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_decimate_fast_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?

UserRevisionLine numberNew 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_fir_decimate_fast_q15.c
emilmont 1:fdd22bb7aa52 9 *
emilmont 2:da51fb522205 10 * Description: Fast Q15 FIR Decimator.
emilmont 1:fdd22bb7aa52 11 *
emilmont 1:fdd22bb7aa52 12 * Target Processor: Cortex-M4/Cortex-M3
emilmont 1:fdd22bb7aa52 13 *
mbed_official 3:7a284390b0ce 14 * Redistribution and use in source and binary forms, with or without
mbed_official 3:7a284390b0ce 15 * modification, are permitted provided that the following conditions
mbed_official 3:7a284390b0ce 16 * are met:
mbed_official 3:7a284390b0ce 17 * - Redistributions of source code must retain the above copyright
mbed_official 3:7a284390b0ce 18 * notice, this list of conditions and the following disclaimer.
mbed_official 3:7a284390b0ce 19 * - Redistributions in binary form must reproduce the above copyright
mbed_official 3:7a284390b0ce 20 * notice, this list of conditions and the following disclaimer in
mbed_official 3:7a284390b0ce 21 * the documentation and/or other materials provided with the
mbed_official 3:7a284390b0ce 22 * distribution.
mbed_official 3:7a284390b0ce 23 * - Neither the name of ARM LIMITED nor the names of its contributors
mbed_official 3:7a284390b0ce 24 * may be used to endorse or promote products derived from this
mbed_official 3:7a284390b0ce 25 * software without specific prior written permission.
mbed_official 3:7a284390b0ce 26 *
mbed_official 3:7a284390b0ce 27 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
mbed_official 3:7a284390b0ce 28 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
mbed_official 3:7a284390b0ce 29 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
mbed_official 3:7a284390b0ce 30 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
mbed_official 3:7a284390b0ce 31 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
mbed_official 3:7a284390b0ce 32 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
mbed_official 3:7a284390b0ce 33 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
mbed_official 3:7a284390b0ce 34 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
mbed_official 3:7a284390b0ce 35 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
mbed_official 3:7a284390b0ce 36 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
mbed_official 3:7a284390b0ce 37 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
mbed_official 3:7a284390b0ce 38 * POSSIBILITY OF SUCH DAMAGE.
emilmont 1:fdd22bb7aa52 39 * -------------------------------------------------------------------- */
emilmont 1:fdd22bb7aa52 40
emilmont 1:fdd22bb7aa52 41 #include "arm_math.h"
emilmont 1:fdd22bb7aa52 42
emilmont 1:fdd22bb7aa52 43 /**
emilmont 1:fdd22bb7aa52 44 * @ingroup groupFilters
emilmont 1:fdd22bb7aa52 45 */
emilmont 1:fdd22bb7aa52 46
emilmont 1:fdd22bb7aa52 47 /**
emilmont 1:fdd22bb7aa52 48 * @addtogroup FIR_decimate
emilmont 1:fdd22bb7aa52 49 * @{
emilmont 1:fdd22bb7aa52 50 */
emilmont 1:fdd22bb7aa52 51
emilmont 1:fdd22bb7aa52 52 /**
emilmont 1:fdd22bb7aa52 53 * @brief Processing function for the Q15 FIR decimator (fast variant) for Cortex-M3 and Cortex-M4.
emilmont 1:fdd22bb7aa52 54 * @param[in] *S points to an instance of the Q15 FIR decimator structure.
emilmont 1:fdd22bb7aa52 55 * @param[in] *pSrc points to the block of input data.
emilmont 1:fdd22bb7aa52 56 * @param[out] *pDst points to the block of output data
emilmont 1:fdd22bb7aa52 57 * @param[in] blockSize number of input samples to process per call.
emilmont 1:fdd22bb7aa52 58 * @return none
emilmont 1:fdd22bb7aa52 59 *
emilmont 1:fdd22bb7aa52 60 * \par Restrictions
emilmont 1:fdd22bb7aa52 61 * If the silicon does not support unaligned memory access enable the macro UNALIGNED_SUPPORT_DISABLE
emilmont 2:da51fb522205 62 * In this case input, output, state buffers should be aligned by 32-bit
emilmont 1:fdd22bb7aa52 63 *
emilmont 1:fdd22bb7aa52 64 * <b>Scaling and Overflow Behavior:</b>
emilmont 1:fdd22bb7aa52 65 * \par
emilmont 1:fdd22bb7aa52 66 * This fast version uses a 32-bit accumulator with 2.30 format.
emilmont 1:fdd22bb7aa52 67 * The accumulator maintains full precision of the intermediate multiplication results but provides only a single guard bit.
emilmont 1:fdd22bb7aa52 68 * Thus, if the accumulator result overflows it wraps around and distorts the result.
emilmont 1:fdd22bb7aa52 69 * In order to avoid overflows completely the input signal must be scaled down by log2(numTaps) bits (log2 is read as log to the base 2).
emilmont 1:fdd22bb7aa52 70 * The 2.30 accumulator is then truncated to 2.15 format and saturated to yield the 1.15 result.
emilmont 1:fdd22bb7aa52 71 *
emilmont 1:fdd22bb7aa52 72 * \par
emilmont 1:fdd22bb7aa52 73 * Refer to the function <code>arm_fir_decimate_q15()</code> for a slower implementation of this function which uses 64-bit accumulation to avoid wrap around distortion.
emilmont 1:fdd22bb7aa52 74 * Both the slow and the fast versions use the same instance structure.
emilmont 1:fdd22bb7aa52 75 * Use the function <code>arm_fir_decimate_init_q15()</code> to initialize the filter structure.
emilmont 1:fdd22bb7aa52 76 */
emilmont 1:fdd22bb7aa52 77
emilmont 1:fdd22bb7aa52 78 #ifndef UNALIGNED_SUPPORT_DISABLE
emilmont 1:fdd22bb7aa52 79
emilmont 1:fdd22bb7aa52 80 void arm_fir_decimate_fast_q15(
emilmont 1:fdd22bb7aa52 81 const arm_fir_decimate_instance_q15 * S,
emilmont 1:fdd22bb7aa52 82 q15_t * pSrc,
emilmont 1:fdd22bb7aa52 83 q15_t * pDst,
emilmont 1:fdd22bb7aa52 84 uint32_t blockSize)
emilmont 1:fdd22bb7aa52 85 {
emilmont 1:fdd22bb7aa52 86 q15_t *pState = S->pState; /* State pointer */
emilmont 1:fdd22bb7aa52 87 q15_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
emilmont 1:fdd22bb7aa52 88 q15_t *pStateCurnt; /* Points to the current sample of the state */
emilmont 1:fdd22bb7aa52 89 q15_t *px; /* Temporary pointer for state buffer */
emilmont 1:fdd22bb7aa52 90 q15_t *pb; /* Temporary pointer coefficient buffer */
emilmont 1:fdd22bb7aa52 91 q31_t x0, x1, c0, c1; /* Temporary variables to hold state and coefficient values */
emilmont 1:fdd22bb7aa52 92 q31_t sum0; /* Accumulators */
emilmont 1:fdd22bb7aa52 93 q31_t acc0, acc1;
emilmont 1:fdd22bb7aa52 94 q15_t *px0, *px1;
emilmont 1:fdd22bb7aa52 95 uint32_t blkCntN3;
emilmont 1:fdd22bb7aa52 96 uint32_t numTaps = S->numTaps; /* Number of taps */
emilmont 1:fdd22bb7aa52 97 uint32_t i, blkCnt, tapCnt, outBlockSize = blockSize / S->M; /* Loop counters */
emilmont 1:fdd22bb7aa52 98
emilmont 1:fdd22bb7aa52 99
emilmont 1:fdd22bb7aa52 100 /* S->pState buffer contains previous frame (numTaps - 1) samples */
emilmont 1:fdd22bb7aa52 101 /* pStateCurnt points to the location where the new input data should be written */
emilmont 1:fdd22bb7aa52 102 pStateCurnt = S->pState + (numTaps - 1u);
emilmont 1:fdd22bb7aa52 103
emilmont 1:fdd22bb7aa52 104
emilmont 1:fdd22bb7aa52 105 /* Total number of output samples to be computed */
emilmont 1:fdd22bb7aa52 106 blkCnt = outBlockSize / 2;
emilmont 1:fdd22bb7aa52 107 blkCntN3 = outBlockSize - (2 * blkCnt);
emilmont 1:fdd22bb7aa52 108
emilmont 1:fdd22bb7aa52 109
emilmont 1:fdd22bb7aa52 110 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 111 {
emilmont 1:fdd22bb7aa52 112 /* Copy decimation factor number of new input samples into the state buffer */
emilmont 1:fdd22bb7aa52 113 i = 2 * S->M;
emilmont 1:fdd22bb7aa52 114
emilmont 1:fdd22bb7aa52 115 do
emilmont 1:fdd22bb7aa52 116 {
emilmont 1:fdd22bb7aa52 117 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 118
emilmont 1:fdd22bb7aa52 119 } while(--i);
emilmont 1:fdd22bb7aa52 120
emilmont 1:fdd22bb7aa52 121 /* Set accumulator to zero */
emilmont 1:fdd22bb7aa52 122 acc0 = 0;
emilmont 1:fdd22bb7aa52 123 acc1 = 0;
emilmont 1:fdd22bb7aa52 124
emilmont 1:fdd22bb7aa52 125 /* Initialize state pointer */
emilmont 1:fdd22bb7aa52 126 px0 = pState;
emilmont 1:fdd22bb7aa52 127
emilmont 1:fdd22bb7aa52 128 px1 = pState + S->M;
emilmont 1:fdd22bb7aa52 129
emilmont 1:fdd22bb7aa52 130
emilmont 1:fdd22bb7aa52 131 /* Initialize coeff pointer */
emilmont 1:fdd22bb7aa52 132 pb = pCoeffs;
emilmont 1:fdd22bb7aa52 133
emilmont 1:fdd22bb7aa52 134 /* Loop unrolling. Process 4 taps at a time. */
emilmont 1:fdd22bb7aa52 135 tapCnt = numTaps >> 2;
emilmont 1:fdd22bb7aa52 136
emilmont 1:fdd22bb7aa52 137 /* Loop over the number of taps. Unroll by a factor of 4.
emilmont 1:fdd22bb7aa52 138 ** Repeat until we've computed numTaps-4 coefficients. */
emilmont 1:fdd22bb7aa52 139 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 140 {
emilmont 1:fdd22bb7aa52 141 /* Read the Read b[numTaps-1] and b[numTaps-2] coefficients */
emilmont 1:fdd22bb7aa52 142 c0 = *__SIMD32(pb)++;
emilmont 1:fdd22bb7aa52 143
emilmont 1:fdd22bb7aa52 144 /* Read x[n-numTaps-1] and x[n-numTaps-2]sample */
emilmont 1:fdd22bb7aa52 145 x0 = *__SIMD32(px0)++;
emilmont 1:fdd22bb7aa52 146
emilmont 1:fdd22bb7aa52 147 x1 = *__SIMD32(px1)++;
emilmont 1:fdd22bb7aa52 148
emilmont 1:fdd22bb7aa52 149 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 150 acc0 = __SMLAD(x0, c0, acc0);
emilmont 1:fdd22bb7aa52 151
emilmont 1:fdd22bb7aa52 152 acc1 = __SMLAD(x1, c0, acc1);
emilmont 1:fdd22bb7aa52 153
emilmont 1:fdd22bb7aa52 154 /* Read the b[numTaps-3] and b[numTaps-4] coefficient */
emilmont 1:fdd22bb7aa52 155 c0 = *__SIMD32(pb)++;
emilmont 1:fdd22bb7aa52 156
emilmont 1:fdd22bb7aa52 157 /* Read x[n-numTaps-2] and x[n-numTaps-3] sample */
emilmont 1:fdd22bb7aa52 158 x0 = *__SIMD32(px0)++;
emilmont 1:fdd22bb7aa52 159
emilmont 1:fdd22bb7aa52 160 x1 = *__SIMD32(px1)++;
emilmont 1:fdd22bb7aa52 161
emilmont 1:fdd22bb7aa52 162 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 163 acc0 = __SMLAD(x0, c0, acc0);
emilmont 1:fdd22bb7aa52 164
emilmont 1:fdd22bb7aa52 165 acc1 = __SMLAD(x1, c0, acc1);
emilmont 1:fdd22bb7aa52 166
emilmont 1:fdd22bb7aa52 167 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 168 tapCnt--;
emilmont 1:fdd22bb7aa52 169 }
emilmont 1:fdd22bb7aa52 170
emilmont 1:fdd22bb7aa52 171 /* If the filter length is not a multiple of 4, compute the remaining filter taps */
emilmont 1:fdd22bb7aa52 172 tapCnt = numTaps % 0x4u;
emilmont 1:fdd22bb7aa52 173
emilmont 1:fdd22bb7aa52 174 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 175 {
emilmont 1:fdd22bb7aa52 176 /* Read coefficients */
emilmont 1:fdd22bb7aa52 177 c0 = *pb++;
emilmont 1:fdd22bb7aa52 178
emilmont 1:fdd22bb7aa52 179 /* Fetch 1 state variable */
emilmont 1:fdd22bb7aa52 180 x0 = *px0++;
emilmont 1:fdd22bb7aa52 181
emilmont 1:fdd22bb7aa52 182 x1 = *px1++;
emilmont 1:fdd22bb7aa52 183
emilmont 1:fdd22bb7aa52 184 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 185 acc0 = __SMLAD(x0, c0, acc0);
emilmont 1:fdd22bb7aa52 186 acc1 = __SMLAD(x1, c0, acc1);
emilmont 1:fdd22bb7aa52 187
emilmont 1:fdd22bb7aa52 188 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 189 tapCnt--;
emilmont 1:fdd22bb7aa52 190 }
emilmont 1:fdd22bb7aa52 191
emilmont 1:fdd22bb7aa52 192 /* Advance the state pointer by the decimation factor
emilmont 1:fdd22bb7aa52 193 * to process the next group of decimation factor number samples */
emilmont 1:fdd22bb7aa52 194 pState = pState + S->M * 2;
emilmont 1:fdd22bb7aa52 195
emilmont 1:fdd22bb7aa52 196 /* Store filter output, smlad returns the values in 2.14 format */
emilmont 1:fdd22bb7aa52 197 /* so downsacle by 15 to get output in 1.15 */
emilmont 1:fdd22bb7aa52 198 *pDst++ = (q15_t) (__SSAT((acc0 >> 15), 16));
emilmont 1:fdd22bb7aa52 199 *pDst++ = (q15_t) (__SSAT((acc1 >> 15), 16));
emilmont 1:fdd22bb7aa52 200
emilmont 1:fdd22bb7aa52 201 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 202 blkCnt--;
emilmont 1:fdd22bb7aa52 203 }
emilmont 1:fdd22bb7aa52 204
emilmont 1:fdd22bb7aa52 205
emilmont 1:fdd22bb7aa52 206
emilmont 1:fdd22bb7aa52 207 while(blkCntN3 > 0u)
emilmont 1:fdd22bb7aa52 208 {
emilmont 1:fdd22bb7aa52 209 /* Copy decimation factor number of new input samples into the state buffer */
emilmont 1:fdd22bb7aa52 210 i = S->M;
emilmont 1:fdd22bb7aa52 211
emilmont 1:fdd22bb7aa52 212 do
emilmont 1:fdd22bb7aa52 213 {
emilmont 1:fdd22bb7aa52 214 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 215
emilmont 1:fdd22bb7aa52 216 } while(--i);
emilmont 1:fdd22bb7aa52 217
emilmont 1:fdd22bb7aa52 218 /*Set sum to zero */
emilmont 1:fdd22bb7aa52 219 sum0 = 0;
emilmont 1:fdd22bb7aa52 220
emilmont 1:fdd22bb7aa52 221 /* Initialize state pointer */
emilmont 1:fdd22bb7aa52 222 px = pState;
emilmont 1:fdd22bb7aa52 223
emilmont 1:fdd22bb7aa52 224 /* Initialize coeff pointer */
emilmont 1:fdd22bb7aa52 225 pb = pCoeffs;
emilmont 1:fdd22bb7aa52 226
emilmont 1:fdd22bb7aa52 227 /* Loop unrolling. Process 4 taps at a time. */
emilmont 1:fdd22bb7aa52 228 tapCnt = numTaps >> 2;
emilmont 1:fdd22bb7aa52 229
emilmont 1:fdd22bb7aa52 230 /* Loop over the number of taps. Unroll by a factor of 4.
emilmont 1:fdd22bb7aa52 231 ** Repeat until we've computed numTaps-4 coefficients. */
emilmont 1:fdd22bb7aa52 232 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 233 {
emilmont 1:fdd22bb7aa52 234 /* Read the Read b[numTaps-1] and b[numTaps-2] coefficients */
emilmont 1:fdd22bb7aa52 235 c0 = *__SIMD32(pb)++;
emilmont 1:fdd22bb7aa52 236
emilmont 1:fdd22bb7aa52 237 /* Read x[n-numTaps-1] and x[n-numTaps-2]sample */
emilmont 1:fdd22bb7aa52 238 x0 = *__SIMD32(px)++;
emilmont 1:fdd22bb7aa52 239
emilmont 1:fdd22bb7aa52 240 /* Read the b[numTaps-3] and b[numTaps-4] coefficient */
emilmont 1:fdd22bb7aa52 241 c1 = *__SIMD32(pb)++;
emilmont 1:fdd22bb7aa52 242
emilmont 1:fdd22bb7aa52 243 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 244 sum0 = __SMLAD(x0, c0, sum0);
emilmont 1:fdd22bb7aa52 245
emilmont 1:fdd22bb7aa52 246 /* Read x[n-numTaps-2] and x[n-numTaps-3] sample */
emilmont 1:fdd22bb7aa52 247 x0 = *__SIMD32(px)++;
emilmont 1:fdd22bb7aa52 248
emilmont 1:fdd22bb7aa52 249 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 250 sum0 = __SMLAD(x0, c1, sum0);
emilmont 1:fdd22bb7aa52 251
emilmont 1:fdd22bb7aa52 252 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 253 tapCnt--;
emilmont 1:fdd22bb7aa52 254 }
emilmont 1:fdd22bb7aa52 255
emilmont 1:fdd22bb7aa52 256 /* If the filter length is not a multiple of 4, compute the remaining filter taps */
emilmont 1:fdd22bb7aa52 257 tapCnt = numTaps % 0x4u;
emilmont 1:fdd22bb7aa52 258
emilmont 1:fdd22bb7aa52 259 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 260 {
emilmont 1:fdd22bb7aa52 261 /* Read coefficients */
emilmont 1:fdd22bb7aa52 262 c0 = *pb++;
emilmont 1:fdd22bb7aa52 263
emilmont 1:fdd22bb7aa52 264 /* Fetch 1 state variable */
emilmont 1:fdd22bb7aa52 265 x0 = *px++;
emilmont 1:fdd22bb7aa52 266
emilmont 1:fdd22bb7aa52 267 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 268 sum0 = __SMLAD(x0, c0, sum0);
emilmont 1:fdd22bb7aa52 269
emilmont 1:fdd22bb7aa52 270 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 271 tapCnt--;
emilmont 1:fdd22bb7aa52 272 }
emilmont 1:fdd22bb7aa52 273
emilmont 1:fdd22bb7aa52 274 /* Advance the state pointer by the decimation factor
emilmont 1:fdd22bb7aa52 275 * to process the next group of decimation factor number samples */
emilmont 1:fdd22bb7aa52 276 pState = pState + S->M;
emilmont 1:fdd22bb7aa52 277
emilmont 1:fdd22bb7aa52 278 /* Store filter output, smlad returns the values in 2.14 format */
emilmont 1:fdd22bb7aa52 279 /* so downsacle by 15 to get output in 1.15 */
emilmont 1:fdd22bb7aa52 280 *pDst++ = (q15_t) (__SSAT((sum0 >> 15), 16));
emilmont 1:fdd22bb7aa52 281
emilmont 1:fdd22bb7aa52 282 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 283 blkCntN3--;
emilmont 1:fdd22bb7aa52 284 }
emilmont 1:fdd22bb7aa52 285
emilmont 1:fdd22bb7aa52 286 /* Processing is complete.
emilmont 1:fdd22bb7aa52 287 ** Now copy the last numTaps - 1 samples to the satrt of the state buffer.
emilmont 1:fdd22bb7aa52 288 ** This prepares the state buffer for the next function call. */
emilmont 1:fdd22bb7aa52 289
emilmont 1:fdd22bb7aa52 290 /* Points to the start of the state buffer */
emilmont 1:fdd22bb7aa52 291 pStateCurnt = S->pState;
emilmont 1:fdd22bb7aa52 292
emilmont 1:fdd22bb7aa52 293 i = (numTaps - 1u) >> 2u;
emilmont 1:fdd22bb7aa52 294
emilmont 1:fdd22bb7aa52 295 /* copy data */
emilmont 1:fdd22bb7aa52 296 while(i > 0u)
emilmont 1:fdd22bb7aa52 297 {
emilmont 1:fdd22bb7aa52 298 *__SIMD32(pStateCurnt)++ = *__SIMD32(pState)++;
emilmont 1:fdd22bb7aa52 299 *__SIMD32(pStateCurnt)++ = *__SIMD32(pState)++;
emilmont 1:fdd22bb7aa52 300
emilmont 1:fdd22bb7aa52 301 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 302 i--;
emilmont 1:fdd22bb7aa52 303 }
emilmont 1:fdd22bb7aa52 304
emilmont 1:fdd22bb7aa52 305 i = (numTaps - 1u) % 0x04u;
emilmont 1:fdd22bb7aa52 306
emilmont 1:fdd22bb7aa52 307 /* copy data */
emilmont 1:fdd22bb7aa52 308 while(i > 0u)
emilmont 1:fdd22bb7aa52 309 {
emilmont 1:fdd22bb7aa52 310 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 311
emilmont 1:fdd22bb7aa52 312 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 313 i--;
emilmont 1:fdd22bb7aa52 314 }
emilmont 1:fdd22bb7aa52 315 }
emilmont 1:fdd22bb7aa52 316
emilmont 1:fdd22bb7aa52 317 #else
emilmont 1:fdd22bb7aa52 318
emilmont 1:fdd22bb7aa52 319
emilmont 1:fdd22bb7aa52 320 void arm_fir_decimate_fast_q15(
emilmont 1:fdd22bb7aa52 321 const arm_fir_decimate_instance_q15 * S,
emilmont 1:fdd22bb7aa52 322 q15_t * pSrc,
emilmont 1:fdd22bb7aa52 323 q15_t * pDst,
emilmont 1:fdd22bb7aa52 324 uint32_t blockSize)
emilmont 1:fdd22bb7aa52 325 {
emilmont 1:fdd22bb7aa52 326 q15_t *pState = S->pState; /* State pointer */
emilmont 1:fdd22bb7aa52 327 q15_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
emilmont 1:fdd22bb7aa52 328 q15_t *pStateCurnt; /* Points to the current sample of the state */
emilmont 1:fdd22bb7aa52 329 q15_t *px; /* Temporary pointer for state buffer */
emilmont 1:fdd22bb7aa52 330 q15_t *pb; /* Temporary pointer coefficient buffer */
emilmont 1:fdd22bb7aa52 331 q15_t x0, x1, c0; /* Temporary variables to hold state and coefficient values */
emilmont 1:fdd22bb7aa52 332 q31_t sum0; /* Accumulators */
emilmont 1:fdd22bb7aa52 333 q31_t acc0, acc1;
emilmont 1:fdd22bb7aa52 334 q15_t *px0, *px1;
emilmont 1:fdd22bb7aa52 335 uint32_t blkCntN3;
emilmont 1:fdd22bb7aa52 336 uint32_t numTaps = S->numTaps; /* Number of taps */
emilmont 1:fdd22bb7aa52 337 uint32_t i, blkCnt, tapCnt, outBlockSize = blockSize / S->M; /* Loop counters */
emilmont 1:fdd22bb7aa52 338
emilmont 1:fdd22bb7aa52 339
emilmont 1:fdd22bb7aa52 340 /* S->pState buffer contains previous frame (numTaps - 1) samples */
emilmont 1:fdd22bb7aa52 341 /* pStateCurnt points to the location where the new input data should be written */
emilmont 1:fdd22bb7aa52 342 pStateCurnt = S->pState + (numTaps - 1u);
emilmont 1:fdd22bb7aa52 343
emilmont 1:fdd22bb7aa52 344
emilmont 1:fdd22bb7aa52 345 /* Total number of output samples to be computed */
emilmont 1:fdd22bb7aa52 346 blkCnt = outBlockSize / 2;
emilmont 1:fdd22bb7aa52 347 blkCntN3 = outBlockSize - (2 * blkCnt);
emilmont 1:fdd22bb7aa52 348
emilmont 1:fdd22bb7aa52 349 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 350 {
emilmont 1:fdd22bb7aa52 351 /* Copy decimation factor number of new input samples into the state buffer */
emilmont 1:fdd22bb7aa52 352 i = 2 * S->M;
emilmont 1:fdd22bb7aa52 353
emilmont 1:fdd22bb7aa52 354 do
emilmont 1:fdd22bb7aa52 355 {
emilmont 1:fdd22bb7aa52 356 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 357
emilmont 1:fdd22bb7aa52 358 } while(--i);
emilmont 1:fdd22bb7aa52 359
emilmont 1:fdd22bb7aa52 360 /* Set accumulator to zero */
emilmont 1:fdd22bb7aa52 361 acc0 = 0;
emilmont 1:fdd22bb7aa52 362 acc1 = 0;
emilmont 1:fdd22bb7aa52 363
emilmont 1:fdd22bb7aa52 364 /* Initialize state pointer */
emilmont 1:fdd22bb7aa52 365 px0 = pState;
emilmont 1:fdd22bb7aa52 366
emilmont 1:fdd22bb7aa52 367 px1 = pState + S->M;
emilmont 1:fdd22bb7aa52 368
emilmont 1:fdd22bb7aa52 369
emilmont 1:fdd22bb7aa52 370 /* Initialize coeff pointer */
emilmont 1:fdd22bb7aa52 371 pb = pCoeffs;
emilmont 1:fdd22bb7aa52 372
emilmont 1:fdd22bb7aa52 373 /* Loop unrolling. Process 4 taps at a time. */
emilmont 1:fdd22bb7aa52 374 tapCnt = numTaps >> 2;
emilmont 1:fdd22bb7aa52 375
emilmont 1:fdd22bb7aa52 376 /* Loop over the number of taps. Unroll by a factor of 4.
emilmont 1:fdd22bb7aa52 377 ** Repeat until we've computed numTaps-4 coefficients. */
emilmont 1:fdd22bb7aa52 378 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 379 {
emilmont 1:fdd22bb7aa52 380 /* Read the Read b[numTaps-1] coefficients */
emilmont 1:fdd22bb7aa52 381 c0 = *pb++;
emilmont 1:fdd22bb7aa52 382
emilmont 1:fdd22bb7aa52 383 /* Read x[n-numTaps-1] for sample 0 and for sample 1 */
emilmont 1:fdd22bb7aa52 384 x0 = *px0++;
emilmont 1:fdd22bb7aa52 385 x1 = *px1++;
emilmont 1:fdd22bb7aa52 386
emilmont 1:fdd22bb7aa52 387 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 388 acc0 += x0 * c0;
emilmont 1:fdd22bb7aa52 389 acc1 += x1 * c0;
emilmont 1:fdd22bb7aa52 390
emilmont 1:fdd22bb7aa52 391 /* Read the b[numTaps-2] coefficient */
emilmont 1:fdd22bb7aa52 392 c0 = *pb++;
emilmont 1:fdd22bb7aa52 393
emilmont 1:fdd22bb7aa52 394 /* Read x[n-numTaps-2] for sample 0 and sample 1 */
emilmont 1:fdd22bb7aa52 395 x0 = *px0++;
emilmont 1:fdd22bb7aa52 396 x1 = *px1++;
emilmont 1:fdd22bb7aa52 397
emilmont 1:fdd22bb7aa52 398 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 399 acc0 += x0 * c0;
emilmont 1:fdd22bb7aa52 400 acc1 += x1 * c0;
emilmont 1:fdd22bb7aa52 401
emilmont 1:fdd22bb7aa52 402 /* Read the b[numTaps-3] coefficients */
emilmont 1:fdd22bb7aa52 403 c0 = *pb++;
emilmont 1:fdd22bb7aa52 404
emilmont 1:fdd22bb7aa52 405 /* Read x[n-numTaps-3] for sample 0 and sample 1 */
emilmont 1:fdd22bb7aa52 406 x0 = *px0++;
emilmont 1:fdd22bb7aa52 407 x1 = *px1++;
emilmont 1:fdd22bb7aa52 408
emilmont 1:fdd22bb7aa52 409 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 410 acc0 += x0 * c0;
emilmont 1:fdd22bb7aa52 411 acc1 += x1 * c0;
emilmont 1:fdd22bb7aa52 412
emilmont 1:fdd22bb7aa52 413 /* Read the b[numTaps-4] coefficient */
emilmont 1:fdd22bb7aa52 414 c0 = *pb++;
emilmont 1:fdd22bb7aa52 415
emilmont 1:fdd22bb7aa52 416 /* Read x[n-numTaps-4] for sample 0 and sample 1 */
emilmont 1:fdd22bb7aa52 417 x0 = *px0++;
emilmont 1:fdd22bb7aa52 418 x1 = *px1++;
emilmont 1:fdd22bb7aa52 419
emilmont 1:fdd22bb7aa52 420 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 421 acc0 += x0 * c0;
emilmont 1:fdd22bb7aa52 422 acc1 += x1 * c0;
emilmont 1:fdd22bb7aa52 423
emilmont 1:fdd22bb7aa52 424 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 425 tapCnt--;
emilmont 1:fdd22bb7aa52 426 }
emilmont 1:fdd22bb7aa52 427
emilmont 1:fdd22bb7aa52 428 /* If the filter length is not a multiple of 4, compute the remaining filter taps */
emilmont 1:fdd22bb7aa52 429 tapCnt = numTaps % 0x4u;
emilmont 1:fdd22bb7aa52 430
emilmont 1:fdd22bb7aa52 431 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 432 {
emilmont 1:fdd22bb7aa52 433 /* Read coefficients */
emilmont 1:fdd22bb7aa52 434 c0 = *pb++;
emilmont 1:fdd22bb7aa52 435
emilmont 1:fdd22bb7aa52 436 /* Fetch 1 state variable */
emilmont 1:fdd22bb7aa52 437 x0 = *px0++;
emilmont 1:fdd22bb7aa52 438 x1 = *px1++;
emilmont 1:fdd22bb7aa52 439
emilmont 1:fdd22bb7aa52 440 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 441 acc0 += x0 * c0;
emilmont 1:fdd22bb7aa52 442 acc1 += x1 * c0;
emilmont 1:fdd22bb7aa52 443
emilmont 1:fdd22bb7aa52 444 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 445 tapCnt--;
emilmont 1:fdd22bb7aa52 446 }
emilmont 1:fdd22bb7aa52 447
emilmont 1:fdd22bb7aa52 448 /* Advance the state pointer by the decimation factor
emilmont 1:fdd22bb7aa52 449 * to process the next group of decimation factor number samples */
emilmont 1:fdd22bb7aa52 450 pState = pState + S->M * 2;
emilmont 1:fdd22bb7aa52 451
emilmont 1:fdd22bb7aa52 452 /* Store filter output, smlad returns the values in 2.14 format */
emilmont 1:fdd22bb7aa52 453 /* so downsacle by 15 to get output in 1.15 */
emilmont 1:fdd22bb7aa52 454
emilmont 1:fdd22bb7aa52 455 *pDst++ = (q15_t) (__SSAT((acc0 >> 15), 16));
emilmont 1:fdd22bb7aa52 456 *pDst++ = (q15_t) (__SSAT((acc1 >> 15), 16));
emilmont 1:fdd22bb7aa52 457
emilmont 1:fdd22bb7aa52 458
emilmont 1:fdd22bb7aa52 459 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 460 blkCnt--;
emilmont 1:fdd22bb7aa52 461 }
emilmont 1:fdd22bb7aa52 462
emilmont 1:fdd22bb7aa52 463 while(blkCntN3 > 0u)
emilmont 1:fdd22bb7aa52 464 {
emilmont 1:fdd22bb7aa52 465 /* Copy decimation factor number of new input samples into the state buffer */
emilmont 1:fdd22bb7aa52 466 i = S->M;
emilmont 1:fdd22bb7aa52 467
emilmont 1:fdd22bb7aa52 468 do
emilmont 1:fdd22bb7aa52 469 {
emilmont 1:fdd22bb7aa52 470 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 471
emilmont 1:fdd22bb7aa52 472 } while(--i);
emilmont 1:fdd22bb7aa52 473
emilmont 1:fdd22bb7aa52 474 /*Set sum to zero */
emilmont 1:fdd22bb7aa52 475 sum0 = 0;
emilmont 1:fdd22bb7aa52 476
emilmont 1:fdd22bb7aa52 477 /* Initialize state pointer */
emilmont 1:fdd22bb7aa52 478 px = pState;
emilmont 1:fdd22bb7aa52 479
emilmont 1:fdd22bb7aa52 480 /* Initialize coeff pointer */
emilmont 1:fdd22bb7aa52 481 pb = pCoeffs;
emilmont 1:fdd22bb7aa52 482
emilmont 1:fdd22bb7aa52 483 /* Loop unrolling. Process 4 taps at a time. */
emilmont 1:fdd22bb7aa52 484 tapCnt = numTaps >> 2;
emilmont 1:fdd22bb7aa52 485
emilmont 1:fdd22bb7aa52 486 /* Loop over the number of taps. Unroll by a factor of 4.
emilmont 1:fdd22bb7aa52 487 ** Repeat until we've computed numTaps-4 coefficients. */
emilmont 1:fdd22bb7aa52 488 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 489 {
emilmont 1:fdd22bb7aa52 490 /* Read the Read b[numTaps-1] coefficients */
emilmont 1:fdd22bb7aa52 491 c0 = *pb++;
emilmont 1:fdd22bb7aa52 492
emilmont 1:fdd22bb7aa52 493 /* Read x[n-numTaps-1] and sample */
emilmont 1:fdd22bb7aa52 494 x0 = *px++;
emilmont 1:fdd22bb7aa52 495
emilmont 1:fdd22bb7aa52 496 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 497 sum0 += x0 * c0;
emilmont 1:fdd22bb7aa52 498
emilmont 1:fdd22bb7aa52 499 /* Read the b[numTaps-2] coefficient */
emilmont 1:fdd22bb7aa52 500 c0 = *pb++;
emilmont 1:fdd22bb7aa52 501
emilmont 1:fdd22bb7aa52 502 /* Read x[n-numTaps-2] and sample */
emilmont 1:fdd22bb7aa52 503 x0 = *px++;
emilmont 1:fdd22bb7aa52 504
emilmont 1:fdd22bb7aa52 505 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 506 sum0 += x0 * c0;
emilmont 1:fdd22bb7aa52 507
emilmont 1:fdd22bb7aa52 508 /* Read the b[numTaps-3] coefficients */
emilmont 1:fdd22bb7aa52 509 c0 = *pb++;
emilmont 1:fdd22bb7aa52 510
emilmont 1:fdd22bb7aa52 511 /* Read x[n-numTaps-3] sample */
emilmont 1:fdd22bb7aa52 512 x0 = *px++;
emilmont 1:fdd22bb7aa52 513
emilmont 1:fdd22bb7aa52 514 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 515 sum0 += x0 * c0;
emilmont 1:fdd22bb7aa52 516
emilmont 1:fdd22bb7aa52 517 /* Read the b[numTaps-4] coefficient */
emilmont 1:fdd22bb7aa52 518 c0 = *pb++;
emilmont 1:fdd22bb7aa52 519
emilmont 1:fdd22bb7aa52 520 /* Read x[n-numTaps-4] sample */
emilmont 1:fdd22bb7aa52 521 x0 = *px++;
emilmont 1:fdd22bb7aa52 522
emilmont 1:fdd22bb7aa52 523 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 524 sum0 += x0 * c0;
emilmont 1:fdd22bb7aa52 525
emilmont 1:fdd22bb7aa52 526 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 527 tapCnt--;
emilmont 1:fdd22bb7aa52 528 }
emilmont 1:fdd22bb7aa52 529
emilmont 1:fdd22bb7aa52 530 /* If the filter length is not a multiple of 4, compute the remaining filter taps */
emilmont 1:fdd22bb7aa52 531 tapCnt = numTaps % 0x4u;
emilmont 1:fdd22bb7aa52 532
emilmont 1:fdd22bb7aa52 533 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 534 {
emilmont 1:fdd22bb7aa52 535 /* Read coefficients */
emilmont 1:fdd22bb7aa52 536 c0 = *pb++;
emilmont 1:fdd22bb7aa52 537
emilmont 1:fdd22bb7aa52 538 /* Fetch 1 state variable */
emilmont 1:fdd22bb7aa52 539 x0 = *px++;
emilmont 1:fdd22bb7aa52 540
emilmont 1:fdd22bb7aa52 541 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 542 sum0 += x0 * c0;
emilmont 1:fdd22bb7aa52 543
emilmont 1:fdd22bb7aa52 544 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 545 tapCnt--;
emilmont 1:fdd22bb7aa52 546 }
emilmont 1:fdd22bb7aa52 547
emilmont 1:fdd22bb7aa52 548 /* Advance the state pointer by the decimation factor
emilmont 1:fdd22bb7aa52 549 * to process the next group of decimation factor number samples */
emilmont 1:fdd22bb7aa52 550 pState = pState + S->M;
emilmont 1:fdd22bb7aa52 551
emilmont 1:fdd22bb7aa52 552 /* Store filter output, smlad returns the values in 2.14 format */
emilmont 1:fdd22bb7aa52 553 /* so downsacle by 15 to get output in 1.15 */
emilmont 1:fdd22bb7aa52 554 *pDst++ = (q15_t) (__SSAT((sum0 >> 15), 16));
emilmont 1:fdd22bb7aa52 555
emilmont 1:fdd22bb7aa52 556 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 557 blkCntN3--;
emilmont 1:fdd22bb7aa52 558 }
emilmont 1:fdd22bb7aa52 559
emilmont 1:fdd22bb7aa52 560 /* Processing is complete.
emilmont 1:fdd22bb7aa52 561 ** Now copy the last numTaps - 1 samples to the satrt of the state buffer.
emilmont 1:fdd22bb7aa52 562 ** This prepares the state buffer for the next function call. */
emilmont 1:fdd22bb7aa52 563
emilmont 1:fdd22bb7aa52 564 /* Points to the start of the state buffer */
emilmont 1:fdd22bb7aa52 565 pStateCurnt = S->pState;
emilmont 1:fdd22bb7aa52 566
emilmont 1:fdd22bb7aa52 567 i = (numTaps - 1u) >> 2u;
emilmont 1:fdd22bb7aa52 568
emilmont 1:fdd22bb7aa52 569 /* copy data */
emilmont 1:fdd22bb7aa52 570 while(i > 0u)
emilmont 1:fdd22bb7aa52 571 {
emilmont 1:fdd22bb7aa52 572 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 573 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 574 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 575 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 576
emilmont 1:fdd22bb7aa52 577 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 578 i--;
emilmont 1:fdd22bb7aa52 579 }
emilmont 1:fdd22bb7aa52 580
emilmont 1:fdd22bb7aa52 581 i = (numTaps - 1u) % 0x04u;
emilmont 1:fdd22bb7aa52 582
emilmont 1:fdd22bb7aa52 583 /* copy data */
emilmont 1:fdd22bb7aa52 584 while(i > 0u)
emilmont 1:fdd22bb7aa52 585 {
emilmont 1:fdd22bb7aa52 586 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 587
emilmont 1:fdd22bb7aa52 588 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 589 i--;
emilmont 1:fdd22bb7aa52 590 }
emilmont 1:fdd22bb7aa52 591 }
emilmont 1:fdd22bb7aa52 592
emilmont 1:fdd22bb7aa52 593
emilmont 2:da51fb522205 594 #endif /* #ifndef UNALIGNED_SUPPORT_DISABLE */
emilmont 1:fdd22bb7aa52 595
emilmont 1:fdd22bb7aa52 596 /**
emilmont 1:fdd22bb7aa52 597 * @} end of FIR_decimate group
emilmont 1:fdd22bb7aa52 598 */