mbed-os 6.10 versione
cmsis_dsp/TransformFunctions/arm_cfft_radix4_q31.c@1:fdd22bb7aa52, 2012-11-28 (annotated)
- Committer:
- emilmont
- Date:
- Wed Nov 28 12:30:09 2012 +0000
- Revision:
- 1:fdd22bb7aa52
- Child:
- 2:da51fb522205
DSP library code
Who changed what in which revision?
User | Revision | Line number | New contents of line |
---|---|---|---|
emilmont | 1:fdd22bb7aa52 | 1 | /* ---------------------------------------------------------------------- |
emilmont | 1:fdd22bb7aa52 | 2 | * Copyright (C) 2010 ARM Limited. All rights reserved. |
emilmont | 1:fdd22bb7aa52 | 3 | * |
emilmont | 1:fdd22bb7aa52 | 4 | * $Date: 15. February 2012 |
emilmont | 1:fdd22bb7aa52 | 5 | * $Revision: V1.1.0 |
emilmont | 1:fdd22bb7aa52 | 6 | * |
emilmont | 1:fdd22bb7aa52 | 7 | * Project: CMSIS DSP Library |
emilmont | 1:fdd22bb7aa52 | 8 | * Title: arm_cfft_radix4_q31.c |
emilmont | 1:fdd22bb7aa52 | 9 | * |
emilmont | 1:fdd22bb7aa52 | 10 | * Description: This file has function definition of Radix-4 FFT & IFFT function and |
emilmont | 1:fdd22bb7aa52 | 11 | * In-place bit reversal using bit reversal table |
emilmont | 1:fdd22bb7aa52 | 12 | * |
emilmont | 1:fdd22bb7aa52 | 13 | * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0 |
emilmont | 1:fdd22bb7aa52 | 14 | * |
emilmont | 1:fdd22bb7aa52 | 15 | * Version 1.1.0 2012/02/15 |
emilmont | 1:fdd22bb7aa52 | 16 | * Updated with more optimizations, bug fixes and minor API changes. |
emilmont | 1:fdd22bb7aa52 | 17 | * |
emilmont | 1:fdd22bb7aa52 | 18 | * Version 1.0.10 2011/7/15 |
emilmont | 1:fdd22bb7aa52 | 19 | * Big Endian support added and Merged M0 and M3/M4 Source code. |
emilmont | 1:fdd22bb7aa52 | 20 | * |
emilmont | 1:fdd22bb7aa52 | 21 | * Version 1.0.3 2010/11/29 |
emilmont | 1:fdd22bb7aa52 | 22 | * Re-organized the CMSIS folders and updated documentation. |
emilmont | 1:fdd22bb7aa52 | 23 | * |
emilmont | 1:fdd22bb7aa52 | 24 | * Version 1.0.2 2010/11/11 |
emilmont | 1:fdd22bb7aa52 | 25 | * Documentation updated. |
emilmont | 1:fdd22bb7aa52 | 26 | * |
emilmont | 1:fdd22bb7aa52 | 27 | * Version 1.0.1 2010/10/05 |
emilmont | 1:fdd22bb7aa52 | 28 | * Production release and review comments incorporated. |
emilmont | 1:fdd22bb7aa52 | 29 | * |
emilmont | 1:fdd22bb7aa52 | 30 | * Version 1.0.0 2010/09/20 |
emilmont | 1:fdd22bb7aa52 | 31 | * Production release and review comments incorporated. |
emilmont | 1:fdd22bb7aa52 | 32 | * |
emilmont | 1:fdd22bb7aa52 | 33 | * Version 0.0.5 2010/04/26 |
emilmont | 1:fdd22bb7aa52 | 34 | * incorporated review comments and updated with latest CMSIS layer |
emilmont | 1:fdd22bb7aa52 | 35 | * |
emilmont | 1:fdd22bb7aa52 | 36 | * Version 0.0.3 2010/03/10 |
emilmont | 1:fdd22bb7aa52 | 37 | * Initial version |
emilmont | 1:fdd22bb7aa52 | 38 | * -------------------------------------------------------------------- */ |
emilmont | 1:fdd22bb7aa52 | 39 | #include "arm_math.h" |
emilmont | 1:fdd22bb7aa52 | 40 | |
emilmont | 1:fdd22bb7aa52 | 41 | |
emilmont | 1:fdd22bb7aa52 | 42 | /** |
emilmont | 1:fdd22bb7aa52 | 43 | * @ingroup groupTransforms |
emilmont | 1:fdd22bb7aa52 | 44 | */ |
emilmont | 1:fdd22bb7aa52 | 45 | |
emilmont | 1:fdd22bb7aa52 | 46 | /** |
emilmont | 1:fdd22bb7aa52 | 47 | * @addtogroup Radix4_CFFT_CIFFT |
emilmont | 1:fdd22bb7aa52 | 48 | * @{ |
emilmont | 1:fdd22bb7aa52 | 49 | */ |
emilmont | 1:fdd22bb7aa52 | 50 | |
emilmont | 1:fdd22bb7aa52 | 51 | /** |
emilmont | 1:fdd22bb7aa52 | 52 | * @details |
emilmont | 1:fdd22bb7aa52 | 53 | * @brief Processing function for the Q31 CFFT/CIFFT. |
emilmont | 1:fdd22bb7aa52 | 54 | * @param[in] *S points to an instance of the Q31 CFFT/CIFFT structure. |
emilmont | 1:fdd22bb7aa52 | 55 | * @param[in, out] *pSrc points to the complex data buffer of size <code>2*fftLen</code>. Processing occurs in-place. |
emilmont | 1:fdd22bb7aa52 | 56 | * @return none. |
emilmont | 1:fdd22bb7aa52 | 57 | * |
emilmont | 1:fdd22bb7aa52 | 58 | * \par Input and output formats: |
emilmont | 1:fdd22bb7aa52 | 59 | * \par |
emilmont | 1:fdd22bb7aa52 | 60 | * Internally input is downscaled by 2 for every stage to avoid saturations inside CFFT/CIFFT process. |
emilmont | 1:fdd22bb7aa52 | 61 | * Hence the output format is different for different FFT sizes. |
emilmont | 1:fdd22bb7aa52 | 62 | * The input and output formats for different FFT sizes and number of bits to upscale are mentioned in the tables below for CFFT and CIFFT: |
emilmont | 1:fdd22bb7aa52 | 63 | * \par |
emilmont | 1:fdd22bb7aa52 | 64 | * \image html CFFTQ31.gif "Input and Output Formats for Q31 CFFT" |
emilmont | 1:fdd22bb7aa52 | 65 | * \image html CIFFTQ31.gif "Input and Output Formats for Q31 CIFFT" |
emilmont | 1:fdd22bb7aa52 | 66 | * |
emilmont | 1:fdd22bb7aa52 | 67 | */ |
emilmont | 1:fdd22bb7aa52 | 68 | |
emilmont | 1:fdd22bb7aa52 | 69 | void arm_cfft_radix4_q31( |
emilmont | 1:fdd22bb7aa52 | 70 | const arm_cfft_radix4_instance_q31 * S, |
emilmont | 1:fdd22bb7aa52 | 71 | q31_t * pSrc) |
emilmont | 1:fdd22bb7aa52 | 72 | { |
emilmont | 1:fdd22bb7aa52 | 73 | if(S->ifftFlag == 1u) |
emilmont | 1:fdd22bb7aa52 | 74 | { |
emilmont | 1:fdd22bb7aa52 | 75 | /* Complex IFFT radix-4 */ |
emilmont | 1:fdd22bb7aa52 | 76 | arm_radix4_butterfly_inverse_q31(pSrc, S->fftLen, S->pTwiddle, |
emilmont | 1:fdd22bb7aa52 | 77 | S->twidCoefModifier); |
emilmont | 1:fdd22bb7aa52 | 78 | } |
emilmont | 1:fdd22bb7aa52 | 79 | else |
emilmont | 1:fdd22bb7aa52 | 80 | { |
emilmont | 1:fdd22bb7aa52 | 81 | /* Complex FFT radix-4 */ |
emilmont | 1:fdd22bb7aa52 | 82 | arm_radix4_butterfly_q31(pSrc, S->fftLen, S->pTwiddle, |
emilmont | 1:fdd22bb7aa52 | 83 | S->twidCoefModifier); |
emilmont | 1:fdd22bb7aa52 | 84 | } |
emilmont | 1:fdd22bb7aa52 | 85 | |
emilmont | 1:fdd22bb7aa52 | 86 | |
emilmont | 1:fdd22bb7aa52 | 87 | if(S->bitReverseFlag == 1u) |
emilmont | 1:fdd22bb7aa52 | 88 | { |
emilmont | 1:fdd22bb7aa52 | 89 | /* Bit Reversal */ |
emilmont | 1:fdd22bb7aa52 | 90 | arm_bitreversal_q31(pSrc, S->fftLen, S->bitRevFactor, S->pBitRevTable); |
emilmont | 1:fdd22bb7aa52 | 91 | } |
emilmont | 1:fdd22bb7aa52 | 92 | |
emilmont | 1:fdd22bb7aa52 | 93 | } |
emilmont | 1:fdd22bb7aa52 | 94 | |
emilmont | 1:fdd22bb7aa52 | 95 | /** |
emilmont | 1:fdd22bb7aa52 | 96 | * @} end of Radix4_CFFT_CIFFT group |
emilmont | 1:fdd22bb7aa52 | 97 | */ |
emilmont | 1:fdd22bb7aa52 | 98 | |
emilmont | 1:fdd22bb7aa52 | 99 | /* |
emilmont | 1:fdd22bb7aa52 | 100 | * Radix-4 FFT algorithm used is : |
emilmont | 1:fdd22bb7aa52 | 101 | * |
emilmont | 1:fdd22bb7aa52 | 102 | * Input real and imaginary data: |
emilmont | 1:fdd22bb7aa52 | 103 | * x(n) = xa + j * ya |
emilmont | 1:fdd22bb7aa52 | 104 | * x(n+N/4 ) = xb + j * yb |
emilmont | 1:fdd22bb7aa52 | 105 | * x(n+N/2 ) = xc + j * yc |
emilmont | 1:fdd22bb7aa52 | 106 | * x(n+3N 4) = xd + j * yd |
emilmont | 1:fdd22bb7aa52 | 107 | * |
emilmont | 1:fdd22bb7aa52 | 108 | * |
emilmont | 1:fdd22bb7aa52 | 109 | * Output real and imaginary data: |
emilmont | 1:fdd22bb7aa52 | 110 | * x(4r) = xa'+ j * ya' |
emilmont | 1:fdd22bb7aa52 | 111 | * x(4r+1) = xb'+ j * yb' |
emilmont | 1:fdd22bb7aa52 | 112 | * x(4r+2) = xc'+ j * yc' |
emilmont | 1:fdd22bb7aa52 | 113 | * x(4r+3) = xd'+ j * yd' |
emilmont | 1:fdd22bb7aa52 | 114 | * |
emilmont | 1:fdd22bb7aa52 | 115 | * |
emilmont | 1:fdd22bb7aa52 | 116 | * Twiddle factors for radix-4 FFT: |
emilmont | 1:fdd22bb7aa52 | 117 | * Wn = co1 + j * (- si1) |
emilmont | 1:fdd22bb7aa52 | 118 | * W2n = co2 + j * (- si2) |
emilmont | 1:fdd22bb7aa52 | 119 | * W3n = co3 + j * (- si3) |
emilmont | 1:fdd22bb7aa52 | 120 | * |
emilmont | 1:fdd22bb7aa52 | 121 | * Butterfly implementation: |
emilmont | 1:fdd22bb7aa52 | 122 | * xa' = xa + xb + xc + xd |
emilmont | 1:fdd22bb7aa52 | 123 | * ya' = ya + yb + yc + yd |
emilmont | 1:fdd22bb7aa52 | 124 | * xb' = (xa+yb-xc-yd)* co1 + (ya-xb-yc+xd)* (si1) |
emilmont | 1:fdd22bb7aa52 | 125 | * yb' = (ya-xb-yc+xd)* co1 - (xa+yb-xc-yd)* (si1) |
emilmont | 1:fdd22bb7aa52 | 126 | * xc' = (xa-xb+xc-xd)* co2 + (ya-yb+yc-yd)* (si2) |
emilmont | 1:fdd22bb7aa52 | 127 | * yc' = (ya-yb+yc-yd)* co2 - (xa-xb+xc-xd)* (si2) |
emilmont | 1:fdd22bb7aa52 | 128 | * xd' = (xa-yb-xc+yd)* co3 + (ya+xb-yc-xd)* (si3) |
emilmont | 1:fdd22bb7aa52 | 129 | * yd' = (ya+xb-yc-xd)* co3 - (xa-yb-xc+yd)* (si3) |
emilmont | 1:fdd22bb7aa52 | 130 | * |
emilmont | 1:fdd22bb7aa52 | 131 | */ |
emilmont | 1:fdd22bb7aa52 | 132 | |
emilmont | 1:fdd22bb7aa52 | 133 | /** |
emilmont | 1:fdd22bb7aa52 | 134 | * @brief Core function for the Q31 CFFT butterfly process. |
emilmont | 1:fdd22bb7aa52 | 135 | * @param[in, out] *pSrc points to the in-place buffer of Q31 data type. |
emilmont | 1:fdd22bb7aa52 | 136 | * @param[in] fftLen length of the FFT. |
emilmont | 1:fdd22bb7aa52 | 137 | * @param[in] *pCoef points to twiddle coefficient buffer. |
emilmont | 1:fdd22bb7aa52 | 138 | * @param[in] twidCoefModifier twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. |
emilmont | 1:fdd22bb7aa52 | 139 | * @return none. |
emilmont | 1:fdd22bb7aa52 | 140 | */ |
emilmont | 1:fdd22bb7aa52 | 141 | |
emilmont | 1:fdd22bb7aa52 | 142 | void arm_radix4_butterfly_q31( |
emilmont | 1:fdd22bb7aa52 | 143 | q31_t * pSrc, |
emilmont | 1:fdd22bb7aa52 | 144 | uint32_t fftLen, |
emilmont | 1:fdd22bb7aa52 | 145 | q31_t * pCoef, |
emilmont | 1:fdd22bb7aa52 | 146 | uint32_t twidCoefModifier) |
emilmont | 1:fdd22bb7aa52 | 147 | { |
emilmont | 1:fdd22bb7aa52 | 148 | uint32_t n1, n2, ia1, ia2, ia3, i0, i1, i2, i3, j, k; |
emilmont | 1:fdd22bb7aa52 | 149 | q31_t t1, t2, r1, r2, s1, s2, co1, co2, co3, si1, si2, si3; |
emilmont | 1:fdd22bb7aa52 | 150 | |
emilmont | 1:fdd22bb7aa52 | 151 | q31_t xa, xb, xc, xd; |
emilmont | 1:fdd22bb7aa52 | 152 | q31_t ya, yb, yc, yd; |
emilmont | 1:fdd22bb7aa52 | 153 | q31_t xa_out, xb_out, xc_out, xd_out; |
emilmont | 1:fdd22bb7aa52 | 154 | q31_t ya_out, yb_out, yc_out, yd_out; |
emilmont | 1:fdd22bb7aa52 | 155 | |
emilmont | 1:fdd22bb7aa52 | 156 | q31_t *ptr1; |
emilmont | 1:fdd22bb7aa52 | 157 | q63_t xaya, xbyb, xcyc, xdyd; |
emilmont | 1:fdd22bb7aa52 | 158 | /* Total process is divided into three stages */ |
emilmont | 1:fdd22bb7aa52 | 159 | |
emilmont | 1:fdd22bb7aa52 | 160 | /* process first stage, middle stages, & last stage */ |
emilmont | 1:fdd22bb7aa52 | 161 | |
emilmont | 1:fdd22bb7aa52 | 162 | |
emilmont | 1:fdd22bb7aa52 | 163 | /* start of first stage process */ |
emilmont | 1:fdd22bb7aa52 | 164 | |
emilmont | 1:fdd22bb7aa52 | 165 | /* Initializations for the first stage */ |
emilmont | 1:fdd22bb7aa52 | 166 | n2 = fftLen; |
emilmont | 1:fdd22bb7aa52 | 167 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 168 | /* n2 = fftLen/4 */ |
emilmont | 1:fdd22bb7aa52 | 169 | n2 >>= 2u; |
emilmont | 1:fdd22bb7aa52 | 170 | i0 = 0u; |
emilmont | 1:fdd22bb7aa52 | 171 | ia1 = 0u; |
emilmont | 1:fdd22bb7aa52 | 172 | |
emilmont | 1:fdd22bb7aa52 | 173 | j = n2; |
emilmont | 1:fdd22bb7aa52 | 174 | |
emilmont | 1:fdd22bb7aa52 | 175 | /* Calculation of first stage */ |
emilmont | 1:fdd22bb7aa52 | 176 | do |
emilmont | 1:fdd22bb7aa52 | 177 | { |
emilmont | 1:fdd22bb7aa52 | 178 | /* index calculation for the input as, */ |
emilmont | 1:fdd22bb7aa52 | 179 | /* pSrc[i0 + 0], pSrc[i0 + fftLen/4], pSrc[i0 + fftLen/2u], pSrc[i0 + 3fftLen/4] */ |
emilmont | 1:fdd22bb7aa52 | 180 | i1 = i0 + n2; |
emilmont | 1:fdd22bb7aa52 | 181 | i2 = i1 + n2; |
emilmont | 1:fdd22bb7aa52 | 182 | i3 = i2 + n2; |
emilmont | 1:fdd22bb7aa52 | 183 | |
emilmont | 1:fdd22bb7aa52 | 184 | /* input is in 1.31(q31) format and provide 4 guard bits for the input */ |
emilmont | 1:fdd22bb7aa52 | 185 | |
emilmont | 1:fdd22bb7aa52 | 186 | /* Butterfly implementation */ |
emilmont | 1:fdd22bb7aa52 | 187 | /* xa + xc */ |
emilmont | 1:fdd22bb7aa52 | 188 | r1 = (pSrc[(2u * i0)] >> 4u) + (pSrc[(2u * i2)] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 189 | /* xa - xc */ |
emilmont | 1:fdd22bb7aa52 | 190 | r2 = (pSrc[2u * i0] >> 4u) - (pSrc[2u * i2] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 191 | |
emilmont | 1:fdd22bb7aa52 | 192 | /* xb + xd */ |
emilmont | 1:fdd22bb7aa52 | 193 | t1 = (pSrc[2u * i1] >> 4u) + (pSrc[2u * i3] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 194 | |
emilmont | 1:fdd22bb7aa52 | 195 | /* ya + yc */ |
emilmont | 1:fdd22bb7aa52 | 196 | s1 = (pSrc[(2u * i0) + 1u] >> 4u) + (pSrc[(2u * i2) + 1u] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 197 | /* ya - yc */ |
emilmont | 1:fdd22bb7aa52 | 198 | s2 = (pSrc[(2u * i0) + 1u] >> 4u) - (pSrc[(2u * i2) + 1u] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 199 | |
emilmont | 1:fdd22bb7aa52 | 200 | /* xa' = xa + xb + xc + xd */ |
emilmont | 1:fdd22bb7aa52 | 201 | pSrc[2u * i0] = (r1 + t1); |
emilmont | 1:fdd22bb7aa52 | 202 | /* (xa + xc) - (xb + xd) */ |
emilmont | 1:fdd22bb7aa52 | 203 | r1 = r1 - t1; |
emilmont | 1:fdd22bb7aa52 | 204 | /* yb + yd */ |
emilmont | 1:fdd22bb7aa52 | 205 | t2 = (pSrc[(2u * i1) + 1u] >> 4u) + (pSrc[(2u * i3) + 1u] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 206 | |
emilmont | 1:fdd22bb7aa52 | 207 | /* ya' = ya + yb + yc + yd */ |
emilmont | 1:fdd22bb7aa52 | 208 | pSrc[(2u * i0) + 1u] = (s1 + t2); |
emilmont | 1:fdd22bb7aa52 | 209 | |
emilmont | 1:fdd22bb7aa52 | 210 | /* (ya + yc) - (yb + yd) */ |
emilmont | 1:fdd22bb7aa52 | 211 | s1 = s1 - t2; |
emilmont | 1:fdd22bb7aa52 | 212 | |
emilmont | 1:fdd22bb7aa52 | 213 | /* yb - yd */ |
emilmont | 1:fdd22bb7aa52 | 214 | t1 = (pSrc[(2u * i1) + 1u] >> 4u) - (pSrc[(2u * i3) + 1u] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 215 | /* xb - xd */ |
emilmont | 1:fdd22bb7aa52 | 216 | t2 = (pSrc[2u * i1] >> 4u) - (pSrc[2u * i3] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 217 | |
emilmont | 1:fdd22bb7aa52 | 218 | /* index calculation for the coefficients */ |
emilmont | 1:fdd22bb7aa52 | 219 | ia2 = 2u * ia1; |
emilmont | 1:fdd22bb7aa52 | 220 | co2 = pCoef[ia2 * 2u]; |
emilmont | 1:fdd22bb7aa52 | 221 | si2 = pCoef[(ia2 * 2u) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 222 | |
emilmont | 1:fdd22bb7aa52 | 223 | /* xc' = (xa-xb+xc-xd)co2 + (ya-yb+yc-yd)(si2) */ |
emilmont | 1:fdd22bb7aa52 | 224 | pSrc[2u * i1] = (((int32_t) (((q63_t) r1 * co2) >> 32)) + |
emilmont | 1:fdd22bb7aa52 | 225 | ((int32_t) (((q63_t) s1 * si2) >> 32))) << 1u; |
emilmont | 1:fdd22bb7aa52 | 226 | |
emilmont | 1:fdd22bb7aa52 | 227 | /* yc' = (ya-yb+yc-yd)co2 - (xa-xb+xc-xd)(si2) */ |
emilmont | 1:fdd22bb7aa52 | 228 | pSrc[(2u * i1) + 1u] = (((int32_t) (((q63_t) s1 * co2) >> 32)) - |
emilmont | 1:fdd22bb7aa52 | 229 | ((int32_t) (((q63_t) r1 * si2) >> 32))) << 1u; |
emilmont | 1:fdd22bb7aa52 | 230 | |
emilmont | 1:fdd22bb7aa52 | 231 | /* (xa - xc) + (yb - yd) */ |
emilmont | 1:fdd22bb7aa52 | 232 | r1 = r2 + t1; |
emilmont | 1:fdd22bb7aa52 | 233 | /* (xa - xc) - (yb - yd) */ |
emilmont | 1:fdd22bb7aa52 | 234 | r2 = r2 - t1; |
emilmont | 1:fdd22bb7aa52 | 235 | |
emilmont | 1:fdd22bb7aa52 | 236 | /* (ya - yc) - (xb - xd) */ |
emilmont | 1:fdd22bb7aa52 | 237 | s1 = s2 - t2; |
emilmont | 1:fdd22bb7aa52 | 238 | /* (ya - yc) + (xb - xd) */ |
emilmont | 1:fdd22bb7aa52 | 239 | s2 = s2 + t2; |
emilmont | 1:fdd22bb7aa52 | 240 | |
emilmont | 1:fdd22bb7aa52 | 241 | co1 = pCoef[ia1 * 2u]; |
emilmont | 1:fdd22bb7aa52 | 242 | si1 = pCoef[(ia1 * 2u) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 243 | |
emilmont | 1:fdd22bb7aa52 | 244 | /* xb' = (xa+yb-xc-yd)co1 + (ya-xb-yc+xd)(si1) */ |
emilmont | 1:fdd22bb7aa52 | 245 | pSrc[2u * i2] = (((int32_t) (((q63_t) r1 * co1) >> 32)) + |
emilmont | 1:fdd22bb7aa52 | 246 | ((int32_t) (((q63_t) s1 * si1) >> 32))) << 1u; |
emilmont | 1:fdd22bb7aa52 | 247 | |
emilmont | 1:fdd22bb7aa52 | 248 | /* yb' = (ya-xb-yc+xd)co1 - (xa+yb-xc-yd)(si1) */ |
emilmont | 1:fdd22bb7aa52 | 249 | pSrc[(2u * i2) + 1u] = (((int32_t) (((q63_t) s1 * co1) >> 32)) - |
emilmont | 1:fdd22bb7aa52 | 250 | ((int32_t) (((q63_t) r1 * si1) >> 32))) << 1u; |
emilmont | 1:fdd22bb7aa52 | 251 | |
emilmont | 1:fdd22bb7aa52 | 252 | /* index calculation for the coefficients */ |
emilmont | 1:fdd22bb7aa52 | 253 | ia3 = 3u * ia1; |
emilmont | 1:fdd22bb7aa52 | 254 | co3 = pCoef[ia3 * 2u]; |
emilmont | 1:fdd22bb7aa52 | 255 | si3 = pCoef[(ia3 * 2u) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 256 | |
emilmont | 1:fdd22bb7aa52 | 257 | /* xd' = (xa-yb-xc+yd)co3 + (ya+xb-yc-xd)(si3) */ |
emilmont | 1:fdd22bb7aa52 | 258 | pSrc[2u * i3] = (((int32_t) (((q63_t) r2 * co3) >> 32)) + |
emilmont | 1:fdd22bb7aa52 | 259 | ((int32_t) (((q63_t) s2 * si3) >> 32))) << 1u; |
emilmont | 1:fdd22bb7aa52 | 260 | |
emilmont | 1:fdd22bb7aa52 | 261 | /* yd' = (ya+xb-yc-xd)co3 - (xa-yb-xc+yd)(si3) */ |
emilmont | 1:fdd22bb7aa52 | 262 | pSrc[(2u * i3) + 1u] = (((int32_t) (((q63_t) s2 * co3) >> 32)) - |
emilmont | 1:fdd22bb7aa52 | 263 | ((int32_t) (((q63_t) r2 * si3) >> 32))) << 1u; |
emilmont | 1:fdd22bb7aa52 | 264 | |
emilmont | 1:fdd22bb7aa52 | 265 | /* Twiddle coefficients index modifier */ |
emilmont | 1:fdd22bb7aa52 | 266 | ia1 = ia1 + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 267 | |
emilmont | 1:fdd22bb7aa52 | 268 | /* Updating input index */ |
emilmont | 1:fdd22bb7aa52 | 269 | i0 = i0 + 1u; |
emilmont | 1:fdd22bb7aa52 | 270 | |
emilmont | 1:fdd22bb7aa52 | 271 | } while(--j); |
emilmont | 1:fdd22bb7aa52 | 272 | |
emilmont | 1:fdd22bb7aa52 | 273 | /* end of first stage process */ |
emilmont | 1:fdd22bb7aa52 | 274 | |
emilmont | 1:fdd22bb7aa52 | 275 | /* data is in 5.27(q27) format */ |
emilmont | 1:fdd22bb7aa52 | 276 | |
emilmont | 1:fdd22bb7aa52 | 277 | |
emilmont | 1:fdd22bb7aa52 | 278 | /* start of Middle stages process */ |
emilmont | 1:fdd22bb7aa52 | 279 | |
emilmont | 1:fdd22bb7aa52 | 280 | |
emilmont | 1:fdd22bb7aa52 | 281 | /* each stage in middle stages provides two down scaling of the input */ |
emilmont | 1:fdd22bb7aa52 | 282 | |
emilmont | 1:fdd22bb7aa52 | 283 | twidCoefModifier <<= 2u; |
emilmont | 1:fdd22bb7aa52 | 284 | |
emilmont | 1:fdd22bb7aa52 | 285 | |
emilmont | 1:fdd22bb7aa52 | 286 | for (k = fftLen / 4u; k > 4u; k >>= 2u) |
emilmont | 1:fdd22bb7aa52 | 287 | { |
emilmont | 1:fdd22bb7aa52 | 288 | /* Initializations for the first stage */ |
emilmont | 1:fdd22bb7aa52 | 289 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 290 | n2 >>= 2u; |
emilmont | 1:fdd22bb7aa52 | 291 | ia1 = 0u; |
emilmont | 1:fdd22bb7aa52 | 292 | |
emilmont | 1:fdd22bb7aa52 | 293 | /* Calculation of first stage */ |
emilmont | 1:fdd22bb7aa52 | 294 | for (j = 0u; j <= (n2 - 1u); j++) |
emilmont | 1:fdd22bb7aa52 | 295 | { |
emilmont | 1:fdd22bb7aa52 | 296 | /* index calculation for the coefficients */ |
emilmont | 1:fdd22bb7aa52 | 297 | ia2 = ia1 + ia1; |
emilmont | 1:fdd22bb7aa52 | 298 | ia3 = ia2 + ia1; |
emilmont | 1:fdd22bb7aa52 | 299 | co1 = pCoef[ia1 * 2u]; |
emilmont | 1:fdd22bb7aa52 | 300 | si1 = pCoef[(ia1 * 2u) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 301 | co2 = pCoef[ia2 * 2u]; |
emilmont | 1:fdd22bb7aa52 | 302 | si2 = pCoef[(ia2 * 2u) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 303 | co3 = pCoef[ia3 * 2u]; |
emilmont | 1:fdd22bb7aa52 | 304 | si3 = pCoef[(ia3 * 2u) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 305 | /* Twiddle coefficients index modifier */ |
emilmont | 1:fdd22bb7aa52 | 306 | ia1 = ia1 + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 307 | |
emilmont | 1:fdd22bb7aa52 | 308 | for (i0 = j; i0 < fftLen; i0 += n1) |
emilmont | 1:fdd22bb7aa52 | 309 | { |
emilmont | 1:fdd22bb7aa52 | 310 | /* index calculation for the input as, */ |
emilmont | 1:fdd22bb7aa52 | 311 | /* pSrc[i0 + 0], pSrc[i0 + fftLen/4], pSrc[i0 + fftLen/2u], pSrc[i0 + 3fftLen/4] */ |
emilmont | 1:fdd22bb7aa52 | 312 | i1 = i0 + n2; |
emilmont | 1:fdd22bb7aa52 | 313 | i2 = i1 + n2; |
emilmont | 1:fdd22bb7aa52 | 314 | i3 = i2 + n2; |
emilmont | 1:fdd22bb7aa52 | 315 | |
emilmont | 1:fdd22bb7aa52 | 316 | /* Butterfly implementation */ |
emilmont | 1:fdd22bb7aa52 | 317 | /* xa + xc */ |
emilmont | 1:fdd22bb7aa52 | 318 | r1 = pSrc[2u * i0] + pSrc[2u * i2]; |
emilmont | 1:fdd22bb7aa52 | 319 | /* xa - xc */ |
emilmont | 1:fdd22bb7aa52 | 320 | r2 = pSrc[2u * i0] - pSrc[2u * i2]; |
emilmont | 1:fdd22bb7aa52 | 321 | |
emilmont | 1:fdd22bb7aa52 | 322 | /* ya + yc */ |
emilmont | 1:fdd22bb7aa52 | 323 | s1 = pSrc[(2u * i0) + 1u] + pSrc[(2u * i2) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 324 | /* ya - yc */ |
emilmont | 1:fdd22bb7aa52 | 325 | s2 = pSrc[(2u * i0) + 1u] - pSrc[(2u * i2) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 326 | |
emilmont | 1:fdd22bb7aa52 | 327 | /* xb + xd */ |
emilmont | 1:fdd22bb7aa52 | 328 | t1 = pSrc[2u * i1] + pSrc[2u * i3]; |
emilmont | 1:fdd22bb7aa52 | 329 | |
emilmont | 1:fdd22bb7aa52 | 330 | /* xa' = xa + xb + xc + xd */ |
emilmont | 1:fdd22bb7aa52 | 331 | pSrc[2u * i0] = (r1 + t1) >> 2u; |
emilmont | 1:fdd22bb7aa52 | 332 | /* xa + xc -(xb + xd) */ |
emilmont | 1:fdd22bb7aa52 | 333 | r1 = r1 - t1; |
emilmont | 1:fdd22bb7aa52 | 334 | |
emilmont | 1:fdd22bb7aa52 | 335 | /* yb + yd */ |
emilmont | 1:fdd22bb7aa52 | 336 | t2 = pSrc[(2u * i1) + 1u] + pSrc[(2u * i3) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 337 | /* ya' = ya + yb + yc + yd */ |
emilmont | 1:fdd22bb7aa52 | 338 | pSrc[(2u * i0) + 1u] = (s1 + t2) >> 2u; |
emilmont | 1:fdd22bb7aa52 | 339 | |
emilmont | 1:fdd22bb7aa52 | 340 | /* (ya + yc) - (yb + yd) */ |
emilmont | 1:fdd22bb7aa52 | 341 | s1 = s1 - t2; |
emilmont | 1:fdd22bb7aa52 | 342 | |
emilmont | 1:fdd22bb7aa52 | 343 | /* (yb - yd) */ |
emilmont | 1:fdd22bb7aa52 | 344 | t1 = pSrc[(2u * i1) + 1u] - pSrc[(2u * i3) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 345 | /* (xb - xd) */ |
emilmont | 1:fdd22bb7aa52 | 346 | t2 = pSrc[2u * i1] - pSrc[2u * i3]; |
emilmont | 1:fdd22bb7aa52 | 347 | |
emilmont | 1:fdd22bb7aa52 | 348 | /* xc' = (xa-xb+xc-xd)co2 + (ya-yb+yc-yd)(si2) */ |
emilmont | 1:fdd22bb7aa52 | 349 | pSrc[2u * i1] = (((int32_t) (((q63_t) r1 * co2) >> 32)) + |
emilmont | 1:fdd22bb7aa52 | 350 | ((int32_t) (((q63_t) s1 * si2) >> 32))) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 351 | |
emilmont | 1:fdd22bb7aa52 | 352 | /* yc' = (ya-yb+yc-yd)co2 - (xa-xb+xc-xd)(si2) */ |
emilmont | 1:fdd22bb7aa52 | 353 | pSrc[(2u * i1) + 1u] = (((int32_t) (((q63_t) s1 * co2) >> 32)) - |
emilmont | 1:fdd22bb7aa52 | 354 | ((int32_t) (((q63_t) r1 * si2) >> 32))) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 355 | |
emilmont | 1:fdd22bb7aa52 | 356 | /* (xa - xc) + (yb - yd) */ |
emilmont | 1:fdd22bb7aa52 | 357 | r1 = r2 + t1; |
emilmont | 1:fdd22bb7aa52 | 358 | /* (xa - xc) - (yb - yd) */ |
emilmont | 1:fdd22bb7aa52 | 359 | r2 = r2 - t1; |
emilmont | 1:fdd22bb7aa52 | 360 | |
emilmont | 1:fdd22bb7aa52 | 361 | /* (ya - yc) - (xb - xd) */ |
emilmont | 1:fdd22bb7aa52 | 362 | s1 = s2 - t2; |
emilmont | 1:fdd22bb7aa52 | 363 | /* (ya - yc) + (xb - xd) */ |
emilmont | 1:fdd22bb7aa52 | 364 | s2 = s2 + t2; |
emilmont | 1:fdd22bb7aa52 | 365 | |
emilmont | 1:fdd22bb7aa52 | 366 | /* xb' = (xa+yb-xc-yd)co1 + (ya-xb-yc+xd)(si1) */ |
emilmont | 1:fdd22bb7aa52 | 367 | pSrc[2u * i2] = (((int32_t) (((q63_t) r1 * co1) >> 32)) + |
emilmont | 1:fdd22bb7aa52 | 368 | ((int32_t) (((q63_t) s1 * si1) >> 32))) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 369 | |
emilmont | 1:fdd22bb7aa52 | 370 | /* yb' = (ya-xb-yc+xd)co1 - (xa+yb-xc-yd)(si1) */ |
emilmont | 1:fdd22bb7aa52 | 371 | pSrc[(2u * i2) + 1u] = (((int32_t) (((q63_t) s1 * co1) >> 32)) - |
emilmont | 1:fdd22bb7aa52 | 372 | ((int32_t) (((q63_t) r1 * si1) >> 32))) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 373 | |
emilmont | 1:fdd22bb7aa52 | 374 | /* xd' = (xa-yb-xc+yd)co3 + (ya+xb-yc-xd)(si3) */ |
emilmont | 1:fdd22bb7aa52 | 375 | pSrc[2u * i3] = (((int32_t) (((q63_t) r2 * co3) >> 32)) + |
emilmont | 1:fdd22bb7aa52 | 376 | ((int32_t) (((q63_t) s2 * si3) >> 32))) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 377 | |
emilmont | 1:fdd22bb7aa52 | 378 | /* yd' = (ya+xb-yc-xd)co3 - (xa-yb-xc+yd)(si3) */ |
emilmont | 1:fdd22bb7aa52 | 379 | pSrc[(2u * i3) + 1u] = (((int32_t) (((q63_t) s2 * co3) >> 32)) - |
emilmont | 1:fdd22bb7aa52 | 380 | ((int32_t) (((q63_t) r2 * si3) >> 32))) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 381 | } |
emilmont | 1:fdd22bb7aa52 | 382 | } |
emilmont | 1:fdd22bb7aa52 | 383 | twidCoefModifier <<= 2u; |
emilmont | 1:fdd22bb7aa52 | 384 | } |
emilmont | 1:fdd22bb7aa52 | 385 | |
emilmont | 1:fdd22bb7aa52 | 386 | /* End of Middle stages process */ |
emilmont | 1:fdd22bb7aa52 | 387 | |
emilmont | 1:fdd22bb7aa52 | 388 | /* data is in 11.21(q21) format for the 1024 point as there are 3 middle stages */ |
emilmont | 1:fdd22bb7aa52 | 389 | /* data is in 9.23(q23) format for the 256 point as there are 2 middle stages */ |
emilmont | 1:fdd22bb7aa52 | 390 | /* data is in 7.25(q25) format for the 64 point as there are 1 middle stage */ |
emilmont | 1:fdd22bb7aa52 | 391 | /* data is in 5.27(q27) format for the 16 point as there are no middle stages */ |
emilmont | 1:fdd22bb7aa52 | 392 | |
emilmont | 1:fdd22bb7aa52 | 393 | |
emilmont | 1:fdd22bb7aa52 | 394 | /* start of Last stage process */ |
emilmont | 1:fdd22bb7aa52 | 395 | /* Initializations for the last stage */ |
emilmont | 1:fdd22bb7aa52 | 396 | j = fftLen >> 2; |
emilmont | 1:fdd22bb7aa52 | 397 | ptr1 = &pSrc[0]; |
emilmont | 1:fdd22bb7aa52 | 398 | |
emilmont | 1:fdd22bb7aa52 | 399 | /* Calculations of last stage */ |
emilmont | 1:fdd22bb7aa52 | 400 | do |
emilmont | 1:fdd22bb7aa52 | 401 | { |
emilmont | 1:fdd22bb7aa52 | 402 | |
emilmont | 1:fdd22bb7aa52 | 403 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 404 | |
emilmont | 1:fdd22bb7aa52 | 405 | /* Read xa (real), ya(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 406 | xaya = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 407 | xa = (q31_t) xaya; |
emilmont | 1:fdd22bb7aa52 | 408 | ya = (q31_t) (xaya >> 32); |
emilmont | 1:fdd22bb7aa52 | 409 | |
emilmont | 1:fdd22bb7aa52 | 410 | /* Read xb (real), yb(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 411 | xbyb = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 412 | xb = (q31_t) xbyb; |
emilmont | 1:fdd22bb7aa52 | 413 | yb = (q31_t) (xbyb >> 32); |
emilmont | 1:fdd22bb7aa52 | 414 | |
emilmont | 1:fdd22bb7aa52 | 415 | /* Read xc (real), yc(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 416 | xcyc = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 417 | xc = (q31_t) xcyc; |
emilmont | 1:fdd22bb7aa52 | 418 | yc = (q31_t) (xcyc >> 32); |
emilmont | 1:fdd22bb7aa52 | 419 | |
emilmont | 1:fdd22bb7aa52 | 420 | /* Read xc (real), yc(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 421 | xdyd = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 422 | xd = (q31_t) xdyd; |
emilmont | 1:fdd22bb7aa52 | 423 | yd = (q31_t) (xdyd >> 32); |
emilmont | 1:fdd22bb7aa52 | 424 | |
emilmont | 1:fdd22bb7aa52 | 425 | #else |
emilmont | 1:fdd22bb7aa52 | 426 | |
emilmont | 1:fdd22bb7aa52 | 427 | /* Read xa (real), ya(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 428 | xaya = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 429 | ya = (q31_t) xaya; |
emilmont | 1:fdd22bb7aa52 | 430 | xa = (q31_t) (xaya >> 32); |
emilmont | 1:fdd22bb7aa52 | 431 | |
emilmont | 1:fdd22bb7aa52 | 432 | /* Read xb (real), yb(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 433 | xbyb = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 434 | yb = (q31_t) xbyb; |
emilmont | 1:fdd22bb7aa52 | 435 | xb = (q31_t) (xbyb >> 32); |
emilmont | 1:fdd22bb7aa52 | 436 | |
emilmont | 1:fdd22bb7aa52 | 437 | /* Read xc (real), yc(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 438 | xcyc = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 439 | yc = (q31_t) xcyc; |
emilmont | 1:fdd22bb7aa52 | 440 | xc = (q31_t) (xcyc >> 32); |
emilmont | 1:fdd22bb7aa52 | 441 | |
emilmont | 1:fdd22bb7aa52 | 442 | /* Read xc (real), yc(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 443 | xdyd = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 444 | yd = (q31_t) xdyd; |
emilmont | 1:fdd22bb7aa52 | 445 | xd = (q31_t) (xdyd >> 32); |
emilmont | 1:fdd22bb7aa52 | 446 | |
emilmont | 1:fdd22bb7aa52 | 447 | |
emilmont | 1:fdd22bb7aa52 | 448 | #endif |
emilmont | 1:fdd22bb7aa52 | 449 | |
emilmont | 1:fdd22bb7aa52 | 450 | /* xa' = xa + xb + xc + xd */ |
emilmont | 1:fdd22bb7aa52 | 451 | xa_out = xa + xb + xc + xd; |
emilmont | 1:fdd22bb7aa52 | 452 | |
emilmont | 1:fdd22bb7aa52 | 453 | /* ya' = ya + yb + yc + yd */ |
emilmont | 1:fdd22bb7aa52 | 454 | ya_out = ya + yb + yc + yd; |
emilmont | 1:fdd22bb7aa52 | 455 | |
emilmont | 1:fdd22bb7aa52 | 456 | /* pointer updation for writing */ |
emilmont | 1:fdd22bb7aa52 | 457 | ptr1 = ptr1 - 8u; |
emilmont | 1:fdd22bb7aa52 | 458 | |
emilmont | 1:fdd22bb7aa52 | 459 | /* writing xa' and ya' */ |
emilmont | 1:fdd22bb7aa52 | 460 | *ptr1++ = xa_out; |
emilmont | 1:fdd22bb7aa52 | 461 | *ptr1++ = ya_out; |
emilmont | 1:fdd22bb7aa52 | 462 | |
emilmont | 1:fdd22bb7aa52 | 463 | xc_out = (xa - xb + xc - xd); |
emilmont | 1:fdd22bb7aa52 | 464 | yc_out = (ya - yb + yc - yd); |
emilmont | 1:fdd22bb7aa52 | 465 | |
emilmont | 1:fdd22bb7aa52 | 466 | /* writing xc' and yc' */ |
emilmont | 1:fdd22bb7aa52 | 467 | *ptr1++ = xc_out; |
emilmont | 1:fdd22bb7aa52 | 468 | *ptr1++ = yc_out; |
emilmont | 1:fdd22bb7aa52 | 469 | |
emilmont | 1:fdd22bb7aa52 | 470 | xb_out = (xa + yb - xc - yd); |
emilmont | 1:fdd22bb7aa52 | 471 | yb_out = (ya - xb - yc + xd); |
emilmont | 1:fdd22bb7aa52 | 472 | |
emilmont | 1:fdd22bb7aa52 | 473 | /* writing xb' and yb' */ |
emilmont | 1:fdd22bb7aa52 | 474 | *ptr1++ = xb_out; |
emilmont | 1:fdd22bb7aa52 | 475 | *ptr1++ = yb_out; |
emilmont | 1:fdd22bb7aa52 | 476 | |
emilmont | 1:fdd22bb7aa52 | 477 | xd_out = (xa - yb - xc + yd); |
emilmont | 1:fdd22bb7aa52 | 478 | yd_out = (ya + xb - yc - xd); |
emilmont | 1:fdd22bb7aa52 | 479 | |
emilmont | 1:fdd22bb7aa52 | 480 | /* writing xd' and yd' */ |
emilmont | 1:fdd22bb7aa52 | 481 | *ptr1++ = xd_out; |
emilmont | 1:fdd22bb7aa52 | 482 | *ptr1++ = yd_out; |
emilmont | 1:fdd22bb7aa52 | 483 | |
emilmont | 1:fdd22bb7aa52 | 484 | |
emilmont | 1:fdd22bb7aa52 | 485 | } while(--j); |
emilmont | 1:fdd22bb7aa52 | 486 | |
emilmont | 1:fdd22bb7aa52 | 487 | /* output is in 11.21(q21) format for the 1024 point */ |
emilmont | 1:fdd22bb7aa52 | 488 | /* output is in 9.23(q23) format for the 256 point */ |
emilmont | 1:fdd22bb7aa52 | 489 | /* output is in 7.25(q25) format for the 64 point */ |
emilmont | 1:fdd22bb7aa52 | 490 | /* output is in 5.27(q27) format for the 16 point */ |
emilmont | 1:fdd22bb7aa52 | 491 | |
emilmont | 1:fdd22bb7aa52 | 492 | /* End of last stage process */ |
emilmont | 1:fdd22bb7aa52 | 493 | |
emilmont | 1:fdd22bb7aa52 | 494 | } |
emilmont | 1:fdd22bb7aa52 | 495 | |
emilmont | 1:fdd22bb7aa52 | 496 | |
emilmont | 1:fdd22bb7aa52 | 497 | /** |
emilmont | 1:fdd22bb7aa52 | 498 | * @brief Core function for the Q31 CIFFT butterfly process. |
emilmont | 1:fdd22bb7aa52 | 499 | * @param[in, out] *pSrc points to the in-place buffer of Q31 data type. |
emilmont | 1:fdd22bb7aa52 | 500 | * @param[in] fftLen length of the FFT. |
emilmont | 1:fdd22bb7aa52 | 501 | * @param[in] *pCoef points to twiddle coefficient buffer. |
emilmont | 1:fdd22bb7aa52 | 502 | * @param[in] twidCoefModifier twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. |
emilmont | 1:fdd22bb7aa52 | 503 | * @return none. |
emilmont | 1:fdd22bb7aa52 | 504 | */ |
emilmont | 1:fdd22bb7aa52 | 505 | |
emilmont | 1:fdd22bb7aa52 | 506 | |
emilmont | 1:fdd22bb7aa52 | 507 | /* |
emilmont | 1:fdd22bb7aa52 | 508 | * Radix-4 IFFT algorithm used is : |
emilmont | 1:fdd22bb7aa52 | 509 | * |
emilmont | 1:fdd22bb7aa52 | 510 | * CIFFT uses same twiddle coefficients as CFFT Function |
emilmont | 1:fdd22bb7aa52 | 511 | * x[k] = x[n] + (j)k * x[n + fftLen/4] + (-1)k * x[n+fftLen/2] + (-j)k * x[n+3*fftLen/4] |
emilmont | 1:fdd22bb7aa52 | 512 | * |
emilmont | 1:fdd22bb7aa52 | 513 | * |
emilmont | 1:fdd22bb7aa52 | 514 | * IFFT is implemented with following changes in equations from FFT |
emilmont | 1:fdd22bb7aa52 | 515 | * |
emilmont | 1:fdd22bb7aa52 | 516 | * Input real and imaginary data: |
emilmont | 1:fdd22bb7aa52 | 517 | * x(n) = xa + j * ya |
emilmont | 1:fdd22bb7aa52 | 518 | * x(n+N/4 ) = xb + j * yb |
emilmont | 1:fdd22bb7aa52 | 519 | * x(n+N/2 ) = xc + j * yc |
emilmont | 1:fdd22bb7aa52 | 520 | * x(n+3N 4) = xd + j * yd |
emilmont | 1:fdd22bb7aa52 | 521 | * |
emilmont | 1:fdd22bb7aa52 | 522 | * |
emilmont | 1:fdd22bb7aa52 | 523 | * Output real and imaginary data: |
emilmont | 1:fdd22bb7aa52 | 524 | * x(4r) = xa'+ j * ya' |
emilmont | 1:fdd22bb7aa52 | 525 | * x(4r+1) = xb'+ j * yb' |
emilmont | 1:fdd22bb7aa52 | 526 | * x(4r+2) = xc'+ j * yc' |
emilmont | 1:fdd22bb7aa52 | 527 | * x(4r+3) = xd'+ j * yd' |
emilmont | 1:fdd22bb7aa52 | 528 | * |
emilmont | 1:fdd22bb7aa52 | 529 | * |
emilmont | 1:fdd22bb7aa52 | 530 | * Twiddle factors for radix-4 IFFT: |
emilmont | 1:fdd22bb7aa52 | 531 | * Wn = co1 + j * (si1) |
emilmont | 1:fdd22bb7aa52 | 532 | * W2n = co2 + j * (si2) |
emilmont | 1:fdd22bb7aa52 | 533 | * W3n = co3 + j * (si3) |
emilmont | 1:fdd22bb7aa52 | 534 | |
emilmont | 1:fdd22bb7aa52 | 535 | * The real and imaginary output values for the radix-4 butterfly are |
emilmont | 1:fdd22bb7aa52 | 536 | * xa' = xa + xb + xc + xd |
emilmont | 1:fdd22bb7aa52 | 537 | * ya' = ya + yb + yc + yd |
emilmont | 1:fdd22bb7aa52 | 538 | * xb' = (xa-yb-xc+yd)* co1 - (ya+xb-yc-xd)* (si1) |
emilmont | 1:fdd22bb7aa52 | 539 | * yb' = (ya+xb-yc-xd)* co1 + (xa-yb-xc+yd)* (si1) |
emilmont | 1:fdd22bb7aa52 | 540 | * xc' = (xa-xb+xc-xd)* co2 - (ya-yb+yc-yd)* (si2) |
emilmont | 1:fdd22bb7aa52 | 541 | * yc' = (ya-yb+yc-yd)* co2 + (xa-xb+xc-xd)* (si2) |
emilmont | 1:fdd22bb7aa52 | 542 | * xd' = (xa+yb-xc-yd)* co3 - (ya-xb-yc+xd)* (si3) |
emilmont | 1:fdd22bb7aa52 | 543 | * yd' = (ya-xb-yc+xd)* co3 + (xa+yb-xc-yd)* (si3) |
emilmont | 1:fdd22bb7aa52 | 544 | * |
emilmont | 1:fdd22bb7aa52 | 545 | */ |
emilmont | 1:fdd22bb7aa52 | 546 | |
emilmont | 1:fdd22bb7aa52 | 547 | void arm_radix4_butterfly_inverse_q31( |
emilmont | 1:fdd22bb7aa52 | 548 | q31_t * pSrc, |
emilmont | 1:fdd22bb7aa52 | 549 | uint32_t fftLen, |
emilmont | 1:fdd22bb7aa52 | 550 | q31_t * pCoef, |
emilmont | 1:fdd22bb7aa52 | 551 | uint32_t twidCoefModifier) |
emilmont | 1:fdd22bb7aa52 | 552 | { |
emilmont | 1:fdd22bb7aa52 | 553 | uint32_t n1, n2, ia1, ia2, ia3, i0, i1, i2, i3, j, k; |
emilmont | 1:fdd22bb7aa52 | 554 | q31_t t1, t2, r1, r2, s1, s2, co1, co2, co3, si1, si2, si3; |
emilmont | 1:fdd22bb7aa52 | 555 | q31_t xa, xb, xc, xd; |
emilmont | 1:fdd22bb7aa52 | 556 | q31_t ya, yb, yc, yd; |
emilmont | 1:fdd22bb7aa52 | 557 | q31_t xa_out, xb_out, xc_out, xd_out; |
emilmont | 1:fdd22bb7aa52 | 558 | q31_t ya_out, yb_out, yc_out, yd_out; |
emilmont | 1:fdd22bb7aa52 | 559 | |
emilmont | 1:fdd22bb7aa52 | 560 | q31_t *ptr1; |
emilmont | 1:fdd22bb7aa52 | 561 | q63_t xaya, xbyb, xcyc, xdyd; |
emilmont | 1:fdd22bb7aa52 | 562 | |
emilmont | 1:fdd22bb7aa52 | 563 | /* input is be 1.31(q31) format for all FFT sizes */ |
emilmont | 1:fdd22bb7aa52 | 564 | /* Total process is divided into three stages */ |
emilmont | 1:fdd22bb7aa52 | 565 | /* process first stage, middle stages, & last stage */ |
emilmont | 1:fdd22bb7aa52 | 566 | |
emilmont | 1:fdd22bb7aa52 | 567 | /* Start of first stage process */ |
emilmont | 1:fdd22bb7aa52 | 568 | |
emilmont | 1:fdd22bb7aa52 | 569 | /* Initializations for the first stage */ |
emilmont | 1:fdd22bb7aa52 | 570 | n2 = fftLen; |
emilmont | 1:fdd22bb7aa52 | 571 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 572 | /* n2 = fftLen/4 */ |
emilmont | 1:fdd22bb7aa52 | 573 | n2 >>= 2u; |
emilmont | 1:fdd22bb7aa52 | 574 | i0 = 0u; |
emilmont | 1:fdd22bb7aa52 | 575 | ia1 = 0u; |
emilmont | 1:fdd22bb7aa52 | 576 | |
emilmont | 1:fdd22bb7aa52 | 577 | j = n2; |
emilmont | 1:fdd22bb7aa52 | 578 | |
emilmont | 1:fdd22bb7aa52 | 579 | do |
emilmont | 1:fdd22bb7aa52 | 580 | { |
emilmont | 1:fdd22bb7aa52 | 581 | |
emilmont | 1:fdd22bb7aa52 | 582 | /* input is in 1.31(q31) format and provide 4 guard bits for the input */ |
emilmont | 1:fdd22bb7aa52 | 583 | |
emilmont | 1:fdd22bb7aa52 | 584 | /* index calculation for the input as, */ |
emilmont | 1:fdd22bb7aa52 | 585 | /* pSrc[i0 + 0], pSrc[i0 + fftLen/4], pSrc[i0 + fftLen/2u], pSrc[i0 + 3fftLen/4] */ |
emilmont | 1:fdd22bb7aa52 | 586 | i1 = i0 + n2; |
emilmont | 1:fdd22bb7aa52 | 587 | i2 = i1 + n2; |
emilmont | 1:fdd22bb7aa52 | 588 | i3 = i2 + n2; |
emilmont | 1:fdd22bb7aa52 | 589 | |
emilmont | 1:fdd22bb7aa52 | 590 | /* Butterfly implementation */ |
emilmont | 1:fdd22bb7aa52 | 591 | /* xa + xc */ |
emilmont | 1:fdd22bb7aa52 | 592 | r1 = (pSrc[2u * i0] >> 4u) + (pSrc[2u * i2] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 593 | /* xa - xc */ |
emilmont | 1:fdd22bb7aa52 | 594 | r2 = (pSrc[2u * i0] >> 4u) - (pSrc[2u * i2] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 595 | |
emilmont | 1:fdd22bb7aa52 | 596 | /* xb + xd */ |
emilmont | 1:fdd22bb7aa52 | 597 | t1 = (pSrc[2u * i1] >> 4u) + (pSrc[2u * i3] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 598 | |
emilmont | 1:fdd22bb7aa52 | 599 | /* ya + yc */ |
emilmont | 1:fdd22bb7aa52 | 600 | s1 = (pSrc[(2u * i0) + 1u] >> 4u) + (pSrc[(2u * i2) + 1u] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 601 | /* ya - yc */ |
emilmont | 1:fdd22bb7aa52 | 602 | s2 = (pSrc[(2u * i0) + 1u] >> 4u) - (pSrc[(2u * i2) + 1u] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 603 | |
emilmont | 1:fdd22bb7aa52 | 604 | /* xa' = xa + xb + xc + xd */ |
emilmont | 1:fdd22bb7aa52 | 605 | pSrc[2u * i0] = (r1 + t1); |
emilmont | 1:fdd22bb7aa52 | 606 | /* (xa + xc) - (xb + xd) */ |
emilmont | 1:fdd22bb7aa52 | 607 | r1 = r1 - t1; |
emilmont | 1:fdd22bb7aa52 | 608 | /* yb + yd */ |
emilmont | 1:fdd22bb7aa52 | 609 | t2 = (pSrc[(2u * i1) + 1u] >> 4u) + (pSrc[(2u * i3) + 1u] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 610 | /* ya' = ya + yb + yc + yd */ |
emilmont | 1:fdd22bb7aa52 | 611 | pSrc[(2u * i0) + 1u] = (s1 + t2); |
emilmont | 1:fdd22bb7aa52 | 612 | |
emilmont | 1:fdd22bb7aa52 | 613 | /* (ya + yc) - (yb + yd) */ |
emilmont | 1:fdd22bb7aa52 | 614 | s1 = s1 - t2; |
emilmont | 1:fdd22bb7aa52 | 615 | |
emilmont | 1:fdd22bb7aa52 | 616 | /* yb - yd */ |
emilmont | 1:fdd22bb7aa52 | 617 | t1 = (pSrc[(2u * i1) + 1u] >> 4u) - (pSrc[(2u * i3) + 1u] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 618 | /* xb - xd */ |
emilmont | 1:fdd22bb7aa52 | 619 | t2 = (pSrc[2u * i1] >> 4u) - (pSrc[2u * i3] >> 4u); |
emilmont | 1:fdd22bb7aa52 | 620 | |
emilmont | 1:fdd22bb7aa52 | 621 | /* index calculation for the coefficients */ |
emilmont | 1:fdd22bb7aa52 | 622 | ia2 = 2u * ia1; |
emilmont | 1:fdd22bb7aa52 | 623 | co2 = pCoef[ia2 * 2u]; |
emilmont | 1:fdd22bb7aa52 | 624 | si2 = pCoef[(ia2 * 2u) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 625 | |
emilmont | 1:fdd22bb7aa52 | 626 | /* xc' = (xa-xb+xc-xd)co2 - (ya-yb+yc-yd)(si2) */ |
emilmont | 1:fdd22bb7aa52 | 627 | pSrc[2u * i1] = (((int32_t) (((q63_t) r1 * co2) >> 32)) - |
emilmont | 1:fdd22bb7aa52 | 628 | ((int32_t) (((q63_t) s1 * si2) >> 32))) << 1u; |
emilmont | 1:fdd22bb7aa52 | 629 | |
emilmont | 1:fdd22bb7aa52 | 630 | /* yc' = (ya-yb+yc-yd)co2 + (xa-xb+xc-xd)(si2) */ |
emilmont | 1:fdd22bb7aa52 | 631 | pSrc[2u * i1 + 1u] = (((int32_t) (((q63_t) s1 * co2) >> 32)) + |
emilmont | 1:fdd22bb7aa52 | 632 | ((int32_t) (((q63_t) r1 * si2) >> 32))) << 1u; |
emilmont | 1:fdd22bb7aa52 | 633 | |
emilmont | 1:fdd22bb7aa52 | 634 | /* (xa - xc) - (yb - yd) */ |
emilmont | 1:fdd22bb7aa52 | 635 | r1 = r2 - t1; |
emilmont | 1:fdd22bb7aa52 | 636 | /* (xa - xc) + (yb - yd) */ |
emilmont | 1:fdd22bb7aa52 | 637 | r2 = r2 + t1; |
emilmont | 1:fdd22bb7aa52 | 638 | |
emilmont | 1:fdd22bb7aa52 | 639 | /* (ya - yc) + (xb - xd) */ |
emilmont | 1:fdd22bb7aa52 | 640 | s1 = s2 + t2; |
emilmont | 1:fdd22bb7aa52 | 641 | /* (ya - yc) - (xb - xd) */ |
emilmont | 1:fdd22bb7aa52 | 642 | s2 = s2 - t2; |
emilmont | 1:fdd22bb7aa52 | 643 | |
emilmont | 1:fdd22bb7aa52 | 644 | co1 = pCoef[ia1 * 2u]; |
emilmont | 1:fdd22bb7aa52 | 645 | si1 = pCoef[(ia1 * 2u) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 646 | |
emilmont | 1:fdd22bb7aa52 | 647 | /* xb' = (xa+yb-xc-yd)co1 - (ya-xb-yc+xd)(si1) */ |
emilmont | 1:fdd22bb7aa52 | 648 | pSrc[2u * i2] = (((int32_t) (((q63_t) r1 * co1) >> 32)) - |
emilmont | 1:fdd22bb7aa52 | 649 | ((int32_t) (((q63_t) s1 * si1) >> 32))) << 1u; |
emilmont | 1:fdd22bb7aa52 | 650 | |
emilmont | 1:fdd22bb7aa52 | 651 | /* yb' = (ya-xb-yc+xd)co1 + (xa+yb-xc-yd)(si1) */ |
emilmont | 1:fdd22bb7aa52 | 652 | pSrc[(2u * i2) + 1u] = (((int32_t) (((q63_t) s1 * co1) >> 32)) + |
emilmont | 1:fdd22bb7aa52 | 653 | ((int32_t) (((q63_t) r1 * si1) >> 32))) << 1u; |
emilmont | 1:fdd22bb7aa52 | 654 | |
emilmont | 1:fdd22bb7aa52 | 655 | /* index calculation for the coefficients */ |
emilmont | 1:fdd22bb7aa52 | 656 | ia3 = 3u * ia1; |
emilmont | 1:fdd22bb7aa52 | 657 | co3 = pCoef[ia3 * 2u]; |
emilmont | 1:fdd22bb7aa52 | 658 | si3 = pCoef[(ia3 * 2u) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 659 | |
emilmont | 1:fdd22bb7aa52 | 660 | /* xd' = (xa-yb-xc+yd)co3 - (ya+xb-yc-xd)(si3) */ |
emilmont | 1:fdd22bb7aa52 | 661 | pSrc[2u * i3] = (((int32_t) (((q63_t) r2 * co3) >> 32)) - |
emilmont | 1:fdd22bb7aa52 | 662 | ((int32_t) (((q63_t) s2 * si3) >> 32))) << 1u; |
emilmont | 1:fdd22bb7aa52 | 663 | |
emilmont | 1:fdd22bb7aa52 | 664 | /* yd' = (ya+xb-yc-xd)co3 + (xa-yb-xc+yd)(si3) */ |
emilmont | 1:fdd22bb7aa52 | 665 | pSrc[(2u * i3) + 1u] = (((int32_t) (((q63_t) s2 * co3) >> 32)) + |
emilmont | 1:fdd22bb7aa52 | 666 | ((int32_t) (((q63_t) r2 * si3) >> 32))) << 1u; |
emilmont | 1:fdd22bb7aa52 | 667 | |
emilmont | 1:fdd22bb7aa52 | 668 | /* Twiddle coefficients index modifier */ |
emilmont | 1:fdd22bb7aa52 | 669 | ia1 = ia1 + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 670 | |
emilmont | 1:fdd22bb7aa52 | 671 | /* Updating input index */ |
emilmont | 1:fdd22bb7aa52 | 672 | i0 = i0 + 1u; |
emilmont | 1:fdd22bb7aa52 | 673 | |
emilmont | 1:fdd22bb7aa52 | 674 | } while(--j); |
emilmont | 1:fdd22bb7aa52 | 675 | |
emilmont | 1:fdd22bb7aa52 | 676 | /* data is in 5.27(q27) format */ |
emilmont | 1:fdd22bb7aa52 | 677 | /* each stage provides two down scaling of the input */ |
emilmont | 1:fdd22bb7aa52 | 678 | |
emilmont | 1:fdd22bb7aa52 | 679 | |
emilmont | 1:fdd22bb7aa52 | 680 | /* Start of Middle stages process */ |
emilmont | 1:fdd22bb7aa52 | 681 | |
emilmont | 1:fdd22bb7aa52 | 682 | twidCoefModifier <<= 2u; |
emilmont | 1:fdd22bb7aa52 | 683 | |
emilmont | 1:fdd22bb7aa52 | 684 | /* Calculation of second stage to excluding last stage */ |
emilmont | 1:fdd22bb7aa52 | 685 | for (k = fftLen / 4u; k > 4u; k >>= 2u) |
emilmont | 1:fdd22bb7aa52 | 686 | { |
emilmont | 1:fdd22bb7aa52 | 687 | /* Initializations for the first stage */ |
emilmont | 1:fdd22bb7aa52 | 688 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 689 | n2 >>= 2u; |
emilmont | 1:fdd22bb7aa52 | 690 | ia1 = 0u; |
emilmont | 1:fdd22bb7aa52 | 691 | |
emilmont | 1:fdd22bb7aa52 | 692 | for (j = 0; j <= (n2 - 1u); j++) |
emilmont | 1:fdd22bb7aa52 | 693 | { |
emilmont | 1:fdd22bb7aa52 | 694 | /* index calculation for the coefficients */ |
emilmont | 1:fdd22bb7aa52 | 695 | ia2 = ia1 + ia1; |
emilmont | 1:fdd22bb7aa52 | 696 | ia3 = ia2 + ia1; |
emilmont | 1:fdd22bb7aa52 | 697 | co1 = pCoef[ia1 * 2u]; |
emilmont | 1:fdd22bb7aa52 | 698 | si1 = pCoef[(ia1 * 2u) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 699 | co2 = pCoef[ia2 * 2u]; |
emilmont | 1:fdd22bb7aa52 | 700 | si2 = pCoef[(ia2 * 2u) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 701 | co3 = pCoef[ia3 * 2u]; |
emilmont | 1:fdd22bb7aa52 | 702 | si3 = pCoef[(ia3 * 2u) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 703 | /* Twiddle coefficients index modifier */ |
emilmont | 1:fdd22bb7aa52 | 704 | ia1 = ia1 + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 705 | |
emilmont | 1:fdd22bb7aa52 | 706 | for (i0 = j; i0 < fftLen; i0 += n1) |
emilmont | 1:fdd22bb7aa52 | 707 | { |
emilmont | 1:fdd22bb7aa52 | 708 | /* index calculation for the input as, */ |
emilmont | 1:fdd22bb7aa52 | 709 | /* pSrc[i0 + 0], pSrc[i0 + fftLen/4], pSrc[i0 + fftLen/2u], pSrc[i0 + 3fftLen/4] */ |
emilmont | 1:fdd22bb7aa52 | 710 | i1 = i0 + n2; |
emilmont | 1:fdd22bb7aa52 | 711 | i2 = i1 + n2; |
emilmont | 1:fdd22bb7aa52 | 712 | i3 = i2 + n2; |
emilmont | 1:fdd22bb7aa52 | 713 | |
emilmont | 1:fdd22bb7aa52 | 714 | /* Butterfly implementation */ |
emilmont | 1:fdd22bb7aa52 | 715 | /* xa + xc */ |
emilmont | 1:fdd22bb7aa52 | 716 | r1 = pSrc[2u * i0] + pSrc[2u * i2]; |
emilmont | 1:fdd22bb7aa52 | 717 | /* xa - xc */ |
emilmont | 1:fdd22bb7aa52 | 718 | r2 = pSrc[2u * i0] - pSrc[2u * i2]; |
emilmont | 1:fdd22bb7aa52 | 719 | |
emilmont | 1:fdd22bb7aa52 | 720 | /* ya + yc */ |
emilmont | 1:fdd22bb7aa52 | 721 | s1 = pSrc[(2u * i0) + 1u] + pSrc[(2u * i2) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 722 | /* ya - yc */ |
emilmont | 1:fdd22bb7aa52 | 723 | s2 = pSrc[(2u * i0) + 1u] - pSrc[(2u * i2) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 724 | |
emilmont | 1:fdd22bb7aa52 | 725 | /* xb + xd */ |
emilmont | 1:fdd22bb7aa52 | 726 | t1 = pSrc[2u * i1] + pSrc[2u * i3]; |
emilmont | 1:fdd22bb7aa52 | 727 | |
emilmont | 1:fdd22bb7aa52 | 728 | /* xa' = xa + xb + xc + xd */ |
emilmont | 1:fdd22bb7aa52 | 729 | pSrc[2u * i0] = (r1 + t1) >> 2u; |
emilmont | 1:fdd22bb7aa52 | 730 | /* xa + xc -(xb + xd) */ |
emilmont | 1:fdd22bb7aa52 | 731 | r1 = r1 - t1; |
emilmont | 1:fdd22bb7aa52 | 732 | /* yb + yd */ |
emilmont | 1:fdd22bb7aa52 | 733 | t2 = pSrc[(2u * i1) + 1u] + pSrc[(2u * i3) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 734 | /* ya' = ya + yb + yc + yd */ |
emilmont | 1:fdd22bb7aa52 | 735 | pSrc[(2u * i0) + 1u] = (s1 + t2) >> 2u; |
emilmont | 1:fdd22bb7aa52 | 736 | |
emilmont | 1:fdd22bb7aa52 | 737 | /* (ya + yc) - (yb + yd) */ |
emilmont | 1:fdd22bb7aa52 | 738 | s1 = s1 - t2; |
emilmont | 1:fdd22bb7aa52 | 739 | |
emilmont | 1:fdd22bb7aa52 | 740 | /* (yb - yd) */ |
emilmont | 1:fdd22bb7aa52 | 741 | t1 = pSrc[(2u * i1) + 1u] - pSrc[(2u * i3) + 1u]; |
emilmont | 1:fdd22bb7aa52 | 742 | /* (xb - xd) */ |
emilmont | 1:fdd22bb7aa52 | 743 | t2 = pSrc[2u * i1] - pSrc[2u * i3]; |
emilmont | 1:fdd22bb7aa52 | 744 | |
emilmont | 1:fdd22bb7aa52 | 745 | /* xc' = (xa-xb+xc-xd)co2 - (ya-yb+yc-yd)(si2) */ |
emilmont | 1:fdd22bb7aa52 | 746 | pSrc[2u * i1] = (((int32_t) (((q63_t) r1 * co2) >> 32u)) - |
emilmont | 1:fdd22bb7aa52 | 747 | ((int32_t) (((q63_t) s1 * si2) >> 32u))) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 748 | |
emilmont | 1:fdd22bb7aa52 | 749 | /* yc' = (ya-yb+yc-yd)co2 + (xa-xb+xc-xd)(si2) */ |
emilmont | 1:fdd22bb7aa52 | 750 | pSrc[(2u * i1) + 1u] = |
emilmont | 1:fdd22bb7aa52 | 751 | (((int32_t) (((q63_t) s1 * co2) >> 32u)) + |
emilmont | 1:fdd22bb7aa52 | 752 | ((int32_t) (((q63_t) r1 * si2) >> 32u))) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 753 | |
emilmont | 1:fdd22bb7aa52 | 754 | /* (xa - xc) - (yb - yd) */ |
emilmont | 1:fdd22bb7aa52 | 755 | r1 = r2 - t1; |
emilmont | 1:fdd22bb7aa52 | 756 | /* (xa - xc) + (yb - yd) */ |
emilmont | 1:fdd22bb7aa52 | 757 | r2 = r2 + t1; |
emilmont | 1:fdd22bb7aa52 | 758 | |
emilmont | 1:fdd22bb7aa52 | 759 | /* (ya - yc) + (xb - xd) */ |
emilmont | 1:fdd22bb7aa52 | 760 | s1 = s2 + t2; |
emilmont | 1:fdd22bb7aa52 | 761 | /* (ya - yc) - (xb - xd) */ |
emilmont | 1:fdd22bb7aa52 | 762 | s2 = s2 - t2; |
emilmont | 1:fdd22bb7aa52 | 763 | |
emilmont | 1:fdd22bb7aa52 | 764 | /* xb' = (xa+yb-xc-yd)co1 - (ya-xb-yc+xd)(si1) */ |
emilmont | 1:fdd22bb7aa52 | 765 | pSrc[2u * i2] = (((int32_t) (((q63_t) r1 * co1) >> 32)) - |
emilmont | 1:fdd22bb7aa52 | 766 | ((int32_t) (((q63_t) s1 * si1) >> 32))) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 767 | |
emilmont | 1:fdd22bb7aa52 | 768 | /* yb' = (ya-xb-yc+xd)co1 + (xa+yb-xc-yd)(si1) */ |
emilmont | 1:fdd22bb7aa52 | 769 | pSrc[(2u * i2) + 1u] = (((int32_t) (((q63_t) s1 * co1) >> 32)) + |
emilmont | 1:fdd22bb7aa52 | 770 | ((int32_t) (((q63_t) r1 * si1) >> 32))) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 771 | |
emilmont | 1:fdd22bb7aa52 | 772 | /* xd' = (xa-yb-xc+yd)co3 - (ya+xb-yc-xd)(si3) */ |
emilmont | 1:fdd22bb7aa52 | 773 | pSrc[(2u * i3)] = (((int32_t) (((q63_t) r2 * co3) >> 32)) - |
emilmont | 1:fdd22bb7aa52 | 774 | ((int32_t) (((q63_t) s2 * si3) >> 32))) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 775 | |
emilmont | 1:fdd22bb7aa52 | 776 | /* yd' = (ya+xb-yc-xd)co3 + (xa-yb-xc+yd)(si3) */ |
emilmont | 1:fdd22bb7aa52 | 777 | pSrc[(2u * i3) + 1u] = (((int32_t) (((q63_t) s2 * co3) >> 32)) + |
emilmont | 1:fdd22bb7aa52 | 778 | ((int32_t) (((q63_t) r2 * si3) >> 32))) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 779 | } |
emilmont | 1:fdd22bb7aa52 | 780 | } |
emilmont | 1:fdd22bb7aa52 | 781 | twidCoefModifier <<= 2u; |
emilmont | 1:fdd22bb7aa52 | 782 | } |
emilmont | 1:fdd22bb7aa52 | 783 | |
emilmont | 1:fdd22bb7aa52 | 784 | /* End of Middle stages process */ |
emilmont | 1:fdd22bb7aa52 | 785 | |
emilmont | 1:fdd22bb7aa52 | 786 | /* data is in 11.21(q21) format for the 1024 point as there are 3 middle stages */ |
emilmont | 1:fdd22bb7aa52 | 787 | /* data is in 9.23(q23) format for the 256 point as there are 2 middle stages */ |
emilmont | 1:fdd22bb7aa52 | 788 | /* data is in 7.25(q25) format for the 64 point as there are 1 middle stage */ |
emilmont | 1:fdd22bb7aa52 | 789 | /* data is in 5.27(q27) format for the 16 point as there are no middle stages */ |
emilmont | 1:fdd22bb7aa52 | 790 | |
emilmont | 1:fdd22bb7aa52 | 791 | |
emilmont | 1:fdd22bb7aa52 | 792 | /* Start of last stage process */ |
emilmont | 1:fdd22bb7aa52 | 793 | |
emilmont | 1:fdd22bb7aa52 | 794 | |
emilmont | 1:fdd22bb7aa52 | 795 | /* Initializations for the last stage */ |
emilmont | 1:fdd22bb7aa52 | 796 | j = fftLen >> 2; |
emilmont | 1:fdd22bb7aa52 | 797 | ptr1 = &pSrc[0]; |
emilmont | 1:fdd22bb7aa52 | 798 | |
emilmont | 1:fdd22bb7aa52 | 799 | /* Calculations of last stage */ |
emilmont | 1:fdd22bb7aa52 | 800 | do |
emilmont | 1:fdd22bb7aa52 | 801 | { |
emilmont | 1:fdd22bb7aa52 | 802 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 803 | /* Read xa (real), ya(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 804 | xaya = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 805 | xa = (q31_t) xaya; |
emilmont | 1:fdd22bb7aa52 | 806 | ya = (q31_t) (xaya >> 32); |
emilmont | 1:fdd22bb7aa52 | 807 | |
emilmont | 1:fdd22bb7aa52 | 808 | /* Read xb (real), yb(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 809 | xbyb = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 810 | xb = (q31_t) xbyb; |
emilmont | 1:fdd22bb7aa52 | 811 | yb = (q31_t) (xbyb >> 32); |
emilmont | 1:fdd22bb7aa52 | 812 | |
emilmont | 1:fdd22bb7aa52 | 813 | /* Read xc (real), yc(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 814 | xcyc = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 815 | xc = (q31_t) xcyc; |
emilmont | 1:fdd22bb7aa52 | 816 | yc = (q31_t) (xcyc >> 32); |
emilmont | 1:fdd22bb7aa52 | 817 | |
emilmont | 1:fdd22bb7aa52 | 818 | /* Read xc (real), yc(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 819 | xdyd = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 820 | xd = (q31_t) xdyd; |
emilmont | 1:fdd22bb7aa52 | 821 | yd = (q31_t) (xdyd >> 32); |
emilmont | 1:fdd22bb7aa52 | 822 | |
emilmont | 1:fdd22bb7aa52 | 823 | #else |
emilmont | 1:fdd22bb7aa52 | 824 | |
emilmont | 1:fdd22bb7aa52 | 825 | /* Read xa (real), ya(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 826 | xaya = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 827 | ya = (q31_t) xaya; |
emilmont | 1:fdd22bb7aa52 | 828 | xa = (q31_t) (xaya >> 32); |
emilmont | 1:fdd22bb7aa52 | 829 | |
emilmont | 1:fdd22bb7aa52 | 830 | /* Read xb (real), yb(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 831 | xbyb = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 832 | yb = (q31_t) xbyb; |
emilmont | 1:fdd22bb7aa52 | 833 | xb = (q31_t) (xbyb >> 32); |
emilmont | 1:fdd22bb7aa52 | 834 | |
emilmont | 1:fdd22bb7aa52 | 835 | /* Read xc (real), yc(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 836 | xcyc = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 837 | yc = (q31_t) xcyc; |
emilmont | 1:fdd22bb7aa52 | 838 | xc = (q31_t) (xcyc >> 32); |
emilmont | 1:fdd22bb7aa52 | 839 | |
emilmont | 1:fdd22bb7aa52 | 840 | /* Read xc (real), yc(imag) input */ |
emilmont | 1:fdd22bb7aa52 | 841 | xdyd = *__SIMD64(ptr1)++; |
emilmont | 1:fdd22bb7aa52 | 842 | yd = (q31_t) xdyd; |
emilmont | 1:fdd22bb7aa52 | 843 | xd = (q31_t) (xdyd >> 32); |
emilmont | 1:fdd22bb7aa52 | 844 | |
emilmont | 1:fdd22bb7aa52 | 845 | |
emilmont | 1:fdd22bb7aa52 | 846 | #endif |
emilmont | 1:fdd22bb7aa52 | 847 | |
emilmont | 1:fdd22bb7aa52 | 848 | /* xa' = xa + xb + xc + xd */ |
emilmont | 1:fdd22bb7aa52 | 849 | xa_out = xa + xb + xc + xd; |
emilmont | 1:fdd22bb7aa52 | 850 | |
emilmont | 1:fdd22bb7aa52 | 851 | /* ya' = ya + yb + yc + yd */ |
emilmont | 1:fdd22bb7aa52 | 852 | ya_out = ya + yb + yc + yd; |
emilmont | 1:fdd22bb7aa52 | 853 | |
emilmont | 1:fdd22bb7aa52 | 854 | /* pointer updation for writing */ |
emilmont | 1:fdd22bb7aa52 | 855 | ptr1 = ptr1 - 8u; |
emilmont | 1:fdd22bb7aa52 | 856 | |
emilmont | 1:fdd22bb7aa52 | 857 | /* writing xa' and ya' */ |
emilmont | 1:fdd22bb7aa52 | 858 | *ptr1++ = xa_out; |
emilmont | 1:fdd22bb7aa52 | 859 | *ptr1++ = ya_out; |
emilmont | 1:fdd22bb7aa52 | 860 | |
emilmont | 1:fdd22bb7aa52 | 861 | xc_out = (xa - xb + xc - xd); |
emilmont | 1:fdd22bb7aa52 | 862 | yc_out = (ya - yb + yc - yd); |
emilmont | 1:fdd22bb7aa52 | 863 | |
emilmont | 1:fdd22bb7aa52 | 864 | /* writing xc' and yc' */ |
emilmont | 1:fdd22bb7aa52 | 865 | *ptr1++ = xc_out; |
emilmont | 1:fdd22bb7aa52 | 866 | *ptr1++ = yc_out; |
emilmont | 1:fdd22bb7aa52 | 867 | |
emilmont | 1:fdd22bb7aa52 | 868 | xb_out = (xa - yb - xc + yd); |
emilmont | 1:fdd22bb7aa52 | 869 | yb_out = (ya + xb - yc - xd); |
emilmont | 1:fdd22bb7aa52 | 870 | |
emilmont | 1:fdd22bb7aa52 | 871 | /* writing xb' and yb' */ |
emilmont | 1:fdd22bb7aa52 | 872 | *ptr1++ = xb_out; |
emilmont | 1:fdd22bb7aa52 | 873 | *ptr1++ = yb_out; |
emilmont | 1:fdd22bb7aa52 | 874 | |
emilmont | 1:fdd22bb7aa52 | 875 | xd_out = (xa + yb - xc - yd); |
emilmont | 1:fdd22bb7aa52 | 876 | yd_out = (ya - xb - yc + xd); |
emilmont | 1:fdd22bb7aa52 | 877 | |
emilmont | 1:fdd22bb7aa52 | 878 | /* writing xd' and yd' */ |
emilmont | 1:fdd22bb7aa52 | 879 | *ptr1++ = xd_out; |
emilmont | 1:fdd22bb7aa52 | 880 | *ptr1++ = yd_out; |
emilmont | 1:fdd22bb7aa52 | 881 | |
emilmont | 1:fdd22bb7aa52 | 882 | |
emilmont | 1:fdd22bb7aa52 | 883 | } while(--j); |
emilmont | 1:fdd22bb7aa52 | 884 | |
emilmont | 1:fdd22bb7aa52 | 885 | /* output is in 11.21(q21) format for the 1024 point */ |
emilmont | 1:fdd22bb7aa52 | 886 | /* output is in 9.23(q23) format for the 256 point */ |
emilmont | 1:fdd22bb7aa52 | 887 | /* output is in 7.25(q25) format for the 64 point */ |
emilmont | 1:fdd22bb7aa52 | 888 | /* output is in 5.27(q27) format for the 16 point */ |
emilmont | 1:fdd22bb7aa52 | 889 | |
emilmont | 1:fdd22bb7aa52 | 890 | /* End of last stage process */ |
emilmont | 1:fdd22bb7aa52 | 891 | } |