mbed-os 6.10 versione
cmsis_dsp/TransformFunctions/arm_cfft_radix2_q15.c@5:ed60e5c5c2a8, 2021-06-21 (annotated)
- Committer:
- michelericcio
- Date:
- Mon Jun 21 10:16:02 2021 +0000
- Revision:
- 5:ed60e5c5c2a8
- Parent:
- 3:7a284390b0ce
mbed-dsp working on ST Nucleo F767ZI and mbed-OS 6
Who changed what in which revision?
User | Revision | Line number | New contents of line |
---|---|---|---|
emilmont | 1:fdd22bb7aa52 | 1 | /* ---------------------------------------------------------------------- |
mbed_official | 3:7a284390b0ce | 2 | * Copyright (C) 2010-2013 ARM Limited. All rights reserved. |
emilmont | 1:fdd22bb7aa52 | 3 | * |
mbed_official | 3:7a284390b0ce | 4 | * $Date: 17. January 2013 |
mbed_official | 3:7a284390b0ce | 5 | * $Revision: V1.4.1 |
emilmont | 1:fdd22bb7aa52 | 6 | * |
emilmont | 2:da51fb522205 | 7 | * Project: CMSIS DSP Library |
emilmont | 2:da51fb522205 | 8 | * Title: arm_cfft_radix2_q15.c |
emilmont | 1:fdd22bb7aa52 | 9 | * |
emilmont | 2:da51fb522205 | 10 | * Description: Radix-2 Decimation in Frequency CFFT & CIFFT Fixed 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 | * |
mbed_official | 3:7a284390b0ce | 15 | * Redistribution and use in source and binary forms, with or without |
mbed_official | 3:7a284390b0ce | 16 | * modification, are permitted provided that the following conditions |
mbed_official | 3:7a284390b0ce | 17 | * are met: |
mbed_official | 3:7a284390b0ce | 18 | * - Redistributions of source code must retain the above copyright |
mbed_official | 3:7a284390b0ce | 19 | * notice, this list of conditions and the following disclaimer. |
mbed_official | 3:7a284390b0ce | 20 | * - Redistributions in binary form must reproduce the above copyright |
mbed_official | 3:7a284390b0ce | 21 | * notice, this list of conditions and the following disclaimer in |
mbed_official | 3:7a284390b0ce | 22 | * the documentation and/or other materials provided with the |
mbed_official | 3:7a284390b0ce | 23 | * distribution. |
mbed_official | 3:7a284390b0ce | 24 | * - Neither the name of ARM LIMITED nor the names of its contributors |
mbed_official | 3:7a284390b0ce | 25 | * may be used to endorse or promote products derived from this |
mbed_official | 3:7a284390b0ce | 26 | * software without specific prior written permission. |
mbed_official | 3:7a284390b0ce | 27 | * |
mbed_official | 3:7a284390b0ce | 28 | * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS |
mbed_official | 3:7a284390b0ce | 29 | * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT |
mbed_official | 3:7a284390b0ce | 30 | * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS |
mbed_official | 3:7a284390b0ce | 31 | * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE |
mbed_official | 3:7a284390b0ce | 32 | * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, |
mbed_official | 3:7a284390b0ce | 33 | * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, |
mbed_official | 3:7a284390b0ce | 34 | * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; |
mbed_official | 3:7a284390b0ce | 35 | * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER |
mbed_official | 3:7a284390b0ce | 36 | * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT |
mbed_official | 3:7a284390b0ce | 37 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN |
mbed_official | 3:7a284390b0ce | 38 | * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE |
mbed_official | 3:7a284390b0ce | 39 | * POSSIBILITY OF SUCH DAMAGE. |
emilmont | 1:fdd22bb7aa52 | 40 | * -------------------------------------------------------------------- */ |
emilmont | 1:fdd22bb7aa52 | 41 | |
emilmont | 1:fdd22bb7aa52 | 42 | #include "arm_math.h" |
emilmont | 1:fdd22bb7aa52 | 43 | |
mbed_official | 3:7a284390b0ce | 44 | void arm_radix2_butterfly_q15( |
mbed_official | 3:7a284390b0ce | 45 | q15_t * pSrc, |
mbed_official | 3:7a284390b0ce | 46 | uint32_t fftLen, |
mbed_official | 3:7a284390b0ce | 47 | q15_t * pCoef, |
mbed_official | 3:7a284390b0ce | 48 | uint16_t twidCoefModifier); |
mbed_official | 3:7a284390b0ce | 49 | |
mbed_official | 3:7a284390b0ce | 50 | void arm_radix2_butterfly_inverse_q15( |
mbed_official | 3:7a284390b0ce | 51 | q15_t * pSrc, |
mbed_official | 3:7a284390b0ce | 52 | uint32_t fftLen, |
mbed_official | 3:7a284390b0ce | 53 | q15_t * pCoef, |
mbed_official | 3:7a284390b0ce | 54 | uint16_t twidCoefModifier); |
mbed_official | 3:7a284390b0ce | 55 | |
mbed_official | 3:7a284390b0ce | 56 | void arm_bitreversal_q15( |
mbed_official | 3:7a284390b0ce | 57 | q15_t * pSrc, |
mbed_official | 3:7a284390b0ce | 58 | uint32_t fftLen, |
mbed_official | 3:7a284390b0ce | 59 | uint16_t bitRevFactor, |
mbed_official | 3:7a284390b0ce | 60 | uint16_t * pBitRevTab); |
mbed_official | 3:7a284390b0ce | 61 | |
emilmont | 1:fdd22bb7aa52 | 62 | /** |
emilmont | 1:fdd22bb7aa52 | 63 | * @ingroup groupTransforms |
emilmont | 1:fdd22bb7aa52 | 64 | */ |
emilmont | 1:fdd22bb7aa52 | 65 | |
emilmont | 1:fdd22bb7aa52 | 66 | /** |
mbed_official | 3:7a284390b0ce | 67 | * @addtogroup ComplexFFT |
emilmont | 1:fdd22bb7aa52 | 68 | * @{ |
emilmont | 1:fdd22bb7aa52 | 69 | */ |
emilmont | 1:fdd22bb7aa52 | 70 | |
emilmont | 1:fdd22bb7aa52 | 71 | /** |
emilmont | 1:fdd22bb7aa52 | 72 | * @details |
emilmont | 1:fdd22bb7aa52 | 73 | * @brief Processing function for the fixed-point CFFT/CIFFT. |
emilmont | 1:fdd22bb7aa52 | 74 | * @param[in] *S points to an instance of the fixed-point CFFT/CIFFT structure. |
emilmont | 1:fdd22bb7aa52 | 75 | * @param[in, out] *pSrc points to the complex data buffer of size <code>2*fftLen</code>. Processing occurs in-place. |
emilmont | 1:fdd22bb7aa52 | 76 | * @return none. |
emilmont | 1:fdd22bb7aa52 | 77 | */ |
emilmont | 1:fdd22bb7aa52 | 78 | |
emilmont | 1:fdd22bb7aa52 | 79 | void arm_cfft_radix2_q15( |
emilmont | 1:fdd22bb7aa52 | 80 | const arm_cfft_radix2_instance_q15 * S, |
emilmont | 1:fdd22bb7aa52 | 81 | q15_t * pSrc) |
emilmont | 1:fdd22bb7aa52 | 82 | { |
emilmont | 1:fdd22bb7aa52 | 83 | |
emilmont | 1:fdd22bb7aa52 | 84 | if(S->ifftFlag == 1u) |
emilmont | 1:fdd22bb7aa52 | 85 | { |
emilmont | 1:fdd22bb7aa52 | 86 | arm_radix2_butterfly_inverse_q15(pSrc, S->fftLen, |
emilmont | 1:fdd22bb7aa52 | 87 | S->pTwiddle, S->twidCoefModifier); |
emilmont | 1:fdd22bb7aa52 | 88 | } |
emilmont | 1:fdd22bb7aa52 | 89 | else |
emilmont | 1:fdd22bb7aa52 | 90 | { |
emilmont | 1:fdd22bb7aa52 | 91 | arm_radix2_butterfly_q15(pSrc, S->fftLen, |
emilmont | 1:fdd22bb7aa52 | 92 | S->pTwiddle, S->twidCoefModifier); |
emilmont | 1:fdd22bb7aa52 | 93 | } |
emilmont | 1:fdd22bb7aa52 | 94 | |
emilmont | 1:fdd22bb7aa52 | 95 | arm_bitreversal_q15(pSrc, S->fftLen, S->bitRevFactor, S->pBitRevTable); |
emilmont | 1:fdd22bb7aa52 | 96 | } |
emilmont | 1:fdd22bb7aa52 | 97 | |
emilmont | 1:fdd22bb7aa52 | 98 | /** |
mbed_official | 3:7a284390b0ce | 99 | * @} end of ComplexFFT group |
emilmont | 1:fdd22bb7aa52 | 100 | */ |
emilmont | 1:fdd22bb7aa52 | 101 | |
emilmont | 1:fdd22bb7aa52 | 102 | void arm_radix2_butterfly_q15( |
emilmont | 1:fdd22bb7aa52 | 103 | q15_t * pSrc, |
emilmont | 1:fdd22bb7aa52 | 104 | uint32_t fftLen, |
emilmont | 1:fdd22bb7aa52 | 105 | q15_t * pCoef, |
emilmont | 1:fdd22bb7aa52 | 106 | uint16_t twidCoefModifier) |
emilmont | 1:fdd22bb7aa52 | 107 | { |
mbed_official | 3:7a284390b0ce | 108 | #ifndef ARM_MATH_CM0_FAMILY |
emilmont | 1:fdd22bb7aa52 | 109 | |
mbed_official | 3:7a284390b0ce | 110 | unsigned i, j, k, l; |
mbed_official | 3:7a284390b0ce | 111 | unsigned n1, n2, ia; |
emilmont | 1:fdd22bb7aa52 | 112 | q15_t in; |
emilmont | 1:fdd22bb7aa52 | 113 | q31_t T, S, R; |
emilmont | 1:fdd22bb7aa52 | 114 | q31_t coeff, out1, out2; |
emilmont | 1:fdd22bb7aa52 | 115 | |
emilmont | 1:fdd22bb7aa52 | 116 | //N = fftLen; |
emilmont | 1:fdd22bb7aa52 | 117 | n2 = fftLen; |
emilmont | 1:fdd22bb7aa52 | 118 | |
emilmont | 1:fdd22bb7aa52 | 119 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 120 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 121 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 122 | |
emilmont | 1:fdd22bb7aa52 | 123 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 124 | for (i = 0; i < n2; i++) |
emilmont | 1:fdd22bb7aa52 | 125 | { |
emilmont | 1:fdd22bb7aa52 | 126 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 127 | |
emilmont | 1:fdd22bb7aa52 | 128 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 129 | |
emilmont | 1:fdd22bb7aa52 | 130 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 131 | |
emilmont | 1:fdd22bb7aa52 | 132 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 133 | in = ((int16_t) (T & 0xFFFF)) >> 2; |
emilmont | 1:fdd22bb7aa52 | 134 | T = ((T >> 2) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 135 | |
emilmont | 1:fdd22bb7aa52 | 136 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 137 | in = ((int16_t) (S & 0xFFFF)) >> 2; |
emilmont | 1:fdd22bb7aa52 | 138 | S = ((S >> 2) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 139 | |
emilmont | 1:fdd22bb7aa52 | 140 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 141 | |
emilmont | 1:fdd22bb7aa52 | 142 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 143 | |
emilmont | 1:fdd22bb7aa52 | 144 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 145 | |
emilmont | 1:fdd22bb7aa52 | 146 | out1 = __SMUAD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 147 | out2 = __SMUSDX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 148 | |
emilmont | 1:fdd22bb7aa52 | 149 | #else |
emilmont | 1:fdd22bb7aa52 | 150 | |
emilmont | 1:fdd22bb7aa52 | 151 | out1 = __SMUSDX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 152 | out2 = __SMUAD(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 153 | |
emilmont | 1:fdd22bb7aa52 | 154 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 155 | |
emilmont | 1:fdd22bb7aa52 | 156 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 157 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 158 | |
emilmont | 1:fdd22bb7aa52 | 159 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 160 | |
emilmont | 1:fdd22bb7aa52 | 161 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 162 | |
emilmont | 1:fdd22bb7aa52 | 163 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 164 | i++; |
emilmont | 1:fdd22bb7aa52 | 165 | l++; |
emilmont | 1:fdd22bb7aa52 | 166 | |
emilmont | 1:fdd22bb7aa52 | 167 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 168 | in = ((int16_t) (T & 0xFFFF)) >> 2; |
emilmont | 1:fdd22bb7aa52 | 169 | T = ((T >> 2) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 170 | |
emilmont | 1:fdd22bb7aa52 | 171 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 172 | in = ((int16_t) (S & 0xFFFF)) >> 2; |
emilmont | 1:fdd22bb7aa52 | 173 | S = ((S >> 2) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 174 | |
emilmont | 1:fdd22bb7aa52 | 175 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 176 | |
emilmont | 1:fdd22bb7aa52 | 177 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 178 | |
emilmont | 1:fdd22bb7aa52 | 179 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 180 | |
emilmont | 1:fdd22bb7aa52 | 181 | out1 = __SMUAD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 182 | out2 = __SMUSDX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 183 | |
emilmont | 1:fdd22bb7aa52 | 184 | #else |
emilmont | 1:fdd22bb7aa52 | 185 | |
emilmont | 1:fdd22bb7aa52 | 186 | out1 = __SMUSDX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 187 | out2 = __SMUAD(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 188 | |
emilmont | 1:fdd22bb7aa52 | 189 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 190 | |
emilmont | 1:fdd22bb7aa52 | 191 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 192 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 193 | |
emilmont | 1:fdd22bb7aa52 | 194 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 195 | |
emilmont | 1:fdd22bb7aa52 | 196 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 197 | |
emilmont | 1:fdd22bb7aa52 | 198 | // loop for stage |
emilmont | 1:fdd22bb7aa52 | 199 | for (k = fftLen / 2; k > 2; k = k >> 1) |
emilmont | 1:fdd22bb7aa52 | 200 | { |
emilmont | 1:fdd22bb7aa52 | 201 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 202 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 203 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 204 | |
emilmont | 1:fdd22bb7aa52 | 205 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 206 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 207 | { |
emilmont | 1:fdd22bb7aa52 | 208 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 209 | |
emilmont | 1:fdd22bb7aa52 | 210 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 211 | |
emilmont | 1:fdd22bb7aa52 | 212 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 213 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 214 | { |
emilmont | 1:fdd22bb7aa52 | 215 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 216 | |
emilmont | 1:fdd22bb7aa52 | 217 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 218 | |
emilmont | 1:fdd22bb7aa52 | 219 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 220 | |
emilmont | 1:fdd22bb7aa52 | 221 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 222 | |
emilmont | 1:fdd22bb7aa52 | 223 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 224 | |
emilmont | 1:fdd22bb7aa52 | 225 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 226 | |
emilmont | 1:fdd22bb7aa52 | 227 | out1 = __SMUAD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 228 | out2 = __SMUSDX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 229 | |
emilmont | 1:fdd22bb7aa52 | 230 | #else |
emilmont | 1:fdd22bb7aa52 | 231 | |
emilmont | 1:fdd22bb7aa52 | 232 | out1 = __SMUSDX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 233 | out2 = __SMUAD(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 234 | |
emilmont | 1:fdd22bb7aa52 | 235 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 236 | |
emilmont | 1:fdd22bb7aa52 | 237 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 238 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 239 | |
emilmont | 1:fdd22bb7aa52 | 240 | i += n1; |
emilmont | 1:fdd22bb7aa52 | 241 | |
emilmont | 1:fdd22bb7aa52 | 242 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 243 | |
emilmont | 1:fdd22bb7aa52 | 244 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 245 | |
emilmont | 1:fdd22bb7aa52 | 246 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 247 | |
emilmont | 1:fdd22bb7aa52 | 248 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 249 | |
emilmont | 1:fdd22bb7aa52 | 250 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 251 | |
emilmont | 1:fdd22bb7aa52 | 252 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 253 | |
emilmont | 1:fdd22bb7aa52 | 254 | out1 = __SMUAD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 255 | out2 = __SMUSDX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 256 | |
emilmont | 1:fdd22bb7aa52 | 257 | #else |
emilmont | 1:fdd22bb7aa52 | 258 | |
emilmont | 1:fdd22bb7aa52 | 259 | out1 = __SMUSDX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 260 | out2 = __SMUAD(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 261 | |
emilmont | 1:fdd22bb7aa52 | 262 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 263 | |
emilmont | 1:fdd22bb7aa52 | 264 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 265 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 266 | |
emilmont | 1:fdd22bb7aa52 | 267 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 268 | |
emilmont | 1:fdd22bb7aa52 | 269 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 270 | |
emilmont | 1:fdd22bb7aa52 | 271 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 272 | } // stages loop end |
emilmont | 1:fdd22bb7aa52 | 273 | |
emilmont | 1:fdd22bb7aa52 | 274 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 275 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 276 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 277 | |
emilmont | 1:fdd22bb7aa52 | 278 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 279 | |
emilmont | 1:fdd22bb7aa52 | 280 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 281 | |
emilmont | 1:fdd22bb7aa52 | 282 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 283 | for (i = 0; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 284 | { |
emilmont | 1:fdd22bb7aa52 | 285 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 286 | |
emilmont | 1:fdd22bb7aa52 | 287 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 288 | |
emilmont | 1:fdd22bb7aa52 | 289 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 290 | |
emilmont | 1:fdd22bb7aa52 | 291 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 292 | |
emilmont | 1:fdd22bb7aa52 | 293 | _SIMD32_OFFSET(pSrc + (2 * i)) = __QADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 294 | |
emilmont | 1:fdd22bb7aa52 | 295 | _SIMD32_OFFSET(pSrc + (2u * l)) = R; |
emilmont | 1:fdd22bb7aa52 | 296 | |
emilmont | 1:fdd22bb7aa52 | 297 | i += n1; |
emilmont | 1:fdd22bb7aa52 | 298 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 299 | |
emilmont | 1:fdd22bb7aa52 | 300 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 301 | |
emilmont | 1:fdd22bb7aa52 | 302 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 303 | |
emilmont | 1:fdd22bb7aa52 | 304 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 305 | |
emilmont | 1:fdd22bb7aa52 | 306 | _SIMD32_OFFSET(pSrc + (2 * i)) = __QADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 307 | |
emilmont | 1:fdd22bb7aa52 | 308 | _SIMD32_OFFSET(pSrc + (2u * l)) = R; |
emilmont | 1:fdd22bb7aa52 | 309 | |
emilmont | 1:fdd22bb7aa52 | 310 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 311 | |
emilmont | 1:fdd22bb7aa52 | 312 | |
emilmont | 1:fdd22bb7aa52 | 313 | #else |
emilmont | 1:fdd22bb7aa52 | 314 | |
mbed_official | 3:7a284390b0ce | 315 | unsigned i, j, k, l; |
mbed_official | 3:7a284390b0ce | 316 | unsigned n1, n2, ia; |
emilmont | 1:fdd22bb7aa52 | 317 | q15_t xt, yt, cosVal, sinVal; |
emilmont | 1:fdd22bb7aa52 | 318 | |
emilmont | 1:fdd22bb7aa52 | 319 | |
emilmont | 1:fdd22bb7aa52 | 320 | //N = fftLen; |
emilmont | 1:fdd22bb7aa52 | 321 | n2 = fftLen; |
emilmont | 1:fdd22bb7aa52 | 322 | |
emilmont | 1:fdd22bb7aa52 | 323 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 324 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 325 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 326 | |
emilmont | 1:fdd22bb7aa52 | 327 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 328 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 329 | { |
emilmont | 1:fdd22bb7aa52 | 330 | cosVal = pCoef[ia * 2]; |
emilmont | 1:fdd22bb7aa52 | 331 | sinVal = pCoef[(ia * 2) + 1]; |
emilmont | 1:fdd22bb7aa52 | 332 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 333 | |
emilmont | 1:fdd22bb7aa52 | 334 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 335 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 336 | { |
emilmont | 1:fdd22bb7aa52 | 337 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 338 | xt = (pSrc[2 * i] >> 2u) - (pSrc[2 * l] >> 2u); |
emilmont | 1:fdd22bb7aa52 | 339 | pSrc[2 * i] = ((pSrc[2 * i] >> 2u) + (pSrc[2 * l] >> 2u)) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 340 | |
emilmont | 1:fdd22bb7aa52 | 341 | yt = (pSrc[2 * i + 1] >> 2u) - (pSrc[2 * l + 1] >> 2u); |
emilmont | 1:fdd22bb7aa52 | 342 | pSrc[2 * i + 1] = |
emilmont | 1:fdd22bb7aa52 | 343 | ((pSrc[2 * l + 1] >> 2u) + (pSrc[2 * i + 1] >> 2u)) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 344 | |
emilmont | 1:fdd22bb7aa52 | 345 | pSrc[2u * l] = (((int16_t) (((q31_t) xt * cosVal) >> 16)) + |
emilmont | 1:fdd22bb7aa52 | 346 | ((int16_t) (((q31_t) yt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 347 | |
emilmont | 1:fdd22bb7aa52 | 348 | pSrc[2u * l + 1u] = (((int16_t) (((q31_t) yt * cosVal) >> 16)) - |
emilmont | 1:fdd22bb7aa52 | 349 | ((int16_t) (((q31_t) xt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 350 | |
emilmont | 1:fdd22bb7aa52 | 351 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 352 | |
emilmont | 1:fdd22bb7aa52 | 353 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 354 | |
emilmont | 1:fdd22bb7aa52 | 355 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 356 | |
emilmont | 1:fdd22bb7aa52 | 357 | // loop for stage |
emilmont | 1:fdd22bb7aa52 | 358 | for (k = fftLen / 2; k > 2; k = k >> 1) |
emilmont | 1:fdd22bb7aa52 | 359 | { |
emilmont | 1:fdd22bb7aa52 | 360 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 361 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 362 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 363 | |
emilmont | 1:fdd22bb7aa52 | 364 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 365 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 366 | { |
emilmont | 1:fdd22bb7aa52 | 367 | cosVal = pCoef[ia * 2]; |
emilmont | 1:fdd22bb7aa52 | 368 | sinVal = pCoef[(ia * 2) + 1]; |
emilmont | 1:fdd22bb7aa52 | 369 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 370 | |
emilmont | 1:fdd22bb7aa52 | 371 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 372 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 373 | { |
emilmont | 1:fdd22bb7aa52 | 374 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 375 | xt = pSrc[2 * i] - pSrc[2 * l]; |
emilmont | 1:fdd22bb7aa52 | 376 | pSrc[2 * i] = (pSrc[2 * i] + pSrc[2 * l]) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 377 | |
emilmont | 1:fdd22bb7aa52 | 378 | yt = pSrc[2 * i + 1] - pSrc[2 * l + 1]; |
emilmont | 1:fdd22bb7aa52 | 379 | pSrc[2 * i + 1] = (pSrc[2 * l + 1] + pSrc[2 * i + 1]) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 380 | |
emilmont | 1:fdd22bb7aa52 | 381 | pSrc[2u * l] = (((int16_t) (((q31_t) xt * cosVal) >> 16)) + |
emilmont | 1:fdd22bb7aa52 | 382 | ((int16_t) (((q31_t) yt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 383 | |
emilmont | 1:fdd22bb7aa52 | 384 | pSrc[2u * l + 1u] = (((int16_t) (((q31_t) yt * cosVal) >> 16)) - |
emilmont | 1:fdd22bb7aa52 | 385 | ((int16_t) (((q31_t) xt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 386 | |
emilmont | 1:fdd22bb7aa52 | 387 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 388 | |
emilmont | 1:fdd22bb7aa52 | 389 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 390 | |
emilmont | 1:fdd22bb7aa52 | 391 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 392 | } // stages loop end |
emilmont | 1:fdd22bb7aa52 | 393 | |
emilmont | 1:fdd22bb7aa52 | 394 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 395 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 396 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 397 | |
emilmont | 1:fdd22bb7aa52 | 398 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 399 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 400 | { |
emilmont | 1:fdd22bb7aa52 | 401 | cosVal = pCoef[ia * 2]; |
emilmont | 1:fdd22bb7aa52 | 402 | sinVal = pCoef[(ia * 2) + 1]; |
emilmont | 1:fdd22bb7aa52 | 403 | |
emilmont | 1:fdd22bb7aa52 | 404 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 405 | |
emilmont | 1:fdd22bb7aa52 | 406 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 407 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 408 | { |
emilmont | 1:fdd22bb7aa52 | 409 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 410 | xt = pSrc[2 * i] - pSrc[2 * l]; |
emilmont | 1:fdd22bb7aa52 | 411 | pSrc[2 * i] = (pSrc[2 * i] + pSrc[2 * l]); |
emilmont | 1:fdd22bb7aa52 | 412 | |
emilmont | 1:fdd22bb7aa52 | 413 | yt = pSrc[2 * i + 1] - pSrc[2 * l + 1]; |
emilmont | 1:fdd22bb7aa52 | 414 | pSrc[2 * i + 1] = (pSrc[2 * l + 1] + pSrc[2 * i + 1]); |
emilmont | 1:fdd22bb7aa52 | 415 | |
emilmont | 1:fdd22bb7aa52 | 416 | pSrc[2u * l] = xt; |
emilmont | 1:fdd22bb7aa52 | 417 | |
emilmont | 1:fdd22bb7aa52 | 418 | pSrc[2u * l + 1u] = yt; |
emilmont | 1:fdd22bb7aa52 | 419 | |
emilmont | 1:fdd22bb7aa52 | 420 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 421 | |
emilmont | 1:fdd22bb7aa52 | 422 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 423 | |
emilmont | 1:fdd22bb7aa52 | 424 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 425 | |
mbed_official | 3:7a284390b0ce | 426 | #endif // #ifndef ARM_MATH_CM0_FAMILY |
emilmont | 1:fdd22bb7aa52 | 427 | |
emilmont | 1:fdd22bb7aa52 | 428 | } |
emilmont | 1:fdd22bb7aa52 | 429 | |
emilmont | 1:fdd22bb7aa52 | 430 | |
emilmont | 1:fdd22bb7aa52 | 431 | void arm_radix2_butterfly_inverse_q15( |
emilmont | 1:fdd22bb7aa52 | 432 | q15_t * pSrc, |
emilmont | 1:fdd22bb7aa52 | 433 | uint32_t fftLen, |
emilmont | 1:fdd22bb7aa52 | 434 | q15_t * pCoef, |
emilmont | 1:fdd22bb7aa52 | 435 | uint16_t twidCoefModifier) |
emilmont | 1:fdd22bb7aa52 | 436 | { |
mbed_official | 3:7a284390b0ce | 437 | #ifndef ARM_MATH_CM0_FAMILY |
emilmont | 1:fdd22bb7aa52 | 438 | |
mbed_official | 3:7a284390b0ce | 439 | unsigned i, j, k, l; |
mbed_official | 3:7a284390b0ce | 440 | unsigned n1, n2, ia; |
emilmont | 1:fdd22bb7aa52 | 441 | q15_t in; |
emilmont | 1:fdd22bb7aa52 | 442 | q31_t T, S, R; |
emilmont | 1:fdd22bb7aa52 | 443 | q31_t coeff, out1, out2; |
emilmont | 1:fdd22bb7aa52 | 444 | |
emilmont | 1:fdd22bb7aa52 | 445 | //N = fftLen; |
emilmont | 1:fdd22bb7aa52 | 446 | n2 = fftLen; |
emilmont | 1:fdd22bb7aa52 | 447 | |
emilmont | 1:fdd22bb7aa52 | 448 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 449 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 450 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 451 | |
emilmont | 1:fdd22bb7aa52 | 452 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 453 | for (i = 0; i < n2; i++) |
emilmont | 1:fdd22bb7aa52 | 454 | { |
emilmont | 1:fdd22bb7aa52 | 455 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 456 | |
emilmont | 1:fdd22bb7aa52 | 457 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 458 | |
emilmont | 1:fdd22bb7aa52 | 459 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 460 | |
emilmont | 1:fdd22bb7aa52 | 461 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 462 | in = ((int16_t) (T & 0xFFFF)) >> 2; |
emilmont | 1:fdd22bb7aa52 | 463 | T = ((T >> 2) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 464 | |
emilmont | 1:fdd22bb7aa52 | 465 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 466 | in = ((int16_t) (S & 0xFFFF)) >> 2; |
emilmont | 1:fdd22bb7aa52 | 467 | S = ((S >> 2) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 468 | |
emilmont | 1:fdd22bb7aa52 | 469 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 470 | |
emilmont | 1:fdd22bb7aa52 | 471 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 472 | |
emilmont | 1:fdd22bb7aa52 | 473 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 474 | |
emilmont | 1:fdd22bb7aa52 | 475 | out1 = __SMUSD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 476 | out2 = __SMUADX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 477 | #else |
emilmont | 1:fdd22bb7aa52 | 478 | |
emilmont | 1:fdd22bb7aa52 | 479 | out1 = __SMUADX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 480 | out2 = __SMUSD(__QSUB(0, coeff), R); |
emilmont | 1:fdd22bb7aa52 | 481 | |
emilmont | 1:fdd22bb7aa52 | 482 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 483 | |
emilmont | 1:fdd22bb7aa52 | 484 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 485 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 486 | |
emilmont | 1:fdd22bb7aa52 | 487 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 488 | |
emilmont | 1:fdd22bb7aa52 | 489 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 490 | |
emilmont | 1:fdd22bb7aa52 | 491 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 492 | i++; |
emilmont | 1:fdd22bb7aa52 | 493 | l++; |
emilmont | 1:fdd22bb7aa52 | 494 | |
emilmont | 1:fdd22bb7aa52 | 495 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 496 | in = ((int16_t) (T & 0xFFFF)) >> 2; |
emilmont | 1:fdd22bb7aa52 | 497 | T = ((T >> 2) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 498 | |
emilmont | 1:fdd22bb7aa52 | 499 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 500 | in = ((int16_t) (S & 0xFFFF)) >> 2; |
emilmont | 1:fdd22bb7aa52 | 501 | S = ((S >> 2) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 502 | |
emilmont | 1:fdd22bb7aa52 | 503 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 504 | |
emilmont | 1:fdd22bb7aa52 | 505 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 506 | |
emilmont | 1:fdd22bb7aa52 | 507 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 508 | |
emilmont | 1:fdd22bb7aa52 | 509 | out1 = __SMUSD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 510 | out2 = __SMUADX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 511 | #else |
emilmont | 1:fdd22bb7aa52 | 512 | |
emilmont | 1:fdd22bb7aa52 | 513 | out1 = __SMUADX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 514 | out2 = __SMUSD(__QSUB(0, coeff), R); |
emilmont | 1:fdd22bb7aa52 | 515 | |
emilmont | 1:fdd22bb7aa52 | 516 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 517 | |
emilmont | 1:fdd22bb7aa52 | 518 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 519 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 520 | |
emilmont | 1:fdd22bb7aa52 | 521 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 522 | |
emilmont | 1:fdd22bb7aa52 | 523 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 524 | |
emilmont | 1:fdd22bb7aa52 | 525 | // loop for stage |
emilmont | 1:fdd22bb7aa52 | 526 | for (k = fftLen / 2; k > 2; k = k >> 1) |
emilmont | 1:fdd22bb7aa52 | 527 | { |
emilmont | 1:fdd22bb7aa52 | 528 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 529 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 530 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 531 | |
emilmont | 1:fdd22bb7aa52 | 532 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 533 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 534 | { |
emilmont | 1:fdd22bb7aa52 | 535 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 536 | |
emilmont | 1:fdd22bb7aa52 | 537 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 538 | |
emilmont | 1:fdd22bb7aa52 | 539 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 540 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 541 | { |
emilmont | 1:fdd22bb7aa52 | 542 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 543 | |
emilmont | 1:fdd22bb7aa52 | 544 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 545 | |
emilmont | 1:fdd22bb7aa52 | 546 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 547 | |
emilmont | 1:fdd22bb7aa52 | 548 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 549 | |
emilmont | 1:fdd22bb7aa52 | 550 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 551 | |
emilmont | 1:fdd22bb7aa52 | 552 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 553 | |
emilmont | 1:fdd22bb7aa52 | 554 | out1 = __SMUSD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 555 | out2 = __SMUADX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 556 | |
emilmont | 1:fdd22bb7aa52 | 557 | #else |
emilmont | 1:fdd22bb7aa52 | 558 | |
emilmont | 1:fdd22bb7aa52 | 559 | out1 = __SMUADX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 560 | out2 = __SMUSD(__QSUB(0, coeff), R); |
emilmont | 1:fdd22bb7aa52 | 561 | |
emilmont | 1:fdd22bb7aa52 | 562 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 563 | |
emilmont | 1:fdd22bb7aa52 | 564 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 565 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 566 | |
emilmont | 1:fdd22bb7aa52 | 567 | i += n1; |
emilmont | 1:fdd22bb7aa52 | 568 | |
emilmont | 1:fdd22bb7aa52 | 569 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 570 | |
emilmont | 1:fdd22bb7aa52 | 571 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 572 | |
emilmont | 1:fdd22bb7aa52 | 573 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 574 | |
emilmont | 1:fdd22bb7aa52 | 575 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 576 | |
emilmont | 1:fdd22bb7aa52 | 577 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 578 | |
emilmont | 1:fdd22bb7aa52 | 579 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 580 | |
emilmont | 1:fdd22bb7aa52 | 581 | out1 = __SMUSD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 582 | out2 = __SMUADX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 583 | #else |
emilmont | 1:fdd22bb7aa52 | 584 | |
emilmont | 1:fdd22bb7aa52 | 585 | out1 = __SMUADX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 586 | out2 = __SMUSD(__QSUB(0, coeff), R); |
emilmont | 1:fdd22bb7aa52 | 587 | |
emilmont | 1:fdd22bb7aa52 | 588 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 589 | |
emilmont | 1:fdd22bb7aa52 | 590 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 591 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 592 | |
emilmont | 1:fdd22bb7aa52 | 593 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 594 | |
emilmont | 1:fdd22bb7aa52 | 595 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 596 | |
emilmont | 1:fdd22bb7aa52 | 597 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 598 | } // stages loop end |
emilmont | 1:fdd22bb7aa52 | 599 | |
emilmont | 1:fdd22bb7aa52 | 600 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 601 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 602 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 603 | |
emilmont | 1:fdd22bb7aa52 | 604 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 605 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 606 | { |
emilmont | 1:fdd22bb7aa52 | 607 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 608 | |
emilmont | 1:fdd22bb7aa52 | 609 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 610 | |
emilmont | 1:fdd22bb7aa52 | 611 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 612 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 613 | { |
emilmont | 1:fdd22bb7aa52 | 614 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 615 | |
emilmont | 1:fdd22bb7aa52 | 616 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 617 | |
emilmont | 1:fdd22bb7aa52 | 618 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 619 | |
emilmont | 1:fdd22bb7aa52 | 620 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 621 | |
emilmont | 1:fdd22bb7aa52 | 622 | _SIMD32_OFFSET(pSrc + (2 * i)) = __QADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 623 | |
emilmont | 1:fdd22bb7aa52 | 624 | _SIMD32_OFFSET(pSrc + (2u * l)) = R; |
emilmont | 1:fdd22bb7aa52 | 625 | |
emilmont | 1:fdd22bb7aa52 | 626 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 627 | |
emilmont | 1:fdd22bb7aa52 | 628 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 629 | |
emilmont | 1:fdd22bb7aa52 | 630 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 631 | |
emilmont | 1:fdd22bb7aa52 | 632 | #else |
emilmont | 1:fdd22bb7aa52 | 633 | |
emilmont | 1:fdd22bb7aa52 | 634 | |
mbed_official | 3:7a284390b0ce | 635 | unsigned i, j, k, l; |
mbed_official | 3:7a284390b0ce | 636 | unsigned n1, n2, ia; |
emilmont | 1:fdd22bb7aa52 | 637 | q15_t xt, yt, cosVal, sinVal; |
emilmont | 1:fdd22bb7aa52 | 638 | |
emilmont | 1:fdd22bb7aa52 | 639 | //N = fftLen; |
emilmont | 1:fdd22bb7aa52 | 640 | n2 = fftLen; |
emilmont | 1:fdd22bb7aa52 | 641 | |
emilmont | 1:fdd22bb7aa52 | 642 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 643 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 644 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 645 | |
emilmont | 1:fdd22bb7aa52 | 646 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 647 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 648 | { |
emilmont | 1:fdd22bb7aa52 | 649 | cosVal = pCoef[ia * 2]; |
emilmont | 1:fdd22bb7aa52 | 650 | sinVal = pCoef[(ia * 2) + 1]; |
emilmont | 1:fdd22bb7aa52 | 651 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 652 | |
emilmont | 1:fdd22bb7aa52 | 653 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 654 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 655 | { |
emilmont | 1:fdd22bb7aa52 | 656 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 657 | xt = (pSrc[2 * i] >> 2u) - (pSrc[2 * l] >> 2u); |
emilmont | 1:fdd22bb7aa52 | 658 | pSrc[2 * i] = ((pSrc[2 * i] >> 2u) + (pSrc[2 * l] >> 2u)) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 659 | |
emilmont | 1:fdd22bb7aa52 | 660 | yt = (pSrc[2 * i + 1] >> 2u) - (pSrc[2 * l + 1] >> 2u); |
emilmont | 1:fdd22bb7aa52 | 661 | pSrc[2 * i + 1] = |
emilmont | 1:fdd22bb7aa52 | 662 | ((pSrc[2 * l + 1] >> 2u) + (pSrc[2 * i + 1] >> 2u)) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 663 | |
emilmont | 1:fdd22bb7aa52 | 664 | pSrc[2u * l] = (((int16_t) (((q31_t) xt * cosVal) >> 16)) - |
emilmont | 1:fdd22bb7aa52 | 665 | ((int16_t) (((q31_t) yt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 666 | |
emilmont | 1:fdd22bb7aa52 | 667 | pSrc[2u * l + 1u] = (((int16_t) (((q31_t) yt * cosVal) >> 16)) + |
emilmont | 1:fdd22bb7aa52 | 668 | ((int16_t) (((q31_t) xt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 669 | |
emilmont | 1:fdd22bb7aa52 | 670 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 671 | |
emilmont | 1:fdd22bb7aa52 | 672 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 673 | |
emilmont | 1:fdd22bb7aa52 | 674 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 675 | |
emilmont | 1:fdd22bb7aa52 | 676 | // loop for stage |
emilmont | 1:fdd22bb7aa52 | 677 | for (k = fftLen / 2; k > 2; k = k >> 1) |
emilmont | 1:fdd22bb7aa52 | 678 | { |
emilmont | 1:fdd22bb7aa52 | 679 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 680 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 681 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 682 | |
emilmont | 1:fdd22bb7aa52 | 683 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 684 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 685 | { |
emilmont | 1:fdd22bb7aa52 | 686 | cosVal = pCoef[ia * 2]; |
emilmont | 1:fdd22bb7aa52 | 687 | sinVal = pCoef[(ia * 2) + 1]; |
emilmont | 1:fdd22bb7aa52 | 688 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 689 | |
emilmont | 1:fdd22bb7aa52 | 690 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 691 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 692 | { |
emilmont | 1:fdd22bb7aa52 | 693 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 694 | xt = pSrc[2 * i] - pSrc[2 * l]; |
emilmont | 1:fdd22bb7aa52 | 695 | pSrc[2 * i] = (pSrc[2 * i] + pSrc[2 * l]) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 696 | |
emilmont | 1:fdd22bb7aa52 | 697 | yt = pSrc[2 * i + 1] - pSrc[2 * l + 1]; |
emilmont | 1:fdd22bb7aa52 | 698 | pSrc[2 * i + 1] = (pSrc[2 * l + 1] + pSrc[2 * i + 1]) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 699 | |
emilmont | 1:fdd22bb7aa52 | 700 | pSrc[2u * l] = (((int16_t) (((q31_t) xt * cosVal) >> 16)) - |
emilmont | 1:fdd22bb7aa52 | 701 | ((int16_t) (((q31_t) yt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 702 | |
emilmont | 1:fdd22bb7aa52 | 703 | pSrc[2u * l + 1u] = (((int16_t) (((q31_t) yt * cosVal) >> 16)) + |
emilmont | 1:fdd22bb7aa52 | 704 | ((int16_t) (((q31_t) xt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 705 | |
emilmont | 1:fdd22bb7aa52 | 706 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 707 | |
emilmont | 1:fdd22bb7aa52 | 708 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 709 | |
emilmont | 1:fdd22bb7aa52 | 710 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 711 | } // stages loop end |
emilmont | 1:fdd22bb7aa52 | 712 | |
emilmont | 1:fdd22bb7aa52 | 713 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 714 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 715 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 716 | |
emilmont | 1:fdd22bb7aa52 | 717 | cosVal = pCoef[ia * 2]; |
emilmont | 1:fdd22bb7aa52 | 718 | sinVal = pCoef[(ia * 2) + 1]; |
emilmont | 1:fdd22bb7aa52 | 719 | |
emilmont | 1:fdd22bb7aa52 | 720 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 721 | |
emilmont | 1:fdd22bb7aa52 | 722 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 723 | for (i = 0; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 724 | { |
emilmont | 1:fdd22bb7aa52 | 725 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 726 | xt = pSrc[2 * i] - pSrc[2 * l]; |
emilmont | 1:fdd22bb7aa52 | 727 | pSrc[2 * i] = (pSrc[2 * i] + pSrc[2 * l]); |
emilmont | 1:fdd22bb7aa52 | 728 | |
emilmont | 1:fdd22bb7aa52 | 729 | yt = pSrc[2 * i + 1] - pSrc[2 * l + 1]; |
emilmont | 1:fdd22bb7aa52 | 730 | pSrc[2 * i + 1] = (pSrc[2 * l + 1] + pSrc[2 * i + 1]); |
emilmont | 1:fdd22bb7aa52 | 731 | |
emilmont | 1:fdd22bb7aa52 | 732 | pSrc[2u * l] = xt; |
emilmont | 1:fdd22bb7aa52 | 733 | |
emilmont | 1:fdd22bb7aa52 | 734 | pSrc[2u * l + 1u] = yt; |
emilmont | 1:fdd22bb7aa52 | 735 | |
emilmont | 1:fdd22bb7aa52 | 736 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 737 | |
emilmont | 1:fdd22bb7aa52 | 738 | |
mbed_official | 3:7a284390b0ce | 739 | #endif // #ifndef ARM_MATH_CM0_FAMILY |
emilmont | 1:fdd22bb7aa52 | 740 | |
emilmont | 1:fdd22bb7aa52 | 741 | } |