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_fast_q15.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_fast_q15.c
emilmont 1:fdd22bb7aa52 9 *
emilmont 1:fdd22bb7aa52 10 * Description: Q15 Fast FIR filter processing function.
emilmont 1:fdd22bb7aa52 11 *
emilmont 1:fdd22bb7aa52 12 * Target Processor: Cortex-M4/Cortex-M3
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 0.0.9 2010/08/16
emilmont 1:fdd22bb7aa52 33 * Initial version
emilmont 1:fdd22bb7aa52 34 *
emilmont 1:fdd22bb7aa52 35 * -------------------------------------------------------------------- */
emilmont 1:fdd22bb7aa52 36
emilmont 1:fdd22bb7aa52 37 #include "arm_math.h"
emilmont 1:fdd22bb7aa52 38
emilmont 1:fdd22bb7aa52 39 /**
emilmont 1:fdd22bb7aa52 40 * @ingroup groupFilters
emilmont 1:fdd22bb7aa52 41 */
emilmont 1:fdd22bb7aa52 42
emilmont 1:fdd22bb7aa52 43 /**
emilmont 1:fdd22bb7aa52 44 * @addtogroup FIR
emilmont 1:fdd22bb7aa52 45 * @{
emilmont 1:fdd22bb7aa52 46 */
emilmont 1:fdd22bb7aa52 47
emilmont 1:fdd22bb7aa52 48 /**
emilmont 1:fdd22bb7aa52 49 * @param[in] *S points to an instance of the Q15 FIR filter structure.
emilmont 1:fdd22bb7aa52 50 * @param[in] *pSrc points to the block of input data.
emilmont 1:fdd22bb7aa52 51 * @param[out] *pDst points to the block of output data.
emilmont 1:fdd22bb7aa52 52 * @param[in] blockSize number of samples to process per call.
emilmont 1:fdd22bb7aa52 53 * @return none.
emilmont 1:fdd22bb7aa52 54 *
emilmont 1:fdd22bb7aa52 55 * <b>Scaling and Overflow Behavior:</b>
emilmont 1:fdd22bb7aa52 56 * \par
emilmont 1:fdd22bb7aa52 57 * This fast version uses a 32-bit accumulator with 2.30 format.
emilmont 1:fdd22bb7aa52 58 * The accumulator maintains full precision of the intermediate multiplication results but provides only a single guard bit.
emilmont 1:fdd22bb7aa52 59 * Thus, if the accumulator result overflows it wraps around and distorts the result.
emilmont 1:fdd22bb7aa52 60 * In order to avoid overflows completely the input signal must be scaled down by log2(numTaps) bits.
emilmont 1:fdd22bb7aa52 61 * The 2.30 accumulator is then truncated to 2.15 format and saturated to yield the 1.15 result.
emilmont 1:fdd22bb7aa52 62 *
emilmont 1:fdd22bb7aa52 63 * \par
emilmont 1:fdd22bb7aa52 64 * Refer to the function <code>arm_fir_q15()</code> for a slower implementation of this function which uses 64-bit accumulation to avoid wrap around distortion. Both the slow and the fast versions use the same instance structure.
emilmont 1:fdd22bb7aa52 65 * Use the function <code>arm_fir_init_q15()</code> to initialize the filter structure.
emilmont 1:fdd22bb7aa52 66 */
emilmont 1:fdd22bb7aa52 67
emilmont 1:fdd22bb7aa52 68 void arm_fir_fast_q15(
emilmont 1:fdd22bb7aa52 69 const arm_fir_instance_q15 * S,
emilmont 1:fdd22bb7aa52 70 q15_t * pSrc,
emilmont 1:fdd22bb7aa52 71 q15_t * pDst,
emilmont 1:fdd22bb7aa52 72 uint32_t blockSize)
emilmont 1:fdd22bb7aa52 73 {
emilmont 1:fdd22bb7aa52 74 q15_t *pState = S->pState; /* State pointer */
emilmont 1:fdd22bb7aa52 75 q15_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
emilmont 1:fdd22bb7aa52 76 q15_t *pStateCurnt; /* Points to the current sample of the state */
emilmont 1:fdd22bb7aa52 77 q31_t acc0, acc1, acc2, acc3; /* Accumulators */
emilmont 1:fdd22bb7aa52 78 q15_t *pb; /* Temporary pointer for coefficient buffer */
emilmont 1:fdd22bb7aa52 79 q15_t *px; /* Temporary q31 pointer for SIMD state buffer accesses */
emilmont 1:fdd22bb7aa52 80 q31_t x0, x1, x2, c0; /* Temporary variables to hold SIMD state and coefficient values */
emilmont 1:fdd22bb7aa52 81 uint32_t numTaps = S->numTaps; /* Number of taps in the filter */
emilmont 1:fdd22bb7aa52 82 uint32_t tapCnt, blkCnt; /* Loop counters */
emilmont 1:fdd22bb7aa52 83
emilmont 1:fdd22bb7aa52 84
emilmont 1:fdd22bb7aa52 85 /* S->pState points to state array which contains previous frame (numTaps - 1) samples */
emilmont 1:fdd22bb7aa52 86 /* pStateCurnt points to the location where the new input data should be written */
emilmont 1:fdd22bb7aa52 87 pStateCurnt = &(S->pState[(numTaps - 1u)]);
emilmont 1:fdd22bb7aa52 88
emilmont 1:fdd22bb7aa52 89 /* Apply loop unrolling and compute 4 output values simultaneously.
emilmont 1:fdd22bb7aa52 90 * The variables acc0 ... acc3 hold output values that are being computed:
emilmont 1:fdd22bb7aa52 91 *
emilmont 1:fdd22bb7aa52 92 * acc0 = b[numTaps-1] * x[n-numTaps-1] + b[numTaps-2] * x[n-numTaps-2] + b[numTaps-3] * x[n-numTaps-3] +...+ b[0] * x[0]
emilmont 1:fdd22bb7aa52 93 * acc1 = b[numTaps-1] * x[n-numTaps] + b[numTaps-2] * x[n-numTaps-1] + b[numTaps-3] * x[n-numTaps-2] +...+ b[0] * x[1]
emilmont 1:fdd22bb7aa52 94 * acc2 = b[numTaps-1] * x[n-numTaps+1] + b[numTaps-2] * x[n-numTaps] + b[numTaps-3] * x[n-numTaps-1] +...+ b[0] * x[2]
emilmont 1:fdd22bb7aa52 95 * acc3 = b[numTaps-1] * x[n-numTaps+2] + b[numTaps-2] * x[n-numTaps+1] + b[numTaps-3] * x[n-numTaps] +...+ b[0] * x[3]
emilmont 1:fdd22bb7aa52 96 */
emilmont 1:fdd22bb7aa52 97
emilmont 1:fdd22bb7aa52 98 blkCnt = blockSize >> 2;
emilmont 1:fdd22bb7aa52 99
emilmont 1:fdd22bb7aa52 100 /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
emilmont 1:fdd22bb7aa52 101 ** a second loop below computes the remaining 1 to 3 samples. */
emilmont 1:fdd22bb7aa52 102 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 103 {
emilmont 1:fdd22bb7aa52 104 /* Copy four new input samples into the state buffer.
emilmont 1:fdd22bb7aa52 105 ** Use 32-bit SIMD to move the 16-bit data. Only requires two copies. */
emilmont 1:fdd22bb7aa52 106 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 107 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 108 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 109 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 110
emilmont 1:fdd22bb7aa52 111
emilmont 1:fdd22bb7aa52 112 /* Set all accumulators to zero */
emilmont 1:fdd22bb7aa52 113 acc0 = 0;
emilmont 1:fdd22bb7aa52 114 acc1 = 0;
emilmont 1:fdd22bb7aa52 115 acc2 = 0;
emilmont 1:fdd22bb7aa52 116 acc3 = 0;
emilmont 1:fdd22bb7aa52 117
emilmont 1:fdd22bb7aa52 118 /* Typecast q15_t pointer to q31_t pointer for state reading in q31_t */
emilmont 1:fdd22bb7aa52 119 px = pState;
emilmont 1:fdd22bb7aa52 120
emilmont 1:fdd22bb7aa52 121 /* Typecast q15_t pointer to q31_t pointer for coefficient reading in q31_t */
emilmont 1:fdd22bb7aa52 122 pb = pCoeffs;
emilmont 1:fdd22bb7aa52 123
emilmont 1:fdd22bb7aa52 124 /* Read the first two samples from the state buffer: x[n-N], x[n-N-1] */
emilmont 1:fdd22bb7aa52 125 x0 = *__SIMD32(px)++;
emilmont 1:fdd22bb7aa52 126
emilmont 1:fdd22bb7aa52 127 /* Read the third and forth samples from the state buffer: x[n-N-2], x[n-N-3] */
emilmont 1:fdd22bb7aa52 128 x2 = *__SIMD32(px)++;
emilmont 1:fdd22bb7aa52 129
emilmont 1:fdd22bb7aa52 130 /* Loop over the number of taps. Unroll by a factor of 4.
emilmont 1:fdd22bb7aa52 131 ** Repeat until we've computed numTaps-(numTaps%4) coefficients. */
emilmont 1:fdd22bb7aa52 132 tapCnt = numTaps >> 2;
emilmont 1:fdd22bb7aa52 133
emilmont 1:fdd22bb7aa52 134 while(tapCnt > 0)
emilmont 1:fdd22bb7aa52 135 {
emilmont 1:fdd22bb7aa52 136 /* Read the first two coefficients using SIMD: b[N] and b[N-1] coefficients */
emilmont 1:fdd22bb7aa52 137 c0 = *__SIMD32(pb)++;
emilmont 1:fdd22bb7aa52 138
emilmont 1:fdd22bb7aa52 139 /* acc0 += b[N] * x[n-N] + b[N-1] * x[n-N-1] */
emilmont 1:fdd22bb7aa52 140 acc0 = __SMLAD(x0, c0, acc0);
emilmont 1:fdd22bb7aa52 141
emilmont 1:fdd22bb7aa52 142 /* acc2 += b[N] * x[n-N-2] + b[N-1] * x[n-N-3] */
emilmont 1:fdd22bb7aa52 143 acc2 = __SMLAD(x2, c0, acc2);
emilmont 1:fdd22bb7aa52 144
emilmont 1:fdd22bb7aa52 145 /* pack x[n-N-1] and x[n-N-2] */
emilmont 1:fdd22bb7aa52 146 #ifndef ARM_MATH_BIG_ENDIAN
emilmont 1:fdd22bb7aa52 147 x1 = __PKHBT(x2, x0, 0);
emilmont 1:fdd22bb7aa52 148 #else
emilmont 1:fdd22bb7aa52 149 x1 = __PKHBT(x0, x2, 0);
emilmont 1:fdd22bb7aa52 150 #endif
emilmont 1:fdd22bb7aa52 151
emilmont 1:fdd22bb7aa52 152 /* Read state x[n-N-4], x[n-N-5] */
emilmont 1:fdd22bb7aa52 153 x0 = _SIMD32_OFFSET(px);
emilmont 1:fdd22bb7aa52 154
emilmont 1:fdd22bb7aa52 155 /* acc1 += b[N] * x[n-N-1] + b[N-1] * x[n-N-2] */
emilmont 1:fdd22bb7aa52 156 acc1 = __SMLADX(x1, c0, acc1);
emilmont 1:fdd22bb7aa52 157
emilmont 1:fdd22bb7aa52 158 /* pack x[n-N-3] and x[n-N-4] */
emilmont 1:fdd22bb7aa52 159 #ifndef ARM_MATH_BIG_ENDIAN
emilmont 1:fdd22bb7aa52 160 x1 = __PKHBT(x0, x2, 0);
emilmont 1:fdd22bb7aa52 161 #else
emilmont 1:fdd22bb7aa52 162 x1 = __PKHBT(x2, x0, 0);
emilmont 1:fdd22bb7aa52 163 #endif
emilmont 1:fdd22bb7aa52 164
emilmont 1:fdd22bb7aa52 165 /* acc3 += b[N] * x[n-N-3] + b[N-1] * x[n-N-4] */
emilmont 1:fdd22bb7aa52 166 acc3 = __SMLADX(x1, c0, acc3);
emilmont 1:fdd22bb7aa52 167
emilmont 1:fdd22bb7aa52 168 /* Read coefficients b[N-2], b[N-3] */
emilmont 1:fdd22bb7aa52 169 c0 = *__SIMD32(pb)++;
emilmont 1:fdd22bb7aa52 170
emilmont 1:fdd22bb7aa52 171 /* acc0 += b[N-2] * x[n-N-2] + b[N-3] * x[n-N-3] */
emilmont 1:fdd22bb7aa52 172 acc0 = __SMLAD(x2, c0, acc0);
emilmont 1:fdd22bb7aa52 173
emilmont 1:fdd22bb7aa52 174 /* Read state x[n-N-6], x[n-N-7] with offset */
emilmont 1:fdd22bb7aa52 175 x2 = _SIMD32_OFFSET(px + 2u);
emilmont 1:fdd22bb7aa52 176
emilmont 1:fdd22bb7aa52 177 /* acc2 += b[N-2] * x[n-N-4] + b[N-3] * x[n-N-5] */
emilmont 1:fdd22bb7aa52 178 acc2 = __SMLAD(x0, c0, acc2);
emilmont 1:fdd22bb7aa52 179
emilmont 1:fdd22bb7aa52 180 /* acc1 += b[N-2] * x[n-N-3] + b[N-3] * x[n-N-4] */
emilmont 1:fdd22bb7aa52 181 acc1 = __SMLADX(x1, c0, acc1);
emilmont 1:fdd22bb7aa52 182
emilmont 1:fdd22bb7aa52 183 /* pack x[n-N-5] and x[n-N-6] */
emilmont 1:fdd22bb7aa52 184 #ifndef ARM_MATH_BIG_ENDIAN
emilmont 1:fdd22bb7aa52 185 x1 = __PKHBT(x2, x0, 0);
emilmont 1:fdd22bb7aa52 186 #else
emilmont 1:fdd22bb7aa52 187 x1 = __PKHBT(x0, x2, 0);
emilmont 1:fdd22bb7aa52 188 #endif
emilmont 1:fdd22bb7aa52 189
emilmont 1:fdd22bb7aa52 190 /* acc3 += b[N-2] * x[n-N-5] + b[N-3] * x[n-N-6] */
emilmont 1:fdd22bb7aa52 191 acc3 = __SMLADX(x1, c0, acc3);
emilmont 1:fdd22bb7aa52 192
emilmont 1:fdd22bb7aa52 193 /* Update state pointer for next state reading */
emilmont 1:fdd22bb7aa52 194 px += 4u;
emilmont 1:fdd22bb7aa52 195
emilmont 1:fdd22bb7aa52 196 /* Decrement tap count */
emilmont 1:fdd22bb7aa52 197 tapCnt--;
emilmont 1:fdd22bb7aa52 198
emilmont 1:fdd22bb7aa52 199 }
emilmont 1:fdd22bb7aa52 200
emilmont 1:fdd22bb7aa52 201 /* If the filter length is not a multiple of 4, compute the remaining filter taps.
emilmont 1:fdd22bb7aa52 202 ** This is always be 2 taps since the filter length is even. */
emilmont 1:fdd22bb7aa52 203 if((numTaps & 0x3u) != 0u)
emilmont 1:fdd22bb7aa52 204 {
emilmont 1:fdd22bb7aa52 205
emilmont 1:fdd22bb7aa52 206 /* Read last two coefficients */
emilmont 1:fdd22bb7aa52 207 c0 = *__SIMD32(pb)++;
emilmont 1:fdd22bb7aa52 208
emilmont 1:fdd22bb7aa52 209 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 210 acc0 = __SMLAD(x0, c0, acc0);
emilmont 1:fdd22bb7aa52 211 acc2 = __SMLAD(x2, c0, acc2);
emilmont 1:fdd22bb7aa52 212
emilmont 1:fdd22bb7aa52 213 /* pack state variables */
emilmont 1:fdd22bb7aa52 214 #ifndef ARM_MATH_BIG_ENDIAN
emilmont 1:fdd22bb7aa52 215 x1 = __PKHBT(x2, x0, 0);
emilmont 1:fdd22bb7aa52 216 #else
emilmont 1:fdd22bb7aa52 217 x1 = __PKHBT(x0, x2, 0);
emilmont 1:fdd22bb7aa52 218 #endif
emilmont 1:fdd22bb7aa52 219
emilmont 1:fdd22bb7aa52 220 /* Read last state variables */
emilmont 1:fdd22bb7aa52 221 x0 = *__SIMD32(px);
emilmont 1:fdd22bb7aa52 222
emilmont 1:fdd22bb7aa52 223 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 224 acc1 = __SMLADX(x1, c0, acc1);
emilmont 1:fdd22bb7aa52 225
emilmont 1:fdd22bb7aa52 226 /* pack state variables */
emilmont 1:fdd22bb7aa52 227 #ifndef ARM_MATH_BIG_ENDIAN
emilmont 1:fdd22bb7aa52 228 x1 = __PKHBT(x0, x2, 0);
emilmont 1:fdd22bb7aa52 229 #else
emilmont 1:fdd22bb7aa52 230 x1 = __PKHBT(x2, x0, 0);
emilmont 1:fdd22bb7aa52 231 #endif
emilmont 1:fdd22bb7aa52 232
emilmont 1:fdd22bb7aa52 233 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 234 acc3 = __SMLADX(x1, c0, acc3);
emilmont 1:fdd22bb7aa52 235 }
emilmont 1:fdd22bb7aa52 236
emilmont 1:fdd22bb7aa52 237 /* The results in the 4 accumulators are in 2.30 format. Convert to 1.15 with saturation.
emilmont 1:fdd22bb7aa52 238 ** Then store the 4 outputs in the destination buffer. */
emilmont 1:fdd22bb7aa52 239
emilmont 1:fdd22bb7aa52 240 #ifndef ARM_MATH_BIG_ENDIAN
emilmont 1:fdd22bb7aa52 241
emilmont 1:fdd22bb7aa52 242 *__SIMD32(pDst)++ =
emilmont 1:fdd22bb7aa52 243 __PKHBT(__SSAT((acc0 >> 15), 16), __SSAT((acc1 >> 15), 16), 16);
emilmont 1:fdd22bb7aa52 244
emilmont 1:fdd22bb7aa52 245 *__SIMD32(pDst)++ =
emilmont 1:fdd22bb7aa52 246 __PKHBT(__SSAT((acc2 >> 15), 16), __SSAT((acc3 >> 15), 16), 16);
emilmont 1:fdd22bb7aa52 247
emilmont 1:fdd22bb7aa52 248 #else
emilmont 1:fdd22bb7aa52 249
emilmont 1:fdd22bb7aa52 250 *__SIMD32(pDst)++ =
emilmont 1:fdd22bb7aa52 251 __PKHBT(__SSAT((acc1 >> 15), 16), __SSAT((acc0 >> 15), 16), 16);
emilmont 1:fdd22bb7aa52 252
emilmont 1:fdd22bb7aa52 253 *__SIMD32(pDst)++ =
emilmont 1:fdd22bb7aa52 254 __PKHBT(__SSAT((acc3 >> 15), 16), __SSAT((acc2 >> 15), 16), 16);
emilmont 1:fdd22bb7aa52 255
emilmont 1:fdd22bb7aa52 256
emilmont 1:fdd22bb7aa52 257 #endif /* #ifndef ARM_MATH_BIG_ENDIAN */
emilmont 1:fdd22bb7aa52 258
emilmont 1:fdd22bb7aa52 259 /* Advance the state pointer by 4 to process the next group of 4 samples */
emilmont 1:fdd22bb7aa52 260 pState = pState + 4u;
emilmont 1:fdd22bb7aa52 261
emilmont 1:fdd22bb7aa52 262 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 263 blkCnt--;
emilmont 1:fdd22bb7aa52 264 }
emilmont 1:fdd22bb7aa52 265
emilmont 1:fdd22bb7aa52 266 /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
emilmont 1:fdd22bb7aa52 267 ** No loop unrolling is used. */
emilmont 1:fdd22bb7aa52 268 blkCnt = blockSize % 0x4u;
emilmont 1:fdd22bb7aa52 269 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 270 {
emilmont 1:fdd22bb7aa52 271 /* Copy two samples into state buffer */
emilmont 1:fdd22bb7aa52 272 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 273
emilmont 1:fdd22bb7aa52 274 /* Set the accumulator to zero */
emilmont 1:fdd22bb7aa52 275 acc0 = 0;
emilmont 1:fdd22bb7aa52 276
emilmont 1:fdd22bb7aa52 277 /* Use SIMD to hold states and coefficients */
emilmont 1:fdd22bb7aa52 278 px = pState;
emilmont 1:fdd22bb7aa52 279 pb = pCoeffs;
emilmont 1:fdd22bb7aa52 280
emilmont 1:fdd22bb7aa52 281 tapCnt = numTaps >> 1u;
emilmont 1:fdd22bb7aa52 282
emilmont 1:fdd22bb7aa52 283 do
emilmont 1:fdd22bb7aa52 284 {
emilmont 1:fdd22bb7aa52 285
emilmont 1:fdd22bb7aa52 286 acc0 += (q31_t) * px++ * *pb++;
emilmont 1:fdd22bb7aa52 287 acc0 += (q31_t) * px++ * *pb++;
emilmont 1:fdd22bb7aa52 288
emilmont 1:fdd22bb7aa52 289 tapCnt--;
emilmont 1:fdd22bb7aa52 290 }
emilmont 1:fdd22bb7aa52 291 while(tapCnt > 0u);
emilmont 1:fdd22bb7aa52 292
emilmont 1:fdd22bb7aa52 293 /* The result is in 2.30 format. Convert to 1.15 with saturation.
emilmont 1:fdd22bb7aa52 294 ** Then store the output in the destination buffer. */
emilmont 1:fdd22bb7aa52 295 *pDst++ = (q15_t) (__SSAT((acc0 >> 15), 16));
emilmont 1:fdd22bb7aa52 296
emilmont 1:fdd22bb7aa52 297 /* Advance state pointer by 1 for the next sample */
emilmont 1:fdd22bb7aa52 298 pState = pState + 1u;
emilmont 1:fdd22bb7aa52 299
emilmont 1:fdd22bb7aa52 300 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 301 blkCnt--;
emilmont 1:fdd22bb7aa52 302 }
emilmont 1:fdd22bb7aa52 303
emilmont 1:fdd22bb7aa52 304 /* Processing is complete.
emilmont 1:fdd22bb7aa52 305 ** Now copy the last numTaps - 1 samples to the satrt of the state buffer.
emilmont 1:fdd22bb7aa52 306 ** This prepares the state buffer for the next function call. */
emilmont 1:fdd22bb7aa52 307
emilmont 1:fdd22bb7aa52 308 /* Points to the start of the state buffer */
emilmont 1:fdd22bb7aa52 309 pStateCurnt = S->pState;
emilmont 1:fdd22bb7aa52 310
emilmont 1:fdd22bb7aa52 311 /* Calculation of count for copying integer writes */
emilmont 1:fdd22bb7aa52 312 tapCnt = (numTaps - 1u) >> 2;
emilmont 1:fdd22bb7aa52 313
emilmont 1:fdd22bb7aa52 314 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 315 {
emilmont 1:fdd22bb7aa52 316 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 317 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 318 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 319 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 320
emilmont 1:fdd22bb7aa52 321 tapCnt--;
emilmont 1:fdd22bb7aa52 322
emilmont 1:fdd22bb7aa52 323 }
emilmont 1:fdd22bb7aa52 324
emilmont 1:fdd22bb7aa52 325 /* Calculation of count for remaining q15_t data */
emilmont 1:fdd22bb7aa52 326 tapCnt = (numTaps - 1u) % 0x4u;
emilmont 1:fdd22bb7aa52 327
emilmont 1:fdd22bb7aa52 328 /* copy remaining data */
emilmont 1:fdd22bb7aa52 329 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 330 {
emilmont 1:fdd22bb7aa52 331 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 332
emilmont 1:fdd22bb7aa52 333 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 334 tapCnt--;
emilmont 1:fdd22bb7aa52 335 }
emilmont 1:fdd22bb7aa52 336
emilmont 1:fdd22bb7aa52 337 }
emilmont 1:fdd22bb7aa52 338
emilmont 1:fdd22bb7aa52 339 /**
emilmont 1:fdd22bb7aa52 340 * @} end of FIR group
emilmont 1:fdd22bb7aa52 341 */