CMSIS DSP library
Dependents: performance_timer Surfboard_ gps2rtty Capstone ... more
Legacy Warning
This is an mbed 2 library. To learn more about mbed OS 5, visit the docs.
cmsis_dsp/TransformFunctions/arm_cfft_radix2_q15.c@5:3762170b6d4d, 2015-11-20 (annotated)
- Committer:
- mbed_official
- Date:
- Fri Nov 20 08:45:18 2015 +0000
- Revision:
- 5:3762170b6d4d
- Parent:
- 3:7a284390b0ce
Synchronized with git revision 2eb940b9a73af188d3004a2575fdfbb05febe62b
Full URL: https://github.com/mbedmicro/mbed/commit/2eb940b9a73af188d3004a2575fdfbb05febe62b/
Added option to build rpc library. closes #1426
Who changed what in which revision?
User | Revision | Line number | New contents of line |
---|---|---|---|
emilmont | 1:fdd22bb7aa52 | 1 | /* ---------------------------------------------------------------------- |
mbed_official | 5:3762170b6d4d | 2 | * Copyright (C) 2010-2014 ARM Limited. All rights reserved. |
emilmont | 1:fdd22bb7aa52 | 3 | * |
mbed_official | 5:3762170b6d4d | 4 | * $Date: 19. March 2015 |
mbed_official | 5:3762170b6d4d | 5 | * $Revision: V.1.4.5 |
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. |
mbed_official | 5:3762170b6d4d | 74 | * @deprecated Do not use this function. It has been superseded by \ref arm_cfft_q15 and will be removed |
emilmont | 1:fdd22bb7aa52 | 75 | * @param[in] *S points to an instance of the fixed-point CFFT/CIFFT structure. |
emilmont | 1:fdd22bb7aa52 | 76 | * @param[in, out] *pSrc points to the complex data buffer of size <code>2*fftLen</code>. Processing occurs in-place. |
emilmont | 1:fdd22bb7aa52 | 77 | * @return none. |
emilmont | 1:fdd22bb7aa52 | 78 | */ |
emilmont | 1:fdd22bb7aa52 | 79 | |
emilmont | 1:fdd22bb7aa52 | 80 | void arm_cfft_radix2_q15( |
emilmont | 1:fdd22bb7aa52 | 81 | const arm_cfft_radix2_instance_q15 * S, |
emilmont | 1:fdd22bb7aa52 | 82 | q15_t * pSrc) |
emilmont | 1:fdd22bb7aa52 | 83 | { |
emilmont | 1:fdd22bb7aa52 | 84 | |
emilmont | 1:fdd22bb7aa52 | 85 | if(S->ifftFlag == 1u) |
emilmont | 1:fdd22bb7aa52 | 86 | { |
emilmont | 1:fdd22bb7aa52 | 87 | arm_radix2_butterfly_inverse_q15(pSrc, S->fftLen, |
emilmont | 1:fdd22bb7aa52 | 88 | S->pTwiddle, S->twidCoefModifier); |
emilmont | 1:fdd22bb7aa52 | 89 | } |
emilmont | 1:fdd22bb7aa52 | 90 | else |
emilmont | 1:fdd22bb7aa52 | 91 | { |
emilmont | 1:fdd22bb7aa52 | 92 | arm_radix2_butterfly_q15(pSrc, S->fftLen, |
emilmont | 1:fdd22bb7aa52 | 93 | S->pTwiddle, S->twidCoefModifier); |
emilmont | 1:fdd22bb7aa52 | 94 | } |
emilmont | 1:fdd22bb7aa52 | 95 | |
emilmont | 1:fdd22bb7aa52 | 96 | arm_bitreversal_q15(pSrc, S->fftLen, S->bitRevFactor, S->pBitRevTable); |
emilmont | 1:fdd22bb7aa52 | 97 | } |
emilmont | 1:fdd22bb7aa52 | 98 | |
emilmont | 1:fdd22bb7aa52 | 99 | /** |
mbed_official | 3:7a284390b0ce | 100 | * @} end of ComplexFFT group |
emilmont | 1:fdd22bb7aa52 | 101 | */ |
emilmont | 1:fdd22bb7aa52 | 102 | |
emilmont | 1:fdd22bb7aa52 | 103 | void arm_radix2_butterfly_q15( |
emilmont | 1:fdd22bb7aa52 | 104 | q15_t * pSrc, |
emilmont | 1:fdd22bb7aa52 | 105 | uint32_t fftLen, |
emilmont | 1:fdd22bb7aa52 | 106 | q15_t * pCoef, |
emilmont | 1:fdd22bb7aa52 | 107 | uint16_t twidCoefModifier) |
emilmont | 1:fdd22bb7aa52 | 108 | { |
mbed_official | 3:7a284390b0ce | 109 | #ifndef ARM_MATH_CM0_FAMILY |
emilmont | 1:fdd22bb7aa52 | 110 | |
mbed_official | 3:7a284390b0ce | 111 | unsigned i, j, k, l; |
mbed_official | 3:7a284390b0ce | 112 | unsigned n1, n2, ia; |
emilmont | 1:fdd22bb7aa52 | 113 | q15_t in; |
emilmont | 1:fdd22bb7aa52 | 114 | q31_t T, S, R; |
emilmont | 1:fdd22bb7aa52 | 115 | q31_t coeff, out1, out2; |
emilmont | 1:fdd22bb7aa52 | 116 | |
emilmont | 1:fdd22bb7aa52 | 117 | //N = fftLen; |
emilmont | 1:fdd22bb7aa52 | 118 | n2 = fftLen; |
emilmont | 1:fdd22bb7aa52 | 119 | |
emilmont | 1:fdd22bb7aa52 | 120 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 121 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 122 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 123 | |
emilmont | 1:fdd22bb7aa52 | 124 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 125 | for (i = 0; i < n2; i++) |
emilmont | 1:fdd22bb7aa52 | 126 | { |
emilmont | 1:fdd22bb7aa52 | 127 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 128 | |
emilmont | 1:fdd22bb7aa52 | 129 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 130 | |
emilmont | 1:fdd22bb7aa52 | 131 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 132 | |
emilmont | 1:fdd22bb7aa52 | 133 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
mbed_official | 5:3762170b6d4d | 134 | in = ((int16_t) (T & 0xFFFF)) >> 1; |
mbed_official | 5:3762170b6d4d | 135 | T = ((T >> 1) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 136 | |
emilmont | 1:fdd22bb7aa52 | 137 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
mbed_official | 5:3762170b6d4d | 138 | in = ((int16_t) (S & 0xFFFF)) >> 1; |
mbed_official | 5:3762170b6d4d | 139 | S = ((S >> 1) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 140 | |
emilmont | 1:fdd22bb7aa52 | 141 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 142 | |
emilmont | 1:fdd22bb7aa52 | 143 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 144 | |
emilmont | 1:fdd22bb7aa52 | 145 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 146 | |
emilmont | 1:fdd22bb7aa52 | 147 | out1 = __SMUAD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 148 | out2 = __SMUSDX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 149 | |
emilmont | 1:fdd22bb7aa52 | 150 | #else |
emilmont | 1:fdd22bb7aa52 | 151 | |
emilmont | 1:fdd22bb7aa52 | 152 | out1 = __SMUSDX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 153 | out2 = __SMUAD(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 154 | |
emilmont | 1:fdd22bb7aa52 | 155 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 156 | |
emilmont | 1:fdd22bb7aa52 | 157 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 158 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 159 | |
emilmont | 1:fdd22bb7aa52 | 160 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 161 | |
emilmont | 1:fdd22bb7aa52 | 162 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 163 | |
emilmont | 1:fdd22bb7aa52 | 164 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 165 | i++; |
emilmont | 1:fdd22bb7aa52 | 166 | l++; |
emilmont | 1:fdd22bb7aa52 | 167 | |
emilmont | 1:fdd22bb7aa52 | 168 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
mbed_official | 5:3762170b6d4d | 169 | in = ((int16_t) (T & 0xFFFF)) >> 1; |
mbed_official | 5:3762170b6d4d | 170 | T = ((T >> 1) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 171 | |
emilmont | 1:fdd22bb7aa52 | 172 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
mbed_official | 5:3762170b6d4d | 173 | in = ((int16_t) (S & 0xFFFF)) >> 1; |
mbed_official | 5:3762170b6d4d | 174 | S = ((S >> 1) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 175 | |
emilmont | 1:fdd22bb7aa52 | 176 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 177 | |
emilmont | 1:fdd22bb7aa52 | 178 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 179 | |
emilmont | 1:fdd22bb7aa52 | 180 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 181 | |
emilmont | 1:fdd22bb7aa52 | 182 | out1 = __SMUAD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 183 | out2 = __SMUSDX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 184 | |
emilmont | 1:fdd22bb7aa52 | 185 | #else |
emilmont | 1:fdd22bb7aa52 | 186 | |
emilmont | 1:fdd22bb7aa52 | 187 | out1 = __SMUSDX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 188 | out2 = __SMUAD(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 189 | |
emilmont | 1:fdd22bb7aa52 | 190 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 191 | |
emilmont | 1:fdd22bb7aa52 | 192 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 193 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 194 | |
emilmont | 1:fdd22bb7aa52 | 195 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 196 | |
emilmont | 1:fdd22bb7aa52 | 197 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 198 | |
emilmont | 1:fdd22bb7aa52 | 199 | // loop for stage |
emilmont | 1:fdd22bb7aa52 | 200 | for (k = fftLen / 2; k > 2; k = k >> 1) |
emilmont | 1:fdd22bb7aa52 | 201 | { |
emilmont | 1:fdd22bb7aa52 | 202 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 203 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 204 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 205 | |
emilmont | 1:fdd22bb7aa52 | 206 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 207 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 208 | { |
emilmont | 1:fdd22bb7aa52 | 209 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 210 | |
emilmont | 1:fdd22bb7aa52 | 211 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 212 | |
emilmont | 1:fdd22bb7aa52 | 213 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 214 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 215 | { |
emilmont | 1:fdd22bb7aa52 | 216 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 217 | |
emilmont | 1:fdd22bb7aa52 | 218 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 219 | |
emilmont | 1:fdd22bb7aa52 | 220 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 221 | |
emilmont | 1:fdd22bb7aa52 | 222 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 223 | |
emilmont | 1:fdd22bb7aa52 | 224 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 225 | |
emilmont | 1:fdd22bb7aa52 | 226 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 227 | |
emilmont | 1:fdd22bb7aa52 | 228 | out1 = __SMUAD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 229 | out2 = __SMUSDX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 230 | |
emilmont | 1:fdd22bb7aa52 | 231 | #else |
emilmont | 1:fdd22bb7aa52 | 232 | |
emilmont | 1:fdd22bb7aa52 | 233 | out1 = __SMUSDX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 234 | out2 = __SMUAD(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 235 | |
emilmont | 1:fdd22bb7aa52 | 236 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 237 | |
emilmont | 1:fdd22bb7aa52 | 238 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 239 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 240 | |
emilmont | 1:fdd22bb7aa52 | 241 | i += n1; |
emilmont | 1:fdd22bb7aa52 | 242 | |
emilmont | 1:fdd22bb7aa52 | 243 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 244 | |
emilmont | 1:fdd22bb7aa52 | 245 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 246 | |
emilmont | 1:fdd22bb7aa52 | 247 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 248 | |
emilmont | 1:fdd22bb7aa52 | 249 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 250 | |
emilmont | 1:fdd22bb7aa52 | 251 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 252 | |
emilmont | 1:fdd22bb7aa52 | 253 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 254 | |
emilmont | 1:fdd22bb7aa52 | 255 | out1 = __SMUAD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 256 | out2 = __SMUSDX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 257 | |
emilmont | 1:fdd22bb7aa52 | 258 | #else |
emilmont | 1:fdd22bb7aa52 | 259 | |
emilmont | 1:fdd22bb7aa52 | 260 | out1 = __SMUSDX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 261 | out2 = __SMUAD(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 262 | |
emilmont | 1:fdd22bb7aa52 | 263 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 264 | |
emilmont | 1:fdd22bb7aa52 | 265 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 266 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 267 | |
emilmont | 1:fdd22bb7aa52 | 268 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 269 | |
emilmont | 1:fdd22bb7aa52 | 270 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 271 | |
emilmont | 1:fdd22bb7aa52 | 272 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 273 | } // stages loop end |
emilmont | 1:fdd22bb7aa52 | 274 | |
emilmont | 1:fdd22bb7aa52 | 275 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 276 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 277 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 278 | |
emilmont | 1:fdd22bb7aa52 | 279 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 280 | |
emilmont | 1:fdd22bb7aa52 | 281 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 282 | |
emilmont | 1:fdd22bb7aa52 | 283 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 284 | for (i = 0; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 285 | { |
emilmont | 1:fdd22bb7aa52 | 286 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 287 | |
emilmont | 1:fdd22bb7aa52 | 288 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 289 | |
emilmont | 1:fdd22bb7aa52 | 290 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 291 | |
emilmont | 1:fdd22bb7aa52 | 292 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 293 | |
emilmont | 1:fdd22bb7aa52 | 294 | _SIMD32_OFFSET(pSrc + (2 * i)) = __QADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 295 | |
emilmont | 1:fdd22bb7aa52 | 296 | _SIMD32_OFFSET(pSrc + (2u * l)) = R; |
emilmont | 1:fdd22bb7aa52 | 297 | |
emilmont | 1:fdd22bb7aa52 | 298 | i += n1; |
emilmont | 1:fdd22bb7aa52 | 299 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 300 | |
emilmont | 1:fdd22bb7aa52 | 301 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 302 | |
emilmont | 1:fdd22bb7aa52 | 303 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 304 | |
emilmont | 1:fdd22bb7aa52 | 305 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 306 | |
emilmont | 1:fdd22bb7aa52 | 307 | _SIMD32_OFFSET(pSrc + (2 * i)) = __QADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 308 | |
emilmont | 1:fdd22bb7aa52 | 309 | _SIMD32_OFFSET(pSrc + (2u * l)) = R; |
emilmont | 1:fdd22bb7aa52 | 310 | |
emilmont | 1:fdd22bb7aa52 | 311 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 312 | |
emilmont | 1:fdd22bb7aa52 | 313 | |
emilmont | 1:fdd22bb7aa52 | 314 | #else |
emilmont | 1:fdd22bb7aa52 | 315 | |
mbed_official | 3:7a284390b0ce | 316 | unsigned i, j, k, l; |
mbed_official | 3:7a284390b0ce | 317 | unsigned n1, n2, ia; |
emilmont | 1:fdd22bb7aa52 | 318 | q15_t xt, yt, cosVal, sinVal; |
emilmont | 1:fdd22bb7aa52 | 319 | |
emilmont | 1:fdd22bb7aa52 | 320 | |
emilmont | 1:fdd22bb7aa52 | 321 | //N = fftLen; |
emilmont | 1:fdd22bb7aa52 | 322 | n2 = fftLen; |
emilmont | 1:fdd22bb7aa52 | 323 | |
emilmont | 1:fdd22bb7aa52 | 324 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 325 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 326 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 327 | |
emilmont | 1:fdd22bb7aa52 | 328 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 329 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 330 | { |
emilmont | 1:fdd22bb7aa52 | 331 | cosVal = pCoef[ia * 2]; |
emilmont | 1:fdd22bb7aa52 | 332 | sinVal = pCoef[(ia * 2) + 1]; |
emilmont | 1:fdd22bb7aa52 | 333 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 334 | |
emilmont | 1:fdd22bb7aa52 | 335 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 336 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 337 | { |
emilmont | 1:fdd22bb7aa52 | 338 | l = i + n2; |
mbed_official | 5:3762170b6d4d | 339 | xt = (pSrc[2 * i] >> 1u) - (pSrc[2 * l] >> 1u); |
mbed_official | 5:3762170b6d4d | 340 | pSrc[2 * i] = ((pSrc[2 * i] >> 1u) + (pSrc[2 * l] >> 1u)) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 341 | |
mbed_official | 5:3762170b6d4d | 342 | yt = (pSrc[2 * i + 1] >> 1u) - (pSrc[2 * l + 1] >> 1u); |
emilmont | 1:fdd22bb7aa52 | 343 | pSrc[2 * i + 1] = |
mbed_official | 5:3762170b6d4d | 344 | ((pSrc[2 * l + 1] >> 1u) + (pSrc[2 * i + 1] >> 1u)) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 345 | |
emilmont | 1:fdd22bb7aa52 | 346 | pSrc[2u * l] = (((int16_t) (((q31_t) xt * cosVal) >> 16)) + |
emilmont | 1:fdd22bb7aa52 | 347 | ((int16_t) (((q31_t) yt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 348 | |
emilmont | 1:fdd22bb7aa52 | 349 | pSrc[2u * l + 1u] = (((int16_t) (((q31_t) yt * cosVal) >> 16)) - |
emilmont | 1:fdd22bb7aa52 | 350 | ((int16_t) (((q31_t) xt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 351 | |
emilmont | 1:fdd22bb7aa52 | 352 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 353 | |
emilmont | 1:fdd22bb7aa52 | 354 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 355 | |
emilmont | 1:fdd22bb7aa52 | 356 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 357 | |
emilmont | 1:fdd22bb7aa52 | 358 | // loop for stage |
emilmont | 1:fdd22bb7aa52 | 359 | for (k = fftLen / 2; k > 2; k = k >> 1) |
emilmont | 1:fdd22bb7aa52 | 360 | { |
emilmont | 1:fdd22bb7aa52 | 361 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 362 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 363 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 364 | |
emilmont | 1:fdd22bb7aa52 | 365 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 366 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 367 | { |
emilmont | 1:fdd22bb7aa52 | 368 | cosVal = pCoef[ia * 2]; |
emilmont | 1:fdd22bb7aa52 | 369 | sinVal = pCoef[(ia * 2) + 1]; |
emilmont | 1:fdd22bb7aa52 | 370 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 371 | |
emilmont | 1:fdd22bb7aa52 | 372 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 373 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 374 | { |
emilmont | 1:fdd22bb7aa52 | 375 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 376 | xt = pSrc[2 * i] - pSrc[2 * l]; |
emilmont | 1:fdd22bb7aa52 | 377 | pSrc[2 * i] = (pSrc[2 * i] + pSrc[2 * l]) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 378 | |
emilmont | 1:fdd22bb7aa52 | 379 | yt = pSrc[2 * i + 1] - pSrc[2 * l + 1]; |
emilmont | 1:fdd22bb7aa52 | 380 | pSrc[2 * i + 1] = (pSrc[2 * l + 1] + pSrc[2 * i + 1]) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 381 | |
emilmont | 1:fdd22bb7aa52 | 382 | pSrc[2u * l] = (((int16_t) (((q31_t) xt * cosVal) >> 16)) + |
emilmont | 1:fdd22bb7aa52 | 383 | ((int16_t) (((q31_t) yt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 384 | |
emilmont | 1:fdd22bb7aa52 | 385 | pSrc[2u * l + 1u] = (((int16_t) (((q31_t) yt * cosVal) >> 16)) - |
emilmont | 1:fdd22bb7aa52 | 386 | ((int16_t) (((q31_t) xt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 387 | |
emilmont | 1:fdd22bb7aa52 | 388 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 389 | |
emilmont | 1:fdd22bb7aa52 | 390 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 391 | |
emilmont | 1:fdd22bb7aa52 | 392 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 393 | } // stages loop end |
emilmont | 1:fdd22bb7aa52 | 394 | |
emilmont | 1:fdd22bb7aa52 | 395 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 396 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 397 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 398 | |
emilmont | 1:fdd22bb7aa52 | 399 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 400 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 401 | { |
emilmont | 1:fdd22bb7aa52 | 402 | cosVal = pCoef[ia * 2]; |
emilmont | 1:fdd22bb7aa52 | 403 | sinVal = pCoef[(ia * 2) + 1]; |
emilmont | 1:fdd22bb7aa52 | 404 | |
emilmont | 1:fdd22bb7aa52 | 405 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 406 | |
emilmont | 1:fdd22bb7aa52 | 407 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 408 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 409 | { |
emilmont | 1:fdd22bb7aa52 | 410 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 411 | xt = pSrc[2 * i] - pSrc[2 * l]; |
emilmont | 1:fdd22bb7aa52 | 412 | pSrc[2 * i] = (pSrc[2 * i] + pSrc[2 * l]); |
emilmont | 1:fdd22bb7aa52 | 413 | |
emilmont | 1:fdd22bb7aa52 | 414 | yt = pSrc[2 * i + 1] - pSrc[2 * l + 1]; |
emilmont | 1:fdd22bb7aa52 | 415 | pSrc[2 * i + 1] = (pSrc[2 * l + 1] + pSrc[2 * i + 1]); |
emilmont | 1:fdd22bb7aa52 | 416 | |
emilmont | 1:fdd22bb7aa52 | 417 | pSrc[2u * l] = xt; |
emilmont | 1:fdd22bb7aa52 | 418 | |
emilmont | 1:fdd22bb7aa52 | 419 | pSrc[2u * l + 1u] = yt; |
emilmont | 1:fdd22bb7aa52 | 420 | |
emilmont | 1:fdd22bb7aa52 | 421 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 422 | |
emilmont | 1:fdd22bb7aa52 | 423 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 424 | |
emilmont | 1:fdd22bb7aa52 | 425 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 426 | |
mbed_official | 3:7a284390b0ce | 427 | #endif // #ifndef ARM_MATH_CM0_FAMILY |
emilmont | 1:fdd22bb7aa52 | 428 | |
emilmont | 1:fdd22bb7aa52 | 429 | } |
emilmont | 1:fdd22bb7aa52 | 430 | |
emilmont | 1:fdd22bb7aa52 | 431 | |
emilmont | 1:fdd22bb7aa52 | 432 | void arm_radix2_butterfly_inverse_q15( |
emilmont | 1:fdd22bb7aa52 | 433 | q15_t * pSrc, |
emilmont | 1:fdd22bb7aa52 | 434 | uint32_t fftLen, |
emilmont | 1:fdd22bb7aa52 | 435 | q15_t * pCoef, |
emilmont | 1:fdd22bb7aa52 | 436 | uint16_t twidCoefModifier) |
emilmont | 1:fdd22bb7aa52 | 437 | { |
mbed_official | 3:7a284390b0ce | 438 | #ifndef ARM_MATH_CM0_FAMILY |
emilmont | 1:fdd22bb7aa52 | 439 | |
mbed_official | 3:7a284390b0ce | 440 | unsigned i, j, k, l; |
mbed_official | 3:7a284390b0ce | 441 | unsigned n1, n2, ia; |
emilmont | 1:fdd22bb7aa52 | 442 | q15_t in; |
emilmont | 1:fdd22bb7aa52 | 443 | q31_t T, S, R; |
emilmont | 1:fdd22bb7aa52 | 444 | q31_t coeff, out1, out2; |
emilmont | 1:fdd22bb7aa52 | 445 | |
emilmont | 1:fdd22bb7aa52 | 446 | //N = fftLen; |
emilmont | 1:fdd22bb7aa52 | 447 | n2 = fftLen; |
emilmont | 1:fdd22bb7aa52 | 448 | |
emilmont | 1:fdd22bb7aa52 | 449 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 450 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 451 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 452 | |
emilmont | 1:fdd22bb7aa52 | 453 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 454 | for (i = 0; i < n2; i++) |
emilmont | 1:fdd22bb7aa52 | 455 | { |
emilmont | 1:fdd22bb7aa52 | 456 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 457 | |
emilmont | 1:fdd22bb7aa52 | 458 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 459 | |
emilmont | 1:fdd22bb7aa52 | 460 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 461 | |
emilmont | 1:fdd22bb7aa52 | 462 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
mbed_official | 5:3762170b6d4d | 463 | in = ((int16_t) (T & 0xFFFF)) >> 1; |
mbed_official | 5:3762170b6d4d | 464 | T = ((T >> 1) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 465 | |
emilmont | 1:fdd22bb7aa52 | 466 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
mbed_official | 5:3762170b6d4d | 467 | in = ((int16_t) (S & 0xFFFF)) >> 1; |
mbed_official | 5:3762170b6d4d | 468 | S = ((S >> 1) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 469 | |
emilmont | 1:fdd22bb7aa52 | 470 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 471 | |
emilmont | 1:fdd22bb7aa52 | 472 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 473 | |
emilmont | 1:fdd22bb7aa52 | 474 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 475 | |
emilmont | 1:fdd22bb7aa52 | 476 | out1 = __SMUSD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 477 | out2 = __SMUADX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 478 | #else |
emilmont | 1:fdd22bb7aa52 | 479 | |
emilmont | 1:fdd22bb7aa52 | 480 | out1 = __SMUADX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 481 | out2 = __SMUSD(__QSUB(0, coeff), R); |
emilmont | 1:fdd22bb7aa52 | 482 | |
emilmont | 1:fdd22bb7aa52 | 483 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 484 | |
emilmont | 1:fdd22bb7aa52 | 485 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 486 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 487 | |
emilmont | 1:fdd22bb7aa52 | 488 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 489 | |
emilmont | 1:fdd22bb7aa52 | 490 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 491 | |
emilmont | 1:fdd22bb7aa52 | 492 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 493 | i++; |
emilmont | 1:fdd22bb7aa52 | 494 | l++; |
emilmont | 1:fdd22bb7aa52 | 495 | |
emilmont | 1:fdd22bb7aa52 | 496 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
mbed_official | 5:3762170b6d4d | 497 | in = ((int16_t) (T & 0xFFFF)) >> 1; |
mbed_official | 5:3762170b6d4d | 498 | T = ((T >> 1) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 499 | |
emilmont | 1:fdd22bb7aa52 | 500 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
mbed_official | 5:3762170b6d4d | 501 | in = ((int16_t) (S & 0xFFFF)) >> 1; |
mbed_official | 5:3762170b6d4d | 502 | S = ((S >> 1) & 0xFFFF0000) | (in & 0xFFFF); |
emilmont | 1:fdd22bb7aa52 | 503 | |
emilmont | 1:fdd22bb7aa52 | 504 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 505 | |
emilmont | 1:fdd22bb7aa52 | 506 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 507 | |
emilmont | 1:fdd22bb7aa52 | 508 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 509 | |
emilmont | 1:fdd22bb7aa52 | 510 | out1 = __SMUSD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 511 | out2 = __SMUADX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 512 | #else |
emilmont | 1:fdd22bb7aa52 | 513 | |
emilmont | 1:fdd22bb7aa52 | 514 | out1 = __SMUADX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 515 | out2 = __SMUSD(__QSUB(0, coeff), R); |
emilmont | 1:fdd22bb7aa52 | 516 | |
emilmont | 1:fdd22bb7aa52 | 517 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 518 | |
emilmont | 1:fdd22bb7aa52 | 519 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 520 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 521 | |
emilmont | 1:fdd22bb7aa52 | 522 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 523 | |
emilmont | 1:fdd22bb7aa52 | 524 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 525 | |
emilmont | 1:fdd22bb7aa52 | 526 | // loop for stage |
emilmont | 1:fdd22bb7aa52 | 527 | for (k = fftLen / 2; k > 2; k = k >> 1) |
emilmont | 1:fdd22bb7aa52 | 528 | { |
emilmont | 1:fdd22bb7aa52 | 529 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 530 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 531 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 532 | |
emilmont | 1:fdd22bb7aa52 | 533 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 534 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 535 | { |
emilmont | 1:fdd22bb7aa52 | 536 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 537 | |
emilmont | 1:fdd22bb7aa52 | 538 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 539 | |
emilmont | 1:fdd22bb7aa52 | 540 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 541 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 542 | { |
emilmont | 1:fdd22bb7aa52 | 543 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 544 | |
emilmont | 1:fdd22bb7aa52 | 545 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 546 | |
emilmont | 1:fdd22bb7aa52 | 547 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 548 | |
emilmont | 1:fdd22bb7aa52 | 549 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 550 | |
emilmont | 1:fdd22bb7aa52 | 551 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 552 | |
emilmont | 1:fdd22bb7aa52 | 553 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 554 | |
emilmont | 1:fdd22bb7aa52 | 555 | out1 = __SMUSD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 556 | out2 = __SMUADX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 557 | |
emilmont | 1:fdd22bb7aa52 | 558 | #else |
emilmont | 1:fdd22bb7aa52 | 559 | |
emilmont | 1:fdd22bb7aa52 | 560 | out1 = __SMUADX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 561 | out2 = __SMUSD(__QSUB(0, coeff), R); |
emilmont | 1:fdd22bb7aa52 | 562 | |
emilmont | 1:fdd22bb7aa52 | 563 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 564 | |
emilmont | 1:fdd22bb7aa52 | 565 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 566 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 567 | |
emilmont | 1:fdd22bb7aa52 | 568 | i += n1; |
emilmont | 1:fdd22bb7aa52 | 569 | |
emilmont | 1:fdd22bb7aa52 | 570 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 571 | |
emilmont | 1:fdd22bb7aa52 | 572 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 573 | |
emilmont | 1:fdd22bb7aa52 | 574 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 575 | |
emilmont | 1:fdd22bb7aa52 | 576 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 577 | |
emilmont | 1:fdd22bb7aa52 | 578 | _SIMD32_OFFSET(pSrc + (2 * i)) = __SHADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 579 | |
emilmont | 1:fdd22bb7aa52 | 580 | #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 581 | |
emilmont | 1:fdd22bb7aa52 | 582 | out1 = __SMUSD(coeff, R) >> 16; |
emilmont | 1:fdd22bb7aa52 | 583 | out2 = __SMUADX(coeff, R); |
emilmont | 1:fdd22bb7aa52 | 584 | #else |
emilmont | 1:fdd22bb7aa52 | 585 | |
emilmont | 1:fdd22bb7aa52 | 586 | out1 = __SMUADX(R, coeff) >> 16u; |
emilmont | 1:fdd22bb7aa52 | 587 | out2 = __SMUSD(__QSUB(0, coeff), R); |
emilmont | 1:fdd22bb7aa52 | 588 | |
emilmont | 1:fdd22bb7aa52 | 589 | #endif // #ifndef ARM_MATH_BIG_ENDIAN |
emilmont | 1:fdd22bb7aa52 | 590 | |
emilmont | 1:fdd22bb7aa52 | 591 | _SIMD32_OFFSET(pSrc + (2u * l)) = |
emilmont | 1:fdd22bb7aa52 | 592 | (q31_t) ((out2) & 0xFFFF0000) | (out1 & 0x0000FFFF); |
emilmont | 1:fdd22bb7aa52 | 593 | |
emilmont | 1:fdd22bb7aa52 | 594 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 595 | |
emilmont | 1:fdd22bb7aa52 | 596 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 597 | |
emilmont | 1:fdd22bb7aa52 | 598 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 599 | } // stages loop end |
emilmont | 1:fdd22bb7aa52 | 600 | |
emilmont | 1:fdd22bb7aa52 | 601 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 602 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 603 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 604 | |
emilmont | 1:fdd22bb7aa52 | 605 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 606 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 607 | { |
emilmont | 1:fdd22bb7aa52 | 608 | coeff = _SIMD32_OFFSET(pCoef + (ia * 2u)); |
emilmont | 1:fdd22bb7aa52 | 609 | |
emilmont | 1:fdd22bb7aa52 | 610 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 611 | |
emilmont | 1:fdd22bb7aa52 | 612 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 613 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 614 | { |
emilmont | 1:fdd22bb7aa52 | 615 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 616 | |
emilmont | 1:fdd22bb7aa52 | 617 | T = _SIMD32_OFFSET(pSrc + (2 * i)); |
emilmont | 1:fdd22bb7aa52 | 618 | |
emilmont | 1:fdd22bb7aa52 | 619 | S = _SIMD32_OFFSET(pSrc + (2 * l)); |
emilmont | 1:fdd22bb7aa52 | 620 | |
emilmont | 1:fdd22bb7aa52 | 621 | R = __QSUB16(T, S); |
emilmont | 1:fdd22bb7aa52 | 622 | |
emilmont | 1:fdd22bb7aa52 | 623 | _SIMD32_OFFSET(pSrc + (2 * i)) = __QADD16(T, S); |
emilmont | 1:fdd22bb7aa52 | 624 | |
emilmont | 1:fdd22bb7aa52 | 625 | _SIMD32_OFFSET(pSrc + (2u * l)) = R; |
emilmont | 1:fdd22bb7aa52 | 626 | |
emilmont | 1:fdd22bb7aa52 | 627 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 628 | |
emilmont | 1:fdd22bb7aa52 | 629 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 630 | |
emilmont | 1:fdd22bb7aa52 | 631 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 632 | |
emilmont | 1:fdd22bb7aa52 | 633 | #else |
emilmont | 1:fdd22bb7aa52 | 634 | |
emilmont | 1:fdd22bb7aa52 | 635 | |
mbed_official | 3:7a284390b0ce | 636 | unsigned i, j, k, l; |
mbed_official | 3:7a284390b0ce | 637 | unsigned n1, n2, ia; |
emilmont | 1:fdd22bb7aa52 | 638 | q15_t xt, yt, cosVal, sinVal; |
emilmont | 1:fdd22bb7aa52 | 639 | |
emilmont | 1:fdd22bb7aa52 | 640 | //N = fftLen; |
emilmont | 1:fdd22bb7aa52 | 641 | n2 = fftLen; |
emilmont | 1:fdd22bb7aa52 | 642 | |
emilmont | 1:fdd22bb7aa52 | 643 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 644 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 645 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 646 | |
emilmont | 1:fdd22bb7aa52 | 647 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 648 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 649 | { |
emilmont | 1:fdd22bb7aa52 | 650 | cosVal = pCoef[ia * 2]; |
emilmont | 1:fdd22bb7aa52 | 651 | sinVal = pCoef[(ia * 2) + 1]; |
emilmont | 1:fdd22bb7aa52 | 652 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 653 | |
emilmont | 1:fdd22bb7aa52 | 654 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 655 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 656 | { |
emilmont | 1:fdd22bb7aa52 | 657 | l = i + n2; |
mbed_official | 5:3762170b6d4d | 658 | xt = (pSrc[2 * i] >> 1u) - (pSrc[2 * l] >> 1u); |
mbed_official | 5:3762170b6d4d | 659 | pSrc[2 * i] = ((pSrc[2 * i] >> 1u) + (pSrc[2 * l] >> 1u)) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 660 | |
mbed_official | 5:3762170b6d4d | 661 | yt = (pSrc[2 * i + 1] >> 1u) - (pSrc[2 * l + 1] >> 1u); |
emilmont | 1:fdd22bb7aa52 | 662 | pSrc[2 * i + 1] = |
mbed_official | 5:3762170b6d4d | 663 | ((pSrc[2 * l + 1] >> 1u) + (pSrc[2 * i + 1] >> 1u)) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 664 | |
emilmont | 1:fdd22bb7aa52 | 665 | pSrc[2u * l] = (((int16_t) (((q31_t) xt * cosVal) >> 16)) - |
emilmont | 1:fdd22bb7aa52 | 666 | ((int16_t) (((q31_t) yt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 667 | |
emilmont | 1:fdd22bb7aa52 | 668 | pSrc[2u * l + 1u] = (((int16_t) (((q31_t) yt * cosVal) >> 16)) + |
emilmont | 1:fdd22bb7aa52 | 669 | ((int16_t) (((q31_t) xt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 670 | |
emilmont | 1:fdd22bb7aa52 | 671 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 672 | |
emilmont | 1:fdd22bb7aa52 | 673 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 674 | |
emilmont | 1:fdd22bb7aa52 | 675 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 676 | |
emilmont | 1:fdd22bb7aa52 | 677 | // loop for stage |
emilmont | 1:fdd22bb7aa52 | 678 | for (k = fftLen / 2; k > 2; k = k >> 1) |
emilmont | 1:fdd22bb7aa52 | 679 | { |
emilmont | 1:fdd22bb7aa52 | 680 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 681 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 682 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 683 | |
emilmont | 1:fdd22bb7aa52 | 684 | // loop for groups |
emilmont | 1:fdd22bb7aa52 | 685 | for (j = 0; j < n2; j++) |
emilmont | 1:fdd22bb7aa52 | 686 | { |
emilmont | 1:fdd22bb7aa52 | 687 | cosVal = pCoef[ia * 2]; |
emilmont | 1:fdd22bb7aa52 | 688 | sinVal = pCoef[(ia * 2) + 1]; |
emilmont | 1:fdd22bb7aa52 | 689 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 690 | |
emilmont | 1:fdd22bb7aa52 | 691 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 692 | for (i = j; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 693 | { |
emilmont | 1:fdd22bb7aa52 | 694 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 695 | xt = pSrc[2 * i] - pSrc[2 * l]; |
emilmont | 1:fdd22bb7aa52 | 696 | pSrc[2 * i] = (pSrc[2 * i] + pSrc[2 * l]) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 697 | |
emilmont | 1:fdd22bb7aa52 | 698 | yt = pSrc[2 * i + 1] - pSrc[2 * l + 1]; |
emilmont | 1:fdd22bb7aa52 | 699 | pSrc[2 * i + 1] = (pSrc[2 * l + 1] + pSrc[2 * i + 1]) >> 1u; |
emilmont | 1:fdd22bb7aa52 | 700 | |
emilmont | 1:fdd22bb7aa52 | 701 | pSrc[2u * l] = (((int16_t) (((q31_t) xt * cosVal) >> 16)) - |
emilmont | 1:fdd22bb7aa52 | 702 | ((int16_t) (((q31_t) yt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 703 | |
emilmont | 1:fdd22bb7aa52 | 704 | pSrc[2u * l + 1u] = (((int16_t) (((q31_t) yt * cosVal) >> 16)) + |
emilmont | 1:fdd22bb7aa52 | 705 | ((int16_t) (((q31_t) xt * sinVal) >> 16))); |
emilmont | 1:fdd22bb7aa52 | 706 | |
emilmont | 1:fdd22bb7aa52 | 707 | } // butterfly loop end |
emilmont | 1:fdd22bb7aa52 | 708 | |
emilmont | 1:fdd22bb7aa52 | 709 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 710 | |
emilmont | 1:fdd22bb7aa52 | 711 | twidCoefModifier = twidCoefModifier << 1u; |
emilmont | 1:fdd22bb7aa52 | 712 | } // stages loop end |
emilmont | 1:fdd22bb7aa52 | 713 | |
emilmont | 1:fdd22bb7aa52 | 714 | n1 = n2; |
emilmont | 1:fdd22bb7aa52 | 715 | n2 = n2 >> 1; |
emilmont | 1:fdd22bb7aa52 | 716 | ia = 0; |
emilmont | 1:fdd22bb7aa52 | 717 | |
emilmont | 1:fdd22bb7aa52 | 718 | cosVal = pCoef[ia * 2]; |
emilmont | 1:fdd22bb7aa52 | 719 | sinVal = pCoef[(ia * 2) + 1]; |
emilmont | 1:fdd22bb7aa52 | 720 | |
emilmont | 1:fdd22bb7aa52 | 721 | ia = ia + twidCoefModifier; |
emilmont | 1:fdd22bb7aa52 | 722 | |
emilmont | 1:fdd22bb7aa52 | 723 | // loop for butterfly |
emilmont | 1:fdd22bb7aa52 | 724 | for (i = 0; i < fftLen; i += n1) |
emilmont | 1:fdd22bb7aa52 | 725 | { |
emilmont | 1:fdd22bb7aa52 | 726 | l = i + n2; |
emilmont | 1:fdd22bb7aa52 | 727 | xt = pSrc[2 * i] - pSrc[2 * l]; |
emilmont | 1:fdd22bb7aa52 | 728 | pSrc[2 * i] = (pSrc[2 * i] + pSrc[2 * l]); |
emilmont | 1:fdd22bb7aa52 | 729 | |
emilmont | 1:fdd22bb7aa52 | 730 | yt = pSrc[2 * i + 1] - pSrc[2 * l + 1]; |
emilmont | 1:fdd22bb7aa52 | 731 | pSrc[2 * i + 1] = (pSrc[2 * l + 1] + pSrc[2 * i + 1]); |
emilmont | 1:fdd22bb7aa52 | 732 | |
emilmont | 1:fdd22bb7aa52 | 733 | pSrc[2u * l] = xt; |
emilmont | 1:fdd22bb7aa52 | 734 | |
emilmont | 1:fdd22bb7aa52 | 735 | pSrc[2u * l + 1u] = yt; |
emilmont | 1:fdd22bb7aa52 | 736 | |
emilmont | 1:fdd22bb7aa52 | 737 | } // groups loop end |
emilmont | 1:fdd22bb7aa52 | 738 | |
emilmont | 1:fdd22bb7aa52 | 739 | |
mbed_official | 3:7a284390b0ce | 740 | #endif // #ifndef ARM_MATH_CM0_FAMILY |
emilmont | 1:fdd22bb7aa52 | 741 | |
emilmont | 1:fdd22bb7aa52 | 742 | } |