[Add] First commit

This commit is contained in:
gaoyang3513
2023-05-18 18:53:00 +08:00
commit 179cffc2c1
6607 changed files with 2163514 additions and 0 deletions

View File

@ -0,0 +1,229 @@
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_bitreversal.c
* Description: Bitreversal functions
*
* $Date: 18. March 2019
* $Revision: V1.6.0
*
* Target Processor: Cortex-M cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2019 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* Licensed under the Apache License, Version 2.0 (the License); you may
* not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an AS IS BASIS, WITHOUT
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "arm_math.h"
#include "arm_common_tables.h"
/**
@brief In-place floating-point bit reversal function.
@param[in,out] pSrc points to in-place floating-point data buffer
@param[in] fftSize length of FFT
@param[in] bitRevFactor bit reversal modifier that supports different size FFTs with the same bit reversal table
@param[in] pBitRevTab points to bit reversal table
@return none
*/
void arm_bitreversal_f32(
float32_t * pSrc,
uint16_t fftSize,
uint16_t bitRevFactor,
const uint16_t * pBitRevTab)
{
uint16_t fftLenBy2, fftLenBy2p1;
uint16_t i, j;
float32_t in;
/* Initializations */
j = 0U;
fftLenBy2 = fftSize >> 1U;
fftLenBy2p1 = (fftSize >> 1U) + 1U;
/* Bit Reversal Implementation */
for (i = 0U; i <= (fftLenBy2 - 2U); i += 2U)
{
if (i < j)
{
/* pSrc[i] <-> pSrc[j]; */
in = pSrc[2U * i];
pSrc[2U * i] = pSrc[2U * j];
pSrc[2U * j] = in;
/* pSrc[i+1U] <-> pSrc[j+1U] */
in = pSrc[(2U * i) + 1U];
pSrc[(2U * i) + 1U] = pSrc[(2U * j) + 1U];
pSrc[(2U * j) + 1U] = in;
/* pSrc[i+fftLenBy2p1] <-> pSrc[j+fftLenBy2p1] */
in = pSrc[2U * (i + fftLenBy2p1)];
pSrc[2U * (i + fftLenBy2p1)] = pSrc[2U * (j + fftLenBy2p1)];
pSrc[2U * (j + fftLenBy2p1)] = in;
/* pSrc[i+fftLenBy2p1+1U] <-> pSrc[j+fftLenBy2p1+1U] */
in = pSrc[(2U * (i + fftLenBy2p1)) + 1U];
pSrc[(2U * (i + fftLenBy2p1)) + 1U] =
pSrc[(2U * (j + fftLenBy2p1)) + 1U];
pSrc[(2U * (j + fftLenBy2p1)) + 1U] = in;
}
/* pSrc[i+1U] <-> pSrc[j+1U] */
in = pSrc[2U * (i + 1U)];
pSrc[2U * (i + 1U)] = pSrc[2U * (j + fftLenBy2)];
pSrc[2U * (j + fftLenBy2)] = in;
/* pSrc[i+2U] <-> pSrc[j+2U] */
in = pSrc[(2U * (i + 1U)) + 1U];
pSrc[(2U * (i + 1U)) + 1U] = pSrc[(2U * (j + fftLenBy2)) + 1U];
pSrc[(2U * (j + fftLenBy2)) + 1U] = in;
/* Reading the index for the bit reversal */
j = *pBitRevTab;
/* Updating the bit reversal index depending on the fft length */
pBitRevTab += bitRevFactor;
}
}
/**
@brief In-place Q31 bit reversal function.
@param[in,out] pSrc points to in-place Q31 data buffer.
@param[in] fftLen length of FFT.
@param[in] bitRevFactor bit reversal modifier that supports different size FFTs with the same bit reversal table
@param[in] pBitRevTab points to bit reversal table
@return none
*/
void arm_bitreversal_q31(
q31_t * pSrc,
uint32_t fftLen,
uint16_t bitRevFactor,
const uint16_t * pBitRevTab)
{
uint32_t fftLenBy2, fftLenBy2p1, i, j;
q31_t in;
/* Initializations */
j = 0U;
fftLenBy2 = fftLen / 2U;
fftLenBy2p1 = (fftLen / 2U) + 1U;
/* Bit Reversal Implementation */
for (i = 0U; i <= (fftLenBy2 - 2U); i += 2U)
{
if (i < j)
{
/* pSrc[i] <-> pSrc[j]; */
in = pSrc[2U * i];
pSrc[2U * i] = pSrc[2U * j];
pSrc[2U * j] = in;
/* pSrc[i+1U] <-> pSrc[j+1U] */
in = pSrc[(2U * i) + 1U];
pSrc[(2U * i) + 1U] = pSrc[(2U * j) + 1U];
pSrc[(2U * j) + 1U] = in;
/* pSrc[i+fftLenBy2p1] <-> pSrc[j+fftLenBy2p1] */
in = pSrc[2U * (i + fftLenBy2p1)];
pSrc[2U * (i + fftLenBy2p1)] = pSrc[2U * (j + fftLenBy2p1)];
pSrc[2U * (j + fftLenBy2p1)] = in;
/* pSrc[i+fftLenBy2p1+1U] <-> pSrc[j+fftLenBy2p1+1U] */
in = pSrc[(2U * (i + fftLenBy2p1)) + 1U];
pSrc[(2U * (i + fftLenBy2p1)) + 1U] =
pSrc[(2U * (j + fftLenBy2p1)) + 1U];
pSrc[(2U * (j + fftLenBy2p1)) + 1U] = in;
}
/* pSrc[i+1U] <-> pSrc[j+1U] */
in = pSrc[2U * (i + 1U)];
pSrc[2U * (i + 1U)] = pSrc[2U * (j + fftLenBy2)];
pSrc[2U * (j + fftLenBy2)] = in;
/* pSrc[i+2U] <-> pSrc[j+2U] */
in = pSrc[(2U * (i + 1U)) + 1U];
pSrc[(2U * (i + 1U)) + 1U] = pSrc[(2U * (j + fftLenBy2)) + 1U];
pSrc[(2U * (j + fftLenBy2)) + 1U] = in;
/* Reading the index for the bit reversal */
j = *pBitRevTab;
/* Updating the bit reversal index depending on the fft length */
pBitRevTab += bitRevFactor;
}
}
/**
@brief In-place Q15 bit reversal function.
@param[in,out] pSrc16 points to in-place Q15 data buffer
@param[in] fftLen length of FFT
@param[in] bitRevFactor bit reversal modifier that supports different size FFTs with the same bit reversal table
@param[in] pBitRevTab points to bit reversal table
@return none
*/
void arm_bitreversal_q15(
q15_t * pSrc16,
uint32_t fftLen,
uint16_t bitRevFactor,
const uint16_t * pBitRevTab)
{
q31_t *pSrc = (q31_t *) pSrc16;
q31_t in;
uint32_t fftLenBy2, fftLenBy2p1;
uint32_t i, j;
/* Initializations */
j = 0U;
fftLenBy2 = fftLen / 2U;
fftLenBy2p1 = (fftLen / 2U) + 1U;
/* Bit Reversal Implementation */
for (i = 0U; i <= (fftLenBy2 - 2U); i += 2U)
{
if (i < j)
{
/* pSrc[i] <-> pSrc[j]; */
/* pSrc[i+1U] <-> pSrc[j+1U] */
in = pSrc[i];
pSrc[i] = pSrc[j];
pSrc[j] = in;
/* pSrc[i + fftLenBy2p1] <-> pSrc[j + fftLenBy2p1]; */
/* pSrc[i + fftLenBy2p1+1U] <-> pSrc[j + fftLenBy2p1+1U] */
in = pSrc[i + fftLenBy2p1];
pSrc[i + fftLenBy2p1] = pSrc[j + fftLenBy2p1];
pSrc[j + fftLenBy2p1] = in;
}
/* pSrc[i+1U] <-> pSrc[j+fftLenBy2]; */
/* pSrc[i+2] <-> pSrc[j+fftLenBy2+1U] */
in = pSrc[i + 1U];
pSrc[i + 1U] = pSrc[j + fftLenBy2];
pSrc[j + fftLenBy2] = in;
/* Reading the index for the bit reversal */
j = *pBitRevTab;
/* Updating the bit reversal index depending on the fft length */
pBitRevTab += bitRevFactor;
}
}

View File

@ -0,0 +1,220 @@
;/* ----------------------------------------------------------------------
; * Project: CMSIS DSP Library
; * Title: arm_bitreversal2.S
; * Description: arm_bitreversal_32 function done in assembly for maximum speed.
; * Called after doing an fft to reorder the output.
; * The function is loop unrolled by 2. arm_bitreversal_16 as well.
; *
; * $Date: 18. March 2019
; * $Revision: V1.5.2
; *
; * Target Processor: Cortex-M cores
; * -------------------------------------------------------------------- */
;/*
; * Copyright (C) 2010-2019 ARM Limited or its affiliates. All rights reserved.
; *
; * SPDX-License-Identifier: Apache-2.0
; *
; * Licensed under the Apache License, Version 2.0 (the License); you may
; * not use this file except in compliance with the License.
; * You may obtain a copy of the License at
; *
; * www.apache.org/licenses/LICENSE-2.0
; *
; * Unless required by applicable law or agreed to in writing, software
; * distributed under the License is distributed on an AS IS BASIS, WITHOUT
; * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
; * See the License for the specific language governing permissions and
; * limitations under the License.
; */
#if defined ( __CC_ARM ) /* Keil */
#define CODESECT AREA ||.text||, CODE, READONLY, ALIGN=2
#define LABEL
#elif defined (__ARMCC_VERSION) && (__ARMCC_VERSION >= 6010050)
#define CODESECT AREA ||.text||, CODE, READONLY, ALIGN=2
#define LABEL
#elif defined ( __IASMARM__ ) /* IAR */
#define CODESECT SECTION `.text`:CODE
#define PROC
#define LABEL
#define ENDP
#define EXPORT PUBLIC
#elif defined ( __CSMC__ ) /* Cosmic */
#define CODESECT switch .text
#define THUMB
#define EXPORT xdef
#define PROC :
#define LABEL :
#define ENDP
#define arm_bitreversal_32 _arm_bitreversal_32
#elif defined ( __TI_ARM__ ) /* TI ARM */
#define THUMB .thumb
#define CODESECT .text
#define EXPORT .global
#define PROC : .asmfunc
#define LABEL :
#define ENDP .endasmfunc
#define END
#elif defined ( __GNUC__ ) /* GCC */
#define THUMB .thumb
#define CODESECT .section .text
#define EXPORT .global
#define PROC :
#define LABEL :
#define ENDP
#define END
.syntax unified
#endif
CODESECT
THUMB
;/**
; @brief In-place bit reversal function.
; @param[in,out] pSrc points to the in-place buffer of unknown 32-bit data type
; @param[in] bitRevLen bit reversal table length
; @param[in] pBitRevTab points to bit reversal table
; @return none
; */
EXPORT arm_bitreversal_32
EXPORT arm_bitreversal_16
#if defined ( __CC_ARM ) /* Keil */
#elif defined (__ARMCC_VERSION) && (__ARMCC_VERSION >= 6010050)
#elif defined ( __IASMARM__ ) /* IAR */
#elif defined ( __CSMC__ ) /* Cosmic */
#elif defined ( __TI_ARM__ ) /* TI ARM */
#elif defined ( __GNUC__ ) /* GCC */
.type arm_bitreversal_16, %function
.type arm_bitreversal_32, %function
#endif
#if defined (ARM_MATH_CM0_FAMILY)
arm_bitreversal_32 PROC
ADDS r3,r1,#1
PUSH {r4-r6}
ADDS r1,r2,#0
LSRS r3,r3,#1
arm_bitreversal_32_0 LABEL
LDRH r2,[r1,#2]
LDRH r6,[r1,#0]
ADD r2,r0,r2
ADD r6,r0,r6
LDR r5,[r2,#0]
LDR r4,[r6,#0]
STR r5,[r6,#0]
STR r4,[r2,#0]
LDR r5,[r2,#4]
LDR r4,[r6,#4]
STR r5,[r6,#4]
STR r4,[r2,#4]
ADDS r1,r1,#4
SUBS r3,r3,#1
BNE arm_bitreversal_32_0
POP {r4-r6}
BX lr
ENDP
arm_bitreversal_16 PROC
ADDS r3,r1,#1
PUSH {r4-r6}
ADDS r1,r2,#0
LSRS r3,r3,#1
arm_bitreversal_16_0 LABEL
LDRH r2,[r1,#2]
LDRH r6,[r1,#0]
LSRS r2,r2,#1
LSRS r6,r6,#1
ADD r2,r0,r2
ADD r6,r0,r6
LDR r5,[r2,#0]
LDR r4,[r6,#0]
STR r5,[r6,#0]
STR r4,[r2,#0]
ADDS r1,r1,#4
SUBS r3,r3,#1
BNE arm_bitreversal_16_0
POP {r4-r6}
BX lr
ENDP
#else
arm_bitreversal_32 PROC
ADDS r3,r1,#1
CMP r3,#1
IT LS
BXLS lr
PUSH {r4-r9}
ADDS r1,r2,#2
LSRS r3,r3,#2
arm_bitreversal_32_0 LABEL ;/* loop unrolled by 2 */
LDRH r8,[r1,#4]
LDRH r9,[r1,#2]
LDRH r2,[r1,#0]
LDRH r12,[r1,#-2]
ADD r8,r0,r8
ADD r9,r0,r9
ADD r2,r0,r2
ADD r12,r0,r12
LDR r7,[r9,#0]
LDR r6,[r8,#0]
LDR r5,[r2,#0]
LDR r4,[r12,#0]
STR r6,[r9,#0]
STR r7,[r8,#0]
STR r5,[r12,#0]
STR r4,[r2,#0]
LDR r7,[r9,#4]
LDR r6,[r8,#4]
LDR r5,[r2,#4]
LDR r4,[r12,#4]
STR r6,[r9,#4]
STR r7,[r8,#4]
STR r5,[r12,#4]
STR r4,[r2,#4]
ADDS r1,r1,#8
SUBS r3,r3,#1
BNE arm_bitreversal_32_0
POP {r4-r9}
BX lr
ENDP
arm_bitreversal_16 PROC
ADDS r3,r1,#1
CMP r3,#1
IT LS
BXLS lr
PUSH {r4-r9}
ADDS r1,r2,#2
LSRS r3,r3,#2
arm_bitreversal_16_0 LABEL ;/* loop unrolled by 2 */
LDRH r8,[r1,#4]
LDRH r9,[r1,#2]
LDRH r2,[r1,#0]
LDRH r12,[r1,#-2]
ADD r8,r0,r8,LSR #1
ADD r9,r0,r9,LSR #1
ADD r2,r0,r2,LSR #1
ADD r12,r0,r12,LSR #1
LDR r7,[r9,#0]
LDR r6,[r8,#0]
LDR r5,[r2,#0]
LDR r4,[r12,#0]
STR r6,[r9,#0]
STR r7,[r8,#0]
STR r5,[r12,#0]
STR r4,[r2,#0]
ADDS r1,r1,#8
SUBS r3,r3,#1
BNE arm_bitreversal_16_0
POP {r4-r9}
BX lr
ENDP
#endif
END

View File

@ -0,0 +1,629 @@
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_cfft_f32.c
* Description: Combined Radix Decimation in Frequency CFFT Floating point processing function
*
* $Date: 18. March 2019
* $Revision: V1.6.0
*
* Target Processor: Cortex-M cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2019 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* Licensed under the Apache License, Version 2.0 (the License); you may
* not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an AS IS BASIS, WITHOUT
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "arm_math.h"
#include "arm_common_tables.h"
extern void arm_radix8_butterfly_f32(
float32_t * pSrc,
uint16_t fftLen,
const float32_t * pCoef,
uint16_t twidCoefModifier);
extern void arm_bitreversal_32(
uint32_t * pSrc,
const uint16_t bitRevLen,
const uint16_t * pBitRevTable);
/**
@ingroup groupTransforms
*/
/**
@defgroup ComplexFFT Complex FFT Functions
@par
The Fast Fourier Transform (FFT) is an efficient algorithm for computing the
Discrete Fourier Transform (DFT). The FFT can be orders of magnitude faster
than the DFT, especially for long lengths.
The algorithms described in this section
operate on complex data. A separate set of functions is devoted to handling
of real sequences.
@par
There are separate algorithms for handling floating-point, Q15, and Q31 data
types. The algorithms available for each data type are described next.
@par
The FFT functions operate in-place. That is, the array holding the input data
will also be used to hold the corresponding result. The input data is complex
and contains <code>2*fftLen</code> interleaved values as shown below.
<pre>{real[0], imag[0], real[1], imag[1], ...} </pre>
The FFT result will be contained in the same array and the frequency domain
values will have the same interleaving.
@par Floating-point
The floating-point complex FFT uses a mixed-radix algorithm. Multiple radix-8
stages are performed along with a single radix-2 or radix-4 stage, as needed.
The algorithm supports lengths of [16, 32, 64, ..., 4096] and each length uses
a different twiddle factor table.
@par
The function uses the standard FFT definition and output values may grow by a
factor of <code>fftLen</code> when computing the forward transform. The
inverse transform includes a scale of <code>1/fftLen</code> as part of the
calculation and this matches the textbook definition of the inverse FFT.
@par
Pre-initialized data structures containing twiddle factors and bit reversal
tables are provided and defined in <code>arm_const_structs.h</code>. Include
this header in your function and then pass one of the constant structures as
an argument to arm_cfft_f32. For example:
@par
<code>arm_cfft_f32(arm_cfft_sR_f32_len64, pSrc, 1, 1)</code>
@par
computes a 64-point inverse complex FFT including bit reversal.
The data structures are treated as constant data and not modified during the
calculation. The same data structure can be reused for multiple transforms
including mixing forward and inverse transforms.
@par
Earlier releases of the library provided separate radix-2 and radix-4
algorithms that operated on floating-point data. These functions are still
provided but are deprecated. The older functions are slower and less general
than the new functions.
@par
An example of initialization of the constants for the arm_cfft_f32 function follows:
@code
const static arm_cfft_instance_f32 *S;
...
switch (length) {
case 16:
S = &arm_cfft_sR_f32_len16;
break;
case 32:
S = &arm_cfft_sR_f32_len32;
break;
case 64:
S = &arm_cfft_sR_f32_len64;
break;
case 128:
S = &arm_cfft_sR_f32_len128;
break;
case 256:
S = &arm_cfft_sR_f32_len256;
break;
case 512:
S = &arm_cfft_sR_f32_len512;
break;
case 1024:
S = &arm_cfft_sR_f32_len1024;
break;
case 2048:
S = &arm_cfft_sR_f32_len2048;
break;
case 4096:
S = &arm_cfft_sR_f32_len4096;
break;
}
@endcode
@par Q15 and Q31
The floating-point complex FFT uses a mixed-radix algorithm. Multiple radix-4
stages are performed along with a single radix-2 stage, as needed.
The algorithm supports lengths of [16, 32, 64, ..., 4096] and each length uses
a different twiddle factor table.
@par
The function uses the standard FFT definition and output values may grow by a
factor of <code>fftLen</code> when computing the forward transform. The
inverse transform includes a scale of <code>1/fftLen</code> as part of the
calculation and this matches the textbook definition of the inverse FFT.
@par
Pre-initialized data structures containing twiddle factors and bit reversal
tables are provided and defined in <code>arm_const_structs.h</code>. Include
this header in your function and then pass one of the constant structures as
an argument to arm_cfft_q31. For example:
@par
<code>arm_cfft_q31(arm_cfft_sR_q31_len64, pSrc, 1, 1)</code>
@par
computes a 64-point inverse complex FFT including bit reversal.
The data structures are treated as constant data and not modified during the
calculation. The same data structure can be reused for multiple transforms
including mixing forward and inverse transforms.
@par
Earlier releases of the library provided separate radix-2 and radix-4
algorithms that operated on floating-point data. These functions are still
provided but are deprecated. The older functions are slower and less general
than the new functions.
@par
An example of initialization of the constants for the arm_cfft_q31 function follows:
@code
const static arm_cfft_instance_q31 *S;
...
switch (length) {
case 16:
S = &arm_cfft_sR_q31_len16;
break;
case 32:
S = &arm_cfft_sR_q31_len32;
break;
case 64:
S = &arm_cfft_sR_q31_len64;
break;
case 128:
S = &arm_cfft_sR_q31_len128;
break;
case 256:
S = &arm_cfft_sR_q31_len256;
break;
case 512:
S = &arm_cfft_sR_q31_len512;
break;
case 1024:
S = &arm_cfft_sR_q31_len1024;
break;
case 2048:
S = &arm_cfft_sR_q31_len2048;
break;
case 4096:
S = &arm_cfft_sR_q31_len4096;
break;
}
@endcode
*/
void arm_cfft_radix8by2_f32 (arm_cfft_instance_f32 * S, float32_t * p1)
{
uint32_t L = S->fftLen;
float32_t * pCol1, * pCol2, * pMid1, * pMid2;
float32_t * p2 = p1 + L;
const float32_t * tw = (float32_t *) S->pTwiddle;
float32_t t1[4], t2[4], t3[4], t4[4], twR, twI;
float32_t m0, m1, m2, m3;
uint32_t l;
pCol1 = p1;
pCol2 = p2;
/* Define new length */
L >>= 1;
/* Initialize mid pointers */
pMid1 = p1 + L;
pMid2 = p2 + L;
/* do two dot Fourier transform */
for (l = L >> 2; l > 0; l-- )
{
t1[0] = p1[0];
t1[1] = p1[1];
t1[2] = p1[2];
t1[3] = p1[3];
t2[0] = p2[0];
t2[1] = p2[1];
t2[2] = p2[2];
t2[3] = p2[3];
t3[0] = pMid1[0];
t3[1] = pMid1[1];
t3[2] = pMid1[2];
t3[3] = pMid1[3];
t4[0] = pMid2[0];
t4[1] = pMid2[1];
t4[2] = pMid2[2];
t4[3] = pMid2[3];
*p1++ = t1[0] + t2[0];
*p1++ = t1[1] + t2[1];
*p1++ = t1[2] + t2[2];
*p1++ = t1[3] + t2[3]; /* col 1 */
t2[0] = t1[0] - t2[0];
t2[1] = t1[1] - t2[1];
t2[2] = t1[2] - t2[2];
t2[3] = t1[3] - t2[3]; /* for col 2 */
*pMid1++ = t3[0] + t4[0];
*pMid1++ = t3[1] + t4[1];
*pMid1++ = t3[2] + t4[2];
*pMid1++ = t3[3] + t4[3]; /* col 1 */
t4[0] = t4[0] - t3[0];
t4[1] = t4[1] - t3[1];
t4[2] = t4[2] - t3[2];
t4[3] = t4[3] - t3[3]; /* for col 2 */
twR = *tw++;
twI = *tw++;
/* multiply by twiddle factors */
m0 = t2[0] * twR;
m1 = t2[1] * twI;
m2 = t2[1] * twR;
m3 = t2[0] * twI;
/* R = R * Tr - I * Ti */
*p2++ = m0 + m1;
/* I = I * Tr + R * Ti */
*p2++ = m2 - m3;
/* use vertical symmetry */
/* 0.9988 - 0.0491i <==> -0.0491 - 0.9988i */
m0 = t4[0] * twI;
m1 = t4[1] * twR;
m2 = t4[1] * twI;
m3 = t4[0] * twR;
*pMid2++ = m0 - m1;
*pMid2++ = m2 + m3;
twR = *tw++;
twI = *tw++;
m0 = t2[2] * twR;
m1 = t2[3] * twI;
m2 = t2[3] * twR;
m3 = t2[2] * twI;
*p2++ = m0 + m1;
*p2++ = m2 - m3;
m0 = t4[2] * twI;
m1 = t4[3] * twR;
m2 = t4[3] * twI;
m3 = t4[2] * twR;
*pMid2++ = m0 - m1;
*pMid2++ = m2 + m3;
}
/* first col */
arm_radix8_butterfly_f32 (pCol1, L, (float32_t *) S->pTwiddle, 2U);
/* second col */
arm_radix8_butterfly_f32 (pCol2, L, (float32_t *) S->pTwiddle, 2U);
}
void arm_cfft_radix8by4_f32 (arm_cfft_instance_f32 * S, float32_t * p1)
{
uint32_t L = S->fftLen >> 1;
float32_t * pCol1, *pCol2, *pCol3, *pCol4, *pEnd1, *pEnd2, *pEnd3, *pEnd4;
const float32_t *tw2, *tw3, *tw4;
float32_t * p2 = p1 + L;
float32_t * p3 = p2 + L;
float32_t * p4 = p3 + L;
float32_t t2[4], t3[4], t4[4], twR, twI;
float32_t p1ap3_0, p1sp3_0, p1ap3_1, p1sp3_1;
float32_t m0, m1, m2, m3;
uint32_t l, twMod2, twMod3, twMod4;
pCol1 = p1; /* points to real values by default */
pCol2 = p2;
pCol3 = p3;
pCol4 = p4;
pEnd1 = p2 - 1; /* points to imaginary values by default */
pEnd2 = p3 - 1;
pEnd3 = p4 - 1;
pEnd4 = pEnd3 + L;
tw2 = tw3 = tw4 = (float32_t *) S->pTwiddle;
L >>= 1;
/* do four dot Fourier transform */
twMod2 = 2;
twMod3 = 4;
twMod4 = 6;
/* TOP */
p1ap3_0 = p1[0] + p3[0];
p1sp3_0 = p1[0] - p3[0];
p1ap3_1 = p1[1] + p3[1];
p1sp3_1 = p1[1] - p3[1];
/* col 2 */
t2[0] = p1sp3_0 + p2[1] - p4[1];
t2[1] = p1sp3_1 - p2[0] + p4[0];
/* col 3 */
t3[0] = p1ap3_0 - p2[0] - p4[0];
t3[1] = p1ap3_1 - p2[1] - p4[1];
/* col 4 */
t4[0] = p1sp3_0 - p2[1] + p4[1];
t4[1] = p1sp3_1 + p2[0] - p4[0];
/* col 1 */
*p1++ = p1ap3_0 + p2[0] + p4[0];
*p1++ = p1ap3_1 + p2[1] + p4[1];
/* Twiddle factors are ones */
*p2++ = t2[0];
*p2++ = t2[1];
*p3++ = t3[0];
*p3++ = t3[1];
*p4++ = t4[0];
*p4++ = t4[1];
tw2 += twMod2;
tw3 += twMod3;
tw4 += twMod4;
for (l = (L - 2) >> 1; l > 0; l-- )
{
/* TOP */
p1ap3_0 = p1[0] + p3[0];
p1sp3_0 = p1[0] - p3[0];
p1ap3_1 = p1[1] + p3[1];
p1sp3_1 = p1[1] - p3[1];
/* col 2 */
t2[0] = p1sp3_0 + p2[1] - p4[1];
t2[1] = p1sp3_1 - p2[0] + p4[0];
/* col 3 */
t3[0] = p1ap3_0 - p2[0] - p4[0];
t3[1] = p1ap3_1 - p2[1] - p4[1];
/* col 4 */
t4[0] = p1sp3_0 - p2[1] + p4[1];
t4[1] = p1sp3_1 + p2[0] - p4[0];
/* col 1 - top */
*p1++ = p1ap3_0 + p2[0] + p4[0];
*p1++ = p1ap3_1 + p2[1] + p4[1];
/* BOTTOM */
p1ap3_1 = pEnd1[-1] + pEnd3[-1];
p1sp3_1 = pEnd1[-1] - pEnd3[-1];
p1ap3_0 = pEnd1[ 0] + pEnd3[0];
p1sp3_0 = pEnd1[ 0] - pEnd3[0];
/* col 2 */
t2[2] = pEnd2[0] - pEnd4[0] + p1sp3_1;
t2[3] = pEnd1[0] - pEnd3[0] - pEnd2[-1] + pEnd4[-1];
/* col 3 */
t3[2] = p1ap3_1 - pEnd2[-1] - pEnd4[-1];
t3[3] = p1ap3_0 - pEnd2[ 0] - pEnd4[ 0];
/* col 4 */
t4[2] = pEnd2[ 0] - pEnd4[ 0] - p1sp3_1;
t4[3] = pEnd4[-1] - pEnd2[-1] - p1sp3_0;
/* col 1 - Bottom */
*pEnd1-- = p1ap3_0 + pEnd2[ 0] + pEnd4[ 0];
*pEnd1-- = p1ap3_1 + pEnd2[-1] + pEnd4[-1];
/* COL 2 */
/* read twiddle factors */
twR = *tw2++;
twI = *tw2++;
/* multiply by twiddle factors */
/* let Z1 = a + i(b), Z2 = c + i(d) */
/* => Z1 * Z2 = (a*c - b*d) + i(b*c + a*d) */
/* Top */
m0 = t2[0] * twR;
m1 = t2[1] * twI;
m2 = t2[1] * twR;
m3 = t2[0] * twI;
*p2++ = m0 + m1;
*p2++ = m2 - m3;
/* use vertical symmetry col 2 */
/* 0.9997 - 0.0245i <==> 0.0245 - 0.9997i */
/* Bottom */
m0 = t2[3] * twI;
m1 = t2[2] * twR;
m2 = t2[2] * twI;
m3 = t2[3] * twR;
*pEnd2-- = m0 - m1;
*pEnd2-- = m2 + m3;
/* COL 3 */
twR = tw3[0];
twI = tw3[1];
tw3 += twMod3;
/* Top */
m0 = t3[0] * twR;
m1 = t3[1] * twI;
m2 = t3[1] * twR;
m3 = t3[0] * twI;
*p3++ = m0 + m1;
*p3++ = m2 - m3;
/* use vertical symmetry col 3 */
/* 0.9988 - 0.0491i <==> -0.9988 - 0.0491i */
/* Bottom */
m0 = -t3[3] * twR;
m1 = t3[2] * twI;
m2 = t3[2] * twR;
m3 = t3[3] * twI;
*pEnd3-- = m0 - m1;
*pEnd3-- = m3 - m2;
/* COL 4 */
twR = tw4[0];
twI = tw4[1];
tw4 += twMod4;
/* Top */
m0 = t4[0] * twR;
m1 = t4[1] * twI;
m2 = t4[1] * twR;
m3 = t4[0] * twI;
*p4++ = m0 + m1;
*p4++ = m2 - m3;
/* use vertical symmetry col 4 */
/* 0.9973 - 0.0736i <==> -0.0736 + 0.9973i */
/* Bottom */
m0 = t4[3] * twI;
m1 = t4[2] * twR;
m2 = t4[2] * twI;
m3 = t4[3] * twR;
*pEnd4-- = m0 - m1;
*pEnd4-- = m2 + m3;
}
/* MIDDLE */
/* Twiddle factors are */
/* 1.0000 0.7071-0.7071i -1.0000i -0.7071-0.7071i */
p1ap3_0 = p1[0] + p3[0];
p1sp3_0 = p1[0] - p3[0];
p1ap3_1 = p1[1] + p3[1];
p1sp3_1 = p1[1] - p3[1];
/* col 2 */
t2[0] = p1sp3_0 + p2[1] - p4[1];
t2[1] = p1sp3_1 - p2[0] + p4[0];
/* col 3 */
t3[0] = p1ap3_0 - p2[0] - p4[0];
t3[1] = p1ap3_1 - p2[1] - p4[1];
/* col 4 */
t4[0] = p1sp3_0 - p2[1] + p4[1];
t4[1] = p1sp3_1 + p2[0] - p4[0];
/* col 1 - Top */
*p1++ = p1ap3_0 + p2[0] + p4[0];
*p1++ = p1ap3_1 + p2[1] + p4[1];
/* COL 2 */
twR = tw2[0];
twI = tw2[1];
m0 = t2[0] * twR;
m1 = t2[1] * twI;
m2 = t2[1] * twR;
m3 = t2[0] * twI;
*p2++ = m0 + m1;
*p2++ = m2 - m3;
/* COL 3 */
twR = tw3[0];
twI = tw3[1];
m0 = t3[0] * twR;
m1 = t3[1] * twI;
m2 = t3[1] * twR;
m3 = t3[0] * twI;
*p3++ = m0 + m1;
*p3++ = m2 - m3;
/* COL 4 */
twR = tw4[0];
twI = tw4[1];
m0 = t4[0] * twR;
m1 = t4[1] * twI;
m2 = t4[1] * twR;
m3 = t4[0] * twI;
*p4++ = m0 + m1;
*p4++ = m2 - m3;
/* first col */
arm_radix8_butterfly_f32 (pCol1, L, (float32_t *) S->pTwiddle, 4U);
/* second col */
arm_radix8_butterfly_f32 (pCol2, L, (float32_t *) S->pTwiddle, 4U);
/* third col */
arm_radix8_butterfly_f32 (pCol3, L, (float32_t *) S->pTwiddle, 4U);
/* fourth col */
arm_radix8_butterfly_f32 (pCol4, L, (float32_t *) S->pTwiddle, 4U);
}
/**
@addtogroup ComplexFFT
@{
*/
/**
@brief Processing function for the floating-point complex FFT.
@param[in] S points to an instance of the floating-point CFFT structure
@param[in,out] p1 points to the complex data buffer of size <code>2*fftLen</code>. Processing occurs in-place
@param[in] ifftFlag flag that selects transform direction
- value = 0: forward transform
- value = 1: inverse transform
@param[in] bitReverseFlag flag that enables / disables bit reversal of output
- value = 0: disables bit reversal of output
- value = 1: enables bit reversal of output
@return none
*/
void arm_cfft_f32(
const arm_cfft_instance_f32 * S,
float32_t * p1,
uint8_t ifftFlag,
uint8_t bitReverseFlag)
{
uint32_t L = S->fftLen, l;
float32_t invL, * pSrc;
if (ifftFlag == 1U)
{
/* Conjugate input data */
pSrc = p1 + 1;
for (l = 0; l < L; l++)
{
*pSrc = -*pSrc;
pSrc += 2;
}
}
switch (L)
{
case 16:
case 128:
case 1024:
arm_cfft_radix8by2_f32 ( (arm_cfft_instance_f32 *) S, p1);
break;
case 32:
case 256:
case 2048:
arm_cfft_radix8by4_f32 ( (arm_cfft_instance_f32 *) S, p1);
break;
case 64:
case 512:
case 4096:
arm_radix8_butterfly_f32 ( p1, L, (float32_t *) S->pTwiddle, 1);
break;
}
if ( bitReverseFlag )
arm_bitreversal_32 ((uint32_t*) p1, S->bitRevLength, S->pBitRevTable);
if (ifftFlag == 1U)
{
invL = 1.0f / (float32_t)L;
/* Conjugate and scale output data */
pSrc = p1;
for (l= 0; l < L; l++)
{
*pSrc++ *= invL ;
*pSrc = -(*pSrc) * invL;
pSrc++;
}
}
}
/**
@} end of ComplexFFT group
*/

View File

@ -0,0 +1,470 @@
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_cfft_radix2_f32.c
* Description: Radix-2 Decimation in Frequency CFFT & CIFFT Floating point processing function
*
* $Date: 18. March 2019
* $Revision: V1.6.0
*
* Target Processor: Cortex-M cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2019 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* Licensed under the Apache License, Version 2.0 (the License); you may
* not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an AS IS BASIS, WITHOUT
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "arm_math.h"
void arm_radix2_butterfly_f32(
float32_t * pSrc,
uint32_t fftLen,
const float32_t * pCoef,
uint16_t twidCoefModifier);
void arm_radix2_butterfly_inverse_f32(
float32_t * pSrc,
uint32_t fftLen,
const float32_t * pCoef,
uint16_t twidCoefModifier,
float32_t onebyfftLen);
extern void arm_bitreversal_f32(
float32_t * pSrc,
uint16_t fftSize,
uint16_t bitRevFactor,
const uint16_t * pBitRevTab);
/**
@ingroup groupTransforms
*/
/**
@addtogroup ComplexFFT
@{
*/
/**
@brief Radix-2 CFFT/CIFFT.
@deprecated Do not use this function. It has been superseded 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)
{
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);
}
}
/**
@} end of ComplexFFT group
*/
/* ----------------------------------------------------------------------
** Internal helper function used by the FFTs
** ------------------------------------------------------------------- */
/**
brief Core function for the floating-point CFFT butterfly process.
param[in,out] pSrc points to in-place buffer of floating-point data type
param[in] fftLen length of the FFT
param[in] pCoef points to 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,
const float32_t * pCoef,
uint16_t 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;
#if defined (ARM_MATH_DSP)
/* Initializations for the first stage */
n2 = fftLen >> 1;
ia = 0;
i = 0;
// loop for groups
for (k = n2; k > 0; k--)
{
cosVal = pCoef[ia * 2];
sinVal = pCoef[(ia * 2) + 1];
/* Twiddle coefficients index modifier */
ia += twidCoefModifier;
/* 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];
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
twidCoefModifier <<= 1U;
// loop for stage
for (k = n2; k > 2; k = k >> 1)
{
n1 = n2;
n2 = n2 >> 1;
ia = 0;
// loop for groups
j = 0;
do
{
cosVal = pCoef[ia * 2];
sinVal = pCoef[(ia * 2) + 1];
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];
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
// 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];
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 /* #if defined (ARM_MATH_DSP) */
n2 = fftLen;
// loop for stage
for (k = fftLen; k > 1; k = k >> 1)
{
n1 = n2;
n2 = n2 >> 1;
ia = 0;
// loop for groups
j = 0;
do
{
cosVal = pCoef[ia * 2];
sinVal = pCoef[(ia * 2) + 1];
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];
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;
}
#endif /* #if defined (ARM_MATH_DSP) */
}
void arm_radix2_butterfly_inverse_f32(
float32_t * pSrc,
uint32_t fftLen,
const float32_t * pCoef,
uint16_t twidCoefModifier,
float32_t onebyfftLen)
{
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;
#if defined (ARM_MATH_DSP)
n2 = fftLen >> 1;
ia = 0;
// loop for groups
for (i = 0; i < n2; i++)
{
cosVal = pCoef[ia * 2];
sinVal = pCoef[(ia * 2) + 1];
ia += twidCoefModifier;
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];
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
twidCoefModifier <<= 1U;
// loop for stage
for (k = fftLen / 2; k > 2; k = k >> 1)
{
n1 = n2;
n2 = n2 >> 1;
ia = 0;
// loop for groups
j = 0;
do
{
cosVal = pCoef[ia * 2];
sinVal = pCoef[(ia * 2) + 1];
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];
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
// 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 /* #if defined (ARM_MATH_DSP) */
n2 = fftLen;
// loop for stage
for (k = fftLen; k > 2; k = k >> 1)
{
n1 = n2;
n2 = n2 >> 1;
ia = 0;
// loop for groups
j = 0;
do
{
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];
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 = twidCoefModifier << 1U;
} // stages loop end
n1 = n2;
n2 = n2 >> 1;
// 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[2 * l] = p2;
pSrc[2 * i + 1] = p1;
pSrc[2 * l + 1] = p3;
} // butterfly loop end
#endif /* #if defined (ARM_MATH_DSP) */
}

File diff suppressed because it is too large Load Diff

View File

@ -0,0 +1,285 @@
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_cfft_radix8_f32.c
* Description: Radix-8 Decimation in Frequency CFFT & CIFFT Floating point processing function
*
* $Date: 18. March 2019
* $Revision: V1.6.0
*
* Target Processor: Cortex-M cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2019 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* Licensed under the Apache License, Version 2.0 (the License); you may
* not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an AS IS BASIS, WITHOUT
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "arm_math.h"
/* ----------------------------------------------------------------------
* Internal helper function used by the FFTs
* -------------------------------------------------------------------- */
/**
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_radix8_butterfly_f32(
float32_t * pSrc,
uint16_t fftLen,
const float32_t * pCoef,
uint16_t twidCoefModifier)
{
uint32_t ia1, ia2, ia3, ia4, ia5, ia6, ia7;
uint32_t i1, i2, i3, i4, i5, i6, i7, i8;
uint32_t id;
uint32_t n1, n2, j;
float32_t r1, r2, r3, r4, r5, r6, r7, r8;
float32_t t1, t2;
float32_t s1, s2, s3, s4, s5, s6, s7, s8;
float32_t p1, p2, p3, p4;
float32_t co2, co3, co4, co5, co6, co7, co8;
float32_t si2, si3, si4, si5, si6, si7, si8;
const float32_t C81 = 0.70710678118f;
n2 = fftLen;
do
{
n1 = n2;
n2 = n2 >> 3;
i1 = 0;
do
{
i2 = i1 + n2;
i3 = i2 + n2;
i4 = i3 + n2;
i5 = i4 + n2;
i6 = i5 + n2;
i7 = i6 + n2;
i8 = i7 + n2;
r1 = pSrc[2 * i1] + pSrc[2 * i5];
r5 = pSrc[2 * i1] - pSrc[2 * i5];
r2 = pSrc[2 * i2] + pSrc[2 * i6];
r6 = pSrc[2 * i2] - pSrc[2 * i6];
r3 = pSrc[2 * i3] + pSrc[2 * i7];
r7 = pSrc[2 * i3] - pSrc[2 * i7];
r4 = pSrc[2 * i4] + pSrc[2 * i8];
r8 = pSrc[2 * i4] - pSrc[2 * i8];
t1 = r1 - r3;
r1 = r1 + r3;
r3 = r2 - r4;
r2 = r2 + r4;
pSrc[2 * i1] = r1 + r2;
pSrc[2 * i5] = r1 - r2;
r1 = pSrc[2 * i1 + 1] + pSrc[2 * i5 + 1];
s5 = pSrc[2 * i1 + 1] - pSrc[2 * i5 + 1];
r2 = pSrc[2 * i2 + 1] + pSrc[2 * i6 + 1];
s6 = pSrc[2 * i2 + 1] - pSrc[2 * i6 + 1];
s3 = pSrc[2 * i3 + 1] + pSrc[2 * i7 + 1];
s7 = pSrc[2 * i3 + 1] - pSrc[2 * i7 + 1];
r4 = pSrc[2 * i4 + 1] + pSrc[2 * i8 + 1];
s8 = pSrc[2 * i4 + 1] - pSrc[2 * i8 + 1];
t2 = r1 - s3;
r1 = r1 + s3;
s3 = r2 - r4;
r2 = r2 + r4;
pSrc[2 * i1 + 1] = r1 + r2;
pSrc[2 * i5 + 1] = r1 - r2;
pSrc[2 * i3] = t1 + s3;
pSrc[2 * i7] = t1 - s3;
pSrc[2 * i3 + 1] = t2 - r3;
pSrc[2 * i7 + 1] = t2 + r3;
r1 = (r6 - r8) * C81;
r6 = (r6 + r8) * C81;
r2 = (s6 - s8) * C81;
s6 = (s6 + s8) * C81;
t1 = r5 - r1;
r5 = r5 + r1;
r8 = r7 - r6;
r7 = r7 + r6;
t2 = s5 - r2;
s5 = s5 + r2;
s8 = s7 - s6;
s7 = s7 + s6;
pSrc[2 * i2] = r5 + s7;
pSrc[2 * i8] = r5 - s7;
pSrc[2 * i6] = t1 + s8;
pSrc[2 * i4] = t1 - s8;
pSrc[2 * i2 + 1] = s5 - r7;
pSrc[2 * i8 + 1] = s5 + r7;
pSrc[2 * i6 + 1] = t2 - r8;
pSrc[2 * i4 + 1] = t2 + r8;
i1 += n1;
} while (i1 < fftLen);
if (n2 < 8)
break;
ia1 = 0;
j = 1;
do
{
/* index calculation for the coefficients */
id = ia1 + twidCoefModifier;
ia1 = id;
ia2 = ia1 + id;
ia3 = ia2 + id;
ia4 = ia3 + id;
ia5 = ia4 + id;
ia6 = ia5 + id;
ia7 = ia6 + id;
co2 = pCoef[2 * ia1];
co3 = pCoef[2 * ia2];
co4 = pCoef[2 * ia3];
co5 = pCoef[2 * ia4];
co6 = pCoef[2 * ia5];
co7 = pCoef[2 * ia6];
co8 = pCoef[2 * ia7];
si2 = pCoef[2 * ia1 + 1];
si3 = pCoef[2 * ia2 + 1];
si4 = pCoef[2 * ia3 + 1];
si5 = pCoef[2 * ia4 + 1];
si6 = pCoef[2 * ia5 + 1];
si7 = pCoef[2 * ia6 + 1];
si8 = pCoef[2 * ia7 + 1];
i1 = j;
do
{
/* index calculation for the input */
i2 = i1 + n2;
i3 = i2 + n2;
i4 = i3 + n2;
i5 = i4 + n2;
i6 = i5 + n2;
i7 = i6 + n2;
i8 = i7 + n2;
r1 = pSrc[2 * i1] + pSrc[2 * i5];
r5 = pSrc[2 * i1] - pSrc[2 * i5];
r2 = pSrc[2 * i2] + pSrc[2 * i6];
r6 = pSrc[2 * i2] - pSrc[2 * i6];
r3 = pSrc[2 * i3] + pSrc[2 * i7];
r7 = pSrc[2 * i3] - pSrc[2 * i7];
r4 = pSrc[2 * i4] + pSrc[2 * i8];
r8 = pSrc[2 * i4] - pSrc[2 * i8];
t1 = r1 - r3;
r1 = r1 + r3;
r3 = r2 - r4;
r2 = r2 + r4;
pSrc[2 * i1] = r1 + r2;
r2 = r1 - r2;
s1 = pSrc[2 * i1 + 1] + pSrc[2 * i5 + 1];
s5 = pSrc[2 * i1 + 1] - pSrc[2 * i5 + 1];
s2 = pSrc[2 * i2 + 1] + pSrc[2 * i6 + 1];
s6 = pSrc[2 * i2 + 1] - pSrc[2 * i6 + 1];
s3 = pSrc[2 * i3 + 1] + pSrc[2 * i7 + 1];
s7 = pSrc[2 * i3 + 1] - pSrc[2 * i7 + 1];
s4 = pSrc[2 * i4 + 1] + pSrc[2 * i8 + 1];
s8 = pSrc[2 * i4 + 1] - pSrc[2 * i8 + 1];
t2 = s1 - s3;
s1 = s1 + s3;
s3 = s2 - s4;
s2 = s2 + s4;
r1 = t1 + s3;
t1 = t1 - s3;
pSrc[2 * i1 + 1] = s1 + s2;
s2 = s1 - s2;
s1 = t2 - r3;
t2 = t2 + r3;
p1 = co5 * r2;
p2 = si5 * s2;
p3 = co5 * s2;
p4 = si5 * r2;
pSrc[2 * i5] = p1 + p2;
pSrc[2 * i5 + 1] = p3 - p4;
p1 = co3 * r1;
p2 = si3 * s1;
p3 = co3 * s1;
p4 = si3 * r1;
pSrc[2 * i3] = p1 + p2;
pSrc[2 * i3 + 1] = p3 - p4;
p1 = co7 * t1;
p2 = si7 * t2;
p3 = co7 * t2;
p4 = si7 * t1;
pSrc[2 * i7] = p1 + p2;
pSrc[2 * i7 + 1] = p3 - p4;
r1 = (r6 - r8) * C81;
r6 = (r6 + r8) * C81;
s1 = (s6 - s8) * C81;
s6 = (s6 + s8) * C81;
t1 = r5 - r1;
r5 = r5 + r1;
r8 = r7 - r6;
r7 = r7 + r6;
t2 = s5 - s1;
s5 = s5 + s1;
s8 = s7 - s6;
s7 = s7 + s6;
r1 = r5 + s7;
r5 = r5 - s7;
r6 = t1 + s8;
t1 = t1 - s8;
s1 = s5 - r7;
s5 = s5 + r7;
s6 = t2 - r8;
t2 = t2 + r8;
p1 = co2 * r1;
p2 = si2 * s1;
p3 = co2 * s1;
p4 = si2 * r1;
pSrc[2 * i2] = p1 + p2;
pSrc[2 * i2 + 1] = p3 - p4;
p1 = co8 * r5;
p2 = si8 * s5;
p3 = co8 * s5;
p4 = si8 * r5;
pSrc[2 * i8] = p1 + p2;
pSrc[2 * i8 + 1] = p3 - p4;
p1 = co6 * r6;
p2 = si6 * s6;
p3 = co6 * s6;
p4 = si6 * r6;
pSrc[2 * i6] = p1 + p2;
pSrc[2 * i6 + 1] = p3 - p4;
p1 = co4 * t1;
p2 = si4 * t2;
p3 = co4 * t2;
p4 = si4 * t1;
pSrc[2 * i4] = p1 + p2;
pSrc[2 * i4 + 1] = p3 - p4;
i1 += n1;
} while (i1 < fftLen);
j++;
} while (j < n2);
twidCoefModifier <<= 3;
} while (n2 > 7);
}

File diff suppressed because it is too large Load Diff

View File

@ -0,0 +1,381 @@
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_const_structs.c
* Description: Constant structs that are initialized for user convenience.
* For example, some can be given as arguments to the arm_cfft_f32() or arm_rfft_f32() functions.
*
* $Date: 27. January 2017
* $Revision: V.1.5.1
*
* Target Processor: Cortex-M cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2017 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* Licensed under the Apache License, Version 2.0 (the License); you may
* not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an AS IS BASIS, WITHOUT
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "arm_const_structs.h"
/* Floating-point structs */
const arm_cfft_instance_f32 arm_cfft_sR_f32_len16 = {
16, twiddleCoef_16, armBitRevIndexTable16, ARMBITREVINDEXTABLE_16_TABLE_LENGTH
};
const arm_cfft_instance_f32 arm_cfft_sR_f32_len32 = {
32, twiddleCoef_32, armBitRevIndexTable32, ARMBITREVINDEXTABLE_32_TABLE_LENGTH
};
const arm_cfft_instance_f32 arm_cfft_sR_f32_len64 = {
64, twiddleCoef_64, armBitRevIndexTable64, ARMBITREVINDEXTABLE_64_TABLE_LENGTH
};
const arm_cfft_instance_f32 arm_cfft_sR_f32_len128 = {
128, twiddleCoef_128, armBitRevIndexTable128, ARMBITREVINDEXTABLE_128_TABLE_LENGTH
};
const arm_cfft_instance_f32 arm_cfft_sR_f32_len256 = {
256, twiddleCoef_256, armBitRevIndexTable256, ARMBITREVINDEXTABLE_256_TABLE_LENGTH
};
const arm_cfft_instance_f32 arm_cfft_sR_f32_len512 = {
512, twiddleCoef_512, armBitRevIndexTable512, ARMBITREVINDEXTABLE_512_TABLE_LENGTH
};
const arm_cfft_instance_f32 arm_cfft_sR_f32_len1024 = {
1024, twiddleCoef_1024, armBitRevIndexTable1024, ARMBITREVINDEXTABLE_1024_TABLE_LENGTH
};
const arm_cfft_instance_f32 arm_cfft_sR_f32_len2048 = {
2048, twiddleCoef_2048, armBitRevIndexTable2048, ARMBITREVINDEXTABLE_2048_TABLE_LENGTH
};
const arm_cfft_instance_f32 arm_cfft_sR_f32_len4096 = {
4096, twiddleCoef_4096, armBitRevIndexTable4096, ARMBITREVINDEXTABLE_4096_TABLE_LENGTH
};
/* Fixed-point structs */
const arm_cfft_instance_q31 arm_cfft_sR_q31_len16 = {
16, twiddleCoef_16_q31, armBitRevIndexTable_fixed_16, ARMBITREVINDEXTABLE_FIXED_16_TABLE_LENGTH
};
const arm_cfft_instance_q31 arm_cfft_sR_q31_len32 = {
32, twiddleCoef_32_q31, armBitRevIndexTable_fixed_32, ARMBITREVINDEXTABLE_FIXED_32_TABLE_LENGTH
};
const arm_cfft_instance_q31 arm_cfft_sR_q31_len64 = {
64, twiddleCoef_64_q31, armBitRevIndexTable_fixed_64, ARMBITREVINDEXTABLE_FIXED_64_TABLE_LENGTH
};
const arm_cfft_instance_q31 arm_cfft_sR_q31_len128 = {
128, twiddleCoef_128_q31, armBitRevIndexTable_fixed_128, ARMBITREVINDEXTABLE_FIXED_128_TABLE_LENGTH
};
const arm_cfft_instance_q31 arm_cfft_sR_q31_len256 = {
256, twiddleCoef_256_q31, armBitRevIndexTable_fixed_256, ARMBITREVINDEXTABLE_FIXED_256_TABLE_LENGTH
};
const arm_cfft_instance_q31 arm_cfft_sR_q31_len512 = {
512, twiddleCoef_512_q31, armBitRevIndexTable_fixed_512, ARMBITREVINDEXTABLE_FIXED_512_TABLE_LENGTH
};
const arm_cfft_instance_q31 arm_cfft_sR_q31_len1024 = {
1024, twiddleCoef_1024_q31, armBitRevIndexTable_fixed_1024, ARMBITREVINDEXTABLE_FIXED_1024_TABLE_LENGTH
};
const arm_cfft_instance_q31 arm_cfft_sR_q31_len2048 = {
2048, twiddleCoef_2048_q31, armBitRevIndexTable_fixed_2048, ARMBITREVINDEXTABLE_FIXED_2048_TABLE_LENGTH
};
const arm_cfft_instance_q31 arm_cfft_sR_q31_len4096 = {
4096, twiddleCoef_4096_q31, armBitRevIndexTable_fixed_4096, ARMBITREVINDEXTABLE_FIXED_4096_TABLE_LENGTH
};
const arm_cfft_instance_q15 arm_cfft_sR_q15_len16 = {
16, twiddleCoef_16_q15, armBitRevIndexTable_fixed_16, ARMBITREVINDEXTABLE_FIXED_16_TABLE_LENGTH
};
const arm_cfft_instance_q15 arm_cfft_sR_q15_len32 = {
32, twiddleCoef_32_q15, armBitRevIndexTable_fixed_32, ARMBITREVINDEXTABLE_FIXED_32_TABLE_LENGTH
};
const arm_cfft_instance_q15 arm_cfft_sR_q15_len64 = {
64, twiddleCoef_64_q15, armBitRevIndexTable_fixed_64, ARMBITREVINDEXTABLE_FIXED_64_TABLE_LENGTH
};
const arm_cfft_instance_q15 arm_cfft_sR_q15_len128 = {
128, twiddleCoef_128_q15, armBitRevIndexTable_fixed_128, ARMBITREVINDEXTABLE_FIXED_128_TABLE_LENGTH
};
const arm_cfft_instance_q15 arm_cfft_sR_q15_len256 = {
256, twiddleCoef_256_q15, armBitRevIndexTable_fixed_256, ARMBITREVINDEXTABLE_FIXED_256_TABLE_LENGTH
};
const arm_cfft_instance_q15 arm_cfft_sR_q15_len512 = {
512, twiddleCoef_512_q15, armBitRevIndexTable_fixed_512, ARMBITREVINDEXTABLE_FIXED_512_TABLE_LENGTH
};
const arm_cfft_instance_q15 arm_cfft_sR_q15_len1024 = {
1024, twiddleCoef_1024_q15, armBitRevIndexTable_fixed_1024, ARMBITREVINDEXTABLE_FIXED_1024_TABLE_LENGTH
};
const arm_cfft_instance_q15 arm_cfft_sR_q15_len2048 = {
2048, twiddleCoef_2048_q15, armBitRevIndexTable_fixed_2048, ARMBITREVINDEXTABLE_FIXED_2048_TABLE_LENGTH
};
const arm_cfft_instance_q15 arm_cfft_sR_q15_len4096 = {
4096, twiddleCoef_4096_q15, armBitRevIndexTable_fixed_4096, ARMBITREVINDEXTABLE_FIXED_4096_TABLE_LENGTH
};
/* Structure for real-value inputs */
/* Floating-point structs */
const arm_rfft_fast_instance_f32 arm_rfft_fast_sR_f32_len32 = {
{ 16, twiddleCoef_32, armBitRevIndexTable32, ARMBITREVINDEXTABLE_16_TABLE_LENGTH },
32U,
(float32_t *)twiddleCoef_rfft_32
};
const arm_rfft_fast_instance_f32 arm_rfft_fast_sR_f32_len64 = {
{ 32, twiddleCoef_32, armBitRevIndexTable32, ARMBITREVINDEXTABLE_32_TABLE_LENGTH },
64U,
(float32_t *)twiddleCoef_rfft_64
};
const arm_rfft_fast_instance_f32 arm_rfft_fast_sR_f32_len128 = {
{ 64, twiddleCoef_64, armBitRevIndexTable64, ARMBITREVINDEXTABLE_64_TABLE_LENGTH },
128U,
(float32_t *)twiddleCoef_rfft_128
};
const arm_rfft_fast_instance_f32 arm_rfft_fast_sR_f32_len256 = {
{ 128, twiddleCoef_128, armBitRevIndexTable128, ARMBITREVINDEXTABLE_128_TABLE_LENGTH },
256U,
(float32_t *)twiddleCoef_rfft_256
};
const arm_rfft_fast_instance_f32 arm_rfft_fast_sR_f32_len512 = {
{ 256, twiddleCoef_256, armBitRevIndexTable256, ARMBITREVINDEXTABLE_256_TABLE_LENGTH },
512U,
(float32_t *)twiddleCoef_rfft_512
};
const arm_rfft_fast_instance_f32 arm_rfft_fast_sR_f32_len1024 = {
{ 512, twiddleCoef_512, armBitRevIndexTable512, ARMBITREVINDEXTABLE_512_TABLE_LENGTH },
1024U,
(float32_t *)twiddleCoef_rfft_1024
};
const arm_rfft_fast_instance_f32 arm_rfft_fast_sR_f32_len2048 = {
{ 1024, twiddleCoef_1024, armBitRevIndexTable1024, ARMBITREVINDEXTABLE_1024_TABLE_LENGTH },
2048U,
(float32_t *)twiddleCoef_rfft_2048
};
const arm_rfft_fast_instance_f32 arm_rfft_fast_sR_f32_len4096 = {
{ 2048, twiddleCoef_2048, armBitRevIndexTable2048, ARMBITREVINDEXTABLE_2048_TABLE_LENGTH },
4096U,
(float32_t *)twiddleCoef_rfft_4096
};
#if 0
/* Fixed-point structs */
/* q31_t */
extern const q31_t realCoefAQ31[8192];
extern const q31_t realCoefBQ31[8192];
const arm_rfft_instance_q31 arm_rfft_sR_q31_len32 = {
32U,
0,
1,
256U,
(q31_t*)realCoefAQ31,
(q31_t*)realCoefBQ31,
&arm_cfft_sR_q31_len16
};
const arm_rfft_instance_q31 arm_rfft_sR_q31_len64 = {
64U,
0,
1,
128U,
(q31_t*)realCoefAQ31,
(q31_t*)realCoefBQ31,
&arm_cfft_sR_q31_len32
};
const arm_rfft_instance_q31 arm_rfft_sR_q31_len128 = {
128U,
0,
1,
64U,
(q31_t*)realCoefAQ31,
(q31_t*)realCoefBQ31,
&arm_cfft_sR_q31_len64
};
const arm_rfft_instance_q31 arm_rfft_sR_q31_len256 = {
256U,
0,
1,
32U,
(q31_t*)realCoefAQ31,
(q31_t*)realCoefBQ31,
&arm_cfft_sR_q31_len128
};
const arm_rfft_instance_q31 arm_rfft_sR_q31_len512 = {
512U,
0,
1,
16U,
(q31_t*)realCoefAQ31,
(q31_t*)realCoefBQ31,
&arm_cfft_sR_q31_len256
};
const arm_rfft_instance_q31 arm_rfft_sR_q31_len1024 = {
1024U,
0,
1,
8U,
(q31_t*)realCoefAQ31,
(q31_t*)realCoefBQ31,
&arm_cfft_sR_q31_len512
};
const arm_rfft_instance_q31 arm_rfft_sR_q31_len2048 = {
2048U,
0,
1,
4U,
(q31_t*)realCoefAQ31,
(q31_t*)realCoefBQ31,
&arm_cfft_sR_q31_len1024
};
const arm_rfft_instance_q31 arm_rfft_sR_q31_len4096 = {
4096U,
0,
1,
2U,
(q31_t*)realCoefAQ31,
(q31_t*)realCoefBQ31,
&arm_cfft_sR_q31_len2048
};
const arm_rfft_instance_q31 arm_rfft_sR_q31_len8192 = {
8192U,
0,
1,
1U,
(q31_t*)realCoefAQ31,
(q31_t*)realCoefBQ31,
&arm_cfft_sR_q31_len4096
};
/* q15_t */
extern const q15_t realCoefAQ15[8192];
extern const q15_t realCoefBQ15[8192];
const arm_rfft_instance_q15 arm_rfft_sR_q15_len32 = {
32U,
0,
1,
256U,
(q15_t*)realCoefAQ15,
(q15_t*)realCoefBQ15,
&arm_cfft_sR_q15_len16
};
const arm_rfft_instance_q15 arm_rfft_sR_q15_len64 = {
64U,
0,
1,
128U,
(q15_t*)realCoefAQ15,
(q15_t*)realCoefBQ15,
&arm_cfft_sR_q15_len32
};
const arm_rfft_instance_q15 arm_rfft_sR_q15_len128 = {
128U,
0,
1,
64U,
(q15_t*)realCoefAQ15,
(q15_t*)realCoefBQ15,
&arm_cfft_sR_q15_len64
};
const arm_rfft_instance_q15 arm_rfft_sR_q15_len256 = {
256U,
0,
1,
32U,
(q15_t*)realCoefAQ15,
(q15_t*)realCoefBQ15,
&arm_cfft_sR_q15_len128
};
const arm_rfft_instance_q15 arm_rfft_sR_q15_len512 = {
512U,
0,
1,
16U,
(q15_t*)realCoefAQ15,
(q15_t*)realCoefBQ15,
&arm_cfft_sR_q15_len256
};
const arm_rfft_instance_q15 arm_rfft_sR_q15_len1024 = {
1024U,
0,
1,
8U,
(q15_t*)realCoefAQ15,
(q15_t*)realCoefBQ15,
&arm_cfft_sR_q15_len512
};
const arm_rfft_instance_q15 arm_rfft_sR_q15_len2048 = {
2048U,
0,
1,
4U,
(q15_t*)realCoefAQ15,
(q15_t*)realCoefBQ15,
&arm_cfft_sR_q15_len1024
};
const arm_rfft_instance_q15 arm_rfft_sR_q15_len4096 = {
4096U,
0,
1,
2U,
(q15_t*)realCoefAQ15,
(q15_t*)realCoefBQ15,
&arm_cfft_sR_q15_len2048
};
const arm_rfft_instance_q15 arm_rfft_sR_q15_len8192 = {
8192U,
0,
1,
1U,
(q15_t*)realCoefAQ15,
(q15_t*)realCoefBQ15,
&arm_cfft_sR_q15_len4096
};
#endif

View File

@ -0,0 +1,122 @@
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_cos_f32.c
* Description: Fast cosine calculation for floating-point values
*
* $Date: 18. March 2019
* $Revision: V1.6.0
*
* Target Processor: Cortex-M cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2019 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* Licensed under the Apache License, Version 2.0 (the License); you may
* not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an AS IS BASIS, WITHOUT
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "arm_math.h"
#include "arm_common_tables.h"
/**
@ingroup groupFastMath
*/
/**
@defgroup cos Cosine
Computes the trigonometric cosine function using a combination of table lookup
and linear interpolation. There are separate functions for
Q15, Q31, and floating-point data types.
The input to the floating-point version is in radians while the
fixed-point Q15 and Q31 have a scaled input with the range
[0 +0.9999] mapping to [0 2*pi). The fixed-point range is chosen so that a
value of 2*pi wraps around to 0.
The implementation is based on table lookup using 256 values together with linear interpolation.
The steps used are:
-# Calculation of the nearest integer table index
-# Compute the fractional portion (fract) of the table index.
-# The final result equals <code>(1.0f-fract)*a + fract*b;</code>
where
<pre>
b = Table[index];
c = Table[index+1];
</pre>
*/
/**
@addtogroup cos
@{
*/
/**
@brief Fast approximation to the trigonometric cosine function for floating-point data.
@param[in] x input value in radians
@return cos(x)
*/
float32_t arm_cos_f32(
float32_t x)
{
float32_t cosVal, fract, in; /* Temporary input, output variables */
uint16_t index; /* Index variable */
float32_t a, b; /* Two nearest output values */
int32_t n;
float32_t findex;
/* input x is in radians */
/* Scale input to [0 1] range from [0 2*PI] , divide input by 2*pi, add 0.25 (pi/2) to read sine table */
in = x * 0.159154943092f + 0.25f;
/* Calculation of floor value of input */
n = (int32_t) in;
/* Make negative values towards -infinity */
if (in < 0.0f)
{
n--;
}
/* Map input value to [0 1] */
in = in - (float32_t) n;
/* Calculation of index of the table */
findex = (float32_t)FAST_MATH_TABLE_SIZE * in;
index = (uint16_t)findex;
/* when "in" is exactly 1, we need to rotate the index down to 0 */
if (index >= FAST_MATH_TABLE_SIZE) {
index = 0;
findex -= (float32_t)FAST_MATH_TABLE_SIZE;
}
/* fractional value calculation */
fract = findex - (float32_t) index;
/* Read two nearest values of input value from the cos table */
a = sinTable_f32[index];
b = sinTable_f32[index+1];
/* Linear interpolation process */
cosVal = (1.0f - fract) * a + fract * b;
/* Return output value */
return (cosVal);
}
/**
@} end of cos group
*/

View File

@ -0,0 +1,146 @@
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_sin_cos_f32.c
* Description: Sine and Cosine calculation for floating-point values
*
* $Date: 18. March 2019
* $Revision: V1.6.0
*
* Target Processor: Cortex-M cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2019 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* Licensed under the Apache License, Version 2.0 (the License); you may
* not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an AS IS BASIS, WITHOUT
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "arm_math.h"
#include "arm_common_tables.h"
/**
@ingroup groupController
*/
/**
@defgroup SinCos Sine Cosine
Computes the trigonometric sine and cosine values using a combination of table lookup
and linear interpolation.
There are separate functions for Q31 and floating-point data types.
The input to the floating-point version is in degrees while the
fixed-point Q31 have a scaled input with the range
[-1 0.9999] mapping to [-180 +180] degrees.
The floating point function also allows values that are out of the usual range. When this happens, the function will
take extra time to adjust the input value to the range of [-180 180].
The result is accurate to 5 digits after the decimal point.
The implementation is based on table lookup using 360 values together with linear interpolation.
The steps used are:
-# Calculation of the nearest integer table index.
-# Compute the fractional portion (fract) of the input.
-# Fetch the value corresponding to \c index from sine table to \c y0 and also value from \c index+1 to \c y1.
-# Sine value is computed as <code> *psinVal = y0 + (fract * (y1 - y0))</code>.
-# Fetch the value corresponding to \c index from cosine table to \c y0 and also value from \c index+1 to \c y1.
-# Cosine value is computed as <code> *pcosVal = y0 + (fract * (y1 - y0))</code>.
*/
/**
@addtogroup SinCos
@{
*/
/**
@brief Floating-point sin_cos function.
@param[in] theta input value in degrees
@param[out] pSinVal points to processed sine output
@param[out] pCosVal points to processed cosine output
@return none
*/
void arm_sin_cos_f32(
float32_t theta,
float32_t * pSinVal,
float32_t * pCosVal)
{
float32_t fract, in; /* Temporary input, output variables */
uint16_t indexS, indexC; /* Index variable */
float32_t f1, f2, d1, d2; /* Two nearest output values */
float32_t Dn, Df;
float32_t temp, findex;
/* input x is in degrees */
/* Scale input, divide input by 360, for cosine add 0.25 (pi/2) to read sine table */
in = theta * 0.00277777777778f;
if (in < 0.0f)
{
in = -in;
}
in = in - (int32_t)in;
/* Calculate the nearest index */
findex = (float32_t)FAST_MATH_TABLE_SIZE * in;
indexS = ((uint16_t)findex) & 0x1ff;
indexC = (indexS + (FAST_MATH_TABLE_SIZE / 4)) & 0x1ff;
/* Calculation of fractional value */
fract = findex - (float32_t) indexS;
/* Read two nearest values of input value from the cos & sin tables */
f1 = sinTable_f32[indexC ];
f2 = sinTable_f32[indexC+1];
d1 = -sinTable_f32[indexS ];
d2 = -sinTable_f32[indexS+1];
temp = (1.0f - fract) * f1 + fract * f2;
Dn = 0.0122718463030f; /* delta between the two points (fixed), in this case 2*pi/FAST_MATH_TABLE_SIZE */
Df = f2 - f1; /* delta between the values of the functions */
temp = Dn * (d1 + d2) - 2 * Df;
temp = fract * temp + (3 * Df - (d2 + 2 * d1) * Dn);
temp = fract * temp + d1 * Dn;
/* Calculation of cosine value */
*pCosVal = fract * temp + f1;
/* Read two nearest values of input value from the cos & sin tables */
f1 = sinTable_f32[indexS ];
f2 = sinTable_f32[indexS+1];
d1 = sinTable_f32[indexC ];
d2 = sinTable_f32[indexC+1];
temp = (1.0f - fract) * f1 + fract * f2;
Df = f2 - f1; // delta between the values of the functions
temp = Dn * (d1 + d2) - 2 * Df;
temp = fract * temp + (3 * Df - (d2 + 2 * d1) * Dn);
temp = fract * temp + d1 * Dn;
/* Calculation of sine value */
*pSinVal = fract * temp + f1;
if (theta < 0.0f)
{
*pSinVal = -*pSinVal;
}
}
/**
@} end of SinCos group
*/

View File

@ -0,0 +1,122 @@
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_sin_f32.c
* Description: Fast sine calculation for floating-point values
*
* $Date: 18. March 2019
* $Revision: V1.6.0
*
* Target Processor: Cortex-M cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2019 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* Licensed under the Apache License, Version 2.0 (the License); you may
* not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an AS IS BASIS, WITHOUT
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "arm_math.h"
#include "arm_common_tables.h"
/**
@ingroup groupFastMath
*/
/**
@defgroup sin Sine
Computes the trigonometric sine function using a combination of table lookup
and linear interpolation. There are separate functions for
Q15, Q31, and floating-point data types.
The input to the floating-point version is in radians while the
fixed-point Q15 and Q31 have a scaled input with the range
[0 +0.9999] mapping to [0 2*pi). The fixed-point range is chosen so that a
value of 2*pi wraps around to 0.
The implementation is based on table lookup using 256 values together with linear interpolation.
The steps used are:
-# Calculation of the nearest integer table index
-# Compute the fractional portion (fract) of the table index.
-# The final result equals <code>(1.0f-fract)*a + fract*b;</code>
where
<pre>
b = Table[index];
c = Table[index+1];
</pre>
*/
/**
@addtogroup sin
@{
*/
/**
@brief Fast approximation to the trigonometric sine function for floating-point data.
@param[in] x input value in radians.
@return sin(x)
*/
float32_t arm_sin_f32(
float32_t x)
{
float32_t sinVal, fract, in; /* Temporary input, output variables */
uint16_t index; /* Index variable */
float32_t a, b; /* Two nearest output values */
int32_t n;
float32_t findex;
/* input x is in radians */
/* Scale input to [0 1] range from [0 2*PI] , divide input by 2*pi */
in = x * 0.159154943092f;
/* Calculation of floor value of input */
n = (int32_t) in;
/* Make negative values towards -infinity */
if (in < 0.0f)
{
n--;
}
/* Map input value to [0 1] */
in = in - (float32_t) n;
/* Calculation of index of the table */
findex = (float32_t)FAST_MATH_TABLE_SIZE * in;
index = (uint16_t)findex;
/* when "in" is exactly 1, we need to rotate the index down to 0 */
if (index >= FAST_MATH_TABLE_SIZE) {
index = 0;
findex -= (float32_t)FAST_MATH_TABLE_SIZE;
}
/* fractional value calculation */
fract = findex - (float32_t) index;
/* Read two nearest values of input value from the sin table */
a = sinTable_f32[index];
b = sinTable_f32[index+1];
/* Linear interpolation process */
sinVal = (1.0f - fract) * a + fract * b;
/* Return output value */
return (sinVal);
}
/**
@} end of sin group
*/