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_lms_norm_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_lms_norm_q15.c
emilmont 1:fdd22bb7aa52 9 *
emilmont 2:da51fb522205 10 * Description: Q15 NLMS filter.
emilmont 1:fdd22bb7aa52 11 *
emilmont 1:fdd22bb7aa52 12 * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
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 LMS_NORM
emilmont 1:fdd22bb7aa52 49 * @{
emilmont 1:fdd22bb7aa52 50 */
emilmont 1:fdd22bb7aa52 51
emilmont 1:fdd22bb7aa52 52 /**
emilmont 1:fdd22bb7aa52 53 * @brief Processing function for Q15 normalized LMS filter.
emilmont 1:fdd22bb7aa52 54 * @param[in] *S points to an instance of the Q15 normalized LMS filter structure.
emilmont 1:fdd22bb7aa52 55 * @param[in] *pSrc points to the block of input data.
emilmont 1:fdd22bb7aa52 56 * @param[in] *pRef points to the block of reference data.
emilmont 1:fdd22bb7aa52 57 * @param[out] *pOut points to the block of output data.
emilmont 1:fdd22bb7aa52 58 * @param[out] *pErr points to the block of error data.
emilmont 1:fdd22bb7aa52 59 * @param[in] blockSize number of samples to process.
emilmont 1:fdd22bb7aa52 60 * @return none.
emilmont 1:fdd22bb7aa52 61 *
emilmont 1:fdd22bb7aa52 62 * <b>Scaling and Overflow Behavior:</b>
emilmont 1:fdd22bb7aa52 63 * \par
emilmont 1:fdd22bb7aa52 64 * The function is implemented using a 64-bit internal accumulator.
emilmont 1:fdd22bb7aa52 65 * Both coefficients and state variables are represented in 1.15 format and
emilmont 1:fdd22bb7aa52 66 * multiplications yield a 2.30 result. The 2.30 intermediate results are
emilmont 1:fdd22bb7aa52 67 * accumulated in a 64-bit accumulator in 34.30 format.
emilmont 1:fdd22bb7aa52 68 * There is no risk of internal overflow with this approach and the full
emilmont 1:fdd22bb7aa52 69 * precision of intermediate multiplications is preserved. After all additions
emilmont 1:fdd22bb7aa52 70 * have been performed, the accumulator is truncated to 34.15 format by
emilmont 1:fdd22bb7aa52 71 * discarding low 15 bits. Lastly, the accumulator is saturated to yield a
emilmont 1:fdd22bb7aa52 72 * result in 1.15 format.
emilmont 1:fdd22bb7aa52 73 *
emilmont 1:fdd22bb7aa52 74 * \par
emilmont 2:da51fb522205 75 * In this filter, filter coefficients are updated for each sample and the updation of filter cofficients are saturted.
emilmont 1:fdd22bb7aa52 76 *
emilmont 1:fdd22bb7aa52 77 */
emilmont 1:fdd22bb7aa52 78
emilmont 1:fdd22bb7aa52 79 void arm_lms_norm_q15(
emilmont 1:fdd22bb7aa52 80 arm_lms_norm_instance_q15 * S,
emilmont 1:fdd22bb7aa52 81 q15_t * pSrc,
emilmont 1:fdd22bb7aa52 82 q15_t * pRef,
emilmont 1:fdd22bb7aa52 83 q15_t * pOut,
emilmont 1:fdd22bb7aa52 84 q15_t * pErr,
emilmont 1:fdd22bb7aa52 85 uint32_t blockSize)
emilmont 1:fdd22bb7aa52 86 {
emilmont 1:fdd22bb7aa52 87 q15_t *pState = S->pState; /* State pointer */
emilmont 1:fdd22bb7aa52 88 q15_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
emilmont 1:fdd22bb7aa52 89 q15_t *pStateCurnt; /* Points to the current sample of the state */
emilmont 1:fdd22bb7aa52 90 q15_t *px, *pb; /* Temporary pointers for state and coefficient buffers */
emilmont 1:fdd22bb7aa52 91 q15_t mu = S->mu; /* Adaptive factor */
emilmont 1:fdd22bb7aa52 92 uint32_t numTaps = S->numTaps; /* Number of filter coefficients in the filter */
emilmont 1:fdd22bb7aa52 93 uint32_t tapCnt, blkCnt; /* Loop counters */
emilmont 1:fdd22bb7aa52 94 q31_t energy; /* Energy of the input */
emilmont 1:fdd22bb7aa52 95 q63_t acc; /* Accumulator */
emilmont 1:fdd22bb7aa52 96 q15_t e = 0, d = 0; /* error, reference data sample */
emilmont 1:fdd22bb7aa52 97 q15_t w = 0, in; /* weight factor and state */
emilmont 1:fdd22bb7aa52 98 q15_t x0; /* temporary variable to hold input sample */
emilmont 1:fdd22bb7aa52 99 //uint32_t shift = (uint32_t) S->postShift + 1u; /* Shift to be applied to the output */
emilmont 1:fdd22bb7aa52 100 q15_t errorXmu, oneByEnergy; /* Temporary variables to store error and mu product and reciprocal of energy */
emilmont 1:fdd22bb7aa52 101 q15_t postShift; /* Post shift to be applied to weight after reciprocal calculation */
emilmont 1:fdd22bb7aa52 102 q31_t coef; /* Teporary variable for coefficient */
emilmont 1:fdd22bb7aa52 103 q31_t acc_l, acc_h;
emilmont 1:fdd22bb7aa52 104 int32_t lShift = (15 - (int32_t) S->postShift); /* Post shift */
emilmont 1:fdd22bb7aa52 105 int32_t uShift = (32 - lShift);
emilmont 1:fdd22bb7aa52 106
emilmont 1:fdd22bb7aa52 107 energy = S->energy;
emilmont 1:fdd22bb7aa52 108 x0 = S->x0;
emilmont 1:fdd22bb7aa52 109
emilmont 1:fdd22bb7aa52 110 /* S->pState points to buffer which contains previous frame (numTaps - 1) samples */
emilmont 1:fdd22bb7aa52 111 /* pStateCurnt points to the location where the new input data should be written */
emilmont 1:fdd22bb7aa52 112 pStateCurnt = &(S->pState[(numTaps - 1u)]);
emilmont 1:fdd22bb7aa52 113
emilmont 1:fdd22bb7aa52 114 /* Loop over blockSize number of values */
emilmont 1:fdd22bb7aa52 115 blkCnt = blockSize;
emilmont 1:fdd22bb7aa52 116
emilmont 1:fdd22bb7aa52 117
mbed_official 3:7a284390b0ce 118 #ifndef ARM_MATH_CM0_FAMILY
emilmont 1:fdd22bb7aa52 119
emilmont 1:fdd22bb7aa52 120 /* Run the below code for Cortex-M4 and Cortex-M3 */
emilmont 1:fdd22bb7aa52 121
emilmont 1:fdd22bb7aa52 122 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 123 {
emilmont 1:fdd22bb7aa52 124 /* Copy the new input sample into the state buffer */
emilmont 1:fdd22bb7aa52 125 *pStateCurnt++ = *pSrc;
emilmont 1:fdd22bb7aa52 126
emilmont 1:fdd22bb7aa52 127 /* Initialize pState pointer */
emilmont 1:fdd22bb7aa52 128 px = pState;
emilmont 1:fdd22bb7aa52 129
emilmont 1:fdd22bb7aa52 130 /* Initialize coeff pointer */
emilmont 1:fdd22bb7aa52 131 pb = (pCoeffs);
emilmont 1:fdd22bb7aa52 132
emilmont 1:fdd22bb7aa52 133 /* Read the sample from input buffer */
emilmont 1:fdd22bb7aa52 134 in = *pSrc++;
emilmont 1:fdd22bb7aa52 135
emilmont 1:fdd22bb7aa52 136 /* Update the energy calculation */
emilmont 1:fdd22bb7aa52 137 energy -= (((q31_t) x0 * (x0)) >> 15);
emilmont 1:fdd22bb7aa52 138 energy += (((q31_t) in * (in)) >> 15);
emilmont 1:fdd22bb7aa52 139
emilmont 1:fdd22bb7aa52 140 /* Set the accumulator to zero */
emilmont 1:fdd22bb7aa52 141 acc = 0;
emilmont 1:fdd22bb7aa52 142
emilmont 1:fdd22bb7aa52 143 /* Loop unrolling. Process 4 taps at a time. */
emilmont 1:fdd22bb7aa52 144 tapCnt = numTaps >> 2;
emilmont 1:fdd22bb7aa52 145
emilmont 1:fdd22bb7aa52 146 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 147 {
emilmont 1:fdd22bb7aa52 148
emilmont 1:fdd22bb7aa52 149 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 150 #ifndef UNALIGNED_SUPPORT_DISABLE
emilmont 1:fdd22bb7aa52 151
emilmont 1:fdd22bb7aa52 152 acc = __SMLALD(*__SIMD32(px)++, (*__SIMD32(pb)++), acc);
emilmont 1:fdd22bb7aa52 153 acc = __SMLALD(*__SIMD32(px)++, (*__SIMD32(pb)++), acc);
emilmont 1:fdd22bb7aa52 154
emilmont 1:fdd22bb7aa52 155 #else
emilmont 1:fdd22bb7aa52 156
emilmont 1:fdd22bb7aa52 157 acc += (((q31_t) * px++ * (*pb++)));
emilmont 1:fdd22bb7aa52 158 acc += (((q31_t) * px++ * (*pb++)));
emilmont 1:fdd22bb7aa52 159 acc += (((q31_t) * px++ * (*pb++)));
emilmont 1:fdd22bb7aa52 160 acc += (((q31_t) * px++ * (*pb++)));
emilmont 1:fdd22bb7aa52 161
emilmont 2:da51fb522205 162 #endif /* #ifndef UNALIGNED_SUPPORT_DISABLE */
emilmont 1:fdd22bb7aa52 163
emilmont 1:fdd22bb7aa52 164 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 165 tapCnt--;
emilmont 1:fdd22bb7aa52 166 }
emilmont 1:fdd22bb7aa52 167
emilmont 1:fdd22bb7aa52 168 /* If the filter length is not a multiple of 4, compute the remaining filter taps */
emilmont 1:fdd22bb7aa52 169 tapCnt = numTaps % 0x4u;
emilmont 1:fdd22bb7aa52 170
emilmont 1:fdd22bb7aa52 171 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 172 {
emilmont 1:fdd22bb7aa52 173 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 174 acc += (((q31_t) * px++ * (*pb++)));
emilmont 1:fdd22bb7aa52 175
emilmont 1:fdd22bb7aa52 176 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 177 tapCnt--;
emilmont 1:fdd22bb7aa52 178 }
emilmont 1:fdd22bb7aa52 179
emilmont 1:fdd22bb7aa52 180 /* Calc lower part of acc */
emilmont 1:fdd22bb7aa52 181 acc_l = acc & 0xffffffff;
emilmont 1:fdd22bb7aa52 182
emilmont 1:fdd22bb7aa52 183 /* Calc upper part of acc */
emilmont 1:fdd22bb7aa52 184 acc_h = (acc >> 32) & 0xffffffff;
emilmont 1:fdd22bb7aa52 185
emilmont 1:fdd22bb7aa52 186 /* Apply shift for lower part of acc and upper part of acc */
emilmont 1:fdd22bb7aa52 187 acc = (uint32_t) acc_l >> lShift | acc_h << uShift;
emilmont 1:fdd22bb7aa52 188
emilmont 1:fdd22bb7aa52 189 /* Converting the result to 1.15 format and saturate the output */
emilmont 1:fdd22bb7aa52 190 acc = __SSAT(acc, 16u);
emilmont 1:fdd22bb7aa52 191
emilmont 1:fdd22bb7aa52 192 /* Store the result from accumulator into the destination buffer. */
emilmont 1:fdd22bb7aa52 193 *pOut++ = (q15_t) acc;
emilmont 1:fdd22bb7aa52 194
emilmont 1:fdd22bb7aa52 195 /* Compute and store error */
emilmont 1:fdd22bb7aa52 196 d = *pRef++;
emilmont 1:fdd22bb7aa52 197 e = d - (q15_t) acc;
emilmont 1:fdd22bb7aa52 198 *pErr++ = e;
emilmont 1:fdd22bb7aa52 199
emilmont 1:fdd22bb7aa52 200 /* Calculation of 1/energy */
emilmont 1:fdd22bb7aa52 201 postShift = arm_recip_q15((q15_t) energy + DELTA_Q15,
emilmont 1:fdd22bb7aa52 202 &oneByEnergy, S->recipTable);
emilmont 1:fdd22bb7aa52 203
emilmont 1:fdd22bb7aa52 204 /* Calculation of e * mu value */
emilmont 1:fdd22bb7aa52 205 errorXmu = (q15_t) (((q31_t) e * mu) >> 15);
emilmont 1:fdd22bb7aa52 206
emilmont 1:fdd22bb7aa52 207 /* Calculation of (e * mu) * (1/energy) value */
emilmont 1:fdd22bb7aa52 208 acc = (((q31_t) errorXmu * oneByEnergy) >> (15 - postShift));
emilmont 1:fdd22bb7aa52 209
emilmont 1:fdd22bb7aa52 210 /* Weighting factor for the normalized version */
emilmont 1:fdd22bb7aa52 211 w = (q15_t) __SSAT((q31_t) acc, 16);
emilmont 1:fdd22bb7aa52 212
emilmont 1:fdd22bb7aa52 213 /* Initialize pState pointer */
emilmont 1:fdd22bb7aa52 214 px = pState;
emilmont 1:fdd22bb7aa52 215
emilmont 1:fdd22bb7aa52 216 /* Initialize coeff pointer */
emilmont 1:fdd22bb7aa52 217 pb = (pCoeffs);
emilmont 1:fdd22bb7aa52 218
emilmont 1:fdd22bb7aa52 219 /* Loop unrolling. Process 4 taps at a time. */
emilmont 1:fdd22bb7aa52 220 tapCnt = numTaps >> 2;
emilmont 1:fdd22bb7aa52 221
emilmont 1:fdd22bb7aa52 222 /* Update filter coefficients */
emilmont 1:fdd22bb7aa52 223 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 224 {
emilmont 1:fdd22bb7aa52 225 coef = *pb + (((q31_t) w * (*px++)) >> 15);
emilmont 1:fdd22bb7aa52 226 *pb++ = (q15_t) __SSAT((coef), 16);
emilmont 1:fdd22bb7aa52 227 coef = *pb + (((q31_t) w * (*px++)) >> 15);
emilmont 1:fdd22bb7aa52 228 *pb++ = (q15_t) __SSAT((coef), 16);
emilmont 1:fdd22bb7aa52 229 coef = *pb + (((q31_t) w * (*px++)) >> 15);
emilmont 1:fdd22bb7aa52 230 *pb++ = (q15_t) __SSAT((coef), 16);
emilmont 1:fdd22bb7aa52 231 coef = *pb + (((q31_t) w * (*px++)) >> 15);
emilmont 1:fdd22bb7aa52 232 *pb++ = (q15_t) __SSAT((coef), 16);
emilmont 1:fdd22bb7aa52 233
emilmont 1:fdd22bb7aa52 234 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 235 tapCnt--;
emilmont 1:fdd22bb7aa52 236 }
emilmont 1:fdd22bb7aa52 237
emilmont 1:fdd22bb7aa52 238 /* If the filter length is not a multiple of 4, compute the remaining filter taps */
emilmont 1:fdd22bb7aa52 239 tapCnt = numTaps % 0x4u;
emilmont 1:fdd22bb7aa52 240
emilmont 1:fdd22bb7aa52 241 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 242 {
emilmont 1:fdd22bb7aa52 243 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 244 coef = *pb + (((q31_t) w * (*px++)) >> 15);
emilmont 1:fdd22bb7aa52 245 *pb++ = (q15_t) __SSAT((coef), 16);
emilmont 1:fdd22bb7aa52 246
emilmont 1:fdd22bb7aa52 247 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 248 tapCnt--;
emilmont 1:fdd22bb7aa52 249 }
emilmont 1:fdd22bb7aa52 250
emilmont 1:fdd22bb7aa52 251 /* Read the sample from state buffer */
emilmont 1:fdd22bb7aa52 252 x0 = *pState;
emilmont 1:fdd22bb7aa52 253
emilmont 1:fdd22bb7aa52 254 /* Advance state pointer by 1 for the next sample */
emilmont 1:fdd22bb7aa52 255 pState = pState + 1u;
emilmont 1:fdd22bb7aa52 256
emilmont 1:fdd22bb7aa52 257 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 258 blkCnt--;
emilmont 1:fdd22bb7aa52 259 }
emilmont 1:fdd22bb7aa52 260
emilmont 1:fdd22bb7aa52 261 /* Save energy and x0 values for the next frame */
emilmont 1:fdd22bb7aa52 262 S->energy = (q15_t) energy;
emilmont 1:fdd22bb7aa52 263 S->x0 = x0;
emilmont 1:fdd22bb7aa52 264
emilmont 1:fdd22bb7aa52 265 /* Processing is complete. Now copy the last numTaps - 1 samples to the
emilmont 1:fdd22bb7aa52 266 satrt of the state buffer. This prepares the state buffer for the
emilmont 1:fdd22bb7aa52 267 next function call. */
emilmont 1:fdd22bb7aa52 268
emilmont 1:fdd22bb7aa52 269 /* Points to the start of the pState buffer */
emilmont 1:fdd22bb7aa52 270 pStateCurnt = S->pState;
emilmont 1:fdd22bb7aa52 271
emilmont 1:fdd22bb7aa52 272 /* Calculation of count for copying integer writes */
emilmont 1:fdd22bb7aa52 273 tapCnt = (numTaps - 1u) >> 2;
emilmont 1:fdd22bb7aa52 274
emilmont 1:fdd22bb7aa52 275 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 276 {
emilmont 1:fdd22bb7aa52 277
emilmont 1:fdd22bb7aa52 278 #ifndef UNALIGNED_SUPPORT_DISABLE
emilmont 1:fdd22bb7aa52 279
emilmont 1:fdd22bb7aa52 280 *__SIMD32(pStateCurnt)++ = *__SIMD32(pState)++;
emilmont 1:fdd22bb7aa52 281 *__SIMD32(pStateCurnt)++ = *__SIMD32(pState)++;
emilmont 1:fdd22bb7aa52 282
emilmont 1:fdd22bb7aa52 283 #else
emilmont 1:fdd22bb7aa52 284
emilmont 1:fdd22bb7aa52 285 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 286 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 287 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 288 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 289
emilmont 1:fdd22bb7aa52 290 #endif
emilmont 1:fdd22bb7aa52 291
emilmont 1:fdd22bb7aa52 292 tapCnt--;
emilmont 1:fdd22bb7aa52 293
emilmont 1:fdd22bb7aa52 294 }
emilmont 1:fdd22bb7aa52 295
emilmont 1:fdd22bb7aa52 296 /* Calculation of count for remaining q15_t data */
emilmont 1:fdd22bb7aa52 297 tapCnt = (numTaps - 1u) % 0x4u;
emilmont 1:fdd22bb7aa52 298
emilmont 1:fdd22bb7aa52 299 /* copy data */
emilmont 1:fdd22bb7aa52 300 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 301 {
emilmont 1:fdd22bb7aa52 302 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 303
emilmont 1:fdd22bb7aa52 304 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 305 tapCnt--;
emilmont 1:fdd22bb7aa52 306 }
emilmont 1:fdd22bb7aa52 307
emilmont 1:fdd22bb7aa52 308 #else
emilmont 1:fdd22bb7aa52 309
emilmont 1:fdd22bb7aa52 310 /* Run the below code for Cortex-M0 */
emilmont 1:fdd22bb7aa52 311
emilmont 1:fdd22bb7aa52 312 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 313 {
emilmont 1:fdd22bb7aa52 314 /* Copy the new input sample into the state buffer */
emilmont 1:fdd22bb7aa52 315 *pStateCurnt++ = *pSrc;
emilmont 1:fdd22bb7aa52 316
emilmont 1:fdd22bb7aa52 317 /* Initialize pState pointer */
emilmont 1:fdd22bb7aa52 318 px = pState;
emilmont 1:fdd22bb7aa52 319
emilmont 1:fdd22bb7aa52 320 /* Initialize pCoeffs pointer */
emilmont 1:fdd22bb7aa52 321 pb = pCoeffs;
emilmont 1:fdd22bb7aa52 322
emilmont 1:fdd22bb7aa52 323 /* Read the sample from input buffer */
emilmont 1:fdd22bb7aa52 324 in = *pSrc++;
emilmont 1:fdd22bb7aa52 325
emilmont 1:fdd22bb7aa52 326 /* Update the energy calculation */
emilmont 1:fdd22bb7aa52 327 energy -= (((q31_t) x0 * (x0)) >> 15);
emilmont 1:fdd22bb7aa52 328 energy += (((q31_t) in * (in)) >> 15);
emilmont 1:fdd22bb7aa52 329
emilmont 1:fdd22bb7aa52 330 /* Set the accumulator to zero */
emilmont 1:fdd22bb7aa52 331 acc = 0;
emilmont 1:fdd22bb7aa52 332
emilmont 1:fdd22bb7aa52 333 /* Loop over numTaps number of values */
emilmont 1:fdd22bb7aa52 334 tapCnt = numTaps;
emilmont 1:fdd22bb7aa52 335
emilmont 1:fdd22bb7aa52 336 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 337 {
emilmont 1:fdd22bb7aa52 338 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 339 acc += (((q31_t) * px++ * (*pb++)));
emilmont 1:fdd22bb7aa52 340
emilmont 1:fdd22bb7aa52 341 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 342 tapCnt--;
emilmont 1:fdd22bb7aa52 343 }
emilmont 1:fdd22bb7aa52 344
emilmont 1:fdd22bb7aa52 345 /* Calc lower part of acc */
emilmont 1:fdd22bb7aa52 346 acc_l = acc & 0xffffffff;
emilmont 1:fdd22bb7aa52 347
emilmont 1:fdd22bb7aa52 348 /* Calc upper part of acc */
emilmont 1:fdd22bb7aa52 349 acc_h = (acc >> 32) & 0xffffffff;
emilmont 1:fdd22bb7aa52 350
emilmont 1:fdd22bb7aa52 351 /* Apply shift for lower part of acc and upper part of acc */
emilmont 1:fdd22bb7aa52 352 acc = (uint32_t) acc_l >> lShift | acc_h << uShift;
emilmont 1:fdd22bb7aa52 353
emilmont 1:fdd22bb7aa52 354 /* Converting the result to 1.15 format and saturate the output */
emilmont 1:fdd22bb7aa52 355 acc = __SSAT(acc, 16u);
emilmont 1:fdd22bb7aa52 356
emilmont 1:fdd22bb7aa52 357 /* Converting the result to 1.15 format */
emilmont 1:fdd22bb7aa52 358 //acc = __SSAT((acc >> (16u - shift)), 16u);
emilmont 1:fdd22bb7aa52 359
emilmont 1:fdd22bb7aa52 360 /* Store the result from accumulator into the destination buffer. */
emilmont 1:fdd22bb7aa52 361 *pOut++ = (q15_t) acc;
emilmont 1:fdd22bb7aa52 362
emilmont 1:fdd22bb7aa52 363 /* Compute and store error */
emilmont 1:fdd22bb7aa52 364 d = *pRef++;
emilmont 1:fdd22bb7aa52 365 e = d - (q15_t) acc;
emilmont 1:fdd22bb7aa52 366 *pErr++ = e;
emilmont 1:fdd22bb7aa52 367
emilmont 1:fdd22bb7aa52 368 /* Calculation of 1/energy */
emilmont 1:fdd22bb7aa52 369 postShift = arm_recip_q15((q15_t) energy + DELTA_Q15,
emilmont 1:fdd22bb7aa52 370 &oneByEnergy, S->recipTable);
emilmont 1:fdd22bb7aa52 371
emilmont 1:fdd22bb7aa52 372 /* Calculation of e * mu value */
emilmont 1:fdd22bb7aa52 373 errorXmu = (q15_t) (((q31_t) e * mu) >> 15);
emilmont 1:fdd22bb7aa52 374
emilmont 1:fdd22bb7aa52 375 /* Calculation of (e * mu) * (1/energy) value */
emilmont 1:fdd22bb7aa52 376 acc = (((q31_t) errorXmu * oneByEnergy) >> (15 - postShift));
emilmont 1:fdd22bb7aa52 377
emilmont 1:fdd22bb7aa52 378 /* Weighting factor for the normalized version */
emilmont 1:fdd22bb7aa52 379 w = (q15_t) __SSAT((q31_t) acc, 16);
emilmont 1:fdd22bb7aa52 380
emilmont 1:fdd22bb7aa52 381 /* Initialize pState pointer */
emilmont 1:fdd22bb7aa52 382 px = pState;
emilmont 1:fdd22bb7aa52 383
emilmont 1:fdd22bb7aa52 384 /* Initialize coeff pointer */
emilmont 1:fdd22bb7aa52 385 pb = (pCoeffs);
emilmont 1:fdd22bb7aa52 386
emilmont 1:fdd22bb7aa52 387 /* Loop over numTaps number of values */
emilmont 1:fdd22bb7aa52 388 tapCnt = numTaps;
emilmont 1:fdd22bb7aa52 389
emilmont 1:fdd22bb7aa52 390 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 391 {
emilmont 1:fdd22bb7aa52 392 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 393 coef = *pb + (((q31_t) w * (*px++)) >> 15);
emilmont 1:fdd22bb7aa52 394 *pb++ = (q15_t) __SSAT((coef), 16);
emilmont 1:fdd22bb7aa52 395
emilmont 1:fdd22bb7aa52 396 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 397 tapCnt--;
emilmont 1:fdd22bb7aa52 398 }
emilmont 1:fdd22bb7aa52 399
emilmont 1:fdd22bb7aa52 400 /* Read the sample from state buffer */
emilmont 1:fdd22bb7aa52 401 x0 = *pState;
emilmont 1:fdd22bb7aa52 402
emilmont 1:fdd22bb7aa52 403 /* Advance state pointer by 1 for the next sample */
emilmont 1:fdd22bb7aa52 404 pState = pState + 1u;
emilmont 1:fdd22bb7aa52 405
emilmont 1:fdd22bb7aa52 406 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 407 blkCnt--;
emilmont 1:fdd22bb7aa52 408 }
emilmont 1:fdd22bb7aa52 409
emilmont 1:fdd22bb7aa52 410 /* Save energy and x0 values for the next frame */
emilmont 1:fdd22bb7aa52 411 S->energy = (q15_t) energy;
emilmont 1:fdd22bb7aa52 412 S->x0 = x0;
emilmont 1:fdd22bb7aa52 413
emilmont 1:fdd22bb7aa52 414 /* Processing is complete. Now copy the last numTaps - 1 samples to the
emilmont 1:fdd22bb7aa52 415 satrt of the state buffer. This prepares the state buffer for the
emilmont 1:fdd22bb7aa52 416 next function call. */
emilmont 1:fdd22bb7aa52 417
emilmont 1:fdd22bb7aa52 418 /* Points to the start of the pState buffer */
emilmont 1:fdd22bb7aa52 419 pStateCurnt = S->pState;
emilmont 1:fdd22bb7aa52 420
emilmont 1:fdd22bb7aa52 421 /* copy (numTaps - 1u) data */
emilmont 1:fdd22bb7aa52 422 tapCnt = (numTaps - 1u);
emilmont 1:fdd22bb7aa52 423
emilmont 1:fdd22bb7aa52 424 /* copy data */
emilmont 1:fdd22bb7aa52 425 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 426 {
emilmont 1:fdd22bb7aa52 427 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 428
emilmont 1:fdd22bb7aa52 429 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 430 tapCnt--;
emilmont 1:fdd22bb7aa52 431 }
emilmont 1:fdd22bb7aa52 432
mbed_official 3:7a284390b0ce 433 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
emilmont 1:fdd22bb7aa52 434
emilmont 1:fdd22bb7aa52 435 }
emilmont 1:fdd22bb7aa52 436
emilmont 1:fdd22bb7aa52 437
emilmont 1:fdd22bb7aa52 438 /**
emilmont 1:fdd22bb7aa52 439 * @} end of LMS_NORM group
emilmont 1:fdd22bb7aa52 440 */