From 96d6da4e252b06dcfdc041e7df23e86161c33007 Mon Sep 17 00:00:00 2001 From: rihab kouki Date: Tue, 28 Jul 2020 11:24:49 +0100 Subject: Official ARM version: v5.6.0 --- DSP/Source/TransformFunctions/arm_rfft_q15.c | 470 ++++++++++++--------------- 1 file changed, 212 insertions(+), 258 deletions(-) (limited to 'DSP/Source/TransformFunctions/arm_rfft_q15.c') diff --git a/DSP/Source/TransformFunctions/arm_rfft_q15.c b/DSP/Source/TransformFunctions/arm_rfft_q15.c index f85cf30..fdc9bab 100644 --- a/DSP/Source/TransformFunctions/arm_rfft_q15.c +++ b/DSP/Source/TransformFunctions/arm_rfft_q15.c @@ -3,13 +3,13 @@ * Title: arm_rfft_q15.c * Description: RFFT & RIFFT Q15 process function * - * $Date: 27. January 2017 - * $Revision: V.1.5.1 + * $Date: 18. March 2019 + * $Revision: V1.6.0 * * Target Processor: Cortex-M cores * -------------------------------------------------------------------- */ /* - * Copyright (C) 2010-2017 ARM Limited or its affiliates. All rights reserved. + * Copyright (C) 2010-2019 ARM Limited or its affiliates. All rights reserved. * * SPDX-License-Identifier: Apache-2.0 * @@ -33,173 +33,161 @@ * -------------------------------------------------------------------- */ void arm_split_rfft_q15( - q15_t * pSrc, - uint32_t fftLen, - q15_t * pATable, - q15_t * pBTable, - q15_t * pDst, - uint32_t modifier); + q15_t * pSrc, + uint32_t fftLen, + const q15_t * pATable, + const q15_t * pBTable, + q15_t * pDst, + uint32_t modifier); void arm_split_rifft_q15( - q15_t * pSrc, - uint32_t fftLen, - q15_t * pATable, - q15_t * pBTable, - q15_t * pDst, - uint32_t modifier); + q15_t * pSrc, + uint32_t fftLen, + const q15_t * pATable, + const q15_t * pBTable, + q15_t * pDst, + uint32_t modifier); /** -* @addtogroup RealFFT -* @{ -*/ + @addtogroup RealFFT + @{ + */ /** -* @brief Processing function for the Q15 RFFT/RIFFT. -* @param[in] *S points to an instance of the Q15 RFFT/RIFFT structure. -* @param[in] *pSrc points to the input buffer. -* @param[out] *pDst points to the output buffer. -* @return none. -* -* \par Input an output formats: -* \par -* Internally input is downscaled by 2 for every stage to avoid saturations inside CFFT/CIFFT process. -* Hence the output format is different for different RFFT sizes. -* The input and output formats for different RFFT sizes and number of bits to upscale are mentioned in the tables below for RFFT and RIFFT: -* \par -* \image html RFFTQ15.gif "Input and Output Formats for Q15 RFFT" -* \par -* \image html RIFFTQ15.gif "Input and Output Formats for Q15 RIFFT" -*/ + @brief Processing function for the Q15 RFFT/RIFFT. + @param[in] S points to an instance of the Q15 RFFT/RIFFT structure + @param[in] pSrc points to input buffer + @param[out] pDst points to output buffer + @return none + + @par Input an output formats + Internally input is downscaled by 2 for every stage to avoid saturations inside CFFT/CIFFT process. + Hence the output format is different for different RFFT sizes. + The input and output formats for different RFFT sizes and number of bits to upscale are mentioned in the tables below for RFFT and RIFFT: + @par + \image html RFFTQ15.gif "Input and Output Formats for Q15 RFFT" + @par + \image html RIFFTQ15.gif "Input and Output Formats for Q15 RIFFT" + */ void arm_rfft_q15( - const arm_rfft_instance_q15 * S, - q15_t * pSrc, - q15_t * pDst) + const arm_rfft_instance_q15 * S, + q15_t * pSrc, + q15_t * pDst) { - const arm_cfft_instance_q15 *S_CFFT = S->pCfft; - uint32_t i; - uint32_t L2 = S->fftLenReal >> 1; - - /* Calculation of RIFFT of input */ - if (S->ifftFlagR == 1U) - { - /* Real IFFT core process */ - arm_split_rifft_q15(pSrc, L2, S->pTwiddleAReal, - S->pTwiddleBReal, pDst, S->twidCoefRModifier); - - /* Complex IFFT process */ - arm_cfft_q15(S_CFFT, pDst, S->ifftFlagR, S->bitReverseFlagR); + const arm_cfft_instance_q15 *S_CFFT = S->pCfft; + uint32_t L2 = S->fftLenReal >> 1U; + uint32_t i; + + /* Calculation of RIFFT of input */ + if (S->ifftFlagR == 1U) + { + /* Real IFFT core process */ + arm_split_rifft_q15 (pSrc, L2, S->pTwiddleAReal, S->pTwiddleBReal, pDst, S->twidCoefRModifier); + + /* Complex IFFT process */ + arm_cfft_q15 (S_CFFT, pDst, S->ifftFlagR, S->bitReverseFlagR); + + for(i = 0; i < S->fftLenReal; i++) + { + pDst[i] = pDst[i] << 1U; + } + } + else + { + /* Calculation of RFFT of input */ + + /* Complex FFT process */ + arm_cfft_q15 (S_CFFT, pSrc, S->ifftFlagR, S->bitReverseFlagR); + + /* Real FFT core process */ + arm_split_rfft_q15 (pSrc, L2, S->pTwiddleAReal, S->pTwiddleBReal, pDst, S->twidCoefRModifier); + } - for(i=0;ifftLenReal;i++) - { - pDst[i] = pDst[i] << 1; - } - } - else - { - /* Calculation of RFFT of input */ - - /* Complex FFT process */ - arm_cfft_q15(S_CFFT, pSrc, S->ifftFlagR, S->bitReverseFlagR); - - /* Real FFT core process */ - arm_split_rfft_q15(pSrc, L2, S->pTwiddleAReal, - S->pTwiddleBReal, pDst, S->twidCoefRModifier); - } } /** -* @} end of RealFFT group -*/ + @} end of RealFFT group + */ /** -* @brief Core Real FFT process -* @param *pSrc points to the input buffer. -* @param fftLen length of FFT. -* @param *pATable points to the A twiddle Coef buffer. -* @param *pBTable points to the B twiddle Coef buffer. -* @param *pDst points to the output buffer. -* @param modifier twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. -* @return none. -* The function implements a Real FFT -*/ + @brief Core Real FFT process + @param[in] pSrc points to input buffer + @param[in] fftLen length of FFT + @param[in] pATable points to twiddle Coef A buffer + @param[in] pBTable points to twiddle Coef B buffer + @param[out] pDst points to output buffer + @param[in] modifier twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table + @return none + + @par + The function implements a Real FFT + */ void arm_split_rfft_q15( - q15_t * pSrc, - uint32_t fftLen, - q15_t * pATable, - q15_t * pBTable, - q15_t * pDst, - uint32_t modifier) -{ - uint32_t i; /* Loop Counter */ - q31_t outR, outI; /* Temporary variables for output */ - q15_t *pCoefA, *pCoefB; /* Temporary pointers for twiddle factors */ - q15_t *pSrc1, *pSrc2; + q15_t * pSrc, + uint32_t fftLen, + const q15_t * pATable, + const q15_t * pBTable, + q15_t * pDst, + uint32_t modifier) +{ + uint32_t i; /* Loop Counter */ + q31_t outR, outI; /* Temporary variables for output */ + const q15_t *pCoefA, *pCoefB; /* Temporary pointers for twiddle factors */ + q15_t *pSrc1, *pSrc2; #if defined (ARM_MATH_DSP) - q15_t *pD1, *pD2; + q15_t *pD1, *pD2; #endif - // pSrc[2U * fftLen] = pSrc[0]; - // pSrc[(2U * fftLen) + 1U] = pSrc[1]; + /* Init coefficient pointers */ + pCoefA = &pATable[modifier * 2]; + pCoefB = &pBTable[modifier * 2]; - pCoefA = &pATable[modifier * 2U]; - pCoefB = &pBTable[modifier * 2U]; - - pSrc1 = &pSrc[2]; - pSrc2 = &pSrc[(2U * fftLen) - 2U]; + pSrc1 = &pSrc[2]; + pSrc2 = &pSrc[(2U * fftLen) - 2U]; #if defined (ARM_MATH_DSP) - /* Run the below code for Cortex-M4 and Cortex-M3 */ i = 1U; pD1 = pDst + 2; pD2 = pDst + (4U * fftLen) - 2; - for(i = fftLen - 1; i > 0; i--) + for (i = fftLen - 1; i > 0; i--) { /* - outR = (pSrc[2 * i] * pATable[2 * i] - pSrc[2 * i + 1] * pATable[2 * i + 1] - + pSrc[2 * n - 2 * i] * pBTable[2 * i] + - pSrc[2 * n - 2 * i + 1] * pBTable[2 * i + 1]); - */ + outR = ( pSrc[2 * i] * pATable[2 * i] + - pSrc[2 * i + 1] * pATable[2 * i + 1] + + pSrc[2 * n - 2 * i] * pBTable[2 * i] + + pSrc[2 * n - 2 * i + 1] * pBTable[2 * i + 1]); - /* outI = (pIn[2 * i + 1] * pATable[2 * i] + pIn[2 * i] * pATable[2 * i + 1] + - pIn[2 * n - 2 * i] * pBTable[2 * i + 1] - - pIn[2 * n - 2 * i + 1] * pBTable[2 * i]); */ + outI = ( pIn[2 * i + 1] * pATable[2 * i] + + pIn[2 * i] * pATable[2 * i + 1] + + pIn[2 * n - 2 * i] * pBTable[2 * i + 1] + - pIn[2 * n - 2 * i + 1] * pBTable[2 * i]) + */ #ifndef ARM_MATH_BIG_ENDIAN - /* pSrc[2 * i] * pATable[2 * i] - pSrc[2 * i + 1] * pATable[2 * i + 1] */ - outR = __SMUSD(*__SIMD32(pSrc1), *__SIMD32(pCoefA)); - + outR = __SMUSD(read_q15x2 (pSrc1), read_q15x2((q15_t *) pCoefA)); #else - /* -(pSrc[2 * i + 1] * pATable[2 * i + 1] - pSrc[2 * i] * pATable[2 * i]) */ - outR = -(__SMUSD(*__SIMD32(pSrc1), *__SIMD32(pCoefA))); - -#endif /* #ifndef ARM_MATH_BIG_ENDIAN */ + outR = -(__SMUSD(read_q15x2 (pSrc1), read_q15x2((q15_t *) pCoefA))); +#endif /* #ifndef ARM_MATH_BIG_ENDIAN */ - /* pSrc[2 * n - 2 * i] * pBTable[2 * i] + - pSrc[2 * n - 2 * i + 1] * pBTable[2 * i + 1]) */ - outR = __SMLAD(*__SIMD32(pSrc2), *__SIMD32(pCoefB), outR) >> 16U; - - /* pIn[2 * n - 2 * i] * pBTable[2 * i + 1] - - pIn[2 * n - 2 * i + 1] * pBTable[2 * i] */ + /* pSrc[2 * n - 2 * i] * pBTable[2 * i] + pSrc[2 * n - 2 * i + 1] * pBTable[2 * i + 1]) */ + outR = __SMLAD(read_q15x2 (pSrc2), read_q15x2((q15_t *) pCoefB), outR) >> 16U; + /* pIn[2 * n - 2 * i] * pBTable[2 * i + 1] - pIn[2 * n - 2 * i + 1] * pBTable[2 * i] */ #ifndef ARM_MATH_BIG_ENDIAN - - outI = __SMUSDX(*__SIMD32(pSrc2)--, *__SIMD32(pCoefB)); - + outI = __SMUSDX(read_q15x2_da (&pSrc2), read_q15x2((q15_t *) pCoefB)); #else - - outI = __SMUSDX(*__SIMD32(pCoefB), *__SIMD32(pSrc2)--); - -#endif /* #ifndef ARM_MATH_BIG_ENDIAN */ + outI = __SMUSDX(read_q15x2 ((q15_t *) pCoefB), read_q15x2_da (&pSrc2)); +#endif /* #ifndef ARM_MATH_BIG_ENDIAN */ /* (pIn[2 * i + 1] * pATable[2 * i] + pIn[2 * i] * pATable[2 * i + 1] */ - outI = __SMLADX(*__SIMD32(pSrc1)++, *__SIMD32(pCoefA), outI); + outI = __SMLADX(read_q15x2_ia (&pSrc1), read_q15x2 ((q15_t *) pCoefA), outI); /* write output */ *pD1++ = (q15_t) outR; @@ -215,23 +203,23 @@ void arm_split_rfft_q15( pCoefA = pCoefA + (2U * modifier); } - pDst[2U * fftLen] = (pSrc[0] - pSrc[1]) >> 1; - pDst[(2U * fftLen) + 1U] = 0; + pDst[2U * fftLen] = (pSrc[0] - pSrc[1]) >> 1U; + pDst[2U * fftLen + 1U] = 0; - pDst[0] = (pSrc[0] + pSrc[1]) >> 1; + pDst[0] = (pSrc[0] + pSrc[1]) >> 1U; pDst[1] = 0; #else - /* Run the below code for Cortex-M0 */ i = 1U; while (i < fftLen) { /* - outR = (pSrc[2 * i] * pATable[2 * i] - pSrc[2 * i + 1] * pATable[2 * i + 1] - + pSrc[2 * n - 2 * i] * pBTable[2 * i] + - pSrc[2 * n - 2 * i + 1] * pBTable[2 * i + 1]); + outR = ( pSrc[2 * i] * pATable[2 * i] + - pSrc[2 * i + 1] * pATable[2 * i + 1] + + pSrc[2 * n - 2 * i] * pBTable[2 * i] + + pSrc[2 * n - 2 * i + 1] * pBTable[2 * i + 1]); */ outR = *pSrc1 * *pCoefA; @@ -239,10 +227,11 @@ void arm_split_rfft_q15( outR = outR + (*pSrc2 * *pCoefB); outR = (outR + (*(pSrc2 + 1) * *(pCoefB + 1))) >> 16; - - /* outI = (pIn[2 * i + 1] * pATable[2 * i] + pIn[2 * i] * pATable[2 * i + 1] + - pIn[2 * n - 2 * i] * pBTable[2 * i + 1] - - pIn[2 * n - 2 * i + 1] * pBTable[2 * i]); + /* + outI = ( pIn[2 * i + 1] * pATable[2 * i] + + pIn[2 * i] * pATable[2 * i + 1] + + pIn[2 * n - 2 * i] * pBTable[2 * i + 1] + - pIn[2 * n - 2 * i + 1] * pBTable[2 * i]); */ outI = *pSrc2 * *(pCoefB + 1); @@ -256,7 +245,7 @@ void arm_split_rfft_q15( /* write output */ pDst[2U * i] = (q15_t) outR; - pDst[(2U * i) + 1U] = outI >> 16U; + pDst[2U * i + 1U] = outI >> 16U; /* write complex conjugate output */ pDst[(4U * fftLen) - (2U * i)] = (q15_t) outR; @@ -270,7 +259,7 @@ void arm_split_rfft_q15( } pDst[2U * fftLen] = (pSrc[0] - pSrc[1]) >> 1; - pDst[(2U * fftLen) + 1U] = 0; + pDst[2U * fftLen + 1U] = 0; pDst[0] = (pSrc[0] + pSrc[1]) >> 1; pDst[1] = 0; @@ -280,147 +269,112 @@ void arm_split_rfft_q15( /** -* @brief Core Real IFFT process -* @param[in] *pSrc points to the input buffer. -* @param[in] fftLen length of FFT. -* @param[in] *pATable points to the twiddle Coef A buffer. -* @param[in] *pBTable points to the twiddle Coef B buffer. -* @param[out] *pDst points to the output buffer. -* @param[in] modifier twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. -* @return none. -* The function implements a Real IFFT -*/ + @brief Core Real IFFT process + @param[in] pSrc points to input buffer + @param[in] fftLen length of FFT + @param[in] pATable points to twiddle Coef A buffer + @param[in] pBTable points to twiddle Coef B buffer + @param[out] pDst points to output buffer + @param[in] modifier twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table + @return none + + @par + The function implements a Real IFFT + */ + void arm_split_rifft_q15( - q15_t * pSrc, - uint32_t fftLen, - q15_t * pATable, - q15_t * pBTable, - q15_t * pDst, - uint32_t modifier) + q15_t * pSrc, + uint32_t fftLen, + const q15_t * pATable, + const q15_t * pBTable, + q15_t * pDst, + uint32_t modifier) { - uint32_t i; /* Loop Counter */ - q31_t outR, outI; /* Temporary variables for output */ - q15_t *pCoefA, *pCoefB; /* Temporary pointers for twiddle factors */ - q15_t *pSrc1, *pSrc2; - q15_t *pDst1 = &pDst[0]; - - pCoefA = &pATable[0]; - pCoefB = &pBTable[0]; - - pSrc1 = &pSrc[0]; - pSrc2 = &pSrc[2U * fftLen]; + uint32_t i; /* Loop Counter */ + q31_t outR, outI; /* Temporary variables for output */ + const q15_t *pCoefA, *pCoefB; /* Temporary pointers for twiddle factors */ + q15_t *pSrc1, *pSrc2; + q15_t *pDst1 = &pDst[0]; + + pCoefA = &pATable[0]; + pCoefB = &pBTable[0]; + + pSrc1 = &pSrc[0]; + pSrc2 = &pSrc[2 * fftLen]; + + i = fftLen; + while (i > 0U) + { + /* + outR = ( pIn[2 * i] * pATable[2 * i] + + pIn[2 * i + 1] * pATable[2 * i + 1] + + pIn[2 * n - 2 * i] * pBTable[2 * i] + - pIn[2 * n - 2 * i + 1] * pBTable[2 * i + 1]); + + outI = ( pIn[2 * i + 1] * pATable[2 * i] + - pIn[2 * i] * pATable[2 * i + 1] + - pIn[2 * n - 2 * i] * pBTable[2 * i + 1] + - pIn[2 * n - 2 * i + 1] * pBTable[2 * i]); + */ #if defined (ARM_MATH_DSP) - /* Run the below code for Cortex-M4 and Cortex-M3 */ - i = fftLen; - - while (i > 0U) - { - /* - outR = (pIn[2 * i] * pATable[2 * i] + pIn[2 * i + 1] * pATable[2 * i + 1] + - pIn[2 * n - 2 * i] * pBTable[2 * i] - - pIn[2 * n - 2 * i + 1] * pBTable[2 * i + 1]); - - outI = (pIn[2 * i + 1] * pATable[2 * i] - pIn[2 * i] * pATable[2 * i + 1] - - pIn[2 * n - 2 * i] * pBTable[2 * i + 1] - - pIn[2 * n - 2 * i + 1] * pBTable[2 * i]); - */ - - #ifndef ARM_MATH_BIG_ENDIAN - - /* pIn[2 * n - 2 * i] * pBTable[2 * i] - - pIn[2 * n - 2 * i + 1] * pBTable[2 * i + 1]) */ - outR = __SMUSD(*__SIMD32(pSrc2), *__SIMD32(pCoefB)); - + /* pIn[2 * n - 2 * i] * pBTable[2 * i] - pIn[2 * n - 2 * i + 1] * pBTable[2 * i + 1]) */ + outR = __SMUSD(read_q15x2(pSrc2), read_q15x2((q15_t *) pCoefB)); #else + /* -(-pIn[2 * n - 2 * i] * pBTable[2 * i] + pIn[2 * n - 2 * i + 1] * pBTable[2 * i + 1])) */ + outR = -(__SMUSD(read_q15x2(pSrc2), read_q15x2((q15_t *) pCoefB))); +#endif /* #ifndef ARM_MATH_BIG_ENDIAN */ - /* -(-pIn[2 * n - 2 * i] * pBTable[2 * i] + - pIn[2 * n - 2 * i + 1] * pBTable[2 * i + 1])) */ - outR = -(__SMUSD(*__SIMD32(pSrc2), *__SIMD32(pCoefB))); - -#endif /* #ifndef ARM_MATH_BIG_ENDIAN */ + /* pIn[2 * i] * pATable[2 * i] + pIn[2 * i + 1] * pATable[2 * i + 1] + pIn[2 * n - 2 * i] * pBTable[2 * i] */ + outR = __SMLAD(read_q15x2(pSrc1), read_q15x2 ((q15_t *) pCoefA), outR) >> 16U; - /* pIn[2 * i] * pATable[2 * i] + pIn[2 * i + 1] * pATable[2 * i + 1] + - pIn[2 * n - 2 * i] * pBTable[2 * i] */ - outR = __SMLAD(*__SIMD32(pSrc1), *__SIMD32(pCoefA), outR) >> 16U; - - /* - -pIn[2 * n - 2 * i] * pBTable[2 * i + 1] + - pIn[2 * n - 2 * i + 1] * pBTable[2 * i] */ - outI = __SMUADX(*__SIMD32(pSrc2)--, *__SIMD32(pCoefB)); - - /* pIn[2 * i + 1] * pATable[2 * i] - pIn[2 * i] * pATable[2 * i + 1] */ + /* -pIn[2 * n - 2 * i] * pBTable[2 * i + 1] + pIn[2 * n - 2 * i + 1] * pBTable[2 * i] */ + outI = __SMUADX(read_q15x2_da (&pSrc2), read_q15x2((q15_t *) pCoefB)); + /* pIn[2 * i + 1] * pATable[2 * i] - pIn[2 * i] * pATable[2 * i + 1] */ #ifndef ARM_MATH_BIG_ENDIAN - - outI = __SMLSDX(*__SIMD32(pCoefA), *__SIMD32(pSrc1)++, -outI); - + outI = __SMLSDX(read_q15x2 ((q15_t *) pCoefA), read_q15x2_ia (&pSrc1), -outI); #else + outI = __SMLSDX(read_q15x2_ia (&pSrc1), read_q15x2 ((q15_t *) pCoefA), -outI); +#endif /* #ifndef ARM_MATH_BIG_ENDIAN */ - outI = __SMLSDX(*__SIMD32(pSrc1)++, *__SIMD32(pCoefA), -outI); - -#endif /* #ifndef ARM_MATH_BIG_ENDIAN */ - /* write output */ - + /* write output */ #ifndef ARM_MATH_BIG_ENDIAN - - *__SIMD32(pDst1)++ = __PKHBT(outR, (outI >> 16U), 16); - + write_q15x2_ia (&pDst1, __PKHBT(outR, (outI >> 16U), 16)); #else + write_q15x2_ia (&pDst1, __PKHBT((outI >> 16U), outR, 16)); +#endif /* #ifndef ARM_MATH_BIG_ENDIAN */ - *__SIMD32(pDst1)++ = __PKHBT((outI >> 16U), outR, 16); -#endif /* #ifndef ARM_MATH_BIG_ENDIAN */ +#else /* #if defined (ARM_MATH_DSP) */ - /* update coefficient pointer */ - pCoefB = pCoefB + (2U * modifier); - pCoefA = pCoefA + (2U * modifier); + outR = *pSrc2 * *pCoefB; + outR = outR - (*(pSrc2 + 1) * *(pCoefB + 1)); + outR = outR + (*pSrc1 * *pCoefA); + outR = (outR + (*(pSrc1 + 1) * *(pCoefA + 1))) >> 16; - i--; - } -#else - /* Run the below code for Cortex-M0 */ - i = fftLen; + outI = *(pSrc1 + 1) * *pCoefA; + outI = outI - (*pSrc1 * *(pCoefA + 1)); + outI = outI - (*pSrc2 * *(pCoefB + 1)); + outI = outI - (*(pSrc2 + 1) * *(pCoefB)); - while (i > 0U) - { - /* - outR = (pIn[2 * i] * pATable[2 * i] + pIn[2 * i + 1] * pATable[2 * i + 1] + - pIn[2 * n - 2 * i] * pBTable[2 * i] - - pIn[2 * n - 2 * i + 1] * pBTable[2 * i + 1]); - */ + /* update input pointers */ + pSrc1 += 2U; + pSrc2 -= 2U; - outR = *pSrc2 * *pCoefB; - outR = outR - (*(pSrc2 + 1) * *(pCoefB + 1)); - outR = outR + (*pSrc1 * *pCoefA); - outR = (outR + (*(pSrc1 + 1) * *(pCoefA + 1))) >> 16; + /* write output */ + *pDst1++ = (q15_t) outR; + *pDst1++ = (q15_t) (outI >> 16); - /* - outI = (pIn[2 * i + 1] * pATable[2 * i] - pIn[2 * i] * pATable[2 * i + 1] - - pIn[2 * n - 2 * i] * pBTable[2 * i + 1] - - pIn[2 * n - 2 * i + 1] * pBTable[2 * i]); - */ - - outI = *(pSrc1 + 1) * *pCoefA; - outI = outI - (*pSrc1 * *(pCoefA + 1)); - outI = outI - (*pSrc2 * *(pCoefB + 1)); - outI = outI - (*(pSrc2 + 1) * *(pCoefB)); - - /* update input pointers */ - pSrc1 += 2U; - pSrc2 -= 2U; +#endif /* #if defined (ARM_MATH_DSP) */ - /* write output */ - *pDst1++ = (q15_t) outR; - *pDst1++ = (q15_t) (outI >> 16); + /* update coefficient pointer */ + pCoefB = pCoefB + (2 * modifier); + pCoefA = pCoefA + (2 * modifier); - /* update coefficient pointer */ - pCoefB = pCoefB + (2U * modifier); - pCoefA = pCoefA + (2U * modifier); + i--; + } - i--; - } -#endif /* #if defined (ARM_MATH_DSP) */ } -- cgit