mbed official / mbed-dsp

CMSIS DSP library

Dependents:   performance_timer Surfboard_ gps2rtty Capstone ... more

Legacy Warning

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

cmsis_dsp/FilteringFunctions/arm_fir_sparse_q15.c@2:da51fb522205, 2013-05-30 (annotated)

Committer:
emilmont
Date:
Thu May 30 17:10:11 2013 +0100
Revision:
2:da51fb522205
Parent:
1:fdd22bb7aa52
Child:
3:7a284390b0ce
Keep "cmsis-dsp" module in synch with its source

Who changed what in which revision?

UserRevisionLine numberNew contents of line
emilmont 1:fdd22bb7aa52 1 /* ----------------------------------------------------------------------
emilmont 1:fdd22bb7aa52 2 * Copyright (C) 2010 ARM Limited. All rights reserved.
emilmont 1:fdd22bb7aa52 3 *
emilmont 1:fdd22bb7aa52 4 * $Date: 15. February 2012
emilmont 2:da51fb522205 5 * $Revision: V1.1.0
emilmont 1:fdd22bb7aa52 6 *
emilmont 2:da51fb522205 7 * Project: CMSIS DSP Library
emilmont 2:da51fb522205 8 * Title: arm_fir_sparse_q15.c
emilmont 1:fdd22bb7aa52 9 *
emilmont 2:da51fb522205 10 * Description: Q15 sparse FIR filter processing function.
emilmont 1:fdd22bb7aa52 11 *
emilmont 1:fdd22bb7aa52 12 * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
emilmont 1:fdd22bb7aa52 13 *
emilmont 1:fdd22bb7aa52 14 * Version 1.1.0 2012/02/15
emilmont 1:fdd22bb7aa52 15 * Updated with more optimizations, bug fixes and minor API changes.
emilmont 1:fdd22bb7aa52 16 *
emilmont 1:fdd22bb7aa52 17 * Version 1.0.10 2011/7/15
emilmont 1:fdd22bb7aa52 18 * Big Endian support added and Merged M0 and M3/M4 Source code.
emilmont 1:fdd22bb7aa52 19 *
emilmont 1:fdd22bb7aa52 20 * Version 1.0.3 2010/11/29
emilmont 1:fdd22bb7aa52 21 * Re-organized the CMSIS folders and updated documentation.
emilmont 1:fdd22bb7aa52 22 *
emilmont 1:fdd22bb7aa52 23 * Version 1.0.2 2010/11/11
emilmont 1:fdd22bb7aa52 24 * Documentation updated.
emilmont 1:fdd22bb7aa52 25 *
emilmont 1:fdd22bb7aa52 26 * Version 1.0.1 2010/10/05
emilmont 1:fdd22bb7aa52 27 * Production release and review comments incorporated.
emilmont 1:fdd22bb7aa52 28 *
emilmont 1:fdd22bb7aa52 29 * Version 1.0.0 2010/09/20
emilmont 1:fdd22bb7aa52 30 * Production release and review comments incorporated
emilmont 1:fdd22bb7aa52 31 *
emilmont 1:fdd22bb7aa52 32 * Version 0.0.7 2010/06/10
emilmont 1:fdd22bb7aa52 33 * Misra-C changes done
emilmont 1:fdd22bb7aa52 34 * ------------------------------------------------------------------- */
emilmont 1:fdd22bb7aa52 35 #include "arm_math.h"
emilmont 1:fdd22bb7aa52 36
emilmont 1:fdd22bb7aa52 37 /**
emilmont 1:fdd22bb7aa52 38 * @addtogroup FIR_Sparse
emilmont 1:fdd22bb7aa52 39 * @{
emilmont 1:fdd22bb7aa52 40 */
emilmont 1:fdd22bb7aa52 41
emilmont 1:fdd22bb7aa52 42 /**
emilmont 1:fdd22bb7aa52 43 * @brief Processing function for the Q15 sparse FIR filter.
emilmont 1:fdd22bb7aa52 44 * @param[in] *S points to an instance of the Q15 sparse FIR structure.
emilmont 1:fdd22bb7aa52 45 * @param[in] *pSrc points to the block of input data.
emilmont 1:fdd22bb7aa52 46 * @param[out] *pDst points to the block of output data
emilmont 1:fdd22bb7aa52 47 * @param[in] *pScratchIn points to a temporary buffer of size blockSize.
emilmont 1:fdd22bb7aa52 48 * @param[in] *pScratchOut points to a temporary buffer of size blockSize.
emilmont 1:fdd22bb7aa52 49 * @param[in] blockSize number of input samples to process per call.
emilmont 1:fdd22bb7aa52 50 * @return none.
emilmont 1:fdd22bb7aa52 51 *
emilmont 1:fdd22bb7aa52 52 * <b>Scaling and Overflow Behavior:</b>
emilmont 1:fdd22bb7aa52 53 * \par
emilmont 1:fdd22bb7aa52 54 * The function is implemented using an internal 32-bit accumulator.
emilmont 1:fdd22bb7aa52 55 * The 1.15 x 1.15 multiplications yield a 2.30 result and these are added to a 2.30 accumulator.
emilmont 1:fdd22bb7aa52 56 * Thus the full precision of the multiplications is maintained but there is only a single guard bit in the accumulator.
emilmont 1:fdd22bb7aa52 57 * If the accumulator result overflows it will wrap around rather than saturate.
emilmont 1:fdd22bb7aa52 58 * After all multiply-accumulates are performed, the 2.30 accumulator is truncated to 2.15 format and then saturated to 1.15 format.
emilmont 1:fdd22bb7aa52 59 * In order to avoid overflows the input signal or coefficients must be scaled down by log2(numTaps) bits.
emilmont 1:fdd22bb7aa52 60 */
emilmont 1:fdd22bb7aa52 61
emilmont 1:fdd22bb7aa52 62
emilmont 1:fdd22bb7aa52 63 void arm_fir_sparse_q15(
emilmont 1:fdd22bb7aa52 64 arm_fir_sparse_instance_q15 * S,
emilmont 1:fdd22bb7aa52 65 q15_t * pSrc,
emilmont 1:fdd22bb7aa52 66 q15_t * pDst,
emilmont 1:fdd22bb7aa52 67 q15_t * pScratchIn,
emilmont 1:fdd22bb7aa52 68 q31_t * pScratchOut,
emilmont 1:fdd22bb7aa52 69 uint32_t blockSize)
emilmont 1:fdd22bb7aa52 70 {
emilmont 1:fdd22bb7aa52 71
emilmont 1:fdd22bb7aa52 72 q15_t *pState = S->pState; /* State pointer */
emilmont 1:fdd22bb7aa52 73 q15_t *pIn = pSrc; /* Working pointer for input */
emilmont 1:fdd22bb7aa52 74 q15_t *pOut = pDst; /* Working pointer for output */
emilmont 1:fdd22bb7aa52 75 q15_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
emilmont 1:fdd22bb7aa52 76 q15_t *px; /* Temporary pointers for scratch buffer */
emilmont 1:fdd22bb7aa52 77 q15_t *pb = pScratchIn; /* Temporary pointers for scratch buffer */
emilmont 1:fdd22bb7aa52 78 q15_t *py = pState; /* Temporary pointers for state buffer */
emilmont 1:fdd22bb7aa52 79 int32_t *pTapDelay = S->pTapDelay; /* Pointer to the array containing offset of the non-zero tap values. */
emilmont 1:fdd22bb7aa52 80 uint32_t delaySize = S->maxDelay + blockSize; /* state length */
emilmont 1:fdd22bb7aa52 81 uint16_t numTaps = S->numTaps; /* Filter order */
emilmont 1:fdd22bb7aa52 82 int32_t readIndex; /* Read index of the state buffer */
emilmont 1:fdd22bb7aa52 83 uint32_t tapCnt, blkCnt; /* loop counters */
emilmont 1:fdd22bb7aa52 84 q15_t coeff = *pCoeffs++; /* Read the first coefficient value */
emilmont 1:fdd22bb7aa52 85 q31_t *pScr2 = pScratchOut; /* Working pointer for pScratchOut */
emilmont 1:fdd22bb7aa52 86
emilmont 1:fdd22bb7aa52 87
emilmont 1:fdd22bb7aa52 88 #ifndef ARM_MATH_CM0
emilmont 1:fdd22bb7aa52 89
emilmont 1:fdd22bb7aa52 90 /* Run the below code for Cortex-M4 and Cortex-M3 */
emilmont 1:fdd22bb7aa52 91
emilmont 1:fdd22bb7aa52 92 q31_t in1, in2; /* Temporary variables */
emilmont 1:fdd22bb7aa52 93
emilmont 1:fdd22bb7aa52 94
emilmont 1:fdd22bb7aa52 95 /* BlockSize of Input samples are copied into the state buffer */
emilmont 1:fdd22bb7aa52 96 /* StateIndex points to the starting position to write in the state buffer */
emilmont 1:fdd22bb7aa52 97 arm_circularWrite_q15(py, delaySize, &S->stateIndex, 1, pIn, 1, blockSize);
emilmont 1:fdd22bb7aa52 98
emilmont 1:fdd22bb7aa52 99 /* Loop over the number of taps. */
emilmont 1:fdd22bb7aa52 100 tapCnt = numTaps;
emilmont 1:fdd22bb7aa52 101
emilmont 1:fdd22bb7aa52 102 /* Read Index, from where the state buffer should be read, is calculated. */
emilmont 1:fdd22bb7aa52 103 readIndex = (S->stateIndex - blockSize) - *pTapDelay++;
emilmont 1:fdd22bb7aa52 104
emilmont 1:fdd22bb7aa52 105 /* Wraparound of readIndex */
emilmont 1:fdd22bb7aa52 106 if(readIndex < 0)
emilmont 1:fdd22bb7aa52 107 {
emilmont 1:fdd22bb7aa52 108 readIndex += (int32_t) delaySize;
emilmont 1:fdd22bb7aa52 109 }
emilmont 1:fdd22bb7aa52 110
emilmont 1:fdd22bb7aa52 111 /* Working pointer for state buffer is updated */
emilmont 1:fdd22bb7aa52 112 py = pState;
emilmont 1:fdd22bb7aa52 113
emilmont 1:fdd22bb7aa52 114 /* blockSize samples are read from the state buffer */
emilmont 1:fdd22bb7aa52 115 arm_circularRead_q15(py, delaySize, &readIndex, 1,
emilmont 1:fdd22bb7aa52 116 pb, pb, blockSize, 1, blockSize);
emilmont 1:fdd22bb7aa52 117
emilmont 1:fdd22bb7aa52 118 /* Working pointer for the scratch buffer of state values */
emilmont 1:fdd22bb7aa52 119 px = pb;
emilmont 1:fdd22bb7aa52 120
emilmont 1:fdd22bb7aa52 121 /* Working pointer for scratch buffer of output values */
emilmont 1:fdd22bb7aa52 122 pScratchOut = pScr2;
emilmont 1:fdd22bb7aa52 123
emilmont 1:fdd22bb7aa52 124 /* Loop over the blockSize. Unroll by a factor of 4.
emilmont 1:fdd22bb7aa52 125 * Compute 4 multiplications at a time. */
emilmont 1:fdd22bb7aa52 126 blkCnt = blockSize >> 2;
emilmont 1:fdd22bb7aa52 127
emilmont 1:fdd22bb7aa52 128 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 129 {
emilmont 1:fdd22bb7aa52 130 /* Perform multiplication and store in the scratch buffer */
emilmont 1:fdd22bb7aa52 131 *pScratchOut++ = ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 132 *pScratchOut++ = ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 133 *pScratchOut++ = ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 134 *pScratchOut++ = ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 135
emilmont 1:fdd22bb7aa52 136 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 137 blkCnt--;
emilmont 1:fdd22bb7aa52 138 }
emilmont 1:fdd22bb7aa52 139
emilmont 1:fdd22bb7aa52 140 /* If the blockSize is not a multiple of 4,
emilmont 1:fdd22bb7aa52 141 * compute the remaining samples */
emilmont 1:fdd22bb7aa52 142 blkCnt = blockSize % 0x4u;
emilmont 1:fdd22bb7aa52 143
emilmont 1:fdd22bb7aa52 144 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 145 {
emilmont 1:fdd22bb7aa52 146 /* Perform multiplication and store in the scratch buffer */
emilmont 1:fdd22bb7aa52 147 *pScratchOut++ = ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 148
emilmont 1:fdd22bb7aa52 149 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 150 blkCnt--;
emilmont 1:fdd22bb7aa52 151 }
emilmont 1:fdd22bb7aa52 152
emilmont 1:fdd22bb7aa52 153 /* Load the coefficient value and
emilmont 1:fdd22bb7aa52 154 * increment the coefficient buffer for the next set of state values */
emilmont 1:fdd22bb7aa52 155 coeff = *pCoeffs++;
emilmont 1:fdd22bb7aa52 156
emilmont 1:fdd22bb7aa52 157 /* Read Index, from where the state buffer should be read, is calculated. */
emilmont 1:fdd22bb7aa52 158 readIndex = (S->stateIndex - blockSize) - *pTapDelay++;
emilmont 1:fdd22bb7aa52 159
emilmont 1:fdd22bb7aa52 160 /* Wraparound of readIndex */
emilmont 1:fdd22bb7aa52 161 if(readIndex < 0)
emilmont 1:fdd22bb7aa52 162 {
emilmont 1:fdd22bb7aa52 163 readIndex += (int32_t) delaySize;
emilmont 1:fdd22bb7aa52 164 }
emilmont 1:fdd22bb7aa52 165
emilmont 1:fdd22bb7aa52 166 /* Loop over the number of taps. */
emilmont 1:fdd22bb7aa52 167 tapCnt = (uint32_t) numTaps - 1u;
emilmont 1:fdd22bb7aa52 168
emilmont 1:fdd22bb7aa52 169 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 170 {
emilmont 1:fdd22bb7aa52 171 /* Working pointer for state buffer is updated */
emilmont 1:fdd22bb7aa52 172 py = pState;
emilmont 1:fdd22bb7aa52 173
emilmont 1:fdd22bb7aa52 174 /* blockSize samples are read from the state buffer */
emilmont 1:fdd22bb7aa52 175 arm_circularRead_q15(py, delaySize, &readIndex, 1,
emilmont 1:fdd22bb7aa52 176 pb, pb, blockSize, 1, blockSize);
emilmont 1:fdd22bb7aa52 177
emilmont 1:fdd22bb7aa52 178 /* Working pointer for the scratch buffer of state values */
emilmont 1:fdd22bb7aa52 179 px = pb;
emilmont 1:fdd22bb7aa52 180
emilmont 1:fdd22bb7aa52 181 /* Working pointer for scratch buffer of output values */
emilmont 1:fdd22bb7aa52 182 pScratchOut = pScr2;
emilmont 1:fdd22bb7aa52 183
emilmont 1:fdd22bb7aa52 184 /* Loop over the blockSize. Unroll by a factor of 4.
emilmont 1:fdd22bb7aa52 185 * Compute 4 MACS at a time. */
emilmont 1:fdd22bb7aa52 186 blkCnt = blockSize >> 2;
emilmont 1:fdd22bb7aa52 187
emilmont 1:fdd22bb7aa52 188 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 189 {
emilmont 1:fdd22bb7aa52 190 /* Perform Multiply-Accumulate */
emilmont 1:fdd22bb7aa52 191 *pScratchOut++ += (q31_t) * px++ * coeff;
emilmont 1:fdd22bb7aa52 192 *pScratchOut++ += (q31_t) * px++ * coeff;
emilmont 1:fdd22bb7aa52 193 *pScratchOut++ += (q31_t) * px++ * coeff;
emilmont 1:fdd22bb7aa52 194 *pScratchOut++ += (q31_t) * px++ * coeff;
emilmont 1:fdd22bb7aa52 195
emilmont 1:fdd22bb7aa52 196 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 197 blkCnt--;
emilmont 1:fdd22bb7aa52 198 }
emilmont 1:fdd22bb7aa52 199
emilmont 1:fdd22bb7aa52 200 /* If the blockSize is not a multiple of 4,
emilmont 1:fdd22bb7aa52 201 * compute the remaining samples */
emilmont 1:fdd22bb7aa52 202 blkCnt = blockSize % 0x4u;
emilmont 1:fdd22bb7aa52 203
emilmont 1:fdd22bb7aa52 204 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 205 {
emilmont 1:fdd22bb7aa52 206 /* Perform Multiply-Accumulate */
emilmont 1:fdd22bb7aa52 207 *pScratchOut++ += (q31_t) * px++ * coeff;
emilmont 1:fdd22bb7aa52 208
emilmont 1:fdd22bb7aa52 209 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 210 blkCnt--;
emilmont 1:fdd22bb7aa52 211 }
emilmont 1:fdd22bb7aa52 212
emilmont 1:fdd22bb7aa52 213 /* Load the coefficient value and
emilmont 1:fdd22bb7aa52 214 * increment the coefficient buffer for the next set of state values */
emilmont 1:fdd22bb7aa52 215 coeff = *pCoeffs++;
emilmont 1:fdd22bb7aa52 216
emilmont 1:fdd22bb7aa52 217 /* Read Index, from where the state buffer should be read, is calculated. */
emilmont 1:fdd22bb7aa52 218 readIndex = (S->stateIndex - blockSize) - *pTapDelay++;
emilmont 1:fdd22bb7aa52 219
emilmont 1:fdd22bb7aa52 220 /* Wraparound of readIndex */
emilmont 1:fdd22bb7aa52 221 if(readIndex < 0)
emilmont 1:fdd22bb7aa52 222 {
emilmont 1:fdd22bb7aa52 223 readIndex += (int32_t) delaySize;
emilmont 1:fdd22bb7aa52 224 }
emilmont 1:fdd22bb7aa52 225
emilmont 1:fdd22bb7aa52 226 /* Decrement the tap loop counter */
emilmont 1:fdd22bb7aa52 227 tapCnt--;
emilmont 1:fdd22bb7aa52 228 }
emilmont 1:fdd22bb7aa52 229
emilmont 1:fdd22bb7aa52 230 /* All the output values are in pScratchOut buffer.
emilmont 1:fdd22bb7aa52 231 Convert them into 1.15 format, saturate and store in the destination buffer. */
emilmont 1:fdd22bb7aa52 232 /* Loop over the blockSize. */
emilmont 1:fdd22bb7aa52 233 blkCnt = blockSize >> 2;
emilmont 1:fdd22bb7aa52 234
emilmont 1:fdd22bb7aa52 235 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 236 {
emilmont 1:fdd22bb7aa52 237 in1 = *pScr2++;
emilmont 1:fdd22bb7aa52 238 in2 = *pScr2++;
emilmont 1:fdd22bb7aa52 239
emilmont 1:fdd22bb7aa52 240 #ifndef ARM_MATH_BIG_ENDIAN
emilmont 1:fdd22bb7aa52 241
emilmont 1:fdd22bb7aa52 242 *__SIMD32(pOut)++ =
emilmont 1:fdd22bb7aa52 243 __PKHBT((q15_t) __SSAT(in1 >> 15, 16), (q15_t) __SSAT(in2 >> 15, 16),
emilmont 1:fdd22bb7aa52 244 16);
emilmont 1:fdd22bb7aa52 245
emilmont 1:fdd22bb7aa52 246 #else
emilmont 1:fdd22bb7aa52 247 *__SIMD32(pOut)++ =
emilmont 1:fdd22bb7aa52 248 __PKHBT((q15_t) __SSAT(in2 >> 15, 16), (q15_t) __SSAT(in1 >> 15, 16),
emilmont 1:fdd22bb7aa52 249 16);
emilmont 1:fdd22bb7aa52 250
emilmont 1:fdd22bb7aa52 251 #endif /* #ifndef ARM_MATH_BIG_ENDIAN */
emilmont 1:fdd22bb7aa52 252
emilmont 1:fdd22bb7aa52 253 in1 = *pScr2++;
emilmont 1:fdd22bb7aa52 254
emilmont 1:fdd22bb7aa52 255 in2 = *pScr2++;
emilmont 1:fdd22bb7aa52 256
emilmont 1:fdd22bb7aa52 257 #ifndef ARM_MATH_BIG_ENDIAN
emilmont 1:fdd22bb7aa52 258
emilmont 1:fdd22bb7aa52 259 *__SIMD32(pOut)++ =
emilmont 1:fdd22bb7aa52 260 __PKHBT((q15_t) __SSAT(in1 >> 15, 16), (q15_t) __SSAT(in2 >> 15, 16),
emilmont 1:fdd22bb7aa52 261 16);
emilmont 1:fdd22bb7aa52 262
emilmont 1:fdd22bb7aa52 263 #else
emilmont 1:fdd22bb7aa52 264
emilmont 1:fdd22bb7aa52 265 *__SIMD32(pOut)++ =
emilmont 1:fdd22bb7aa52 266 __PKHBT((q15_t) __SSAT(in2 >> 15, 16), (q15_t) __SSAT(in1 >> 15, 16),
emilmont 1:fdd22bb7aa52 267 16);
emilmont 1:fdd22bb7aa52 268
emilmont 1:fdd22bb7aa52 269 #endif /* #ifndef ARM_MATH_BIG_ENDIAN */
emilmont 1:fdd22bb7aa52 270
emilmont 1:fdd22bb7aa52 271
emilmont 1:fdd22bb7aa52 272 blkCnt--;
emilmont 1:fdd22bb7aa52 273
emilmont 1:fdd22bb7aa52 274 }
emilmont 1:fdd22bb7aa52 275
emilmont 1:fdd22bb7aa52 276 /* If the blockSize is not a multiple of 4,
emilmont 1:fdd22bb7aa52 277 remaining samples are processed in the below loop */
emilmont 1:fdd22bb7aa52 278 blkCnt = blockSize % 0x4u;
emilmont 1:fdd22bb7aa52 279
emilmont 1:fdd22bb7aa52 280 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 281 {
emilmont 1:fdd22bb7aa52 282 *pOut++ = (q15_t) __SSAT(*pScr2++ >> 15, 16);
emilmont 1:fdd22bb7aa52 283 blkCnt--;
emilmont 1:fdd22bb7aa52 284 }
emilmont 1:fdd22bb7aa52 285
emilmont 1:fdd22bb7aa52 286 #else
emilmont 1:fdd22bb7aa52 287
emilmont 1:fdd22bb7aa52 288 /* Run the below code for Cortex-M0 */
emilmont 1:fdd22bb7aa52 289
emilmont 1:fdd22bb7aa52 290 /* BlockSize of Input samples are copied into the state buffer */
emilmont 1:fdd22bb7aa52 291 /* StateIndex points to the starting position to write in the state buffer */
emilmont 1:fdd22bb7aa52 292 arm_circularWrite_q15(py, delaySize, &S->stateIndex, 1, pIn, 1, blockSize);
emilmont 1:fdd22bb7aa52 293
emilmont 1:fdd22bb7aa52 294 /* Loop over the number of taps. */
emilmont 1:fdd22bb7aa52 295 tapCnt = numTaps;
emilmont 1:fdd22bb7aa52 296
emilmont 1:fdd22bb7aa52 297 /* Read Index, from where the state buffer should be read, is calculated. */
emilmont 1:fdd22bb7aa52 298 readIndex = (S->stateIndex - blockSize) - *pTapDelay++;
emilmont 1:fdd22bb7aa52 299
emilmont 1:fdd22bb7aa52 300 /* Wraparound of readIndex */
emilmont 1:fdd22bb7aa52 301 if(readIndex < 0)
emilmont 1:fdd22bb7aa52 302 {
emilmont 1:fdd22bb7aa52 303 readIndex += (int32_t) delaySize;
emilmont 1:fdd22bb7aa52 304 }
emilmont 1:fdd22bb7aa52 305
emilmont 1:fdd22bb7aa52 306 /* Working pointer for state buffer is updated */
emilmont 1:fdd22bb7aa52 307 py = pState;
emilmont 1:fdd22bb7aa52 308
emilmont 1:fdd22bb7aa52 309 /* blockSize samples are read from the state buffer */
emilmont 1:fdd22bb7aa52 310 arm_circularRead_q15(py, delaySize, &readIndex, 1,
emilmont 1:fdd22bb7aa52 311 pb, pb, blockSize, 1, blockSize);
emilmont 1:fdd22bb7aa52 312
emilmont 1:fdd22bb7aa52 313 /* Working pointer for the scratch buffer of state values */
emilmont 1:fdd22bb7aa52 314 px = pb;
emilmont 1:fdd22bb7aa52 315
emilmont 1:fdd22bb7aa52 316 /* Working pointer for scratch buffer of output values */
emilmont 1:fdd22bb7aa52 317 pScratchOut = pScr2;
emilmont 1:fdd22bb7aa52 318
emilmont 1:fdd22bb7aa52 319 blkCnt = blockSize;
emilmont 1:fdd22bb7aa52 320
emilmont 1:fdd22bb7aa52 321 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 322 {
emilmont 1:fdd22bb7aa52 323 /* Perform multiplication and store in the scratch buffer */
emilmont 1:fdd22bb7aa52 324 *pScratchOut++ = ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 325
emilmont 1:fdd22bb7aa52 326 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 327 blkCnt--;
emilmont 1:fdd22bb7aa52 328 }
emilmont 1:fdd22bb7aa52 329
emilmont 1:fdd22bb7aa52 330 /* Load the coefficient value and
emilmont 1:fdd22bb7aa52 331 * increment the coefficient buffer for the next set of state values */
emilmont 1:fdd22bb7aa52 332 coeff = *pCoeffs++;
emilmont 1:fdd22bb7aa52 333
emilmont 1:fdd22bb7aa52 334 /* Read Index, from where the state buffer should be read, is calculated. */
emilmont 1:fdd22bb7aa52 335 readIndex = (S->stateIndex - blockSize) - *pTapDelay++;
emilmont 1:fdd22bb7aa52 336
emilmont 1:fdd22bb7aa52 337 /* Wraparound of readIndex */
emilmont 1:fdd22bb7aa52 338 if(readIndex < 0)
emilmont 1:fdd22bb7aa52 339 {
emilmont 1:fdd22bb7aa52 340 readIndex += (int32_t) delaySize;
emilmont 1:fdd22bb7aa52 341 }
emilmont 1:fdd22bb7aa52 342
emilmont 1:fdd22bb7aa52 343 /* Loop over the number of taps. */
emilmont 1:fdd22bb7aa52 344 tapCnt = (uint32_t) numTaps - 1u;
emilmont 1:fdd22bb7aa52 345
emilmont 1:fdd22bb7aa52 346 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 347 {
emilmont 1:fdd22bb7aa52 348 /* Working pointer for state buffer is updated */
emilmont 1:fdd22bb7aa52 349 py = pState;
emilmont 1:fdd22bb7aa52 350
emilmont 1:fdd22bb7aa52 351 /* blockSize samples are read from the state buffer */
emilmont 1:fdd22bb7aa52 352 arm_circularRead_q15(py, delaySize, &readIndex, 1,
emilmont 1:fdd22bb7aa52 353 pb, pb, blockSize, 1, blockSize);
emilmont 1:fdd22bb7aa52 354
emilmont 1:fdd22bb7aa52 355 /* Working pointer for the scratch buffer of state values */
emilmont 1:fdd22bb7aa52 356 px = pb;
emilmont 1:fdd22bb7aa52 357
emilmont 1:fdd22bb7aa52 358 /* Working pointer for scratch buffer of output values */
emilmont 1:fdd22bb7aa52 359 pScratchOut = pScr2;
emilmont 1:fdd22bb7aa52 360
emilmont 1:fdd22bb7aa52 361 blkCnt = blockSize;
emilmont 1:fdd22bb7aa52 362
emilmont 1:fdd22bb7aa52 363 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 364 {
emilmont 1:fdd22bb7aa52 365 /* Perform Multiply-Accumulate */
emilmont 1:fdd22bb7aa52 366 *pScratchOut++ += (q31_t) * px++ * coeff;
emilmont 1:fdd22bb7aa52 367
emilmont 1:fdd22bb7aa52 368 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 369 blkCnt--;
emilmont 1:fdd22bb7aa52 370 }
emilmont 1:fdd22bb7aa52 371
emilmont 1:fdd22bb7aa52 372 /* Load the coefficient value and
emilmont 1:fdd22bb7aa52 373 * increment the coefficient buffer for the next set of state values */
emilmont 1:fdd22bb7aa52 374 coeff = *pCoeffs++;
emilmont 1:fdd22bb7aa52 375
emilmont 1:fdd22bb7aa52 376 /* Read Index, from where the state buffer should be read, is calculated. */
emilmont 1:fdd22bb7aa52 377 readIndex = (S->stateIndex - blockSize) - *pTapDelay++;
emilmont 1:fdd22bb7aa52 378
emilmont 1:fdd22bb7aa52 379 /* Wraparound of readIndex */
emilmont 1:fdd22bb7aa52 380 if(readIndex < 0)
emilmont 1:fdd22bb7aa52 381 {
emilmont 1:fdd22bb7aa52 382 readIndex += (int32_t) delaySize;
emilmont 1:fdd22bb7aa52 383 }
emilmont 1:fdd22bb7aa52 384
emilmont 1:fdd22bb7aa52 385 /* Decrement the tap loop counter */
emilmont 1:fdd22bb7aa52 386 tapCnt--;
emilmont 1:fdd22bb7aa52 387 }
emilmont 1:fdd22bb7aa52 388
emilmont 1:fdd22bb7aa52 389 /* All the output values are in pScratchOut buffer.
emilmont 1:fdd22bb7aa52 390 Convert them into 1.15 format, saturate and store in the destination buffer. */
emilmont 1:fdd22bb7aa52 391 /* Loop over the blockSize. */
emilmont 1:fdd22bb7aa52 392 blkCnt = blockSize;
emilmont 1:fdd22bb7aa52 393
emilmont 1:fdd22bb7aa52 394 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 395 {
emilmont 1:fdd22bb7aa52 396 *pOut++ = (q15_t) __SSAT(*pScr2++ >> 15, 16);
emilmont 1:fdd22bb7aa52 397 blkCnt--;
emilmont 1:fdd22bb7aa52 398 }
emilmont 1:fdd22bb7aa52 399
emilmont 1:fdd22bb7aa52 400 #endif /* #ifndef ARM_MATH_CM0 */
emilmont 1:fdd22bb7aa52 401
emilmont 1:fdd22bb7aa52 402 }
emilmont 1:fdd22bb7aa52 403
emilmont 1:fdd22bb7aa52 404 /**
emilmont 1:fdd22bb7aa52 405 * @} end of FIR_Sparse group
emilmont 1:fdd22bb7aa52 406 */