mbed official
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.
Diff: cmsis_dsp/TransformFunctions/arm_cfft_radix2_f32.c
- Revision:
- 3:7a284390b0ce
- Parent:
- 2:da51fb522205
- Child:
- 5:3762170b6d4d
--- a/cmsis_dsp/TransformFunctions/arm_cfft_radix2_f32.c Thu May 30 17:10:11 2013 +0100
+++ b/cmsis_dsp/TransformFunctions/arm_cfft_radix2_f32.c Fri Nov 08 13:45:10 2013 +0000
@@ -1,8 +1,8 @@
/* ----------------------------------------------------------------------
-* Copyright (C) 2010 ARM Limited. All rights reserved.
+* Copyright (C) 2010-2013 ARM Limited. All rights reserved.
*
-* $Date: 15. February 2012
-* $Revision: V1.1.0
+* $Date: 17. January 2013
+* $Revision: V1.4.1
*
* Project: CMSIS DSP Library
* Title: arm_cfft_radix2_f32.c
@@ -12,169 +12,103 @@
*
* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
*
-* Version 1.1.0 2012/02/15
-* Updated with more optimizations, bug fixes and minor API changes.
-*
-* Version 1.0.3 2010/11/29
-* Initial version
+* Redistribution and use in source and binary forms, with or without
+* modification, are permitted provided that the following conditions
+* are met:
+* - Redistributions of source code must retain the above copyright
+* notice, this list of conditions and the following disclaimer.
+* - Redistributions in binary form must reproduce the above copyright
+* notice, this list of conditions and the following disclaimer in
+* the documentation and/or other materials provided with the
+* distribution.
+* - Neither the name of ARM LIMITED nor the names of its contributors
+* may be used to endorse or promote products derived from this
+* software without specific prior written permission.
+*
+* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
+* FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
+* COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
+* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
+* BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+* CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
+* ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+* POSSIBILITY OF SUCH DAMAGE.
* -------------------------------------------------------------------- */
#include "arm_math.h"
+void arm_radix2_butterfly_f32(
+ float32_t * pSrc,
+ uint32_t fftLen,
+ float32_t * pCoef,
+ uint16_t twidCoefModifier);
+
+void arm_radix2_butterfly_inverse_f32(
+ float32_t * pSrc,
+ uint32_t fftLen,
+ float32_t * pCoef,
+ uint16_t twidCoefModifier,
+ float32_t onebyfftLen);
+
+extern void arm_bitreversal_f32(
+ float32_t * pSrc,
+ uint16_t fftSize,
+ uint16_t bitRevFactor,
+ uint16_t * pBitRevTab);
+
/**
- * @ingroup groupTransforms
- */
+* @ingroup groupTransforms
+*/
+
+/**
+* @addtogroup ComplexFFT
+* @{
+*/
/**
- * @defgroup Radix2_CFFT_CIFFT Radix-2 Complex FFT Functions
- *
- * \par
- * 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).
- * Computational complexity of CFFT reduces drastically when compared to DFT.
- * \par
- * This set of functions implements CFFT/CIFFT
- * for Q15, Q31, and floating-point data types. The functions operates on in-place buffer which uses same buffer for input and output.
- * Complex input is stored in input buffer in an interleaved fashion.
- *
- * \par
- * The functions operate on blocks of input and output data and each call to the function processes
- * <code>2*fftLen</code> samples through the transform. <code>pSrc</code> points to In-place arrays containing <code>2*fftLen</code> values.
- * \par
- * 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.
- * <pre> {real[0], imag[0], real[1], imag[1],..} </pre>
- *
- * \par Lengths supported by the transform:
- * \par
- * Internally, the function utilize a radix-2 decimation in frequency(DIF) algorithm
- * and the size of the FFT supported are of the lengths [16, 32, 64, 128, 256, 512, 1024, 2048, 4096].
- *
- *
- * \par Algorithm:
- *
- * <b>Complex Fast Fourier Transform:</b>
- * \par
- * Input real and imaginary data:
- * <pre>
- * x(n) = xa + j * ya
- * x(n+N/2 ) = xb + j * yb
- * </pre>
- * where N is length of FFT
- * \par
- * Output real and imaginary data:
- * <pre>
- * X(2r) = xa'+ j * ya'
- * X(2r+1) = xb'+ j * yb'
- * </pre>
- * \par
- * Twiddle factors for radix-2 FFT:
- * <pre>
- * Wn = cosVal + j * (- sinVal)
- * </pre>
- *
- * \par
- * \image html CFFT_Radix2.gif "Radix-2 Decimation-in Frequency Complex Fast Fourier Transform"
- *
- * \par
- * Output from Radix-2 CFFT Results in Digit reversal order. Interchange middle two branches of every butterfly results in Bit reversed output.
- * \par
- * <b> Butterfly CFFT equations:</b>
- * <pre>
- * xa' = xa + xb
- * ya' = ya + yb
- * xb' = (xa-xb)* cosVal + (ya-yb) * sinVal
- * yb' = (ya-yb)* cosVal - (xa-xb) * sinVal
- * </pre>
- *
- *
- * <b>Complex Inverse Fast Fourier Transform:</b>
- * \par
- * CIFFT uses same twiddle factor table as CFFT with modifications in the design equation as shown below.
- *
- * \par
- * <b> Modified Butterfly CIFFT equations:</b>
- * <pre>
- * xa' = xa + xb
- * ya' = ya + yb
- * xb' = (xa-xb)* cosVal - (ya-yb) * sinVal
- * yb' = (ya-yb)* cosVal + (xa-xb) * sinVal
- * </pre>
- *
- * \par Instance Structure
- * A separate instance structure must be defined for each Instance but the twiddle factors and bit reversal tables can be reused.
- * There are separate instance structure declarations for each of the 3 supported data types.
- *
- * \par Initialization Functions
- * There is also an associated initialization function for each data type.
- * The initialization function performs the following operations:
- * - Sets the values of the internal structure fields.
- * - Initializes twiddle factor table and bit reversal table pointers
- * \par
- * Use of the initialization function is optional.
- * However, if the initialization function is used, then the instance structure cannot be placed into a const data section.
- * To place an instance structure into a const data section, the instance structure must be manually initialized.
- * Manually initialize the instance structure as follows:
- * <pre>
- *arm_cfft_radix2_instance_f32 S = {fftLen, ifftFlag, bitReverseFlag, pTwiddle, pBitRevTable, twidCoefModifier, bitRevFactor, onebyfftLen};
- *arm_cfft_radix2_instance_q31 S = {fftLen, ifftFlag, bitReverseFlag, pTwiddle, pBitRevTable, twidCoefModifier, bitRevFactor};
- *arm_cfft_radix2_instance_q15 S = {fftLen, ifftFlag, bitReverseFlag, pTwiddle, pBitRevTable, twidCoefModifier, bitRevFactor};
- * </pre>
- * \par
- * 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);
- * <code>bitReverseFlag</code> Flag for selection of output order(Set bitReverseFlag to output in normal order otherwise output in bit reversed order);
- * <code>pTwiddle</code>points to array of twiddle coefficients; <code>pBitRevTable</code> points to the array of bit reversal table.
- * <code>twidCoefModifier</code> modifier for twiddle factor table which supports all FFT lengths with same table;
- * <code>pBitRevTable</code> modifier for bit reversal table which supports all FFT lengths with same table.
- * <code>onebyfftLen</code> value of 1/fftLen to calculate CIFFT;
- *
- * \par Fixed-Point Behavior
- * Care must be taken when using the fixed-point versions of the CFFT/CIFFT function.
- * Refer to the function specific documentation below for usage guidelines.
- */
-
-
-/**
- * @addtogroup Radix2_CFFT_CIFFT
- * @{
- */
-
-/**
- * @details
- * @brief Processing function for the floating-point Radix-2 CFFT/CIFFT.
- * @param[in] *S points to an instance of the floating-point Radix-2 CFFT/CIFFT structure.
- * @param[in, out] *pSrc points to the complex data buffer of size <code>2*fftLen</code>. Processing occurs in-place.
- * @return none.
- */
+* @details
+* @brief Radix-2 CFFT/CIFFT.
+* @deprecated Do not use this function. It has been superceded by \ref arm_cfft_f32 and will be removed
+* in the future.
+* @param[in] *S points to an instance of the floating-point Radix-2 CFFT/CIFFT structure.
+* @param[in, out] *pSrc points to the complex data buffer of size <code>2*fftLen</code>. Processing occurs in-place.
+* @return none.
+*/
void arm_cfft_radix2_f32(
- const arm_cfft_radix2_instance_f32 * S,
- float32_t * pSrc)
+const arm_cfft_radix2_instance_f32 * S,
+float32_t * pSrc)
{
- if(S->ifftFlag == 1u)
- {
- /* Complex IFFT radix-2 */
- arm_radix2_butterfly_inverse_f32(pSrc, S->fftLen, S->pTwiddle,
- S->twidCoefModifier, S->onebyfftLen);
- }
- else
- {
- /* Complex FFT radix-2 */
- arm_radix2_butterfly_f32(pSrc, S->fftLen, S->pTwiddle,
- S->twidCoefModifier);
- }
+ if(S->ifftFlag == 1u)
+ {
+ /* Complex IFFT radix-2 */
+ arm_radix2_butterfly_inverse_f32(pSrc, S->fftLen, S->pTwiddle,
+ S->twidCoefModifier, S->onebyfftLen);
+ }
+ else
+ {
+ /* Complex FFT radix-2 */
+ arm_radix2_butterfly_f32(pSrc, S->fftLen, S->pTwiddle,
+ S->twidCoefModifier);
+ }
- if(S->bitReverseFlag == 1u)
- {
- /* Bit Reversal */
- arm_bitreversal_f32(pSrc, S->fftLen, S->bitRevFactor, S->pBitRevTable);
- }
+ if(S->bitReverseFlag == 1u)
+ {
+ /* Bit Reversal */
+ arm_bitreversal_f32(pSrc, S->fftLen, S->bitRevFactor, S->pBitRevTable);
+ }
}
/**
- * @} end of Radix2_CFFT_CIFFT group
- */
+* @} end of ComplexFFT group
+*/
@@ -183,329 +117,369 @@
** ------------------------------------------------------------------- */
/*
- * @brief Core function for the floating-point CFFT butterfly process.
- * @param[in, out] *pSrc points to the in-place buffer of floating-point data type.
- * @param[in] fftLen length of the FFT.
- * @param[in] *pCoef points to the twiddle coefficient buffer.
- * @param[in] twidCoefModifier twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table.
- * @return none.
- */
+* @brief Core function for the floating-point CFFT butterfly process.
+* @param[in, out] *pSrc points to the in-place buffer of floating-point data type.
+* @param[in] fftLen length of the FFT.
+* @param[in] *pCoef points to the twiddle coefficient buffer.
+* @param[in] twidCoefModifier twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table.
+* @return none.
+*/
void arm_radix2_butterfly_f32(
- float32_t * pSrc,
- uint32_t fftLen,
- float32_t * pCoef,
- uint16_t twidCoefModifier)
+float32_t * pSrc,
+uint32_t fftLen,
+float32_t * pCoef,
+uint16_t twidCoefModifier)
{
- int i, j, k, l;
- int n1, n2, ia;
- float32_t xt, yt, cosVal, sinVal;
-
-#ifndef ARM_MATH_CM0
-
- /* Initializations for the first stage */
- n2 = fftLen;
-
- n1 = n2;
- n2 = n2 >> 1;
- ia = 0;
+ uint32_t i, j, k, l;
+ uint32_t n1, n2, ia;
+ float32_t xt, yt, cosVal, sinVal;
+ float32_t p0, p1, p2, p3;
+ float32_t a0, a1;
- // loop for groups
- for (i = 0; i < n2; i++)
- {
- cosVal = pCoef[ia * 2];
- sinVal = pCoef[(ia * 2) + 1];
-
- /* Twiddle coefficients index modifier */
- ia = ia + twidCoefModifier;
-
- /* index calculation for the input as, */
- /* pSrc[i + 0], pSrc[i + fftLen/1] */
- l = i + n2;
+#ifndef ARM_MATH_CM0_FAMILY
- /* Butterfly implementation */
- xt = pSrc[2 * i] - pSrc[2 * l];
- pSrc[2 * i] = pSrc[2 * i] + pSrc[2 * l];
-
- yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
- pSrc[2 * i + 1] = pSrc[2 * l + 1] + pSrc[2 * i + 1];
-
- pSrc[2u * l] = xt * cosVal + yt * sinVal;
-
- pSrc[2u * l + 1u] = yt * cosVal - xt * sinVal;
-
- } // groups loop end
+ /* Initializations for the first stage */
+ n2 = fftLen >> 1;
+ ia = 0;
+ i = 0;
- twidCoefModifier = twidCoefModifier << 1u;
-
- // loop for stage
- for (k = fftLen / 2; k > 2; k = k >> 1)
- {
- n1 = n2;
- n2 = n2 >> 1;
- ia = 0;
-
- // loop for groups
- for (j = 0; j < n2; j++)
- {
+ // loop for groups
+ for (k = n2; k > 0; k--)
+ {
cosVal = pCoef[ia * 2];
sinVal = pCoef[(ia * 2) + 1];
- ia = ia + twidCoefModifier;
+
+ /* Twiddle coefficients index modifier */
+ ia += twidCoefModifier;
- // loop for butterfly
- for (i = j; i < fftLen; i += n1)
- {
- l = i + n2;
- xt = pSrc[2 * i] - pSrc[2 * l];
- pSrc[2 * i] = pSrc[2 * i] + pSrc[2 * l];
+ /* index calculation for the input as, */
+ /* pSrc[i + 0], pSrc[i + fftLen/1] */
+ l = i + n2;
+
+ /* Butterfly implementation */
+ a0 = pSrc[2 * i] + pSrc[2 * l];
+ xt = pSrc[2 * i] - pSrc[2 * l];
- yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
- pSrc[2 * i + 1] = pSrc[2 * l + 1] + pSrc[2 * i + 1];
-
- pSrc[2u * l] = xt * cosVal + yt * sinVal;
-
- pSrc[2u * l + 1u] = yt * cosVal - xt * sinVal;
+ yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
+ a1 = pSrc[2 * l + 1] + pSrc[2 * i + 1];
+
+ p0 = xt * cosVal;
+ p1 = yt * sinVal;
+ p2 = yt * cosVal;
+ p3 = xt * sinVal;
+
+ pSrc[2 * i] = a0;
+ pSrc[2 * i + 1] = a1;
+
+ pSrc[2 * l] = p0 + p1;
+ pSrc[2 * l + 1] = p2 - p3;
+
+ i++;
+ } // groups loop end
- } // butterfly loop end
+ twidCoefModifier <<= 1u;
- } // groups loop end
-
- twidCoefModifier = twidCoefModifier << 1u;
- } // stages loop end
+ // loop for stage
+ for (k = n2; k > 2; k = k >> 1)
+ {
+ n1 = n2;
+ n2 = n2 >> 1;
+ ia = 0;
- n1 = n2;
- n2 = n2 >> 1;
- ia = 0;
+ // loop for groups
+ j = 0;
+ do
+ {
+ cosVal = pCoef[ia * 2];
+ sinVal = pCoef[(ia * 2) + 1];
+ ia += twidCoefModifier;
- cosVal = pCoef[ia * 2];
- sinVal = pCoef[(ia * 2) + 1];
- ia = ia + twidCoefModifier;
+ // loop for butterfly
+ i = j;
+ do
+ {
+ l = i + n2;
+ a0 = pSrc[2 * i] + pSrc[2 * l];
+ xt = pSrc[2 * i] - pSrc[2 * l];
- // loop for butterfly
- for (i = 0; i < fftLen; i += n1)
- {
- l = i + n2;
- xt = pSrc[2 * i] - pSrc[2 * l];
- pSrc[2 * i] = (pSrc[2 * i] + pSrc[2 * l]);
+ yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
+ a1 = pSrc[2 * l + 1] + pSrc[2 * i + 1];
+
+ p0 = xt * cosVal;
+ p1 = yt * sinVal;
+ p2 = yt * cosVal;
+ p3 = xt * sinVal;
+
+ pSrc[2 * i] = a0;
+ pSrc[2 * i + 1] = a1;
+
+ pSrc[2 * l] = p0 + p1;
+ pSrc[2 * l + 1] = p2 - p3;
+
+ i += n1;
+ } while( i < fftLen ); // butterfly loop end
+ j++;
+ } while( j < n2); // groups loop end
+ twidCoefModifier <<= 1u;
+ } // stages loop end
- yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
- pSrc[2 * i + 1] = (pSrc[2 * l + 1] + pSrc[2 * i + 1]);
+ // loop for butterfly
+ for (i = 0; i < fftLen; i += 2)
+ {
+ a0 = pSrc[2 * i] + pSrc[2 * i + 2];
+ xt = pSrc[2 * i] - pSrc[2 * i + 2];
- pSrc[2u * l] = xt;
-
- pSrc[2u * l + 1u] = yt;
-
- } // groups loop end
+ yt = pSrc[2 * i + 1] - pSrc[2 * i + 3];
+ a1 = pSrc[2 * i + 3] + pSrc[2 * i + 1];
+
+ pSrc[2 * i] = a0;
+ pSrc[2 * i + 1] = a1;
+ pSrc[2 * i + 2] = xt;
+ pSrc[2 * i + 3] = yt;
+ } // groups loop end
#else
-
- //N = fftLen;
- n2 = fftLen;
+
+ n2 = fftLen;
- // loop for stage
- for (k = fftLen; k > 1; k = k >> 1)
- {
- n1 = n2;
- n2 = n2 >> 1;
- ia = 0;
+ // loop for stage
+ for (k = fftLen; k > 1; k = k >> 1)
+ {
+ n1 = n2;
+ n2 = n2 >> 1;
+ ia = 0;
- // loop for groups
- for (j = 0; j < n2; j++)
- {
- cosVal = pCoef[ia * 2];
- sinVal = pCoef[(ia * 2) + 1];
- ia = ia + twidCoefModifier;
+ // loop for groups
+ j = 0;
+ do
+ {
+ cosVal = pCoef[ia * 2];
+ sinVal = pCoef[(ia * 2) + 1];
+ ia += twidCoefModifier;
- // loop for butterfly
- for (i = j; i < fftLen; i += n1)
- {
- l = i + n2;
- xt = pSrc[2 * i] - pSrc[2 * l];
- pSrc[2 * i] = pSrc[2 * i] + pSrc[2 * l];
-
- yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
- pSrc[2 * i + 1] = pSrc[2 * l + 1] + pSrc[2 * i + 1];
+ // loop for butterfly
+ i = j;
+ do
+ {
+ l = i + n2;
+ a0 = pSrc[2 * i] + pSrc[2 * l];
+ xt = pSrc[2 * i] - pSrc[2 * l];
- pSrc[2 * l] = (cosVal * xt + sinVal * yt); // >> 15;
- pSrc[2 * l + 1] = (cosVal * yt - sinVal * xt); // >> 15;
+ yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
+ a1 = pSrc[2 * l + 1] + pSrc[2 * i + 1];
+
+ p0 = xt * cosVal;
+ p1 = yt * sinVal;
+ p2 = yt * cosVal;
+ p3 = xt * sinVal;
+
+ pSrc[2 * i] = a0;
+ pSrc[2 * i + 1] = a1;
+
+ pSrc[2 * l] = p0 + p1;
+ pSrc[2 * l + 1] = p2 - p3;
+
+ i += n1;
+ } while(i < fftLen);
+ j++;
+ } while(j < n2);
+ twidCoefModifier <<= 1u;
+ }
- }
- }
- twidCoefModifier = twidCoefModifier << 1u;
- }
-
-#endif // #ifndef ARM_MATH_CM0
+#endif // #ifndef ARM_MATH_CM0_FAMILY
}
void arm_radix2_butterfly_inverse_f32(
- float32_t * pSrc,
- uint32_t fftLen,
- float32_t * pCoef,
- uint16_t twidCoefModifier,
- float32_t onebyfftLen)
+float32_t * pSrc,
+uint32_t fftLen,
+float32_t * pCoef,
+uint16_t twidCoefModifier,
+float32_t onebyfftLen)
{
- int i, j, k, l;
- int n1, n2, ia;
- float32_t xt, yt, cosVal, sinVal;
-
-#ifndef ARM_MATH_CM0
-
- //N = fftLen;
- n2 = fftLen;
-
- n1 = n2;
- n2 = n2 >> 1;
- ia = 0;
-
- // loop for groups
- for (i = 0; i < n2; i++)
- {
- cosVal = pCoef[ia * 2];
- sinVal = pCoef[(ia * 2) + 1];
- ia = ia + twidCoefModifier;
+ uint32_t i, j, k, l;
+ uint32_t n1, n2, ia;
+ float32_t xt, yt, cosVal, sinVal;
+ float32_t p0, p1, p2, p3;
+ float32_t a0, a1;
- l = i + n2;
- xt = pSrc[2 * i] - pSrc[2 * l];
- pSrc[2 * i] = pSrc[2 * i] + pSrc[2 * l];
-
- yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
- pSrc[2 * i + 1] = pSrc[2 * l + 1] + pSrc[2 * i + 1];
-
- pSrc[2u * l] = xt * cosVal - yt * sinVal;
-
- pSrc[2u * l + 1u] = yt * cosVal + xt * sinVal;
-
- } // groups loop end
+#ifndef ARM_MATH_CM0_FAMILY
- twidCoefModifier = twidCoefModifier << 1u;
+ n2 = fftLen >> 1;
+ ia = 0;
- // loop for stage
- for (k = fftLen / 2; k > 2; k = k >> 1)
- {
- n1 = n2;
- n2 = n2 >> 1;
- ia = 0;
-
- // loop for groups
- for (j = 0; j < n2; j++)
- {
+ // loop for groups
+ for (i = 0; i < n2; i++)
+ {
cosVal = pCoef[ia * 2];
sinVal = pCoef[(ia * 2) + 1];
- ia = ia + twidCoefModifier;
+ ia += twidCoefModifier;
- // loop for butterfly
- for (i = j; i < fftLen; i += n1)
- {
- l = i + n2;
- xt = pSrc[2 * i] - pSrc[2 * l];
- pSrc[2 * i] = pSrc[2 * i] + pSrc[2 * l];
+ l = i + n2;
+ a0 = pSrc[2 * i] + pSrc[2 * l];
+ xt = pSrc[2 * i] - pSrc[2 * l];
- yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
- pSrc[2 * i + 1] = pSrc[2 * l + 1] + pSrc[2 * i + 1];
-
- pSrc[2u * l] = xt * cosVal - yt * sinVal;
-
- pSrc[2u * l + 1u] = yt * cosVal + xt * sinVal;
+ yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
+ a1 = pSrc[2 * l + 1] + pSrc[2 * i + 1];
+
+ p0 = xt * cosVal;
+ p1 = yt * sinVal;
+ p2 = yt * cosVal;
+ p3 = xt * sinVal;
+
+ pSrc[2 * i] = a0;
+ pSrc[2 * i + 1] = a1;
+
+ pSrc[2 * l] = p0 - p1;
+ pSrc[2 * l + 1] = p2 + p3;
+ } // groups loop end
- } // butterfly loop end
+ twidCoefModifier <<= 1u;
- } // groups loop end
+ // loop for stage
+ for (k = fftLen / 2; k > 2; k = k >> 1)
+ {
+ n1 = n2;
+ n2 = n2 >> 1;
+ ia = 0;
- twidCoefModifier = twidCoefModifier << 1u;
- } // stages loop end
+ // loop for groups
+ j = 0;
+ do
+ {
+ cosVal = pCoef[ia * 2];
+ sinVal = pCoef[(ia * 2) + 1];
+ ia += twidCoefModifier;
- n1 = n2;
- n2 = n2 >> 1;
- ia = 0;
+ // loop for butterfly
+ i = j;
+ do
+ {
+ l = i + n2;
+ a0 = pSrc[2 * i] + pSrc[2 * l];
+ xt = pSrc[2 * i] - pSrc[2 * l];
- cosVal = pCoef[ia * 2];
- sinVal = pCoef[(ia * 2) + 1];
- ia = ia + twidCoefModifier;
+ yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
+ a1 = pSrc[2 * l + 1] + pSrc[2 * i + 1];
+
+ p0 = xt * cosVal;
+ p1 = yt * sinVal;
+ p2 = yt * cosVal;
+ p3 = xt * sinVal;
+
+ pSrc[2 * i] = a0;
+ pSrc[2 * i + 1] = a1;
+
+ pSrc[2 * l] = p0 - p1;
+ pSrc[2 * l + 1] = p2 + p3;
- // loop for butterfly
- for (i = 0; i < fftLen; i += n1)
- {
- l = i + n2;
- xt = pSrc[2 * i] - pSrc[2 * l];
- pSrc[2 * i] = (pSrc[2 * i] + pSrc[2 * l]) * onebyfftLen;
+ i += n1;
+ } while( i < fftLen ); // butterfly loop end
+ j++;
+ } while(j < n2); // groups loop end
+
+ twidCoefModifier <<= 1u;
+ } // stages loop end
- yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
- pSrc[2 * i + 1] = (pSrc[2 * l + 1] + pSrc[2 * i + 1]) * onebyfftLen;
-
- pSrc[2u * l] = xt * onebyfftLen;
-
- pSrc[2u * l + 1u] = yt * onebyfftLen;
-
- } // butterfly loop end
+ // loop for butterfly
+ for (i = 0; i < fftLen; i += 2)
+ {
+ a0 = pSrc[2 * i] + pSrc[2 * i + 2];
+ xt = pSrc[2 * i] - pSrc[2 * i + 2];
+
+ a1 = pSrc[2 * i + 3] + pSrc[2 * i + 1];
+ yt = pSrc[2 * i + 1] - pSrc[2 * i + 3];
+
+ p0 = a0 * onebyfftLen;
+ p2 = xt * onebyfftLen;
+ p1 = a1 * onebyfftLen;
+ p3 = yt * onebyfftLen;
+
+ pSrc[2 * i] = p0;
+ pSrc[2 * i + 1] = p1;
+ pSrc[2 * i + 2] = p2;
+ pSrc[2 * i + 3] = p3;
+ } // butterfly loop end
#else
- //N = fftLen;
- n2 = fftLen;
+ n2 = fftLen;
- // loop for stage
- for (k = fftLen; k > 2; k = k >> 1)
- {
- n1 = n2;
- n2 = n2 >> 1;
- ia = 0;
+ // loop for stage
+ for (k = fftLen; k > 2; k = k >> 1)
+ {
+ n1 = n2;
+ n2 = n2 >> 1;
+ ia = 0;
- // loop for groups
- for (j = 0; j < n2; j++)
- {
- cosVal = pCoef[ia * 2];
- sinVal = pCoef[(ia * 2) + 1];
- ia = ia + twidCoefModifier;
-
- // loop for butterfly
- for (i = j; i < fftLen; i += n1)
+ // loop for groups
+ j = 0;
+ do
{
- l = i + n2;
- xt = pSrc[2 * i] - pSrc[2 * l];
- pSrc[2 * i] = pSrc[2 * i] + pSrc[2 * l];
+ cosVal = pCoef[ia * 2];
+ sinVal = pCoef[(ia * 2) + 1];
+ ia = ia + twidCoefModifier;
- yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
- pSrc[2 * i + 1] = pSrc[2 * l + 1] + pSrc[2 * i + 1];
-
- pSrc[2u * l] = xt * cosVal - yt * sinVal;
-
- pSrc[2u * l + 1u] = yt * cosVal + xt * sinVal;
+ // loop for butterfly
+ i = j;
+ do
+ {
+ l = i + n2;
+ a0 = pSrc[2 * i] + pSrc[2 * l];
+ xt = pSrc[2 * i] - pSrc[2 * l];
- } // butterfly loop end
-
- } // groups loop end
-
- twidCoefModifier = twidCoefModifier << 1u;
- } // stages loop end
+ yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
+ a1 = pSrc[2 * l + 1] + pSrc[2 * i + 1];
+
+ p0 = xt * cosVal;
+ p1 = yt * sinVal;
+ p2 = yt * cosVal;
+ p3 = xt * sinVal;
+
+ pSrc[2 * i] = a0;
+ pSrc[2 * i + 1] = a1;
+
+ pSrc[2 * l] = p0 - p1;
+ pSrc[2 * l + 1] = p2 + p3;
+
+ i += n1;
+ } while( i < fftLen ); // butterfly loop end
+ j++;
+ } while( j < n2 ); // groups loop end
- n1 = n2;
- n2 = n2 >> 1;
- ia = 0;
+ twidCoefModifier = twidCoefModifier << 1u;
+ } // stages loop end
- cosVal = pCoef[ia * 2];
- sinVal = pCoef[(ia * 2) + 1];
- ia = ia + twidCoefModifier;
+ n1 = n2;
+ n2 = n2 >> 1;
- // loop for butterfly
- for (i = 0; i < fftLen; i += n1)
- {
- l = i + n2;
- xt = pSrc[2 * i] - pSrc[2 * l];
- pSrc[2 * i] = (pSrc[2 * i] + pSrc[2 * l]) * onebyfftLen;
+ // loop for butterfly
+ for (i = 0; i < fftLen; i += n1)
+ {
+ l = i + n2;
+
+ a0 = pSrc[2 * i] + pSrc[2 * l];
+ xt = pSrc[2 * i] - pSrc[2 * l];
+
+ a1 = pSrc[2 * l + 1] + pSrc[2 * i + 1];
+ yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
+
+ p0 = a0 * onebyfftLen;
+ p2 = xt * onebyfftLen;
+ p1 = a1 * onebyfftLen;
+ p3 = yt * onebyfftLen;
+
+ pSrc[2 * i] = p0;
+ pSrc[2u * l] = p2;
+
+ pSrc[2 * i + 1] = p1;
+ pSrc[2u * l + 1u] = p3;
+ } // butterfly loop end
- yt = pSrc[2 * i + 1] - pSrc[2 * l + 1];
- pSrc[2 * i + 1] = (pSrc[2 * l + 1] + pSrc[2 * i + 1]) * onebyfftLen;
-
- pSrc[2u * l] = xt * onebyfftLen;
-
- pSrc[2u * l + 1u] = yt * onebyfftLen;
-
- } // butterfly loop end
-
-#endif // #ifndef ARM_MATH_CM0
+#endif // #ifndef ARM_MATH_CM0_FAMILY
}