mbed official / mbed-dsp

CMSIS DSP library

Dependents:   performance_timer Surfboard_ gps2rtty Capstone ... more

Legacy Warning

This is an mbed 2 library. To learn more about mbed OS 5, visit the docs.

cmsis_dsp/TransformFunctions/arm_cfft_radix4_f32.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?

UserRevisionLine numberNew 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_f32.c
emilmont 1:fdd22bb7aa52 9 *
emilmont 1:fdd22bb7aa52 10 * Description: Radix-4 Decimation in Frequency CFFT & CIFFT Floating point processing function
emilmont 1:fdd22bb7aa52 11 *
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
emilmont 1:fdd22bb7aa52 40 #include "arm_math.h"
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 * @defgroup Radix4_CFFT_CIFFT Radix-4 Complex FFT Functions
emilmont 1:fdd22bb7aa52 48 *
emilmont 1:fdd22bb7aa52 49 * \par
emilmont 1:fdd22bb7aa52 50 * Complex Fast Fourier Transform(CFFT) and Complex Inverse Fast Fourier Transform(CIFFT) is an efficient algorithm to compute Discrete Fourier Transform(DFT) and Inverse Discrete Fourier Transform(IDFT).
emilmont 1:fdd22bb7aa52 51 * Computational complexity of CFFT reduces drastically when compared to DFT.
emilmont 1:fdd22bb7aa52 52 * \par
emilmont 1:fdd22bb7aa52 53 * This set of functions implements CFFT/CIFFT
emilmont 1:fdd22bb7aa52 54 * for Q15, Q31, and floating-point data types. The functions operates on in-place buffer which uses same buffer for input and output.
emilmont 1:fdd22bb7aa52 55 * Complex input is stored in input buffer in an interleaved fashion.
emilmont 1:fdd22bb7aa52 56 *
emilmont 1:fdd22bb7aa52 57 * \par
emilmont 1:fdd22bb7aa52 58 * The functions operate on blocks of input and output data and each call to the function processes
emilmont 1:fdd22bb7aa52 59 * <code>2*fftLen</code> samples through the transform. <code>pSrc</code> points to In-place arrays containing <code>2*fftLen</code> values.
emilmont 1:fdd22bb7aa52 60 * \par
emilmont 1:fdd22bb7aa52 61 * The <code>pSrc</code> points to the array of in-place buffer of size <code>2*fftLen</code> and inputs and outputs are stored in an interleaved fashion as shown below.
emilmont 1:fdd22bb7aa52 62 * <pre> {real[0], imag[0], real[1], imag[1],..} </pre>
emilmont 1:fdd22bb7aa52 63 *
emilmont 1:fdd22bb7aa52 64 * \par Lengths supported by the transform:
emilmont 1:fdd22bb7aa52 65 * \par
emilmont 1:fdd22bb7aa52 66 * Internally, the function utilize a radix-4 decimation in frequency(DIF) algorithm
emilmont 1:fdd22bb7aa52 67 * and the size of the FFT supported are of the lengths [16, 64, 256, 1024].
emilmont 1:fdd22bb7aa52 68 *
emilmont 1:fdd22bb7aa52 69 *
emilmont 1:fdd22bb7aa52 70 * \par Algorithm:
emilmont 1:fdd22bb7aa52 71 *
emilmont 1:fdd22bb7aa52 72 * <b>Complex Fast Fourier Transform:</b>
emilmont 1:fdd22bb7aa52 73 * \par
emilmont 1:fdd22bb7aa52 74 * Input real and imaginary data:
emilmont 1:fdd22bb7aa52 75 * <pre>
emilmont 1:fdd22bb7aa52 76 * x(n) = xa + j * ya
emilmont 1:fdd22bb7aa52 77 * x(n+N/4 ) = xb + j * yb
emilmont 1:fdd22bb7aa52 78 * x(n+N/2 ) = xc + j * yc
emilmont 1:fdd22bb7aa52 79 * x(n+3N 4) = xd + j * yd
emilmont 1:fdd22bb7aa52 80 * </pre>
emilmont 1:fdd22bb7aa52 81 * where N is length of FFT
emilmont 1:fdd22bb7aa52 82 * \par
emilmont 1:fdd22bb7aa52 83 * Output real and imaginary data:
emilmont 1:fdd22bb7aa52 84 * <pre>
emilmont 1:fdd22bb7aa52 85 * X(4r) = xa'+ j * ya'
emilmont 1:fdd22bb7aa52 86 * X(4r+1) = xb'+ j * yb'
emilmont 1:fdd22bb7aa52 87 * X(4r+2) = xc'+ j * yc'
emilmont 1:fdd22bb7aa52 88 * X(4r+3) = xd'+ j * yd'
emilmont 1:fdd22bb7aa52 89 * </pre>
emilmont 1:fdd22bb7aa52 90 * \par
emilmont 1:fdd22bb7aa52 91 * Twiddle factors for radix-4 FFT:
emilmont 1:fdd22bb7aa52 92 * <pre>
emilmont 1:fdd22bb7aa52 93 * Wn = co1 + j * (- si1)
emilmont 1:fdd22bb7aa52 94 * W2n = co2 + j * (- si2)
emilmont 1:fdd22bb7aa52 95 * W3n = co3 + j * (- si3)
emilmont 1:fdd22bb7aa52 96 * </pre>
emilmont 1:fdd22bb7aa52 97 *
emilmont 1:fdd22bb7aa52 98 * \par
emilmont 1:fdd22bb7aa52 99 * \image html CFFT.gif "Radix-4 Decimation-in Frequency Complex Fast Fourier Transform"
emilmont 1:fdd22bb7aa52 100 *
emilmont 1:fdd22bb7aa52 101 * \par
emilmont 1:fdd22bb7aa52 102 * Output from Radix-4 CFFT Results in Digit reversal order. Interchange middle two branches of every butterfly results in Bit reversed output.
emilmont 1:fdd22bb7aa52 103 * \par
emilmont 1:fdd22bb7aa52 104 * <b> Butterfly CFFT equations:</b>
emilmont 1:fdd22bb7aa52 105 * <pre>
emilmont 1:fdd22bb7aa52 106 * xa' = xa + xb + xc + xd
emilmont 1:fdd22bb7aa52 107 * ya' = ya + yb + yc + yd
emilmont 1:fdd22bb7aa52 108 * xc' = (xa+yb-xc-yd)* co1 + (ya-xb-yc+xd)* (si1)
emilmont 1:fdd22bb7aa52 109 * yc' = (ya-xb-yc+xd)* co1 - (xa+yb-xc-yd)* (si1)
emilmont 1:fdd22bb7aa52 110 * xb' = (xa-xb+xc-xd)* co2 + (ya-yb+yc-yd)* (si2)
emilmont 1:fdd22bb7aa52 111 * yb' = (ya-yb+yc-yd)* co2 - (xa-xb+xc-xd)* (si2)
emilmont 1:fdd22bb7aa52 112 * xd' = (xa-yb-xc+yd)* co3 + (ya+xb-yc-xd)* (si3)
emilmont 1:fdd22bb7aa52 113 * yd' = (ya+xb-yc-xd)* co3 - (xa-yb-xc+yd)* (si3)
emilmont 1:fdd22bb7aa52 114 * </pre>
emilmont 1:fdd22bb7aa52 115 *
emilmont 1:fdd22bb7aa52 116 *
emilmont 1:fdd22bb7aa52 117 * <b>Complex Inverse Fast Fourier Transform:</b>
emilmont 1:fdd22bb7aa52 118 * \par
emilmont 1:fdd22bb7aa52 119 * CIFFT uses same twiddle factor table as CFFT with modifications in the design equation as shown below.
emilmont 1:fdd22bb7aa52 120 *
emilmont 1:fdd22bb7aa52 121 * \par
emilmont 1:fdd22bb7aa52 122 * <b> Modified Butterfly CIFFT equations:</b>
emilmont 1:fdd22bb7aa52 123 * <pre>
emilmont 1:fdd22bb7aa52 124 * xa' = xa + xb + xc + xd
emilmont 1:fdd22bb7aa52 125 * ya' = ya + yb + yc + yd
emilmont 1:fdd22bb7aa52 126 * xc' = (xa-yb-xc+yd)* co1 - (ya+xb-yc-xd)* (si1)
emilmont 1:fdd22bb7aa52 127 * yc' = (ya+xb-yc-xd)* co1 + (xa-yb-xc+yd)* (si1)
emilmont 1:fdd22bb7aa52 128 * xb' = (xa-xb+xc-xd)* co2 - (ya-yb+yc-yd)* (si2)
emilmont 1:fdd22bb7aa52 129 * yb' = (ya-yb+yc-yd)* co2 + (xa-xb+xc-xd)* (si2)
emilmont 1:fdd22bb7aa52 130 * xd' = (xa+yb-xc-yd)* co3 - (ya-xb-yc+xd)* (si3)
emilmont 1:fdd22bb7aa52 131 * yd' = (ya-xb-yc+xd)* co3 + (xa+yb-xc-yd)* (si3)
emilmont 1:fdd22bb7aa52 132 * </pre>
emilmont 1:fdd22bb7aa52 133 *
emilmont 1:fdd22bb7aa52 134 * \par Instance Structure
emilmont 1:fdd22bb7aa52 135 * A separate instance structure must be defined for each Instance but the twiddle factors and bit reversal tables can be reused.
emilmont 1:fdd22bb7aa52 136 * There are separate instance structure declarations for each of the 3 supported data types.
emilmont 1:fdd22bb7aa52 137 *
emilmont 1:fdd22bb7aa52 138 * \par Initialization Functions
emilmont 1:fdd22bb7aa52 139 * There is also an associated initialization function for each data type.
emilmont 1:fdd22bb7aa52 140 * The initialization function performs the following operations:
emilmont 1:fdd22bb7aa52 141 * - Sets the values of the internal structure fields.
emilmont 1:fdd22bb7aa52 142 * - Initializes twiddle factor table and bit reversal table pointers
emilmont 1:fdd22bb7aa52 143 * \par
emilmont 1:fdd22bb7aa52 144 * Use of the initialization function is optional.
emilmont 1:fdd22bb7aa52 145 * However, if the initialization function is used, then the instance structure cannot be placed into a const data section.
emilmont 1:fdd22bb7aa52 146 * To place an instance structure into a const data section, the instance structure must be manually initialized.
emilmont 1:fdd22bb7aa52 147 * Manually initialize the instance structure as follows:
emilmont 1:fdd22bb7aa52 148 * <pre>
emilmont 1:fdd22bb7aa52 149 *arm_cfft_radix4_instance_f32 S = {fftLen, ifftFlag, bitReverseFlag, pTwiddle, pBitRevTable, twidCoefModifier, bitRevFactor, onebyfftLen};
emilmont 1:fdd22bb7aa52 150 *arm_cfft_radix4_instance_q31 S = {fftLen, ifftFlag, bitReverseFlag, pTwiddle, pBitRevTable, twidCoefModifier, bitRevFactor};
emilmont 1:fdd22bb7aa52 151 *arm_cfft_radix4_instance_q15 S = {fftLen, ifftFlag, bitReverseFlag, pTwiddle, pBitRevTable, twidCoefModifier, bitRevFactor};
emilmont 1:fdd22bb7aa52 152 * </pre>
emilmont 1:fdd22bb7aa52 153 * \par
emilmont 1:fdd22bb7aa52 154 * where <code>fftLen</code> length of CFFT/CIFFT; <code>ifftFlag</code> Flag for selection of CFFT or CIFFT(Set ifftFlag to calculate CIFFT otherwise calculates CFFT);
emilmont 1:fdd22bb7aa52 155 * <code>bitReverseFlag</code> Flag for selection of output order(Set bitReverseFlag to output in normal order otherwise output in bit reversed order);
emilmont 1:fdd22bb7aa52 156 * <code>pTwiddle</code>points to array of twiddle coefficients; <code>pBitRevTable</code> points to the array of bit reversal table.
emilmont 1:fdd22bb7aa52 157 * <code>twidCoefModifier</code> modifier for twiddle factor table which supports all FFT lengths with same table;
emilmont 1:fdd22bb7aa52 158 * <code>pBitRevTable</code> modifier for bit reversal table which supports all FFT lengths with same table.
emilmont 1:fdd22bb7aa52 159 * <code>onebyfftLen</code> value of 1/fftLen to calculate CIFFT;
emilmont 1:fdd22bb7aa52 160 *
emilmont 1:fdd22bb7aa52 161 * \par Fixed-Point Behavior
emilmont 1:fdd22bb7aa52 162 * Care must be taken when using the fixed-point versions of the CFFT/CIFFT function.
emilmont 1:fdd22bb7aa52 163 * Refer to the function specific documentation below for usage guidelines.
emilmont 1:fdd22bb7aa52 164 */
emilmont 1:fdd22bb7aa52 165
emilmont 1:fdd22bb7aa52 166
emilmont 1:fdd22bb7aa52 167 /**
emilmont 1:fdd22bb7aa52 168 * @addtogroup Radix4_CFFT_CIFFT
emilmont 1:fdd22bb7aa52 169 * @{
emilmont 1:fdd22bb7aa52 170 */
emilmont 1:fdd22bb7aa52 171
emilmont 1:fdd22bb7aa52 172 /**
emilmont 1:fdd22bb7aa52 173 * @details
emilmont 1:fdd22bb7aa52 174 * @brief Processing function for the floating-point Radix-4 CFFT/CIFFT.
emilmont 1:fdd22bb7aa52 175 * @param[in] *S points to an instance of the floating-point Radix-4 CFFT/CIFFT structure.
emilmont 1:fdd22bb7aa52 176 * @param[in, out] *pSrc points to the complex data buffer of size <code>2*fftLen</code>. Processing occurs in-place.
emilmont 1:fdd22bb7aa52 177 * @return none.
emilmont 1:fdd22bb7aa52 178 */
emilmont 1:fdd22bb7aa52 179
emilmont 1:fdd22bb7aa52 180 void arm_cfft_radix4_f32(
emilmont 1:fdd22bb7aa52 181 const arm_cfft_radix4_instance_f32 * S,
emilmont 1:fdd22bb7aa52 182 float32_t * pSrc)
emilmont 1:fdd22bb7aa52 183 {
emilmont 1:fdd22bb7aa52 184
emilmont 1:fdd22bb7aa52 185 if(S->ifftFlag == 1u)
emilmont 1:fdd22bb7aa52 186 {
emilmont 1:fdd22bb7aa52 187 /* Complex IFFT radix-4 */
emilmont 1:fdd22bb7aa52 188 arm_radix4_butterfly_inverse_f32(pSrc, S->fftLen, S->pTwiddle,
emilmont 1:fdd22bb7aa52 189 S->twidCoefModifier, S->onebyfftLen);
emilmont 1:fdd22bb7aa52 190 }
emilmont 1:fdd22bb7aa52 191 else
emilmont 1:fdd22bb7aa52 192 {
emilmont 1:fdd22bb7aa52 193 /* Complex FFT radix-4 */
emilmont 1:fdd22bb7aa52 194 arm_radix4_butterfly_f32(pSrc, S->fftLen, S->pTwiddle,
emilmont 1:fdd22bb7aa52 195 S->twidCoefModifier);
emilmont 1:fdd22bb7aa52 196 }
emilmont 1:fdd22bb7aa52 197
emilmont 1:fdd22bb7aa52 198 if(S->bitReverseFlag == 1u)
emilmont 1:fdd22bb7aa52 199 {
emilmont 1:fdd22bb7aa52 200 /* Bit Reversal */
emilmont 1:fdd22bb7aa52 201 arm_bitreversal_f32(pSrc, S->fftLen, S->bitRevFactor, S->pBitRevTable);
emilmont 1:fdd22bb7aa52 202 }
emilmont 1:fdd22bb7aa52 203
emilmont 1:fdd22bb7aa52 204 }
emilmont 1:fdd22bb7aa52 205
emilmont 1:fdd22bb7aa52 206
emilmont 1:fdd22bb7aa52 207 /**
emilmont 1:fdd22bb7aa52 208 * @} end of Radix4_CFFT_CIFFT group
emilmont 1:fdd22bb7aa52 209 */
emilmont 1:fdd22bb7aa52 210
emilmont 1:fdd22bb7aa52 211
emilmont 1:fdd22bb7aa52 212 /* ----------------------------------------------------------------------
emilmont 1:fdd22bb7aa52 213 ** Internal helper function used by the FFTs
emilmont 1:fdd22bb7aa52 214 ** ------------------------------------------------------------------- */
emilmont 1:fdd22bb7aa52 215
emilmont 1:fdd22bb7aa52 216 /*
emilmont 1:fdd22bb7aa52 217 * @brief Core function for the floating-point CFFT butterfly process.
emilmont 1:fdd22bb7aa52 218 * @param[in, out] *pSrc points to the in-place buffer of floating-point data type.
emilmont 1:fdd22bb7aa52 219 * @param[in] fftLen length of the FFT.
emilmont 1:fdd22bb7aa52 220 * @param[in] *pCoef points to the twiddle coefficient buffer.
emilmont 1:fdd22bb7aa52 221 * @param[in] twidCoefModifier twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table.
emilmont 1:fdd22bb7aa52 222 * @return none.
emilmont 1:fdd22bb7aa52 223 */
emilmont 1:fdd22bb7aa52 224
emilmont 1:fdd22bb7aa52 225 void arm_radix4_butterfly_f32(
emilmont 1:fdd22bb7aa52 226 float32_t * pSrc,
emilmont 1:fdd22bb7aa52 227 uint16_t fftLen,
emilmont 1:fdd22bb7aa52 228 float32_t * pCoef,
emilmont 1:fdd22bb7aa52 229 uint16_t twidCoefModifier)
emilmont 1:fdd22bb7aa52 230 {
emilmont 1:fdd22bb7aa52 231
emilmont 1:fdd22bb7aa52 232 float32_t co1, co2, co3, si1, si2, si3;
emilmont 1:fdd22bb7aa52 233 uint32_t ia1, ia2, ia3;
emilmont 1:fdd22bb7aa52 234 uint32_t i0, i1, i2, i3;
emilmont 1:fdd22bb7aa52 235 uint32_t n1, n2, j, k;
emilmont 1:fdd22bb7aa52 236
emilmont 1:fdd22bb7aa52 237 #ifndef ARM_MATH_CM0
emilmont 1:fdd22bb7aa52 238
emilmont 1:fdd22bb7aa52 239 /* Run the below code for Cortex-M4 and Cortex-M3 */
emilmont 1:fdd22bb7aa52 240
emilmont 1:fdd22bb7aa52 241 float32_t xaIn, yaIn, xbIn, ybIn, xcIn, ycIn, xdIn, ydIn;
emilmont 1:fdd22bb7aa52 242 float32_t Xaplusc, Xbplusd, Yaplusc, Ybplusd, Xaminusc, Xbminusd, Yaminusc,
emilmont 1:fdd22bb7aa52 243 Ybminusd;
emilmont 1:fdd22bb7aa52 244 float32_t Xb12C_out, Yb12C_out, Xc12C_out, Yc12C_out, Xd12C_out, Yd12C_out;
emilmont 1:fdd22bb7aa52 245 float32_t Xb12_out, Yb12_out, Xc12_out, Yc12_out, Xd12_out, Yd12_out;
emilmont 1:fdd22bb7aa52 246 float32_t *ptr1;
emilmont 1:fdd22bb7aa52 247
emilmont 1:fdd22bb7aa52 248 /* Initializations for the first stage */
emilmont 1:fdd22bb7aa52 249 n2 = fftLen;
emilmont 1:fdd22bb7aa52 250 n1 = n2;
emilmont 1:fdd22bb7aa52 251
emilmont 1:fdd22bb7aa52 252 /* n2 = fftLen/4 */
emilmont 1:fdd22bb7aa52 253 n2 >>= 2u;
emilmont 1:fdd22bb7aa52 254 i0 = 0u;
emilmont 1:fdd22bb7aa52 255 ia1 = 0u;
emilmont 1:fdd22bb7aa52 256
emilmont 1:fdd22bb7aa52 257 j = n2;
emilmont 1:fdd22bb7aa52 258
emilmont 1:fdd22bb7aa52 259 /* Calculation of first stage */
emilmont 1:fdd22bb7aa52 260 do
emilmont 1:fdd22bb7aa52 261 {
emilmont 1:fdd22bb7aa52 262 /* index calculation for the input as, */
emilmont 1:fdd22bb7aa52 263 /* pSrc[i0 + 0], pSrc[i0 + fftLen/4], pSrc[i0 + fftLen/2], pSrc[i0 + 3fftLen/4] */
emilmont 1:fdd22bb7aa52 264 i1 = i0 + n2;
emilmont 1:fdd22bb7aa52 265 i2 = i1 + n2;
emilmont 1:fdd22bb7aa52 266 i3 = i2 + n2;
emilmont 1:fdd22bb7aa52 267
emilmont 1:fdd22bb7aa52 268 xaIn = pSrc[(2u * i0)];
emilmont 1:fdd22bb7aa52 269 yaIn = pSrc[(2u * i0) + 1u];
emilmont 1:fdd22bb7aa52 270
emilmont 1:fdd22bb7aa52 271 xcIn = pSrc[(2u * i2)];
emilmont 1:fdd22bb7aa52 272 ycIn = pSrc[(2u * i2) + 1u];
emilmont 1:fdd22bb7aa52 273
emilmont 1:fdd22bb7aa52 274 xbIn = pSrc[(2u * i1)];
emilmont 1:fdd22bb7aa52 275 ybIn = pSrc[(2u * i1) + 1u];
emilmont 1:fdd22bb7aa52 276
emilmont 1:fdd22bb7aa52 277 xdIn = pSrc[(2u * i3)];
emilmont 1:fdd22bb7aa52 278 ydIn = pSrc[(2u * i3) + 1u];
emilmont 1:fdd22bb7aa52 279
emilmont 1:fdd22bb7aa52 280 /* xa + xc */
emilmont 1:fdd22bb7aa52 281 Xaplusc = xaIn + xcIn;
emilmont 1:fdd22bb7aa52 282 /* xb + xd */
emilmont 1:fdd22bb7aa52 283 Xbplusd = xbIn + xdIn;
emilmont 1:fdd22bb7aa52 284 /* ya + yc */
emilmont 1:fdd22bb7aa52 285 Yaplusc = yaIn + ycIn;
emilmont 1:fdd22bb7aa52 286 /* yb + yd */
emilmont 1:fdd22bb7aa52 287 Ybplusd = ybIn + ydIn;
emilmont 1:fdd22bb7aa52 288
emilmont 1:fdd22bb7aa52 289 /* index calculation for the coefficients */
emilmont 1:fdd22bb7aa52 290 ia2 = ia1 + ia1;
emilmont 1:fdd22bb7aa52 291 co2 = pCoef[ia2 * 2u];
emilmont 1:fdd22bb7aa52 292 si2 = pCoef[(ia2 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 293
emilmont 1:fdd22bb7aa52 294 /* xa - xc */
emilmont 1:fdd22bb7aa52 295 Xaminusc = xaIn - xcIn;
emilmont 1:fdd22bb7aa52 296 /* xb - xd */
emilmont 1:fdd22bb7aa52 297 Xbminusd = xbIn - xdIn;
emilmont 1:fdd22bb7aa52 298 /* ya - yc */
emilmont 1:fdd22bb7aa52 299 Yaminusc = yaIn - ycIn;
emilmont 1:fdd22bb7aa52 300 /* yb + yd */
emilmont 1:fdd22bb7aa52 301 Ybminusd = ybIn - ydIn;
emilmont 1:fdd22bb7aa52 302
emilmont 1:fdd22bb7aa52 303 /* xa' = xa + xb + xc + xd */
emilmont 1:fdd22bb7aa52 304 pSrc[(2u * i0)] = Xaplusc + Xbplusd;
emilmont 1:fdd22bb7aa52 305 /* ya' = ya + yb + yc + yd */
emilmont 1:fdd22bb7aa52 306 pSrc[(2u * i0) + 1u] = Yaplusc + Ybplusd;
emilmont 1:fdd22bb7aa52 307
emilmont 1:fdd22bb7aa52 308 /* (xa - xc) + (yb - yd) */
emilmont 1:fdd22bb7aa52 309 Xb12C_out = (Xaminusc + Ybminusd);
emilmont 1:fdd22bb7aa52 310 /* (ya - yc) + (xb - xd) */
emilmont 1:fdd22bb7aa52 311 Yb12C_out = (Yaminusc - Xbminusd);
emilmont 1:fdd22bb7aa52 312 /* (xa + xc) - (xb + xd) */
emilmont 1:fdd22bb7aa52 313 Xc12C_out = (Xaplusc - Xbplusd);
emilmont 1:fdd22bb7aa52 314 /* (ya + yc) - (yb + yd) */
emilmont 1:fdd22bb7aa52 315 Yc12C_out = (Yaplusc - Ybplusd);
emilmont 1:fdd22bb7aa52 316 /* (xa - xc) - (yb - yd) */
emilmont 1:fdd22bb7aa52 317 Xd12C_out = (Xaminusc - Ybminusd);
emilmont 1:fdd22bb7aa52 318 /* (ya - yc) + (xb - xd) */
emilmont 1:fdd22bb7aa52 319 Yd12C_out = (Xbminusd + Yaminusc);
emilmont 1:fdd22bb7aa52 320
emilmont 1:fdd22bb7aa52 321 co1 = pCoef[ia1 * 2u];
emilmont 1:fdd22bb7aa52 322 si1 = pCoef[(ia1 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 323
emilmont 1:fdd22bb7aa52 324 /* index calculation for the coefficients */
emilmont 1:fdd22bb7aa52 325 ia3 = ia2 + ia1;
emilmont 1:fdd22bb7aa52 326 co3 = pCoef[ia3 * 2u];
emilmont 1:fdd22bb7aa52 327 si3 = pCoef[(ia3 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 328
emilmont 1:fdd22bb7aa52 329 Xb12_out = Xb12C_out * co1;
emilmont 1:fdd22bb7aa52 330 Yb12_out = Yb12C_out * co1;
emilmont 1:fdd22bb7aa52 331 Xc12_out = Xc12C_out * co2;
emilmont 1:fdd22bb7aa52 332 Yc12_out = Yc12C_out * co2;
emilmont 1:fdd22bb7aa52 333 Xd12_out = Xd12C_out * co3;
emilmont 1:fdd22bb7aa52 334 Yd12_out = Yd12C_out * co3;
emilmont 1:fdd22bb7aa52 335
emilmont 1:fdd22bb7aa52 336 /* xb' = (xa+yb-xc-yd)co1 + (ya-xb-yc+xd)(si1) */
emilmont 1:fdd22bb7aa52 337 Xb12_out += Yb12C_out * si1;
emilmont 1:fdd22bb7aa52 338 /* yb' = (ya-xb-yc+xd)co1 - (xa+yb-xc-yd)(si1) */
emilmont 1:fdd22bb7aa52 339 Yb12_out -= Xb12C_out * si1;
emilmont 1:fdd22bb7aa52 340 /* xc' = (xa-xb+xc-xd)co2 + (ya-yb+yc-yd)(si2) */
emilmont 1:fdd22bb7aa52 341 Xc12_out += Yc12C_out * si2;
emilmont 1:fdd22bb7aa52 342 /* yc' = (ya-yb+yc-yd)co2 - (xa-xb+xc-xd)(si2) */
emilmont 1:fdd22bb7aa52 343 Yc12_out -= Xc12C_out * si2;
emilmont 1:fdd22bb7aa52 344 /* xd' = (xa-yb-xc+yd)co3 + (ya+xb-yc-xd)(si3) */
emilmont 1:fdd22bb7aa52 345 Xd12_out += Yd12C_out * si3;
emilmont 1:fdd22bb7aa52 346 /* yd' = (ya+xb-yc-xd)co3 - (xa-yb-xc+yd)(si3) */
emilmont 1:fdd22bb7aa52 347 Yd12_out -= Xd12C_out * si3;
emilmont 1:fdd22bb7aa52 348
emilmont 1:fdd22bb7aa52 349
emilmont 1:fdd22bb7aa52 350 /* xc' = (xa-xb+xc-xd)co2 + (ya-yb+yc-yd)(si2) */
emilmont 1:fdd22bb7aa52 351 pSrc[2u * i1] = Xc12_out;
emilmont 1:fdd22bb7aa52 352
emilmont 1:fdd22bb7aa52 353 /* yc' = (ya-yb+yc-yd)co2 - (xa-xb+xc-xd)(si2) */
emilmont 1:fdd22bb7aa52 354 pSrc[(2u * i1) + 1u] = Yc12_out;
emilmont 1:fdd22bb7aa52 355
emilmont 1:fdd22bb7aa52 356 /* xb' = (xa+yb-xc-yd)co1 + (ya-xb-yc+xd)(si1) */
emilmont 1:fdd22bb7aa52 357 pSrc[2u * i2] = Xb12_out;
emilmont 1:fdd22bb7aa52 358
emilmont 1:fdd22bb7aa52 359 /* yb' = (ya-xb-yc+xd)co1 - (xa+yb-xc-yd)(si1) */
emilmont 1:fdd22bb7aa52 360 pSrc[(2u * i2) + 1u] = Yb12_out;
emilmont 1:fdd22bb7aa52 361
emilmont 1:fdd22bb7aa52 362 /* xd' = (xa-yb-xc+yd)co3 + (ya+xb-yc-xd)(si3) */
emilmont 1:fdd22bb7aa52 363 pSrc[2u * i3] = Xd12_out;
emilmont 1:fdd22bb7aa52 364
emilmont 1:fdd22bb7aa52 365 /* yd' = (ya+xb-yc-xd)co3 - (xa-yb-xc+yd)(si3) */
emilmont 1:fdd22bb7aa52 366 pSrc[(2u * i3) + 1u] = Yd12_out;
emilmont 1:fdd22bb7aa52 367
emilmont 1:fdd22bb7aa52 368 /* Twiddle coefficients index modifier */
emilmont 1:fdd22bb7aa52 369 ia1 = ia1 + twidCoefModifier;
emilmont 1:fdd22bb7aa52 370
emilmont 1:fdd22bb7aa52 371 /* Updating input index */
emilmont 1:fdd22bb7aa52 372 i0 = i0 + 1u;
emilmont 1:fdd22bb7aa52 373
emilmont 1:fdd22bb7aa52 374 }
emilmont 1:fdd22bb7aa52 375 while(--j);
emilmont 1:fdd22bb7aa52 376
emilmont 1:fdd22bb7aa52 377 twidCoefModifier <<= 2u;
emilmont 1:fdd22bb7aa52 378
emilmont 1:fdd22bb7aa52 379 /* Calculation of second stage to excluding last stage */
emilmont 1:fdd22bb7aa52 380 for (k = fftLen / 4; k > 4u; k >>= 2u)
emilmont 1:fdd22bb7aa52 381 {
emilmont 1:fdd22bb7aa52 382 /* Initializations for the first stage */
emilmont 1:fdd22bb7aa52 383 n1 = n2;
emilmont 1:fdd22bb7aa52 384 n2 >>= 2u;
emilmont 1:fdd22bb7aa52 385 ia1 = 0u;
emilmont 1:fdd22bb7aa52 386
emilmont 1:fdd22bb7aa52 387 /* Calculation of first stage */
emilmont 1:fdd22bb7aa52 388 for (j = 0u; j <= (n2 - 1u); j++)
emilmont 1:fdd22bb7aa52 389 {
emilmont 1:fdd22bb7aa52 390 /* index calculation for the coefficients */
emilmont 1:fdd22bb7aa52 391 ia2 = ia1 + ia1;
emilmont 1:fdd22bb7aa52 392 ia3 = ia2 + ia1;
emilmont 1:fdd22bb7aa52 393 co1 = pCoef[ia1 * 2u];
emilmont 1:fdd22bb7aa52 394 si1 = pCoef[(ia1 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 395 co2 = pCoef[ia2 * 2u];
emilmont 1:fdd22bb7aa52 396 si2 = pCoef[(ia2 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 397 co3 = pCoef[ia3 * 2u];
emilmont 1:fdd22bb7aa52 398 si3 = pCoef[(ia3 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 399
emilmont 1:fdd22bb7aa52 400 /* Twiddle coefficients index modifier */
emilmont 1:fdd22bb7aa52 401 ia1 = ia1 + twidCoefModifier;
emilmont 1:fdd22bb7aa52 402
emilmont 1:fdd22bb7aa52 403 for (i0 = j; i0 < fftLen; i0 += n1)
emilmont 1:fdd22bb7aa52 404 {
emilmont 1:fdd22bb7aa52 405 /* index calculation for the input as, */
emilmont 1:fdd22bb7aa52 406 /* pSrc[i0 + 0], pSrc[i0 + fftLen/4], pSrc[i0 + fftLen/2], pSrc[i0 + 3fftLen/4] */
emilmont 1:fdd22bb7aa52 407 i1 = i0 + n2;
emilmont 1:fdd22bb7aa52 408 i2 = i1 + n2;
emilmont 1:fdd22bb7aa52 409 i3 = i2 + n2;
emilmont 1:fdd22bb7aa52 410
emilmont 1:fdd22bb7aa52 411 xaIn = pSrc[(2u * i0)];
emilmont 1:fdd22bb7aa52 412 yaIn = pSrc[(2u * i0) + 1u];
emilmont 1:fdd22bb7aa52 413
emilmont 1:fdd22bb7aa52 414 xbIn = pSrc[(2u * i1)];
emilmont 1:fdd22bb7aa52 415 ybIn = pSrc[(2u * i1) + 1u];
emilmont 1:fdd22bb7aa52 416
emilmont 1:fdd22bb7aa52 417 xcIn = pSrc[(2u * i2)];
emilmont 1:fdd22bb7aa52 418 ycIn = pSrc[(2u * i2) + 1u];
emilmont 1:fdd22bb7aa52 419
emilmont 1:fdd22bb7aa52 420 xdIn = pSrc[(2u * i3)];
emilmont 1:fdd22bb7aa52 421 ydIn = pSrc[(2u * i3) + 1u];
emilmont 1:fdd22bb7aa52 422
emilmont 1:fdd22bb7aa52 423 /* xa - xc */
emilmont 1:fdd22bb7aa52 424 Xaminusc = xaIn - xcIn;
emilmont 1:fdd22bb7aa52 425 /* (xb - xd) */
emilmont 1:fdd22bb7aa52 426 Xbminusd = xbIn - xdIn;
emilmont 1:fdd22bb7aa52 427 /* ya - yc */
emilmont 1:fdd22bb7aa52 428 Yaminusc = yaIn - ycIn;
emilmont 1:fdd22bb7aa52 429 /* (yb - yd) */
emilmont 1:fdd22bb7aa52 430 Ybminusd = ybIn - ydIn;
emilmont 1:fdd22bb7aa52 431
emilmont 1:fdd22bb7aa52 432 /* xa + xc */
emilmont 1:fdd22bb7aa52 433 Xaplusc = xaIn + xcIn;
emilmont 1:fdd22bb7aa52 434 /* xb + xd */
emilmont 1:fdd22bb7aa52 435 Xbplusd = xbIn + xdIn;
emilmont 1:fdd22bb7aa52 436 /* ya + yc */
emilmont 1:fdd22bb7aa52 437 Yaplusc = yaIn + ycIn;
emilmont 1:fdd22bb7aa52 438 /* yb + yd */
emilmont 1:fdd22bb7aa52 439 Ybplusd = ybIn + ydIn;
emilmont 1:fdd22bb7aa52 440
emilmont 1:fdd22bb7aa52 441 /* (xa - xc) + (yb - yd) */
emilmont 1:fdd22bb7aa52 442 Xb12C_out = (Xaminusc + Ybminusd);
emilmont 1:fdd22bb7aa52 443 /* (ya - yc) - (xb - xd) */
emilmont 1:fdd22bb7aa52 444 Yb12C_out = (Yaminusc - Xbminusd);
emilmont 1:fdd22bb7aa52 445 /* xa + xc -(xb + xd) */
emilmont 1:fdd22bb7aa52 446 Xc12C_out = (Xaplusc - Xbplusd);
emilmont 1:fdd22bb7aa52 447 /* (ya + yc) - (yb + yd) */
emilmont 1:fdd22bb7aa52 448 Yc12C_out = (Yaplusc - Ybplusd);
emilmont 1:fdd22bb7aa52 449 /* (xa - xc) - (yb - yd) */
emilmont 1:fdd22bb7aa52 450 Xd12C_out = (Xaminusc - Ybminusd);
emilmont 1:fdd22bb7aa52 451 /* (ya - yc) + (xb - xd) */
emilmont 1:fdd22bb7aa52 452 Yd12C_out = (Xbminusd + Yaminusc);
emilmont 1:fdd22bb7aa52 453
emilmont 1:fdd22bb7aa52 454 pSrc[(2u * i0)] = Xaplusc + Xbplusd;
emilmont 1:fdd22bb7aa52 455 pSrc[(2u * i0) + 1u] = Yaplusc + Ybplusd;
emilmont 1:fdd22bb7aa52 456
emilmont 1:fdd22bb7aa52 457 Xb12_out = Xb12C_out * co1;
emilmont 1:fdd22bb7aa52 458 Yb12_out = Yb12C_out * co1;
emilmont 1:fdd22bb7aa52 459 Xc12_out = Xc12C_out * co2;
emilmont 1:fdd22bb7aa52 460 Yc12_out = Yc12C_out * co2;
emilmont 1:fdd22bb7aa52 461 Xd12_out = Xd12C_out * co3;
emilmont 1:fdd22bb7aa52 462 Yd12_out = Yd12C_out * co3;
emilmont 1:fdd22bb7aa52 463
emilmont 1:fdd22bb7aa52 464 /* xb' = (xa+yb-xc-yd)co1 + (ya-xb-yc+xd)(si1) */
emilmont 1:fdd22bb7aa52 465 Xb12_out += Yb12C_out * si1;
emilmont 1:fdd22bb7aa52 466 /* yb' = (ya-xb-yc+xd)co1 - (xa+yb-xc-yd)(si1) */
emilmont 1:fdd22bb7aa52 467 Yb12_out -= Xb12C_out * si1;
emilmont 1:fdd22bb7aa52 468 /* xc' = (xa-xb+xc-xd)co2 + (ya-yb+yc-yd)(si2) */
emilmont 1:fdd22bb7aa52 469 Xc12_out += Yc12C_out * si2;
emilmont 1:fdd22bb7aa52 470 /* yc' = (ya-yb+yc-yd)co2 - (xa-xb+xc-xd)(si2) */
emilmont 1:fdd22bb7aa52 471 Yc12_out -= Xc12C_out * si2;
emilmont 1:fdd22bb7aa52 472 /* xd' = (xa-yb-xc+yd)co3 + (ya+xb-yc-xd)(si3) */
emilmont 1:fdd22bb7aa52 473 Xd12_out += Yd12C_out * si3;
emilmont 1:fdd22bb7aa52 474 /* yd' = (ya+xb-yc-xd)co3 - (xa-yb-xc+yd)(si3) */
emilmont 1:fdd22bb7aa52 475 Yd12_out -= Xd12C_out * si3;
emilmont 1:fdd22bb7aa52 476
emilmont 1:fdd22bb7aa52 477 /* xc' = (xa-xb+xc-xd)co2 + (ya-yb+yc-yd)(si2) */
emilmont 1:fdd22bb7aa52 478 pSrc[2u * i1] = Xc12_out;
emilmont 1:fdd22bb7aa52 479
emilmont 1:fdd22bb7aa52 480 /* yc' = (ya-yb+yc-yd)co2 - (xa-xb+xc-xd)(si2) */
emilmont 1:fdd22bb7aa52 481 pSrc[(2u * i1) + 1u] = Yc12_out;
emilmont 1:fdd22bb7aa52 482
emilmont 1:fdd22bb7aa52 483 /* xb' = (xa+yb-xc-yd)co1 + (ya-xb-yc+xd)(si1) */
emilmont 1:fdd22bb7aa52 484 pSrc[2u * i2] = Xb12_out;
emilmont 1:fdd22bb7aa52 485
emilmont 1:fdd22bb7aa52 486 /* yb' = (ya-xb-yc+xd)co1 - (xa+yb-xc-yd)(si1) */
emilmont 1:fdd22bb7aa52 487 pSrc[(2u * i2) + 1u] = Yb12_out;
emilmont 1:fdd22bb7aa52 488
emilmont 1:fdd22bb7aa52 489 /* xd' = (xa-yb-xc+yd)co3 + (ya+xb-yc-xd)(si3) */
emilmont 1:fdd22bb7aa52 490 pSrc[2u * i3] = Xd12_out;
emilmont 1:fdd22bb7aa52 491
emilmont 1:fdd22bb7aa52 492 /* yd' = (ya+xb-yc-xd)co3 - (xa-yb-xc+yd)(si3) */
emilmont 1:fdd22bb7aa52 493 pSrc[(2u * i3) + 1u] = Yd12_out;
emilmont 1:fdd22bb7aa52 494
emilmont 1:fdd22bb7aa52 495 }
emilmont 1:fdd22bb7aa52 496 }
emilmont 1:fdd22bb7aa52 497 twidCoefModifier <<= 2u;
emilmont 1:fdd22bb7aa52 498 }
emilmont 1:fdd22bb7aa52 499
emilmont 1:fdd22bb7aa52 500 j = fftLen >> 2;
emilmont 1:fdd22bb7aa52 501 ptr1 = &pSrc[0];
emilmont 1:fdd22bb7aa52 502
emilmont 1:fdd22bb7aa52 503 /* Calculations of last stage */
emilmont 1:fdd22bb7aa52 504 do
emilmont 1:fdd22bb7aa52 505 {
emilmont 1:fdd22bb7aa52 506
emilmont 1:fdd22bb7aa52 507 xaIn = ptr1[0];
emilmont 1:fdd22bb7aa52 508 xcIn = ptr1[4];
emilmont 1:fdd22bb7aa52 509 yaIn = ptr1[1];
emilmont 1:fdd22bb7aa52 510 ycIn = ptr1[5];
emilmont 1:fdd22bb7aa52 511
emilmont 1:fdd22bb7aa52 512 /* xa + xc */
emilmont 1:fdd22bb7aa52 513 Xaplusc = xaIn + xcIn;
emilmont 1:fdd22bb7aa52 514
emilmont 1:fdd22bb7aa52 515 xbIn = ptr1[2];
emilmont 1:fdd22bb7aa52 516
emilmont 1:fdd22bb7aa52 517 /* xa - xc */
emilmont 1:fdd22bb7aa52 518 Xaminusc = xaIn - xcIn;
emilmont 1:fdd22bb7aa52 519
emilmont 1:fdd22bb7aa52 520 xdIn = ptr1[6];
emilmont 1:fdd22bb7aa52 521
emilmont 1:fdd22bb7aa52 522 /* ya + yc */
emilmont 1:fdd22bb7aa52 523 Yaplusc = yaIn + ycIn;
emilmont 1:fdd22bb7aa52 524
emilmont 1:fdd22bb7aa52 525 ybIn = ptr1[3];
emilmont 1:fdd22bb7aa52 526
emilmont 1:fdd22bb7aa52 527 /* ya - yc */
emilmont 1:fdd22bb7aa52 528 Yaminusc = yaIn - ycIn;
emilmont 1:fdd22bb7aa52 529
emilmont 1:fdd22bb7aa52 530 ydIn = ptr1[7];
emilmont 1:fdd22bb7aa52 531
emilmont 1:fdd22bb7aa52 532 /* xb + xd */
emilmont 1:fdd22bb7aa52 533 Xbplusd = xbIn + xdIn;
emilmont 1:fdd22bb7aa52 534
emilmont 1:fdd22bb7aa52 535 /* yb + yd */
emilmont 1:fdd22bb7aa52 536 Ybplusd = ybIn + ydIn;
emilmont 1:fdd22bb7aa52 537
emilmont 1:fdd22bb7aa52 538 /* xa' = xa + xb + xc + xd */
emilmont 1:fdd22bb7aa52 539 ptr1[0] = (Xaplusc + Xbplusd);
emilmont 1:fdd22bb7aa52 540
emilmont 1:fdd22bb7aa52 541 /* (xb-xd) */
emilmont 1:fdd22bb7aa52 542 Xbminusd = xbIn - xdIn;
emilmont 1:fdd22bb7aa52 543
emilmont 1:fdd22bb7aa52 544 /* ya' = ya + yb + yc + yd */
emilmont 1:fdd22bb7aa52 545 ptr1[1] = (Yaplusc + Ybplusd);
emilmont 1:fdd22bb7aa52 546
emilmont 1:fdd22bb7aa52 547 /* (yb-yd) */
emilmont 1:fdd22bb7aa52 548 Ybminusd = ybIn - ydIn;
emilmont 1:fdd22bb7aa52 549
emilmont 1:fdd22bb7aa52 550 /* xc' = (xa-xb+xc-xd) */
emilmont 1:fdd22bb7aa52 551 ptr1[2] = (Xaplusc - Xbplusd);
emilmont 1:fdd22bb7aa52 552 /* yc' = (ya-yb+yc-yd) */
emilmont 1:fdd22bb7aa52 553 ptr1[3] = (Yaplusc - Ybplusd);
emilmont 1:fdd22bb7aa52 554 /* xb' = (xa+yb-xc-yd) */
emilmont 1:fdd22bb7aa52 555 ptr1[4] = (Xaminusc + Ybminusd);
emilmont 1:fdd22bb7aa52 556 /* yb' = (ya-xb-yc+xd) */
emilmont 1:fdd22bb7aa52 557 ptr1[5] = (Yaminusc - Xbminusd);
emilmont 1:fdd22bb7aa52 558 /* xd' = (xa-yb-xc+yd)) */
emilmont 1:fdd22bb7aa52 559 ptr1[6] = (Xaminusc - Ybminusd);
emilmont 1:fdd22bb7aa52 560 /* yd' = (ya+xb-yc-xd) */
emilmont 1:fdd22bb7aa52 561 ptr1[7] = (Xbminusd + Yaminusc);
emilmont 1:fdd22bb7aa52 562
emilmont 1:fdd22bb7aa52 563 /* increment pointer by 8 */
emilmont 1:fdd22bb7aa52 564 ptr1 = ptr1 + 8u;
emilmont 1:fdd22bb7aa52 565
emilmont 1:fdd22bb7aa52 566 } while(--j);
emilmont 1:fdd22bb7aa52 567
emilmont 1:fdd22bb7aa52 568 #else
emilmont 1:fdd22bb7aa52 569
emilmont 1:fdd22bb7aa52 570 float32_t t1, t2, r1, r2, s1, s2;
emilmont 1:fdd22bb7aa52 571
emilmont 1:fdd22bb7aa52 572 /* Run the below code for Cortex-M0 */
emilmont 1:fdd22bb7aa52 573
emilmont 1:fdd22bb7aa52 574 /* Initializations for the fft calculation */
emilmont 1:fdd22bb7aa52 575 n2 = fftLen;
emilmont 1:fdd22bb7aa52 576 n1 = n2;
emilmont 1:fdd22bb7aa52 577 for (k = fftLen; k > 1u; k >>= 2u)
emilmont 1:fdd22bb7aa52 578 {
emilmont 1:fdd22bb7aa52 579 /* Initializations for the fft calculation */
emilmont 1:fdd22bb7aa52 580 n1 = n2;
emilmont 1:fdd22bb7aa52 581 n2 >>= 2u;
emilmont 1:fdd22bb7aa52 582 ia1 = 0u;
emilmont 1:fdd22bb7aa52 583
emilmont 1:fdd22bb7aa52 584 /* FFT Calculation */
emilmont 1:fdd22bb7aa52 585 for (j = 0u; j <= (n2 - 1u); j++)
emilmont 1:fdd22bb7aa52 586 {
emilmont 1:fdd22bb7aa52 587 /* index calculation for the coefficients */
emilmont 1:fdd22bb7aa52 588 ia2 = ia1 + ia1;
emilmont 1:fdd22bb7aa52 589 ia3 = ia2 + ia1;
emilmont 1:fdd22bb7aa52 590 co1 = pCoef[ia1 * 2u];
emilmont 1:fdd22bb7aa52 591 si1 = pCoef[(ia1 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 592 co2 = pCoef[ia2 * 2u];
emilmont 1:fdd22bb7aa52 593 si2 = pCoef[(ia2 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 594 co3 = pCoef[ia3 * 2u];
emilmont 1:fdd22bb7aa52 595 si3 = pCoef[(ia3 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 596
emilmont 1:fdd22bb7aa52 597 /* Twiddle coefficients index modifier */
emilmont 1:fdd22bb7aa52 598 ia1 = ia1 + twidCoefModifier;
emilmont 1:fdd22bb7aa52 599
emilmont 1:fdd22bb7aa52 600 for (i0 = j; i0 < fftLen; i0 += n1)
emilmont 1:fdd22bb7aa52 601 {
emilmont 1:fdd22bb7aa52 602 /* index calculation for the input as, */
emilmont 1:fdd22bb7aa52 603 /* pSrc[i0 + 0], pSrc[i0 + fftLen/4], pSrc[i0 + fftLen/2], pSrc[i0 + 3fftLen/4] */
emilmont 1:fdd22bb7aa52 604 i1 = i0 + n2;
emilmont 1:fdd22bb7aa52 605 i2 = i1 + n2;
emilmont 1:fdd22bb7aa52 606 i3 = i2 + n2;
emilmont 1:fdd22bb7aa52 607
emilmont 1:fdd22bb7aa52 608 /* xa + xc */
emilmont 1:fdd22bb7aa52 609 r1 = pSrc[(2u * i0)] + pSrc[(2u * i2)];
emilmont 1:fdd22bb7aa52 610
emilmont 1:fdd22bb7aa52 611 /* xa - xc */
emilmont 1:fdd22bb7aa52 612 r2 = pSrc[(2u * i0)] - pSrc[(2u * i2)];
emilmont 1:fdd22bb7aa52 613
emilmont 1:fdd22bb7aa52 614 /* ya + yc */
emilmont 1:fdd22bb7aa52 615 s1 = pSrc[(2u * i0) + 1u] + pSrc[(2u * i2) + 1u];
emilmont 1:fdd22bb7aa52 616
emilmont 1:fdd22bb7aa52 617 /* ya - yc */
emilmont 1:fdd22bb7aa52 618 s2 = pSrc[(2u * i0) + 1u] - pSrc[(2u * i2) + 1u];
emilmont 1:fdd22bb7aa52 619
emilmont 1:fdd22bb7aa52 620 /* xb + xd */
emilmont 1:fdd22bb7aa52 621 t1 = pSrc[2u * i1] + pSrc[2u * i3];
emilmont 1:fdd22bb7aa52 622
emilmont 1:fdd22bb7aa52 623 /* xa' = xa + xb + xc + xd */
emilmont 1:fdd22bb7aa52 624 pSrc[2u * i0] = r1 + t1;
emilmont 1:fdd22bb7aa52 625
emilmont 1:fdd22bb7aa52 626 /* xa + xc -(xb + xd) */
emilmont 1:fdd22bb7aa52 627 r1 = r1 - t1;
emilmont 1:fdd22bb7aa52 628
emilmont 1:fdd22bb7aa52 629 /* yb + yd */
emilmont 1:fdd22bb7aa52 630 t2 = pSrc[(2u * i1) + 1u] + pSrc[(2u * i3) + 1u];
emilmont 1:fdd22bb7aa52 631
emilmont 1:fdd22bb7aa52 632 /* ya' = ya + yb + yc + yd */
emilmont 1:fdd22bb7aa52 633 pSrc[(2u * i0) + 1u] = s1 + t2;
emilmont 1:fdd22bb7aa52 634
emilmont 1:fdd22bb7aa52 635 /* (ya + yc) - (yb + yd) */
emilmont 1:fdd22bb7aa52 636 s1 = s1 - t2;
emilmont 1:fdd22bb7aa52 637
emilmont 1:fdd22bb7aa52 638 /* (yb - yd) */
emilmont 1:fdd22bb7aa52 639 t1 = pSrc[(2u * i1) + 1u] - pSrc[(2u * i3) + 1u];
emilmont 1:fdd22bb7aa52 640
emilmont 1:fdd22bb7aa52 641 /* (xb - xd) */
emilmont 1:fdd22bb7aa52 642 t2 = pSrc[2u * i1] - pSrc[2u * i3];
emilmont 1:fdd22bb7aa52 643
emilmont 1:fdd22bb7aa52 644 /* xc' = (xa-xb+xc-xd)co2 + (ya-yb+yc-yd)(si2) */
emilmont 1:fdd22bb7aa52 645 pSrc[2u * i1] = (r1 * co2) + (s1 * si2);
emilmont 1:fdd22bb7aa52 646
emilmont 1:fdd22bb7aa52 647 /* yc' = (ya-yb+yc-yd)co2 - (xa-xb+xc-xd)(si2) */
emilmont 1:fdd22bb7aa52 648 pSrc[(2u * i1) + 1u] = (s1 * co2) - (r1 * si2);
emilmont 1:fdd22bb7aa52 649
emilmont 1:fdd22bb7aa52 650 /* (xa - xc) + (yb - yd) */
emilmont 1:fdd22bb7aa52 651 r1 = r2 + t1;
emilmont 1:fdd22bb7aa52 652
emilmont 1:fdd22bb7aa52 653 /* (xa - xc) - (yb - yd) */
emilmont 1:fdd22bb7aa52 654 r2 = r2 - t1;
emilmont 1:fdd22bb7aa52 655
emilmont 1:fdd22bb7aa52 656 /* (ya - yc) - (xb - xd) */
emilmont 1:fdd22bb7aa52 657 s1 = s2 - t2;
emilmont 1:fdd22bb7aa52 658
emilmont 1:fdd22bb7aa52 659 /* (ya - yc) + (xb - xd) */
emilmont 1:fdd22bb7aa52 660 s2 = s2 + t2;
emilmont 1:fdd22bb7aa52 661
emilmont 1:fdd22bb7aa52 662 /* xb' = (xa+yb-xc-yd)co1 + (ya-xb-yc+xd)(si1) */
emilmont 1:fdd22bb7aa52 663 pSrc[2u * i2] = (r1 * co1) + (s1 * si1);
emilmont 1:fdd22bb7aa52 664
emilmont 1:fdd22bb7aa52 665 /* yb' = (ya-xb-yc+xd)co1 - (xa+yb-xc-yd)(si1) */
emilmont 1:fdd22bb7aa52 666 pSrc[(2u * i2) + 1u] = (s1 * co1) - (r1 * si1);
emilmont 1:fdd22bb7aa52 667
emilmont 1:fdd22bb7aa52 668 /* xd' = (xa-yb-xc+yd)co3 + (ya+xb-yc-xd)(si3) */
emilmont 1:fdd22bb7aa52 669 pSrc[2u * i3] = (r2 * co3) + (s2 * si3);
emilmont 1:fdd22bb7aa52 670
emilmont 1:fdd22bb7aa52 671 /* yd' = (ya+xb-yc-xd)co3 - (xa-yb-xc+yd)(si3) */
emilmont 1:fdd22bb7aa52 672 pSrc[(2u * i3) + 1u] = (s2 * co3) - (r2 * si3);
emilmont 1:fdd22bb7aa52 673 }
emilmont 1:fdd22bb7aa52 674 }
emilmont 1:fdd22bb7aa52 675 twidCoefModifier <<= 2u;
emilmont 1:fdd22bb7aa52 676 }
emilmont 1:fdd22bb7aa52 677
emilmont 1:fdd22bb7aa52 678 #endif /* #ifndef ARM_MATH_CM0 */
emilmont 1:fdd22bb7aa52 679
emilmont 1:fdd22bb7aa52 680 }
emilmont 1:fdd22bb7aa52 681
emilmont 1:fdd22bb7aa52 682 /*
emilmont 1:fdd22bb7aa52 683 * @brief Core function for the floating-point CIFFT butterfly process.
emilmont 1:fdd22bb7aa52 684 * @param[in, out] *pSrc points to the in-place buffer of floating-point data type.
emilmont 1:fdd22bb7aa52 685 * @param[in] fftLen length of the FFT.
emilmont 1:fdd22bb7aa52 686 * @param[in] *pCoef points to twiddle coefficient buffer.
emilmont 1:fdd22bb7aa52 687 * @param[in] twidCoefModifier twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table.
emilmont 1:fdd22bb7aa52 688 * @param[in] onebyfftLen value of 1/fftLen.
emilmont 1:fdd22bb7aa52 689 * @return none.
emilmont 1:fdd22bb7aa52 690 */
emilmont 1:fdd22bb7aa52 691
emilmont 1:fdd22bb7aa52 692 void arm_radix4_butterfly_inverse_f32(
emilmont 1:fdd22bb7aa52 693 float32_t * pSrc,
emilmont 1:fdd22bb7aa52 694 uint16_t fftLen,
emilmont 1:fdd22bb7aa52 695 float32_t * pCoef,
emilmont 1:fdd22bb7aa52 696 uint16_t twidCoefModifier,
emilmont 1:fdd22bb7aa52 697 float32_t onebyfftLen)
emilmont 1:fdd22bb7aa52 698 {
emilmont 1:fdd22bb7aa52 699 float32_t co1, co2, co3, si1, si2, si3;
emilmont 1:fdd22bb7aa52 700 uint32_t ia1, ia2, ia3;
emilmont 1:fdd22bb7aa52 701 uint32_t i0, i1, i2, i3;
emilmont 1:fdd22bb7aa52 702 uint32_t n1, n2, j, k;
emilmont 1:fdd22bb7aa52 703
emilmont 1:fdd22bb7aa52 704 #ifndef ARM_MATH_CM0
emilmont 1:fdd22bb7aa52 705
emilmont 1:fdd22bb7aa52 706 float32_t xaIn, yaIn, xbIn, ybIn, xcIn, ycIn, xdIn, ydIn;
emilmont 1:fdd22bb7aa52 707 float32_t Xaplusc, Xbplusd, Yaplusc, Ybplusd, Xaminusc, Xbminusd, Yaminusc,
emilmont 1:fdd22bb7aa52 708 Ybminusd;
emilmont 1:fdd22bb7aa52 709 float32_t Xb12C_out, Yb12C_out, Xc12C_out, Yc12C_out, Xd12C_out, Yd12C_out;
emilmont 1:fdd22bb7aa52 710 float32_t Xb12_out, Yb12_out, Xc12_out, Yc12_out, Xd12_out, Yd12_out;
emilmont 1:fdd22bb7aa52 711 float32_t *ptr1;
emilmont 1:fdd22bb7aa52 712
emilmont 1:fdd22bb7aa52 713
emilmont 1:fdd22bb7aa52 714 /* Initializations for the first stage */
emilmont 1:fdd22bb7aa52 715 n2 = fftLen;
emilmont 1:fdd22bb7aa52 716 n1 = n2;
emilmont 1:fdd22bb7aa52 717
emilmont 1:fdd22bb7aa52 718 /* n2 = fftLen/4 */
emilmont 1:fdd22bb7aa52 719 n2 >>= 2u;
emilmont 1:fdd22bb7aa52 720 i0 = 0u;
emilmont 1:fdd22bb7aa52 721 ia1 = 0u;
emilmont 1:fdd22bb7aa52 722
emilmont 1:fdd22bb7aa52 723 j = n2;
emilmont 1:fdd22bb7aa52 724
emilmont 1:fdd22bb7aa52 725 /* Calculation of first stage */
emilmont 1:fdd22bb7aa52 726 do
emilmont 1:fdd22bb7aa52 727 {
emilmont 1:fdd22bb7aa52 728 /* index calculation for the input as, */
emilmont 1:fdd22bb7aa52 729 /* pSrc[i0 + 0], pSrc[i0 + fftLen/4], pSrc[i0 + fftLen/2], pSrc[i0 + 3fftLen/4] */
emilmont 1:fdd22bb7aa52 730 i1 = i0 + n2;
emilmont 1:fdd22bb7aa52 731 i2 = i1 + n2;
emilmont 1:fdd22bb7aa52 732 i3 = i2 + n2;
emilmont 1:fdd22bb7aa52 733
emilmont 1:fdd22bb7aa52 734 /* Butterfly implementation */
emilmont 1:fdd22bb7aa52 735 xaIn = pSrc[(2u * i0)];
emilmont 1:fdd22bb7aa52 736 yaIn = pSrc[(2u * i0) + 1u];
emilmont 1:fdd22bb7aa52 737
emilmont 1:fdd22bb7aa52 738 xcIn = pSrc[(2u * i2)];
emilmont 1:fdd22bb7aa52 739 ycIn = pSrc[(2u * i2) + 1u];
emilmont 1:fdd22bb7aa52 740
emilmont 1:fdd22bb7aa52 741 xbIn = pSrc[(2u * i1)];
emilmont 1:fdd22bb7aa52 742 ybIn = pSrc[(2u * i1) + 1u];
emilmont 1:fdd22bb7aa52 743
emilmont 1:fdd22bb7aa52 744 xdIn = pSrc[(2u * i3)];
emilmont 1:fdd22bb7aa52 745 ydIn = pSrc[(2u * i3) + 1u];
emilmont 1:fdd22bb7aa52 746
emilmont 1:fdd22bb7aa52 747 /* xa + xc */
emilmont 1:fdd22bb7aa52 748 Xaplusc = xaIn + xcIn;
emilmont 1:fdd22bb7aa52 749 /* xb + xd */
emilmont 1:fdd22bb7aa52 750 Xbplusd = xbIn + xdIn;
emilmont 1:fdd22bb7aa52 751 /* ya + yc */
emilmont 1:fdd22bb7aa52 752 Yaplusc = yaIn + ycIn;
emilmont 1:fdd22bb7aa52 753 /* yb + yd */
emilmont 1:fdd22bb7aa52 754 Ybplusd = ybIn + ydIn;
emilmont 1:fdd22bb7aa52 755
emilmont 1:fdd22bb7aa52 756 /* index calculation for the coefficients */
emilmont 1:fdd22bb7aa52 757 ia2 = ia1 + ia1;
emilmont 1:fdd22bb7aa52 758 co2 = pCoef[ia2 * 2u];
emilmont 1:fdd22bb7aa52 759 si2 = pCoef[(ia2 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 760
emilmont 1:fdd22bb7aa52 761 /* xa - xc */
emilmont 1:fdd22bb7aa52 762 Xaminusc = xaIn - xcIn;
emilmont 1:fdd22bb7aa52 763 /* xb - xd */
emilmont 1:fdd22bb7aa52 764 Xbminusd = xbIn - xdIn;
emilmont 1:fdd22bb7aa52 765 /* ya - yc */
emilmont 1:fdd22bb7aa52 766 Yaminusc = yaIn - ycIn;
emilmont 1:fdd22bb7aa52 767 /* yb - yd */
emilmont 1:fdd22bb7aa52 768 Ybminusd = ybIn - ydIn;
emilmont 1:fdd22bb7aa52 769
emilmont 1:fdd22bb7aa52 770 /* xa' = xa + xb + xc + xd */
emilmont 1:fdd22bb7aa52 771 pSrc[(2u * i0)] = Xaplusc + Xbplusd;
emilmont 1:fdd22bb7aa52 772
emilmont 1:fdd22bb7aa52 773 /* ya' = ya + yb + yc + yd */
emilmont 1:fdd22bb7aa52 774 pSrc[(2u * i0) + 1u] = Yaplusc + Ybplusd;
emilmont 1:fdd22bb7aa52 775
emilmont 1:fdd22bb7aa52 776 /* (xa - xc) - (yb - yd) */
emilmont 1:fdd22bb7aa52 777 Xb12C_out = (Xaminusc - Ybminusd);
emilmont 1:fdd22bb7aa52 778 /* (ya - yc) + (xb - xd) */
emilmont 1:fdd22bb7aa52 779 Yb12C_out = (Yaminusc + Xbminusd);
emilmont 1:fdd22bb7aa52 780 /* (xa + xc) - (xb + xd) */
emilmont 1:fdd22bb7aa52 781 Xc12C_out = (Xaplusc - Xbplusd);
emilmont 1:fdd22bb7aa52 782 /* (ya + yc) - (yb + yd) */
emilmont 1:fdd22bb7aa52 783 Yc12C_out = (Yaplusc - Ybplusd);
emilmont 1:fdd22bb7aa52 784 /* (xa - xc) + (yb - yd) */
emilmont 1:fdd22bb7aa52 785 Xd12C_out = (Xaminusc + Ybminusd);
emilmont 1:fdd22bb7aa52 786 /* (ya - yc) - (xb - xd) */
emilmont 1:fdd22bb7aa52 787 Yd12C_out = (Yaminusc - Xbminusd);
emilmont 1:fdd22bb7aa52 788
emilmont 1:fdd22bb7aa52 789 co1 = pCoef[ia1 * 2u];
emilmont 1:fdd22bb7aa52 790 si1 = pCoef[(ia1 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 791
emilmont 1:fdd22bb7aa52 792 /* index calculation for the coefficients */
emilmont 1:fdd22bb7aa52 793 ia3 = ia2 + ia1;
emilmont 1:fdd22bb7aa52 794 co3 = pCoef[ia3 * 2u];
emilmont 1:fdd22bb7aa52 795 si3 = pCoef[(ia3 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 796
emilmont 1:fdd22bb7aa52 797 Xb12_out = Xb12C_out * co1;
emilmont 1:fdd22bb7aa52 798 Yb12_out = Yb12C_out * co1;
emilmont 1:fdd22bb7aa52 799 Xc12_out = Xc12C_out * co2;
emilmont 1:fdd22bb7aa52 800 Yc12_out = Yc12C_out * co2;
emilmont 1:fdd22bb7aa52 801 Xd12_out = Xd12C_out * co3;
emilmont 1:fdd22bb7aa52 802 Yd12_out = Yd12C_out * co3;
emilmont 1:fdd22bb7aa52 803
emilmont 1:fdd22bb7aa52 804 /* xb' = (xa+yb-xc-yd)co1 - (ya-xb-yc+xd)(si1) */
emilmont 1:fdd22bb7aa52 805 Xb12_out -= Yb12C_out * si1;
emilmont 1:fdd22bb7aa52 806 /* yb' = (ya-xb-yc+xd)co1 + (xa+yb-xc-yd)(si1) */
emilmont 1:fdd22bb7aa52 807 Yb12_out += Xb12C_out * si1;
emilmont 1:fdd22bb7aa52 808 /* xc' = (xa-xb+xc-xd)co2 - (ya-yb+yc-yd)(si2) */
emilmont 1:fdd22bb7aa52 809 Xc12_out -= Yc12C_out * si2;
emilmont 1:fdd22bb7aa52 810 /* yc' = (ya-yb+yc-yd)co2 + (xa-xb+xc-xd)(si2) */
emilmont 1:fdd22bb7aa52 811 Yc12_out += Xc12C_out * si2;
emilmont 1:fdd22bb7aa52 812 /* xd' = (xa-yb-xc+yd)co3 - (ya+xb-yc-xd)(si3) */
emilmont 1:fdd22bb7aa52 813 Xd12_out -= Yd12C_out * si3;
emilmont 1:fdd22bb7aa52 814 /* yd' = (ya+xb-yc-xd)co3 + (xa-yb-xc+yd)(si3) */
emilmont 1:fdd22bb7aa52 815 Yd12_out += Xd12C_out * si3;
emilmont 1:fdd22bb7aa52 816
emilmont 1:fdd22bb7aa52 817 /* xc' = (xa-xb+xc-xd)co2 - (ya-yb+yc-yd)(si2) */
emilmont 1:fdd22bb7aa52 818 pSrc[2u * i1] = Xc12_out;
emilmont 1:fdd22bb7aa52 819
emilmont 1:fdd22bb7aa52 820 /* yc' = (ya-yb+yc-yd)co2 + (xa-xb+xc-xd)(si2) */
emilmont 1:fdd22bb7aa52 821 pSrc[(2u * i1) + 1u] = Yc12_out;
emilmont 1:fdd22bb7aa52 822
emilmont 1:fdd22bb7aa52 823 /* xb' = (xa+yb-xc-yd)co1 - (ya-xb-yc+xd)(si1) */
emilmont 1:fdd22bb7aa52 824 pSrc[2u * i2] = Xb12_out;
emilmont 1:fdd22bb7aa52 825
emilmont 1:fdd22bb7aa52 826 /* yb' = (ya-xb-yc+xd)co1 + (xa+yb-xc-yd)(si1) */
emilmont 1:fdd22bb7aa52 827 pSrc[(2u * i2) + 1u] = Yb12_out;
emilmont 1:fdd22bb7aa52 828
emilmont 1:fdd22bb7aa52 829 /* xd' = (xa-yb-xc+yd)co3 - (ya+xb-yc-xd)(si3) */
emilmont 1:fdd22bb7aa52 830 pSrc[2u * i3] = Xd12_out;
emilmont 1:fdd22bb7aa52 831
emilmont 1:fdd22bb7aa52 832 /* yd' = (ya+xb-yc-xd)co3 + (xa-yb-xc+yd)(si3) */
emilmont 1:fdd22bb7aa52 833 pSrc[(2u * i3) + 1u] = Yd12_out;
emilmont 1:fdd22bb7aa52 834
emilmont 1:fdd22bb7aa52 835 /* Twiddle coefficients index modifier */
emilmont 1:fdd22bb7aa52 836 ia1 = ia1 + twidCoefModifier;
emilmont 1:fdd22bb7aa52 837
emilmont 1:fdd22bb7aa52 838 /* Updating input index */
emilmont 1:fdd22bb7aa52 839 i0 = i0 + 1u;
emilmont 1:fdd22bb7aa52 840
emilmont 1:fdd22bb7aa52 841 } while(--j);
emilmont 1:fdd22bb7aa52 842
emilmont 1:fdd22bb7aa52 843 twidCoefModifier <<= 2u;
emilmont 1:fdd22bb7aa52 844
emilmont 1:fdd22bb7aa52 845 /* Calculation of second stage to excluding last stage */
emilmont 1:fdd22bb7aa52 846 for (k = fftLen / 4; k > 4u; k >>= 2u)
emilmont 1:fdd22bb7aa52 847 {
emilmont 1:fdd22bb7aa52 848 /* Initializations for the first stage */
emilmont 1:fdd22bb7aa52 849 n1 = n2;
emilmont 1:fdd22bb7aa52 850 n2 >>= 2u;
emilmont 1:fdd22bb7aa52 851 ia1 = 0u;
emilmont 1:fdd22bb7aa52 852
emilmont 1:fdd22bb7aa52 853 /* Calculation of first stage */
emilmont 1:fdd22bb7aa52 854 for (j = 0u; j <= (n2 - 1u); j++)
emilmont 1:fdd22bb7aa52 855 {
emilmont 1:fdd22bb7aa52 856 /* index calculation for the coefficients */
emilmont 1:fdd22bb7aa52 857 ia2 = ia1 + ia1;
emilmont 1:fdd22bb7aa52 858 ia3 = ia2 + ia1;
emilmont 1:fdd22bb7aa52 859 co1 = pCoef[ia1 * 2u];
emilmont 1:fdd22bb7aa52 860 si1 = pCoef[(ia1 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 861 co2 = pCoef[ia2 * 2u];
emilmont 1:fdd22bb7aa52 862 si2 = pCoef[(ia2 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 863 co3 = pCoef[ia3 * 2u];
emilmont 1:fdd22bb7aa52 864 si3 = pCoef[(ia3 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 865
emilmont 1:fdd22bb7aa52 866 /* Twiddle coefficients index modifier */
emilmont 1:fdd22bb7aa52 867 ia1 = ia1 + twidCoefModifier;
emilmont 1:fdd22bb7aa52 868
emilmont 1:fdd22bb7aa52 869 for (i0 = j; i0 < fftLen; i0 += n1)
emilmont 1:fdd22bb7aa52 870 {
emilmont 1:fdd22bb7aa52 871 /* index calculation for the input as, */
emilmont 1:fdd22bb7aa52 872 /* pSrc[i0 + 0], pSrc[i0 + fftLen/4], pSrc[i0 + fftLen/2], pSrc[i0 + 3fftLen/4] */
emilmont 1:fdd22bb7aa52 873 i1 = i0 + n2;
emilmont 1:fdd22bb7aa52 874 i2 = i1 + n2;
emilmont 1:fdd22bb7aa52 875 i3 = i2 + n2;
emilmont 1:fdd22bb7aa52 876
emilmont 1:fdd22bb7aa52 877 xaIn = pSrc[(2u * i0)];
emilmont 1:fdd22bb7aa52 878 yaIn = pSrc[(2u * i0) + 1u];
emilmont 1:fdd22bb7aa52 879
emilmont 1:fdd22bb7aa52 880 xbIn = pSrc[(2u * i1)];
emilmont 1:fdd22bb7aa52 881 ybIn = pSrc[(2u * i1) + 1u];
emilmont 1:fdd22bb7aa52 882
emilmont 1:fdd22bb7aa52 883 xcIn = pSrc[(2u * i2)];
emilmont 1:fdd22bb7aa52 884 ycIn = pSrc[(2u * i2) + 1u];
emilmont 1:fdd22bb7aa52 885
emilmont 1:fdd22bb7aa52 886 xdIn = pSrc[(2u * i3)];
emilmont 1:fdd22bb7aa52 887 ydIn = pSrc[(2u * i3) + 1u];
emilmont 1:fdd22bb7aa52 888
emilmont 1:fdd22bb7aa52 889 /* xa - xc */
emilmont 1:fdd22bb7aa52 890 Xaminusc = xaIn - xcIn;
emilmont 1:fdd22bb7aa52 891 /* (xb - xd) */
emilmont 1:fdd22bb7aa52 892 Xbminusd = xbIn - xdIn;
emilmont 1:fdd22bb7aa52 893 /* ya - yc */
emilmont 1:fdd22bb7aa52 894 Yaminusc = yaIn - ycIn;
emilmont 1:fdd22bb7aa52 895 /* (yb - yd) */
emilmont 1:fdd22bb7aa52 896 Ybminusd = ybIn - ydIn;
emilmont 1:fdd22bb7aa52 897
emilmont 1:fdd22bb7aa52 898 /* xa + xc */
emilmont 1:fdd22bb7aa52 899 Xaplusc = xaIn + xcIn;
emilmont 1:fdd22bb7aa52 900 /* xb + xd */
emilmont 1:fdd22bb7aa52 901 Xbplusd = xbIn + xdIn;
emilmont 1:fdd22bb7aa52 902 /* ya + yc */
emilmont 1:fdd22bb7aa52 903 Yaplusc = yaIn + ycIn;
emilmont 1:fdd22bb7aa52 904 /* yb + yd */
emilmont 1:fdd22bb7aa52 905 Ybplusd = ybIn + ydIn;
emilmont 1:fdd22bb7aa52 906
emilmont 1:fdd22bb7aa52 907 /* (xa - xc) - (yb - yd) */
emilmont 1:fdd22bb7aa52 908 Xb12C_out = (Xaminusc - Ybminusd);
emilmont 1:fdd22bb7aa52 909 /* (ya - yc) + (xb - xd) */
emilmont 1:fdd22bb7aa52 910 Yb12C_out = (Yaminusc + Xbminusd);
emilmont 1:fdd22bb7aa52 911 /* xa + xc -(xb + xd) */
emilmont 1:fdd22bb7aa52 912 Xc12C_out = (Xaplusc - Xbplusd);
emilmont 1:fdd22bb7aa52 913 /* (ya + yc) - (yb + yd) */
emilmont 1:fdd22bb7aa52 914 Yc12C_out = (Yaplusc - Ybplusd);
emilmont 1:fdd22bb7aa52 915 /* (xa - xc) + (yb - yd) */
emilmont 1:fdd22bb7aa52 916 Xd12C_out = (Xaminusc + Ybminusd);
emilmont 1:fdd22bb7aa52 917 /* (ya - yc) - (xb - xd) */
emilmont 1:fdd22bb7aa52 918 Yd12C_out = (Yaminusc - Xbminusd);
emilmont 1:fdd22bb7aa52 919
emilmont 1:fdd22bb7aa52 920 pSrc[(2u * i0)] = Xaplusc + Xbplusd;
emilmont 1:fdd22bb7aa52 921 pSrc[(2u * i0) + 1u] = Yaplusc + Ybplusd;
emilmont 1:fdd22bb7aa52 922
emilmont 1:fdd22bb7aa52 923 Xb12_out = Xb12C_out * co1;
emilmont 1:fdd22bb7aa52 924 Yb12_out = Yb12C_out * co1;
emilmont 1:fdd22bb7aa52 925 Xc12_out = Xc12C_out * co2;
emilmont 1:fdd22bb7aa52 926 Yc12_out = Yc12C_out * co2;
emilmont 1:fdd22bb7aa52 927 Xd12_out = Xd12C_out * co3;
emilmont 1:fdd22bb7aa52 928 Yd12_out = Yd12C_out * co3;
emilmont 1:fdd22bb7aa52 929
emilmont 1:fdd22bb7aa52 930 /* xb' = (xa+yb-xc-yd)co1 - (ya-xb-yc+xd)(si1) */
emilmont 1:fdd22bb7aa52 931 Xb12_out -= Yb12C_out * si1;
emilmont 1:fdd22bb7aa52 932 /* yb' = (ya-xb-yc+xd)co1 + (xa+yb-xc-yd)(si1) */
emilmont 1:fdd22bb7aa52 933 Yb12_out += Xb12C_out * si1;
emilmont 1:fdd22bb7aa52 934 /* xc' = (xa-xb+xc-xd)co2 - (ya-yb+yc-yd)(si2) */
emilmont 1:fdd22bb7aa52 935 Xc12_out -= Yc12C_out * si2;
emilmont 1:fdd22bb7aa52 936 /* yc' = (ya-yb+yc-yd)co2 + (xa-xb+xc-xd)(si2) */
emilmont 1:fdd22bb7aa52 937 Yc12_out += Xc12C_out * si2;
emilmont 1:fdd22bb7aa52 938 /* xd' = (xa-yb-xc+yd)co3 - (ya+xb-yc-xd)(si3) */
emilmont 1:fdd22bb7aa52 939 Xd12_out -= Yd12C_out * si3;
emilmont 1:fdd22bb7aa52 940 /* yd' = (ya+xb-yc-xd)co3 + (xa-yb-xc+yd)(si3) */
emilmont 1:fdd22bb7aa52 941 Yd12_out += Xd12C_out * si3;
emilmont 1:fdd22bb7aa52 942
emilmont 1:fdd22bb7aa52 943 /* xc' = (xa-xb+xc-xd)co2 - (ya-yb+yc-yd)(si2) */
emilmont 1:fdd22bb7aa52 944 pSrc[2u * i1] = Xc12_out;
emilmont 1:fdd22bb7aa52 945
emilmont 1:fdd22bb7aa52 946 /* yc' = (ya-yb+yc-yd)co2 + (xa-xb+xc-xd)(si2) */
emilmont 1:fdd22bb7aa52 947 pSrc[(2u * i1) + 1u] = Yc12_out;
emilmont 1:fdd22bb7aa52 948
emilmont 1:fdd22bb7aa52 949 /* xb' = (xa+yb-xc-yd)co1 - (ya-xb-yc+xd)(si1) */
emilmont 1:fdd22bb7aa52 950 pSrc[2u * i2] = Xb12_out;
emilmont 1:fdd22bb7aa52 951
emilmont 1:fdd22bb7aa52 952 /* yb' = (ya-xb-yc+xd)co1 + (xa+yb-xc-yd)(si1) */
emilmont 1:fdd22bb7aa52 953 pSrc[(2u * i2) + 1u] = Yb12_out;
emilmont 1:fdd22bb7aa52 954
emilmont 1:fdd22bb7aa52 955 /* xd' = (xa-yb-xc+yd)co3 - (ya+xb-yc-xd)(si3) */
emilmont 1:fdd22bb7aa52 956 pSrc[2u * i3] = Xd12_out;
emilmont 1:fdd22bb7aa52 957
emilmont 1:fdd22bb7aa52 958 /* yd' = (ya+xb-yc-xd)co3 + (xa-yb-xc+yd)(si3) */
emilmont 1:fdd22bb7aa52 959 pSrc[(2u * i3) + 1u] = Yd12_out;
emilmont 1:fdd22bb7aa52 960
emilmont 1:fdd22bb7aa52 961 }
emilmont 1:fdd22bb7aa52 962 }
emilmont 1:fdd22bb7aa52 963 twidCoefModifier <<= 2u;
emilmont 1:fdd22bb7aa52 964 }
emilmont 1:fdd22bb7aa52 965 /* Initializations of last stage */
emilmont 1:fdd22bb7aa52 966
emilmont 1:fdd22bb7aa52 967 j = fftLen >> 2;
emilmont 1:fdd22bb7aa52 968 ptr1 = &pSrc[0];
emilmont 1:fdd22bb7aa52 969
emilmont 1:fdd22bb7aa52 970 /* Calculations of last stage */
emilmont 1:fdd22bb7aa52 971 do
emilmont 1:fdd22bb7aa52 972 {
emilmont 1:fdd22bb7aa52 973
emilmont 1:fdd22bb7aa52 974 xaIn = ptr1[0];
emilmont 1:fdd22bb7aa52 975 xcIn = ptr1[4];
emilmont 1:fdd22bb7aa52 976 yaIn = ptr1[1];
emilmont 1:fdd22bb7aa52 977 ycIn = ptr1[5];
emilmont 1:fdd22bb7aa52 978
emilmont 1:fdd22bb7aa52 979 /* Butterfly implementation */
emilmont 1:fdd22bb7aa52 980 /* xa + xc */
emilmont 1:fdd22bb7aa52 981 Xaplusc = xaIn + xcIn;
emilmont 1:fdd22bb7aa52 982
emilmont 1:fdd22bb7aa52 983 xbIn = ptr1[2];
emilmont 1:fdd22bb7aa52 984
emilmont 1:fdd22bb7aa52 985 /* xa - xc */
emilmont 1:fdd22bb7aa52 986 Xaminusc = xaIn - xcIn;
emilmont 1:fdd22bb7aa52 987
emilmont 1:fdd22bb7aa52 988 xdIn = ptr1[6];
emilmont 1:fdd22bb7aa52 989
emilmont 1:fdd22bb7aa52 990 /* ya + yc */
emilmont 1:fdd22bb7aa52 991 Yaplusc = yaIn + ycIn;
emilmont 1:fdd22bb7aa52 992
emilmont 1:fdd22bb7aa52 993 ybIn = ptr1[3];
emilmont 1:fdd22bb7aa52 994
emilmont 1:fdd22bb7aa52 995 /* ya - yc */
emilmont 1:fdd22bb7aa52 996 Yaminusc = yaIn - ycIn;
emilmont 1:fdd22bb7aa52 997
emilmont 1:fdd22bb7aa52 998 ydIn = ptr1[7];
emilmont 1:fdd22bb7aa52 999
emilmont 1:fdd22bb7aa52 1000 /* xc + xd */
emilmont 1:fdd22bb7aa52 1001 Xbplusd = xbIn + xdIn;
emilmont 1:fdd22bb7aa52 1002
emilmont 1:fdd22bb7aa52 1003 /* yb + yd */
emilmont 1:fdd22bb7aa52 1004 Ybplusd = ybIn + ydIn;
emilmont 1:fdd22bb7aa52 1005
emilmont 1:fdd22bb7aa52 1006 /* xa' = xa + xb + xc + xd */
emilmont 1:fdd22bb7aa52 1007 ptr1[0] = (Xaplusc + Xbplusd) * onebyfftLen;
emilmont 1:fdd22bb7aa52 1008
emilmont 1:fdd22bb7aa52 1009 /* (xb-xd) */
emilmont 1:fdd22bb7aa52 1010 Xbminusd = xbIn - xdIn;
emilmont 1:fdd22bb7aa52 1011
emilmont 1:fdd22bb7aa52 1012 /* ya' = ya + yb + yc + yd */
emilmont 1:fdd22bb7aa52 1013 ptr1[1] = (Yaplusc + Ybplusd) * onebyfftLen;
emilmont 1:fdd22bb7aa52 1014
emilmont 1:fdd22bb7aa52 1015 /* (yb-yd) */
emilmont 1:fdd22bb7aa52 1016 Ybminusd = ybIn - ydIn;
emilmont 1:fdd22bb7aa52 1017
emilmont 1:fdd22bb7aa52 1018 /* xc' = (xa-xb+xc-xd) * onebyfftLen */
emilmont 1:fdd22bb7aa52 1019 ptr1[2] = (Xaplusc - Xbplusd) * onebyfftLen;
emilmont 1:fdd22bb7aa52 1020
emilmont 1:fdd22bb7aa52 1021 /* yc' = (ya-yb+yc-yd) * onebyfftLen */
emilmont 1:fdd22bb7aa52 1022 ptr1[3] = (Yaplusc - Ybplusd) * onebyfftLen;
emilmont 1:fdd22bb7aa52 1023
emilmont 1:fdd22bb7aa52 1024 /* xb' = (xa-yb-xc+yd) * onebyfftLen */
emilmont 1:fdd22bb7aa52 1025 ptr1[4] = (Xaminusc - Ybminusd) * onebyfftLen;
emilmont 1:fdd22bb7aa52 1026
emilmont 1:fdd22bb7aa52 1027 /* yb' = (ya+xb-yc-xd) * onebyfftLen */
emilmont 1:fdd22bb7aa52 1028 ptr1[5] = (Yaminusc + Xbminusd) * onebyfftLen;
emilmont 1:fdd22bb7aa52 1029
emilmont 1:fdd22bb7aa52 1030 /* xd' = (xa-yb-xc+yd) * onebyfftLen */
emilmont 1:fdd22bb7aa52 1031 ptr1[6] = (Xaminusc + Ybminusd) * onebyfftLen;
emilmont 1:fdd22bb7aa52 1032
emilmont 1:fdd22bb7aa52 1033 /* yd' = (ya-xb-yc+xd) * onebyfftLen */
emilmont 1:fdd22bb7aa52 1034 ptr1[7] = (Yaminusc - Xbminusd) * onebyfftLen;
emilmont 1:fdd22bb7aa52 1035
emilmont 1:fdd22bb7aa52 1036 /* increment source pointer by 8 for next calculations */
emilmont 1:fdd22bb7aa52 1037 ptr1 = ptr1 + 8u;
emilmont 1:fdd22bb7aa52 1038
emilmont 1:fdd22bb7aa52 1039 } while(--j);
emilmont 1:fdd22bb7aa52 1040
emilmont 1:fdd22bb7aa52 1041 #else
emilmont 1:fdd22bb7aa52 1042
emilmont 1:fdd22bb7aa52 1043 float32_t t1, t2, r1, r2, s1, s2;
emilmont 1:fdd22bb7aa52 1044
emilmont 1:fdd22bb7aa52 1045 /* Run the below code for Cortex-M0 */
emilmont 1:fdd22bb7aa52 1046
emilmont 1:fdd22bb7aa52 1047 /* Initializations for the first stage */
emilmont 1:fdd22bb7aa52 1048 n2 = fftLen;
emilmont 1:fdd22bb7aa52 1049 n1 = n2;
emilmont 1:fdd22bb7aa52 1050
emilmont 1:fdd22bb7aa52 1051 /* Calculation of first stage */
emilmont 1:fdd22bb7aa52 1052 for (k = fftLen; k > 4u; k >>= 2u)
emilmont 1:fdd22bb7aa52 1053 {
emilmont 1:fdd22bb7aa52 1054 /* Initializations for the first stage */
emilmont 1:fdd22bb7aa52 1055 n1 = n2;
emilmont 1:fdd22bb7aa52 1056 n2 >>= 2u;
emilmont 1:fdd22bb7aa52 1057 ia1 = 0u;
emilmont 1:fdd22bb7aa52 1058
emilmont 1:fdd22bb7aa52 1059 /* Calculation of first stage */
emilmont 1:fdd22bb7aa52 1060 for (j = 0u; j <= (n2 - 1u); j++)
emilmont 1:fdd22bb7aa52 1061 {
emilmont 1:fdd22bb7aa52 1062 /* index calculation for the coefficients */
emilmont 1:fdd22bb7aa52 1063 ia2 = ia1 + ia1;
emilmont 1:fdd22bb7aa52 1064 ia3 = ia2 + ia1;
emilmont 1:fdd22bb7aa52 1065 co1 = pCoef[ia1 * 2u];
emilmont 1:fdd22bb7aa52 1066 si1 = pCoef[(ia1 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 1067 co2 = pCoef[ia2 * 2u];
emilmont 1:fdd22bb7aa52 1068 si2 = pCoef[(ia2 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 1069 co3 = pCoef[ia3 * 2u];
emilmont 1:fdd22bb7aa52 1070 si3 = pCoef[(ia3 * 2u) + 1u];
emilmont 1:fdd22bb7aa52 1071
emilmont 1:fdd22bb7aa52 1072 /* Twiddle coefficients index modifier */
emilmont 1:fdd22bb7aa52 1073 ia1 = ia1 + twidCoefModifier;
emilmont 1:fdd22bb7aa52 1074
emilmont 1:fdd22bb7aa52 1075 for (i0 = j; i0 < fftLen; i0 += n1)
emilmont 1:fdd22bb7aa52 1076 {
emilmont 1:fdd22bb7aa52 1077 /* index calculation for the input as, */
emilmont 1:fdd22bb7aa52 1078 /* pSrc[i0 + 0], pSrc[i0 + fftLen/4], pSrc[i0 + fftLen/2], pSrc[i0 + 3fftLen/4] */
emilmont 1:fdd22bb7aa52 1079 i1 = i0 + n2;
emilmont 1:fdd22bb7aa52 1080 i2 = i1 + n2;
emilmont 1:fdd22bb7aa52 1081 i3 = i2 + n2;
emilmont 1:fdd22bb7aa52 1082
emilmont 1:fdd22bb7aa52 1083 /* xa + xc */
emilmont 1:fdd22bb7aa52 1084 r1 = pSrc[(2u * i0)] + pSrc[(2u * i2)];
emilmont 1:fdd22bb7aa52 1085
emilmont 1:fdd22bb7aa52 1086 /* xa - xc */
emilmont 1:fdd22bb7aa52 1087 r2 = pSrc[(2u * i0)] - pSrc[(2u * i2)];
emilmont 1:fdd22bb7aa52 1088
emilmont 1:fdd22bb7aa52 1089 /* ya + yc */
emilmont 1:fdd22bb7aa52 1090 s1 = pSrc[(2u * i0) + 1u] + pSrc[(2u * i2) + 1u];
emilmont 1:fdd22bb7aa52 1091
emilmont 1:fdd22bb7aa52 1092 /* ya - yc */
emilmont 1:fdd22bb7aa52 1093 s2 = pSrc[(2u * i0) + 1u] - pSrc[(2u * i2) + 1u];
emilmont 1:fdd22bb7aa52 1094
emilmont 1:fdd22bb7aa52 1095 /* xb + xd */
emilmont 1:fdd22bb7aa52 1096 t1 = pSrc[2u * i1] + pSrc[2u * i3];
emilmont 1:fdd22bb7aa52 1097
emilmont 1:fdd22bb7aa52 1098 /* xa' = xa + xb + xc + xd */
emilmont 1:fdd22bb7aa52 1099 pSrc[2u * i0] = r1 + t1;
emilmont 1:fdd22bb7aa52 1100
emilmont 1:fdd22bb7aa52 1101 /* xa + xc -(xb + xd) */
emilmont 1:fdd22bb7aa52 1102 r1 = r1 - t1;
emilmont 1:fdd22bb7aa52 1103
emilmont 1:fdd22bb7aa52 1104 /* yb + yd */
emilmont 1:fdd22bb7aa52 1105 t2 = pSrc[(2u * i1) + 1u] + pSrc[(2u * i3) + 1u];
emilmont 1:fdd22bb7aa52 1106
emilmont 1:fdd22bb7aa52 1107 /* ya' = ya + yb + yc + yd */
emilmont 1:fdd22bb7aa52 1108 pSrc[(2u * i0) + 1u] = s1 + t2;
emilmont 1:fdd22bb7aa52 1109
emilmont 1:fdd22bb7aa52 1110 /* (ya + yc) - (yb + yd) */
emilmont 1:fdd22bb7aa52 1111 s1 = s1 - t2;
emilmont 1:fdd22bb7aa52 1112
emilmont 1:fdd22bb7aa52 1113 /* (yb - yd) */
emilmont 1:fdd22bb7aa52 1114 t1 = pSrc[(2u * i1) + 1u] - pSrc[(2u * i3) + 1u];
emilmont 1:fdd22bb7aa52 1115
emilmont 1:fdd22bb7aa52 1116 /* (xb - xd) */
emilmont 1:fdd22bb7aa52 1117 t2 = pSrc[2u * i1] - pSrc[2u * i3];
emilmont 1:fdd22bb7aa52 1118
emilmont 1:fdd22bb7aa52 1119 /* xc' = (xa-xb+xc-xd)co2 - (ya-yb+yc-yd)(si2) */
emilmont 1:fdd22bb7aa52 1120 pSrc[2u * i1] = (r1 * co2) - (s1 * si2);
emilmont 1:fdd22bb7aa52 1121
emilmont 1:fdd22bb7aa52 1122 /* yc' = (ya-yb+yc-yd)co2 + (xa-xb+xc-xd)(si2) */
emilmont 1:fdd22bb7aa52 1123 pSrc[(2u * i1) + 1u] = (s1 * co2) + (r1 * si2);
emilmont 1:fdd22bb7aa52 1124
emilmont 1:fdd22bb7aa52 1125 /* (xa - xc) - (yb - yd) */
emilmont 1:fdd22bb7aa52 1126 r1 = r2 - t1;
emilmont 1:fdd22bb7aa52 1127
emilmont 1:fdd22bb7aa52 1128 /* (xa - xc) + (yb - yd) */
emilmont 1:fdd22bb7aa52 1129 r2 = r2 + t1;
emilmont 1:fdd22bb7aa52 1130
emilmont 1:fdd22bb7aa52 1131 /* (ya - yc) + (xb - xd) */
emilmont 1:fdd22bb7aa52 1132 s1 = s2 + t2;
emilmont 1:fdd22bb7aa52 1133
emilmont 1:fdd22bb7aa52 1134 /* (ya - yc) - (xb - xd) */
emilmont 1:fdd22bb7aa52 1135 s2 = s2 - t2;
emilmont 1:fdd22bb7aa52 1136
emilmont 1:fdd22bb7aa52 1137 /* xb' = (xa+yb-xc-yd)co1 - (ya-xb-yc+xd)(si1) */
emilmont 1:fdd22bb7aa52 1138 pSrc[2u * i2] = (r1 * co1) - (s1 * si1);
emilmont 1:fdd22bb7aa52 1139
emilmont 1:fdd22bb7aa52 1140 /* yb' = (ya-xb-yc+xd)co1 + (xa+yb-xc-yd)(si1) */
emilmont 1:fdd22bb7aa52 1141 pSrc[(2u * i2) + 1u] = (s1 * co1) + (r1 * si1);
emilmont 1:fdd22bb7aa52 1142
emilmont 1:fdd22bb7aa52 1143 /* xd' = (xa-yb-xc+yd)co3 - (ya+xb-yc-xd)(si3) */
emilmont 1:fdd22bb7aa52 1144 pSrc[2u * i3] = (r2 * co3) - (s2 * si3);
emilmont 1:fdd22bb7aa52 1145
emilmont 1:fdd22bb7aa52 1146 /* yd' = (ya+xb-yc-xd)co3 + (xa-yb-xc+yd)(si3) */
emilmont 1:fdd22bb7aa52 1147 pSrc[(2u * i3) + 1u] = (s2 * co3) + (r2 * si3);
emilmont 1:fdd22bb7aa52 1148 }
emilmont 1:fdd22bb7aa52 1149 }
emilmont 1:fdd22bb7aa52 1150 twidCoefModifier <<= 2u;
emilmont 1:fdd22bb7aa52 1151 }
emilmont 1:fdd22bb7aa52 1152 /* Initializations of last stage */
emilmont 1:fdd22bb7aa52 1153 n1 = n2;
emilmont 1:fdd22bb7aa52 1154 n2 >>= 2u;
emilmont 1:fdd22bb7aa52 1155
emilmont 1:fdd22bb7aa52 1156 /* Calculations of last stage */
emilmont 1:fdd22bb7aa52 1157 for (i0 = 0u; i0 <= (fftLen - n1); i0 += n1)
emilmont 1:fdd22bb7aa52 1158 {
emilmont 1:fdd22bb7aa52 1159 /* index calculation for the input as, */
emilmont 1:fdd22bb7aa52 1160 /* pSrc[i0 + 0], pSrc[i0 + fftLen/4], pSrc[i0 + fftLen/2], pSrc[i0 + 3fftLen/4] */
emilmont 1:fdd22bb7aa52 1161 i1 = i0 + n2;
emilmont 1:fdd22bb7aa52 1162 i2 = i1 + n2;
emilmont 1:fdd22bb7aa52 1163 i3 = i2 + n2;
emilmont 1:fdd22bb7aa52 1164
emilmont 1:fdd22bb7aa52 1165 /* Butterfly implementation */
emilmont 1:fdd22bb7aa52 1166 /* xa + xc */
emilmont 1:fdd22bb7aa52 1167 r1 = pSrc[2u * i0] + pSrc[2u * i2];
emilmont 1:fdd22bb7aa52 1168
emilmont 1:fdd22bb7aa52 1169 /* xa - xc */
emilmont 1:fdd22bb7aa52 1170 r2 = pSrc[2u * i0] - pSrc[2u * i2];
emilmont 1:fdd22bb7aa52 1171
emilmont 1:fdd22bb7aa52 1172 /* ya + yc */
emilmont 1:fdd22bb7aa52 1173 s1 = pSrc[(2u * i0) + 1u] + pSrc[(2u * i2) + 1u];
emilmont 1:fdd22bb7aa52 1174
emilmont 1:fdd22bb7aa52 1175 /* ya - yc */
emilmont 1:fdd22bb7aa52 1176 s2 = pSrc[(2u * i0) + 1u] - pSrc[(2u * i2) + 1u];
emilmont 1:fdd22bb7aa52 1177
emilmont 1:fdd22bb7aa52 1178 /* xc + xd */
emilmont 1:fdd22bb7aa52 1179 t1 = pSrc[2u * i1] + pSrc[2u * i3];
emilmont 1:fdd22bb7aa52 1180
emilmont 1:fdd22bb7aa52 1181 /* xa' = xa + xb + xc + xd */
emilmont 1:fdd22bb7aa52 1182 pSrc[2u * i0] = (r1 + t1) * onebyfftLen;
emilmont 1:fdd22bb7aa52 1183
emilmont 1:fdd22bb7aa52 1184 /* (xa + xb) - (xc + xd) */
emilmont 1:fdd22bb7aa52 1185 r1 = r1 - t1;
emilmont 1:fdd22bb7aa52 1186
emilmont 1:fdd22bb7aa52 1187 /* yb + yd */
emilmont 1:fdd22bb7aa52 1188 t2 = pSrc[(2u * i1) + 1u] + pSrc[(2u * i3) + 1u];
emilmont 1:fdd22bb7aa52 1189
emilmont 1:fdd22bb7aa52 1190 /* ya' = ya + yb + yc + yd */
emilmont 1:fdd22bb7aa52 1191 pSrc[(2u * i0) + 1u] = (s1 + t2) * onebyfftLen;
emilmont 1:fdd22bb7aa52 1192
emilmont 1:fdd22bb7aa52 1193 /* (ya + yc) - (yb + yd) */
emilmont 1:fdd22bb7aa52 1194 s1 = s1 - t2;
emilmont 1:fdd22bb7aa52 1195
emilmont 1:fdd22bb7aa52 1196 /* (yb-yd) */
emilmont 1:fdd22bb7aa52 1197 t1 = pSrc[(2u * i1) + 1u] - pSrc[(2u * i3) + 1u];
emilmont 1:fdd22bb7aa52 1198
emilmont 1:fdd22bb7aa52 1199 /* (xb-xd) */
emilmont 1:fdd22bb7aa52 1200 t2 = pSrc[2u * i1] - pSrc[2u * i3];
emilmont 1:fdd22bb7aa52 1201
emilmont 1:fdd22bb7aa52 1202 /* xc' = (xa-xb+xc-xd)co2 - (ya-yb+yc-yd)(si2) */
emilmont 1:fdd22bb7aa52 1203 pSrc[2u * i1] = r1 * onebyfftLen;
emilmont 1:fdd22bb7aa52 1204
emilmont 1:fdd22bb7aa52 1205 /* yc' = (ya-yb+yc-yd)co2 + (xa-xb+xc-xd)(si2) */
emilmont 1:fdd22bb7aa52 1206 pSrc[(2u * i1) + 1u] = s1 * onebyfftLen;
emilmont 1:fdd22bb7aa52 1207
emilmont 1:fdd22bb7aa52 1208
emilmont 1:fdd22bb7aa52 1209 /* (xa - xc) - (yb-yd) */
emilmont 1:fdd22bb7aa52 1210 r1 = r2 - t1;
emilmont 1:fdd22bb7aa52 1211
emilmont 1:fdd22bb7aa52 1212 /* (xa - xc) + (yb-yd) */
emilmont 1:fdd22bb7aa52 1213 r2 = r2 + t1;
emilmont 1:fdd22bb7aa52 1214
emilmont 1:fdd22bb7aa52 1215 /* (ya - yc) + (xb-xd) */
emilmont 1:fdd22bb7aa52 1216 s1 = s2 + t2;
emilmont 1:fdd22bb7aa52 1217
emilmont 1:fdd22bb7aa52 1218 /* (ya - yc) - (xb-xd) */
emilmont 1:fdd22bb7aa52 1219 s2 = s2 - t2;
emilmont 1:fdd22bb7aa52 1220
emilmont 1:fdd22bb7aa52 1221 /* xb' = (xa+yb-xc-yd)co1 - (ya-xb-yc+xd)(si1) */
emilmont 1:fdd22bb7aa52 1222 pSrc[2u * i2] = r1 * onebyfftLen;
emilmont 1:fdd22bb7aa52 1223
emilmont 1:fdd22bb7aa52 1224 /* yb' = (ya-xb-yc+xd)co1 + (xa+yb-xc-yd)(si1) */
emilmont 1:fdd22bb7aa52 1225 pSrc[(2u * i2) + 1u] = s1 * onebyfftLen;
emilmont 1:fdd22bb7aa52 1226
emilmont 1:fdd22bb7aa52 1227 /* xd' = (xa-yb-xc+yd)co3 - (ya+xb-yc-xd)(si3) */
emilmont 1:fdd22bb7aa52 1228 pSrc[2u * i3] = r2 * onebyfftLen;
emilmont 1:fdd22bb7aa52 1229
emilmont 1:fdd22bb7aa52 1230 /* yd' = (ya+xb-yc-xd)co3 + (xa-yb-xc+yd)(si3) */
emilmont 1:fdd22bb7aa52 1231 pSrc[(2u * i3) + 1u] = s2 * onebyfftLen;
emilmont 1:fdd22bb7aa52 1232 }
emilmont 1:fdd22bb7aa52 1233
emilmont 1:fdd22bb7aa52 1234 #endif /* #ifndef ARM_MATH_CM0 */
emilmont 1:fdd22bb7aa52 1235
emilmont 1:fdd22bb7aa52 1236 }