* Data Types ****************************************************************/ + +/* Constants *****************************************************************/ + +/* Function / Class Declarations *********************************************/ + +/** + * \brief Vector function : Sum up complex power + * + * Description : ret = sum( re{X[i]} * re{X[i]} + im{X[i]} * + * im{X[i]} ), i=0,...,n-1 ret is scaled by outScaleFactor + * + * \param const FIXP_DPK x[] + * Input: complex vector of the length n + * + * \param int scaleMode + * Input: choose static or dynamic scaling + * (SUM_UP_DYNAMIC_SCALE/SUM_UP_STATIC_SCALE) + * + * \param int inScaleFactor + * Input: determine headroom bits for the complex input vector + * + * \param int outScaleFactor + * Output: complete scaling in energy calculation + * + * \return FIXP_DBL ret + */ +FIXP_DBL sumUpCplxPow2(const FIXP_DPK *const x, const INT scaleMode, + const INT inScaleFactor, INT *const outScaleFactor, + const INT n); + +/** + * \brief Vector function : Sum up complex power + * + * Description : ret = sum( re{X[i][j]} * re{X[i][]} + + * im{X[i][]} * im{X[i][]} ), i=sDim1,...,nDim1-1 i=sDim2,...,nDim2-1 ret is + * scaled by outScaleFactor + * + * \param const FIXP_DPK x[] + * Input: complex vector of the length n + * + * \param int scaleMode + * Input: choose static or dynamic scaling + * (SUM_UP_DYNAMIC_SCALE/SUM_UP_STATIC_SCALE) + * + * \param int inScaleFactor + * Input: determine headroom bits for the complex input vector + * + * \param int outScaleFactor + * Output: complete scaling in energy calculation + * + * \param int sDim1 + * Input: start index for loop counter in dimension 1 + * + * \param int nDim1 + * Input: loop counter in dimension 1 + * + * \param int sDim2 + * Input: start index for loop counter in dimension 2 + * + * \param int nDim2 + * Input: loop counter in dimension 2 + * + * \return FIXP_DBL ret + */ +FIXP_DBL sumUpCplxPow2Dim2(const FIXP_DPK *const *const x, const INT scaleMode, + const INT inScaleFactor, INT *const outScaleFactor, + const INT sDim1, const INT nDim1, const INT sDim2, + const INT nDim2); + +/** + * \brief Vector function : Z[i] = X[i], i=0,...,n-1 + * + * Description : re{Z[i]} = re{X[i]}, i=0,...,n-1 + * im{Z[i]} = im{X[i]}, i=0,...,n-1 + * + * Copy complex vector X[] to complex vector Z[]. + * It is allowed to overlay X[] with Z[]. + * + * \param FIXP_DPK Z[] + * Output: vector of the length n + * + * \param const FIXP_DPK X[] + * Input: vector of the length n + * + * \param int n + * Input: length of vector Z[] and X[] + * + * \return void + */ +void copyCplxVec(FIXP_DPK *const Z, const FIXP_DPK *const X, const INT n); + +/** + * \brief Vector function : Z[i] = a, i=0,...,n-1 + * + * Description : re{Z[i]} = a, i=0,...,n-1 + * im{Z[i]} = a, i=0,...,n-1 + * + * Set real and imaginary part of the complex value Z to a. + * + * \param FIPX_DPK Z[] + * Output: vector of the length n + * + * \param const FIXP_DBL a + * Input: constant value + * + * \param int n + * Input: length of vector Z[] + * + * \return void + */ +void setCplxVec(FIXP_DPK *const Z, const FIXP_DBL a, const INT n); + +/** + * \brief Vector function : Calculate complex-valued result of complex + * scalar product + * + * Description : re{Z} = sum( re{X[i]} * re{Y[i]} + im{X[i]} * + * im{Y[i]}, i=0,...,n-1 ) im{Z} = sum( im{X[i]} * re{Y[i]} - re{X[i]} * + * im{Y[i]}, i=0,...,n-1 ) + * + * The function returns the complex-valued result of the complex + * scalar product at the address of Z. The result is scaled by scaleZ. + * + * \param FIXP_DPK *Z + * Output: pointer to Z + * + * \param const FIXP_DPK *const *const X + * Input: vector of the length n + * + * \param const FIXP_DPK *const *const Y + * Input: vector of the length n + * + * \param int scaleX + * Input: scalefactor of vector X[] + * + * \param int scaleY + * Input: scalefactor of vector Y[] + * + * \param int scaleZ + * Output: scalefactor of vector Z[] + * + * \param int sDim1 + * Input: start index for loop counter in dimension 1 + * + * \param int nDim1 + * Input: loop counter in dimension 1 + * + * \param int sDim2 + * Input: start index for loop counter in dimension 2 + * + * \param int nDim2 + * Input: loop counter in dimension 2 + * + * \return void + */ +void cplx_cplxScalarProduct(FIXP_DPK *const Z, const FIXP_DPK *const *const X, + const FIXP_DPK *const *const Y, const INT scaleX, + const INT scaleY, INT *const scaleZ, + const INT sDim1, const INT nDim1, const INT sDim2, + const INT nDim2); + +/** + * \brief Vector function : Calculate correlation + * + * Description : z[i] = pr12[i] / sqrt(p1[i]*p2[i]) , + * i=0,...,n-1 + * + * \param FIXP_DBL z[] + * Output: vector of length n + * + * \param const FIXP_DBL pr12[] + * Input: vector of the length n + * + * \param const FIXP_DBL p1[] + * Input: vector of the length n + * + * \param const FIXP_DBL p2[] + * Input: vector of the length n + * + * \param int n + * Input: length of vector pr12[], p1[] and p2[] + * + * \return void + */ +void FDKcalcCorrelationVec(FIXP_DBL *const z, const FIXP_DBL *const pr12, + const FIXP_DBL *const p1, const FIXP_DBL *const p2, + const INT n); + +/** + * \brief Vector function : Calculate coherence + * + * Description : z[i] = sqrt( (p12r[i]*p12r[i] + + * p12i[i]*p12i[i]) / (p1[i]*p2[i]) ), i=0,...,n-1 + * + * \param FIXP_DBL z[] + * Output: vector of length n + * + * \param const FIXP_DBL p12r[] + * Input: vector of the length n + * + * \param const FIXP_DBL p12i[] + * Input: vector of the length n + * + * \param const FIXP_DBL p1[] + * Input: vector of the length n + * + * \param const FIXP_DBL p2[] + * Input: vector of the length n + * + * \param int scaleP12[] + * Input: scalefactor of p12r and p12i + * + * \param int scaleP + * Input: scalefactor of p1 and p2 + * + * \param int n + * Input: length of vector p12r[], p12i[], p1[] and p2[] + * + * \return void + */ +void calcCoherenceVec(FIXP_DBL *const z, const FIXP_DBL *const p12r, + const FIXP_DBL *const p12i, const FIXP_DBL *const p1, + const FIXP_DBL *const p2, const INT scaleP12, + const INT scaleP, const INT n); + +/** + * \brief Vector function : Z[j][i] = a[pb] * X[j][i] + b[pb] * + * Y[j][i], j=0,...,nHybridBands-1; i=startTimeSlot,...,nTimeSlots-1; + * pb=0,...,nParameterBands-1 + * + * Description : re{Z[j][i]} = a[pb] * re{X[j][i]} + b[pb] * + * re{Y[j][i]}, j=0,...,nHybridBands-1; i=startTimeSlot,...,nTimeSlots-1; + * pb=0,...,nParameterBands-1 im{Z[j][i]} = a[pb] * im{X[j][i]} + b[pb] * + * im{Y[j][i]}, j=0,...,nHybridBands-1; + * i=startTimeSlot,...,nTimeSlots-1; pb=0,...,nParameterBands-1 + * + * It is allowed to overlay X[] or Y[] with Z[]. The scalefactor + * of channel 1 is updated with the common scalefactor of channel 1 and + * channel 2. + * + * \param FIXP_DPK **Z + * Output: vector of the length nHybridBands*nTimeSlots + * + * \param const FIXP_DBL *a + * Input: vector of length nParameterBands + * + * \param const FIXP_DPK **X + * Input: vector of the length nHybridBands*nTimeSlots + * + * \param const FIXP_DBL *b + * Input: vector of length nParameterBands + * + * \param const FIXP_DPK **Y + * Input: vector of the length nHybridBands*nTimeSlots + * + * \param int scale + * Input: scale of vector a and b + * + * \param int *scaleCh1 + * Input: scale of ch1 + * + * \param int scaleCh2 + * Input: scale of ch2 + * + * \param UCHAR *pParameterBand2HybridBandOffset + * Input: vector of length nParameterBands + * + * \param int nTimeSlots + * Input: number of time slots + * + * \param int startTimeSlot + * Input: start time slot + * + * \return void + */ +void addWeightedCplxVec(FIXP_DPK *const *const Z, const FIXP_DBL *const a, + const FIXP_DPK *const *const X, const FIXP_DBL *const b, + const FIXP_DPK *const *const Y, const INT scale, + INT *const scaleCh1, const INT scaleCh2, + const UCHAR *const pParameterBand2HybridBandOffset, + const INT nParameterBands, const INT nTimeSlots, + const INT startTimeSlot); + +/** + * \brief Vector function : Calculate the headroom of a complex vector + * in a parameter band grid + * + * \param FIXP_DPK **x + * Input: pointer to complex input vector + * + * \param UCHAR *pParameterBand2HybridBandOffset + * Input: pointer to hybrid band offsets + * + * \param int *outScaleFactor + * Input: pointer to ouput scalefactor + * + * \param int startTimeSlot + * Input: start time slot + * + * \param int nTimeSlots + * Input: number of time slot + * + * \param int nParamBands + * Input: number of parameter bands + * + * \return void + */ +void FDKcalcPbScaleFactor(const FIXP_DPK *const *const x, + const UCHAR *const pParameterBand2HybridBandOffset, + INT *const outScaleFactor, const INT startTimeSlot, + const INT nTimeSlots, const INT nParamBands); + +/** + * \brief Vector function : Calculate the common headroom of two + * sparate vectors + * + * \param FIXP_DBL *x + * Input: pointer to first input vector + * + * \param FIXP_DBL *y + * Input: pointer to second input vector + * + * \param int n + * Input: number of samples + * + * \return int headromm in bits + */ +INT FDKcalcScaleFactor(const FIXP_DBL *const x, const FIXP_DBL *const y, + const INT n); + +/** + * \brief Vector function : Calculate the headroom of a complex vector + * + * \param FIXP_DPK *x + * Input: pointer to complex input vector + * + * \param INT startBand + * Input: start band + * + * \param INT bands + * Input: number of bands + * + * \return int headromm in bits + */ +INT FDKcalcScaleFactorDPK(const FIXP_DPK *RESTRICT x, const INT startBand, + const INT bands); + +/* Function / Class Definition ***********************************************/ +template <class T> +inline void FDKmemcpy_flex(T *const dst, const INT dstStride, + const T *const src, const INT srcStride, + const INT nSamples) { + int i; + + for (i = 0; i < nSamples; i++) { + dst[i * dstStride] = src[i * srcStride]; + } +} + +template <class T> +inline void FDKmemset_flex(T *const x, const T c, const INT nSamples) { + int i; + + for (i = 0; i < nSamples; i++) { + x[i] = c; + } +} + +#endif /* SACENC_VECTORFUNCTIONS_H */ |