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_q7.c@1:fdd22bb7aa52, 2012-11-28 (annotated)

Committer:
emilmont
Date:
Wed Nov 28 12:30:09 2012 +0000
Revision:
1:fdd22bb7aa52
Child:
2:da51fb522205
DSP library code

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 1:fdd22bb7aa52 5 * $Revision: V1.1.0
emilmont 1:fdd22bb7aa52 6 *
emilmont 1:fdd22bb7aa52 7 * Project: CMSIS DSP Library
emilmont 1:fdd22bb7aa52 8 * Title: arm_fir_sparse_q7.c
emilmont 1:fdd22bb7aa52 9 *
emilmont 1:fdd22bb7aa52 10 * Description: Q7 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 /**
emilmont 1:fdd22bb7aa52 39 * @ingroup groupFilters
emilmont 1:fdd22bb7aa52 40 */
emilmont 1:fdd22bb7aa52 41
emilmont 1:fdd22bb7aa52 42 /**
emilmont 1:fdd22bb7aa52 43 * @addtogroup FIR_Sparse
emilmont 1:fdd22bb7aa52 44 * @{
emilmont 1:fdd22bb7aa52 45 */
emilmont 1:fdd22bb7aa52 46
emilmont 1:fdd22bb7aa52 47
emilmont 1:fdd22bb7aa52 48 /**
emilmont 1:fdd22bb7aa52 49 * @brief Processing function for the Q7 sparse FIR filter.
emilmont 1:fdd22bb7aa52 50 * @param[in] *S points to an instance of the Q7 sparse FIR structure.
emilmont 1:fdd22bb7aa52 51 * @param[in] *pSrc points to the block of input data.
emilmont 1:fdd22bb7aa52 52 * @param[out] *pDst points to the block of output data
emilmont 1:fdd22bb7aa52 53 * @param[in] *pScratchIn points to a temporary buffer of size blockSize.
emilmont 1:fdd22bb7aa52 54 * @param[in] *pScratchOut points to a temporary buffer of size blockSize.
emilmont 1:fdd22bb7aa52 55 * @param[in] blockSize number of input samples to process per call.
emilmont 1:fdd22bb7aa52 56 * @return none.
emilmont 1:fdd22bb7aa52 57 *
emilmont 1:fdd22bb7aa52 58 * <b>Scaling and Overflow Behavior:</b>
emilmont 1:fdd22bb7aa52 59 * \par
emilmont 1:fdd22bb7aa52 60 * The function is implemented using a 32-bit internal accumulator.
emilmont 1:fdd22bb7aa52 61 * Both coefficients and state variables are represented in 1.7 format and multiplications yield a 2.14 result.
emilmont 1:fdd22bb7aa52 62 * The 2.14 intermediate results are accumulated in a 32-bit accumulator in 18.14 format.
emilmont 1:fdd22bb7aa52 63 * There is no risk of internal overflow with this approach and the full precision of intermediate multiplications is preserved.
emilmont 1:fdd22bb7aa52 64 * The accumulator is then converted to 18.7 format by discarding the low 7 bits.
emilmont 1:fdd22bb7aa52 65 * Finally, the result is truncated to 1.7 format.
emilmont 1:fdd22bb7aa52 66 */
emilmont 1:fdd22bb7aa52 67
emilmont 1:fdd22bb7aa52 68 void arm_fir_sparse_q7(
emilmont 1:fdd22bb7aa52 69 arm_fir_sparse_instance_q7 * S,
emilmont 1:fdd22bb7aa52 70 q7_t * pSrc,
emilmont 1:fdd22bb7aa52 71 q7_t * pDst,
emilmont 1:fdd22bb7aa52 72 q7_t * pScratchIn,
emilmont 1:fdd22bb7aa52 73 q31_t * pScratchOut,
emilmont 1:fdd22bb7aa52 74 uint32_t blockSize)
emilmont 1:fdd22bb7aa52 75 {
emilmont 1:fdd22bb7aa52 76
emilmont 1:fdd22bb7aa52 77 q7_t *pState = S->pState; /* State pointer */
emilmont 1:fdd22bb7aa52 78 q7_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
emilmont 1:fdd22bb7aa52 79 q7_t *px; /* Scratch buffer pointer */
emilmont 1:fdd22bb7aa52 80 q7_t *py = pState; /* Temporary pointers for state buffer */
emilmont 1:fdd22bb7aa52 81 q7_t *pb = pScratchIn; /* Temporary pointers for scratch buffer */
emilmont 1:fdd22bb7aa52 82 q7_t *pOut = pDst; /* Destination pointer */
emilmont 1:fdd22bb7aa52 83 int32_t *pTapDelay = S->pTapDelay; /* Pointer to the array containing offset of the non-zero tap values. */
emilmont 1:fdd22bb7aa52 84 uint32_t delaySize = S->maxDelay + blockSize; /* state length */
emilmont 1:fdd22bb7aa52 85 uint16_t numTaps = S->numTaps; /* Filter order */
emilmont 1:fdd22bb7aa52 86 int32_t readIndex; /* Read index of the state buffer */
emilmont 1:fdd22bb7aa52 87 uint32_t tapCnt, blkCnt; /* loop counters */
emilmont 1:fdd22bb7aa52 88 q7_t coeff = *pCoeffs++; /* Read the coefficient value */
emilmont 1:fdd22bb7aa52 89 q31_t *pScr2 = pScratchOut; /* Working pointer for scratch buffer of output values */
emilmont 1:fdd22bb7aa52 90 q31_t in;
emilmont 1:fdd22bb7aa52 91
emilmont 1:fdd22bb7aa52 92
emilmont 1:fdd22bb7aa52 93 #ifndef ARM_MATH_CM0
emilmont 1:fdd22bb7aa52 94
emilmont 1:fdd22bb7aa52 95 /* Run the below code for Cortex-M4 and Cortex-M3 */
emilmont 1:fdd22bb7aa52 96
emilmont 1:fdd22bb7aa52 97 q7_t in1, in2, in3, in4;
emilmont 1:fdd22bb7aa52 98
emilmont 1:fdd22bb7aa52 99 /* BlockSize of Input samples are copied into the state buffer */
emilmont 1:fdd22bb7aa52 100 /* StateIndex points to the starting position to write in the state buffer */
emilmont 1:fdd22bb7aa52 101 arm_circularWrite_q7(py, (int32_t) delaySize, &S->stateIndex, 1, pSrc, 1,
emilmont 1:fdd22bb7aa52 102 blockSize);
emilmont 1:fdd22bb7aa52 103
emilmont 1:fdd22bb7aa52 104 /* Loop over the number of taps. */
emilmont 1:fdd22bb7aa52 105 tapCnt = numTaps;
emilmont 1:fdd22bb7aa52 106
emilmont 1:fdd22bb7aa52 107 /* Read Index, from where the state buffer should be read, is calculated. */
emilmont 1:fdd22bb7aa52 108 readIndex = ((int32_t) S->stateIndex - (int32_t) blockSize) - *pTapDelay++;
emilmont 1:fdd22bb7aa52 109
emilmont 1:fdd22bb7aa52 110 /* Wraparound of readIndex */
emilmont 1:fdd22bb7aa52 111 if(readIndex < 0)
emilmont 1:fdd22bb7aa52 112 {
emilmont 1:fdd22bb7aa52 113 readIndex += (int32_t) delaySize;
emilmont 1:fdd22bb7aa52 114 }
emilmont 1:fdd22bb7aa52 115
emilmont 1:fdd22bb7aa52 116 /* Working pointer for state buffer is updated */
emilmont 1:fdd22bb7aa52 117 py = pState;
emilmont 1:fdd22bb7aa52 118
emilmont 1:fdd22bb7aa52 119 /* blockSize samples are read from the state buffer */
emilmont 1:fdd22bb7aa52 120 arm_circularRead_q7(py, (int32_t) delaySize, &readIndex, 1, pb, pb,
emilmont 1:fdd22bb7aa52 121 (int32_t) blockSize, 1, blockSize);
emilmont 1:fdd22bb7aa52 122
emilmont 1:fdd22bb7aa52 123 /* Working pointer for the scratch buffer of state values */
emilmont 1:fdd22bb7aa52 124 px = pb;
emilmont 1:fdd22bb7aa52 125
emilmont 1:fdd22bb7aa52 126 /* Working pointer for scratch buffer of output values */
emilmont 1:fdd22bb7aa52 127 pScratchOut = pScr2;
emilmont 1:fdd22bb7aa52 128
emilmont 1:fdd22bb7aa52 129 /* Loop over the blockSize. Unroll by a factor of 4.
emilmont 1:fdd22bb7aa52 130 * Compute 4 multiplications at a time. */
emilmont 1:fdd22bb7aa52 131 blkCnt = blockSize >> 2;
emilmont 1:fdd22bb7aa52 132
emilmont 1:fdd22bb7aa52 133 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 134 {
emilmont 1:fdd22bb7aa52 135 /* Perform multiplication and store in the scratch buffer */
emilmont 1:fdd22bb7aa52 136 *pScratchOut++ = ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 137 *pScratchOut++ = ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 138 *pScratchOut++ = ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 139 *pScratchOut++ = ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 140
emilmont 1:fdd22bb7aa52 141 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 142 blkCnt--;
emilmont 1:fdd22bb7aa52 143 }
emilmont 1:fdd22bb7aa52 144
emilmont 1:fdd22bb7aa52 145 /* If the blockSize is not a multiple of 4,
emilmont 1:fdd22bb7aa52 146 * compute the remaining samples */
emilmont 1:fdd22bb7aa52 147 blkCnt = blockSize % 0x4u;
emilmont 1:fdd22bb7aa52 148
emilmont 1:fdd22bb7aa52 149 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 150 {
emilmont 1:fdd22bb7aa52 151 /* Perform multiplication and store in the scratch buffer */
emilmont 1:fdd22bb7aa52 152 *pScratchOut++ = ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 153
emilmont 1:fdd22bb7aa52 154 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 155 blkCnt--;
emilmont 1:fdd22bb7aa52 156 }
emilmont 1:fdd22bb7aa52 157
emilmont 1:fdd22bb7aa52 158 /* Load the coefficient value and
emilmont 1:fdd22bb7aa52 159 * increment the coefficient buffer for the next set of state values */
emilmont 1:fdd22bb7aa52 160 coeff = *pCoeffs++;
emilmont 1:fdd22bb7aa52 161
emilmont 1:fdd22bb7aa52 162 /* Read Index, from where the state buffer should be read, is calculated. */
emilmont 1:fdd22bb7aa52 163 readIndex = ((int32_t) S->stateIndex - (int32_t) blockSize) - *pTapDelay++;
emilmont 1:fdd22bb7aa52 164
emilmont 1:fdd22bb7aa52 165 /* Wraparound of readIndex */
emilmont 1:fdd22bb7aa52 166 if(readIndex < 0)
emilmont 1:fdd22bb7aa52 167 {
emilmont 1:fdd22bb7aa52 168 readIndex += (int32_t) delaySize;
emilmont 1:fdd22bb7aa52 169 }
emilmont 1:fdd22bb7aa52 170
emilmont 1:fdd22bb7aa52 171 /* Loop over the number of taps. */
emilmont 1:fdd22bb7aa52 172 tapCnt = (uint32_t) numTaps - 1u;
emilmont 1:fdd22bb7aa52 173
emilmont 1:fdd22bb7aa52 174 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 175 {
emilmont 1:fdd22bb7aa52 176 /* Working pointer for state buffer is updated */
emilmont 1:fdd22bb7aa52 177 py = pState;
emilmont 1:fdd22bb7aa52 178
emilmont 1:fdd22bb7aa52 179 /* blockSize samples are read from the state buffer */
emilmont 1:fdd22bb7aa52 180 arm_circularRead_q7(py, (int32_t) delaySize, &readIndex, 1, pb, pb,
emilmont 1:fdd22bb7aa52 181 (int32_t) blockSize, 1, blockSize);
emilmont 1:fdd22bb7aa52 182
emilmont 1:fdd22bb7aa52 183 /* Working pointer for the scratch buffer of state values */
emilmont 1:fdd22bb7aa52 184 px = pb;
emilmont 1:fdd22bb7aa52 185
emilmont 1:fdd22bb7aa52 186 /* Working pointer for scratch buffer of output values */
emilmont 1:fdd22bb7aa52 187 pScratchOut = pScr2;
emilmont 1:fdd22bb7aa52 188
emilmont 1:fdd22bb7aa52 189 /* Loop over the blockSize. Unroll by a factor of 4.
emilmont 1:fdd22bb7aa52 190 * Compute 4 MACS at a time. */
emilmont 1:fdd22bb7aa52 191 blkCnt = blockSize >> 2;
emilmont 1:fdd22bb7aa52 192
emilmont 1:fdd22bb7aa52 193 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 194 {
emilmont 1:fdd22bb7aa52 195 /* Perform Multiply-Accumulate */
emilmont 1:fdd22bb7aa52 196 in = *pScratchOut + ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 197 *pScratchOut++ = in;
emilmont 1:fdd22bb7aa52 198 in = *pScratchOut + ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 199 *pScratchOut++ = in;
emilmont 1:fdd22bb7aa52 200 in = *pScratchOut + ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 201 *pScratchOut++ = in;
emilmont 1:fdd22bb7aa52 202 in = *pScratchOut + ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 203 *pScratchOut++ = in;
emilmont 1:fdd22bb7aa52 204
emilmont 1:fdd22bb7aa52 205 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 206 blkCnt--;
emilmont 1:fdd22bb7aa52 207 }
emilmont 1:fdd22bb7aa52 208
emilmont 1:fdd22bb7aa52 209 /* If the blockSize is not a multiple of 4,
emilmont 1:fdd22bb7aa52 210 * compute the remaining samples */
emilmont 1:fdd22bb7aa52 211 blkCnt = blockSize % 0x4u;
emilmont 1:fdd22bb7aa52 212
emilmont 1:fdd22bb7aa52 213 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 214 {
emilmont 1:fdd22bb7aa52 215 /* Perform Multiply-Accumulate */
emilmont 1:fdd22bb7aa52 216 in = *pScratchOut + ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 217 *pScratchOut++ = in;
emilmont 1:fdd22bb7aa52 218
emilmont 1:fdd22bb7aa52 219 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 220 blkCnt--;
emilmont 1:fdd22bb7aa52 221 }
emilmont 1:fdd22bb7aa52 222
emilmont 1:fdd22bb7aa52 223 /* Load the coefficient value and
emilmont 1:fdd22bb7aa52 224 * increment the coefficient buffer for the next set of state values */
emilmont 1:fdd22bb7aa52 225 coeff = *pCoeffs++;
emilmont 1:fdd22bb7aa52 226
emilmont 1:fdd22bb7aa52 227 /* Read Index, from where the state buffer should be read, is calculated. */
emilmont 1:fdd22bb7aa52 228 readIndex = ((int32_t) S->stateIndex -
emilmont 1:fdd22bb7aa52 229 (int32_t) blockSize) - *pTapDelay++;
emilmont 1:fdd22bb7aa52 230
emilmont 1:fdd22bb7aa52 231 /* Wraparound of readIndex */
emilmont 1:fdd22bb7aa52 232 if(readIndex < 0)
emilmont 1:fdd22bb7aa52 233 {
emilmont 1:fdd22bb7aa52 234 readIndex += (int32_t) delaySize;
emilmont 1:fdd22bb7aa52 235 }
emilmont 1:fdd22bb7aa52 236
emilmont 1:fdd22bb7aa52 237 /* Decrement the tap loop counter */
emilmont 1:fdd22bb7aa52 238 tapCnt--;
emilmont 1:fdd22bb7aa52 239 }
emilmont 1:fdd22bb7aa52 240
emilmont 1:fdd22bb7aa52 241 /* All the output values are in pScratchOut buffer.
emilmont 1:fdd22bb7aa52 242 Convert them into 1.15 format, saturate and store in the destination buffer. */
emilmont 1:fdd22bb7aa52 243 /* Loop over the blockSize. */
emilmont 1:fdd22bb7aa52 244 blkCnt = blockSize >> 2;
emilmont 1:fdd22bb7aa52 245
emilmont 1:fdd22bb7aa52 246 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 247 {
emilmont 1:fdd22bb7aa52 248 in1 = (q7_t) __SSAT(*pScr2++ >> 7, 8);
emilmont 1:fdd22bb7aa52 249 in2 = (q7_t) __SSAT(*pScr2++ >> 7, 8);
emilmont 1:fdd22bb7aa52 250 in3 = (q7_t) __SSAT(*pScr2++ >> 7, 8);
emilmont 1:fdd22bb7aa52 251 in4 = (q7_t) __SSAT(*pScr2++ >> 7, 8);
emilmont 1:fdd22bb7aa52 252
emilmont 1:fdd22bb7aa52 253 *__SIMD32(pOut)++ = __PACKq7(in1, in2, in3, in4);
emilmont 1:fdd22bb7aa52 254
emilmont 1:fdd22bb7aa52 255 /* Decrement the blockSize loop counter */
emilmont 1:fdd22bb7aa52 256 blkCnt--;
emilmont 1:fdd22bb7aa52 257 }
emilmont 1:fdd22bb7aa52 258
emilmont 1:fdd22bb7aa52 259 /* If the blockSize is not a multiple of 4,
emilmont 1:fdd22bb7aa52 260 remaining samples are processed in the below loop */
emilmont 1:fdd22bb7aa52 261 blkCnt = blockSize % 0x4u;
emilmont 1:fdd22bb7aa52 262
emilmont 1:fdd22bb7aa52 263 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 264 {
emilmont 1:fdd22bb7aa52 265 *pOut++ = (q7_t) __SSAT(*pScr2++ >> 7, 8);
emilmont 1:fdd22bb7aa52 266
emilmont 1:fdd22bb7aa52 267 /* Decrement the blockSize loop counter */
emilmont 1:fdd22bb7aa52 268 blkCnt--;
emilmont 1:fdd22bb7aa52 269 }
emilmont 1:fdd22bb7aa52 270
emilmont 1:fdd22bb7aa52 271 #else
emilmont 1:fdd22bb7aa52 272
emilmont 1:fdd22bb7aa52 273 /* Run the below code for Cortex-M0 */
emilmont 1:fdd22bb7aa52 274
emilmont 1:fdd22bb7aa52 275 /* BlockSize of Input samples are copied into the state buffer */
emilmont 1:fdd22bb7aa52 276 /* StateIndex points to the starting position to write in the state buffer */
emilmont 1:fdd22bb7aa52 277 arm_circularWrite_q7(py, (int32_t) delaySize, &S->stateIndex, 1, pSrc, 1,
emilmont 1:fdd22bb7aa52 278 blockSize);
emilmont 1:fdd22bb7aa52 279
emilmont 1:fdd22bb7aa52 280 /* Loop over the number of taps. */
emilmont 1:fdd22bb7aa52 281 tapCnt = numTaps;
emilmont 1:fdd22bb7aa52 282
emilmont 1:fdd22bb7aa52 283 /* Read Index, from where the state buffer should be read, is calculated. */
emilmont 1:fdd22bb7aa52 284 readIndex = ((int32_t) S->stateIndex - (int32_t) blockSize) - *pTapDelay++;
emilmont 1:fdd22bb7aa52 285
emilmont 1:fdd22bb7aa52 286 /* Wraparound of readIndex */
emilmont 1:fdd22bb7aa52 287 if(readIndex < 0)
emilmont 1:fdd22bb7aa52 288 {
emilmont 1:fdd22bb7aa52 289 readIndex += (int32_t) delaySize;
emilmont 1:fdd22bb7aa52 290 }
emilmont 1:fdd22bb7aa52 291
emilmont 1:fdd22bb7aa52 292 /* Working pointer for state buffer is updated */
emilmont 1:fdd22bb7aa52 293 py = pState;
emilmont 1:fdd22bb7aa52 294
emilmont 1:fdd22bb7aa52 295 /* blockSize samples are read from the state buffer */
emilmont 1:fdd22bb7aa52 296 arm_circularRead_q7(py, (int32_t) delaySize, &readIndex, 1, pb, pb,
emilmont 1:fdd22bb7aa52 297 (int32_t) blockSize, 1, blockSize);
emilmont 1:fdd22bb7aa52 298
emilmont 1:fdd22bb7aa52 299 /* Working pointer for the scratch buffer of state values */
emilmont 1:fdd22bb7aa52 300 px = pb;
emilmont 1:fdd22bb7aa52 301
emilmont 1:fdd22bb7aa52 302 /* Working pointer for scratch buffer of output values */
emilmont 1:fdd22bb7aa52 303 pScratchOut = pScr2;
emilmont 1:fdd22bb7aa52 304
emilmont 1:fdd22bb7aa52 305 /* Loop over the blockSize */
emilmont 1:fdd22bb7aa52 306 blkCnt = blockSize;
emilmont 1:fdd22bb7aa52 307
emilmont 1:fdd22bb7aa52 308 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 309 {
emilmont 1:fdd22bb7aa52 310 /* Perform multiplication and store in the scratch buffer */
emilmont 1:fdd22bb7aa52 311 *pScratchOut++ = ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 312
emilmont 1:fdd22bb7aa52 313 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 314 blkCnt--;
emilmont 1:fdd22bb7aa52 315 }
emilmont 1:fdd22bb7aa52 316
emilmont 1:fdd22bb7aa52 317 /* Load the coefficient value and
emilmont 1:fdd22bb7aa52 318 * increment the coefficient buffer for the next set of state values */
emilmont 1:fdd22bb7aa52 319 coeff = *pCoeffs++;
emilmont 1:fdd22bb7aa52 320
emilmont 1:fdd22bb7aa52 321 /* Read Index, from where the state buffer should be read, is calculated. */
emilmont 1:fdd22bb7aa52 322 readIndex = ((int32_t) S->stateIndex - (int32_t) blockSize) - *pTapDelay++;
emilmont 1:fdd22bb7aa52 323
emilmont 1:fdd22bb7aa52 324 /* Wraparound of readIndex */
emilmont 1:fdd22bb7aa52 325 if(readIndex < 0)
emilmont 1:fdd22bb7aa52 326 {
emilmont 1:fdd22bb7aa52 327 readIndex += (int32_t) delaySize;
emilmont 1:fdd22bb7aa52 328 }
emilmont 1:fdd22bb7aa52 329
emilmont 1:fdd22bb7aa52 330 /* Loop over the number of taps. */
emilmont 1:fdd22bb7aa52 331 tapCnt = (uint32_t) numTaps - 1u;
emilmont 1:fdd22bb7aa52 332
emilmont 1:fdd22bb7aa52 333 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 334 {
emilmont 1:fdd22bb7aa52 335 /* Working pointer for state buffer is updated */
emilmont 1:fdd22bb7aa52 336 py = pState;
emilmont 1:fdd22bb7aa52 337
emilmont 1:fdd22bb7aa52 338 /* blockSize samples are read from the state buffer */
emilmont 1:fdd22bb7aa52 339 arm_circularRead_q7(py, (int32_t) delaySize, &readIndex, 1, pb, pb,
emilmont 1:fdd22bb7aa52 340 (int32_t) blockSize, 1, blockSize);
emilmont 1:fdd22bb7aa52 341
emilmont 1:fdd22bb7aa52 342 /* Working pointer for the scratch buffer of state values */
emilmont 1:fdd22bb7aa52 343 px = pb;
emilmont 1:fdd22bb7aa52 344
emilmont 1:fdd22bb7aa52 345 /* Working pointer for scratch buffer of output values */
emilmont 1:fdd22bb7aa52 346 pScratchOut = pScr2;
emilmont 1:fdd22bb7aa52 347
emilmont 1:fdd22bb7aa52 348 /* Loop over the blockSize */
emilmont 1:fdd22bb7aa52 349 blkCnt = blockSize;
emilmont 1:fdd22bb7aa52 350
emilmont 1:fdd22bb7aa52 351 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 352 {
emilmont 1:fdd22bb7aa52 353 /* Perform Multiply-Accumulate */
emilmont 1:fdd22bb7aa52 354 in = *pScratchOut + ((q31_t) * px++ * coeff);
emilmont 1:fdd22bb7aa52 355 *pScratchOut++ = in;
emilmont 1:fdd22bb7aa52 356
emilmont 1:fdd22bb7aa52 357 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 358 blkCnt--;
emilmont 1:fdd22bb7aa52 359 }
emilmont 1:fdd22bb7aa52 360
emilmont 1:fdd22bb7aa52 361 /* Load the coefficient value and
emilmont 1:fdd22bb7aa52 362 * increment the coefficient buffer for the next set of state values */
emilmont 1:fdd22bb7aa52 363 coeff = *pCoeffs++;
emilmont 1:fdd22bb7aa52 364
emilmont 1:fdd22bb7aa52 365 /* Read Index, from where the state buffer should be read, is calculated. */
emilmont 1:fdd22bb7aa52 366 readIndex =
emilmont 1:fdd22bb7aa52 367 ((int32_t) S->stateIndex - (int32_t) blockSize) - *pTapDelay++;
emilmont 1:fdd22bb7aa52 368
emilmont 1:fdd22bb7aa52 369 /* Wraparound of readIndex */
emilmont 1:fdd22bb7aa52 370 if(readIndex < 0)
emilmont 1:fdd22bb7aa52 371 {
emilmont 1:fdd22bb7aa52 372 readIndex += (int32_t) delaySize;
emilmont 1:fdd22bb7aa52 373 }
emilmont 1:fdd22bb7aa52 374
emilmont 1:fdd22bb7aa52 375 /* Decrement the tap loop counter */
emilmont 1:fdd22bb7aa52 376 tapCnt--;
emilmont 1:fdd22bb7aa52 377 }
emilmont 1:fdd22bb7aa52 378
emilmont 1:fdd22bb7aa52 379 /* All the output values are in pScratchOut buffer.
emilmont 1:fdd22bb7aa52 380 Convert them into 1.15 format, saturate and store in the destination buffer. */
emilmont 1:fdd22bb7aa52 381 /* Loop over the blockSize. */
emilmont 1:fdd22bb7aa52 382 blkCnt = blockSize;
emilmont 1:fdd22bb7aa52 383
emilmont 1:fdd22bb7aa52 384 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 385 {
emilmont 1:fdd22bb7aa52 386 *pOut++ = (q7_t) __SSAT(*pScr2++ >> 7, 8);
emilmont 1:fdd22bb7aa52 387
emilmont 1:fdd22bb7aa52 388 /* Decrement the blockSize loop counter */
emilmont 1:fdd22bb7aa52 389 blkCnt--;
emilmont 1:fdd22bb7aa52 390 }
emilmont 1:fdd22bb7aa52 391
emilmont 1:fdd22bb7aa52 392 #endif /* #ifndef ARM_MATH_CM0 */
emilmont 1:fdd22bb7aa52 393
emilmont 1:fdd22bb7aa52 394 }
emilmont 1:fdd22bb7aa52 395
emilmont 1:fdd22bb7aa52 396 /**
emilmont 1:fdd22bb7aa52 397 * @} end of FIR_Sparse group
emilmont 1:fdd22bb7aa52 398 */