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_lattice_q15.c@2:da51fb522205, 2013-05-30 (annotated)

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

Who changed what in which revision?

UserRevisionLine numberNew contents of line
emilmont 1:fdd22bb7aa52 1 /* ----------------------------------------------------------------------
emilmont 1:fdd22bb7aa52 2 * Copyright (C) 2010 ARM Limited. All rights reserved.
emilmont 1:fdd22bb7aa52 3 *
emilmont 1:fdd22bb7aa52 4 * $Date: 15. February 2012
emilmont 2:da51fb522205 5 * $Revision: V1.1.0
emilmont 1:fdd22bb7aa52 6 *
emilmont 2:da51fb522205 7 * Project: CMSIS DSP Library
emilmont 2:da51fb522205 8 * Title: arm_fir_lattice_q15.c
emilmont 1:fdd22bb7aa52 9 *
emilmont 2:da51fb522205 10 * Description: Q15 FIR lattice 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
emilmont 1:fdd22bb7aa52 36 #include "arm_math.h"
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_Lattice
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 Q15 FIR lattice filter.
emilmont 1:fdd22bb7aa52 50 * @param[in] *S points to an instance of the Q15 FIR lattice 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] blockSize number of samples to process.
emilmont 1:fdd22bb7aa52 54 * @return none.
emilmont 1:fdd22bb7aa52 55 */
emilmont 1:fdd22bb7aa52 56
emilmont 1:fdd22bb7aa52 57 void arm_fir_lattice_q15(
emilmont 1:fdd22bb7aa52 58 const arm_fir_lattice_instance_q15 * S,
emilmont 1:fdd22bb7aa52 59 q15_t * pSrc,
emilmont 1:fdd22bb7aa52 60 q15_t * pDst,
emilmont 1:fdd22bb7aa52 61 uint32_t blockSize)
emilmont 1:fdd22bb7aa52 62 {
emilmont 1:fdd22bb7aa52 63 q15_t *pState; /* State pointer */
emilmont 1:fdd22bb7aa52 64 q15_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
emilmont 1:fdd22bb7aa52 65 q15_t *px; /* temporary state pointer */
emilmont 1:fdd22bb7aa52 66 q15_t *pk; /* temporary coefficient pointer */
emilmont 1:fdd22bb7aa52 67
emilmont 1:fdd22bb7aa52 68
emilmont 1:fdd22bb7aa52 69 #ifndef ARM_MATH_CM0
emilmont 1:fdd22bb7aa52 70
emilmont 1:fdd22bb7aa52 71 /* Run the below code for Cortex-M4 and Cortex-M3 */
emilmont 1:fdd22bb7aa52 72
emilmont 1:fdd22bb7aa52 73 q31_t fcurnt1, fnext1, gcurnt1 = 0, gnext1; /* temporary variables for first sample in loop unrolling */
emilmont 1:fdd22bb7aa52 74 q31_t fcurnt2, fnext2, gnext2; /* temporary variables for second sample in loop unrolling */
emilmont 1:fdd22bb7aa52 75 q31_t fcurnt3, fnext3, gnext3; /* temporary variables for third sample in loop unrolling */
emilmont 1:fdd22bb7aa52 76 q31_t fcurnt4, fnext4, gnext4; /* temporary variables for fourth sample in loop unrolling */
emilmont 1:fdd22bb7aa52 77 uint32_t numStages = S->numStages; /* Number of stages in the filter */
emilmont 1:fdd22bb7aa52 78 uint32_t blkCnt, stageCnt; /* temporary variables for counts */
emilmont 1:fdd22bb7aa52 79
emilmont 1:fdd22bb7aa52 80 pState = &S->pState[0];
emilmont 1:fdd22bb7aa52 81
emilmont 1:fdd22bb7aa52 82 blkCnt = blockSize >> 2u;
emilmont 1:fdd22bb7aa52 83
emilmont 1:fdd22bb7aa52 84 /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
emilmont 1:fdd22bb7aa52 85 ** a second loop below computes the remaining 1 to 3 samples. */
emilmont 1:fdd22bb7aa52 86 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 87 {
emilmont 1:fdd22bb7aa52 88
emilmont 1:fdd22bb7aa52 89 /* Read two samples from input buffer */
emilmont 1:fdd22bb7aa52 90 /* f0(n) = x(n) */
emilmont 1:fdd22bb7aa52 91 fcurnt1 = *pSrc++;
emilmont 1:fdd22bb7aa52 92 fcurnt2 = *pSrc++;
emilmont 1:fdd22bb7aa52 93
emilmont 1:fdd22bb7aa52 94 /* Initialize coeff pointer */
emilmont 1:fdd22bb7aa52 95 pk = (pCoeffs);
emilmont 1:fdd22bb7aa52 96
emilmont 1:fdd22bb7aa52 97 /* Initialize state pointer */
emilmont 1:fdd22bb7aa52 98 px = pState;
emilmont 1:fdd22bb7aa52 99
emilmont 1:fdd22bb7aa52 100 /* Read g0(n-1) from state */
emilmont 1:fdd22bb7aa52 101 gcurnt1 = *px;
emilmont 1:fdd22bb7aa52 102
emilmont 1:fdd22bb7aa52 103 /* Process first sample for first tap */
emilmont 1:fdd22bb7aa52 104 /* f1(n) = f0(n) + K1 * g0(n-1) */
emilmont 1:fdd22bb7aa52 105 fnext1 = (q31_t) ((gcurnt1 * (*pk)) >> 15u) + fcurnt1;
emilmont 1:fdd22bb7aa52 106 fnext1 = __SSAT(fnext1, 16);
emilmont 1:fdd22bb7aa52 107
emilmont 1:fdd22bb7aa52 108 /* g1(n) = f0(n) * K1 + g0(n-1) */
emilmont 1:fdd22bb7aa52 109 gnext1 = (q31_t) ((fcurnt1 * (*pk)) >> 15u) + gcurnt1;
emilmont 1:fdd22bb7aa52 110 gnext1 = __SSAT(gnext1, 16);
emilmont 1:fdd22bb7aa52 111
emilmont 1:fdd22bb7aa52 112 /* Process second sample for first tap */
emilmont 1:fdd22bb7aa52 113 /* for sample 2 processing */
emilmont 1:fdd22bb7aa52 114 fnext2 = (q31_t) ((fcurnt1 * (*pk)) >> 15u) + fcurnt2;
emilmont 1:fdd22bb7aa52 115 fnext2 = __SSAT(fnext2, 16);
emilmont 1:fdd22bb7aa52 116
emilmont 1:fdd22bb7aa52 117 gnext2 = (q31_t) ((fcurnt2 * (*pk)) >> 15u) + fcurnt1;
emilmont 1:fdd22bb7aa52 118 gnext2 = __SSAT(gnext2, 16);
emilmont 1:fdd22bb7aa52 119
emilmont 1:fdd22bb7aa52 120
emilmont 1:fdd22bb7aa52 121 /* Read next two samples from input buffer */
emilmont 1:fdd22bb7aa52 122 /* f0(n+2) = x(n+2) */
emilmont 1:fdd22bb7aa52 123 fcurnt3 = *pSrc++;
emilmont 1:fdd22bb7aa52 124 fcurnt4 = *pSrc++;
emilmont 1:fdd22bb7aa52 125
emilmont 1:fdd22bb7aa52 126 /* Copy only last input samples into the state buffer
emilmont 1:fdd22bb7aa52 127 which is used for next four samples processing */
emilmont 1:fdd22bb7aa52 128 *px++ = (q15_t) fcurnt4;
emilmont 1:fdd22bb7aa52 129
emilmont 1:fdd22bb7aa52 130 /* Process third sample for first tap */
emilmont 1:fdd22bb7aa52 131 fnext3 = (q31_t) ((fcurnt2 * (*pk)) >> 15u) + fcurnt3;
emilmont 1:fdd22bb7aa52 132 fnext3 = __SSAT(fnext3, 16);
emilmont 1:fdd22bb7aa52 133 gnext3 = (q31_t) ((fcurnt3 * (*pk)) >> 15u) + fcurnt2;
emilmont 1:fdd22bb7aa52 134 gnext3 = __SSAT(gnext3, 16);
emilmont 1:fdd22bb7aa52 135
emilmont 1:fdd22bb7aa52 136 /* Process fourth sample for first tap */
emilmont 1:fdd22bb7aa52 137 fnext4 = (q31_t) ((fcurnt3 * (*pk)) >> 15u) + fcurnt4;
emilmont 1:fdd22bb7aa52 138 fnext4 = __SSAT(fnext4, 16);
emilmont 1:fdd22bb7aa52 139 gnext4 = (q31_t) ((fcurnt4 * (*pk++)) >> 15u) + fcurnt3;
emilmont 1:fdd22bb7aa52 140 gnext4 = __SSAT(gnext4, 16);
emilmont 1:fdd22bb7aa52 141
emilmont 1:fdd22bb7aa52 142 /* Update of f values for next coefficient set processing */
emilmont 1:fdd22bb7aa52 143 fcurnt1 = fnext1;
emilmont 1:fdd22bb7aa52 144 fcurnt2 = fnext2;
emilmont 1:fdd22bb7aa52 145 fcurnt3 = fnext3;
emilmont 1:fdd22bb7aa52 146 fcurnt4 = fnext4;
emilmont 1:fdd22bb7aa52 147
emilmont 1:fdd22bb7aa52 148
emilmont 1:fdd22bb7aa52 149 /* Loop unrolling. Process 4 taps at a time . */
emilmont 1:fdd22bb7aa52 150 stageCnt = (numStages - 1u) >> 2;
emilmont 1:fdd22bb7aa52 151
emilmont 1:fdd22bb7aa52 152
emilmont 1:fdd22bb7aa52 153 /* Loop over the number of taps. Unroll by a factor of 4.
emilmont 1:fdd22bb7aa52 154 ** Repeat until we've computed numStages-3 coefficients. */
emilmont 1:fdd22bb7aa52 155
emilmont 1:fdd22bb7aa52 156 /* Process 2nd, 3rd, 4th and 5th taps ... here */
emilmont 1:fdd22bb7aa52 157 while(stageCnt > 0u)
emilmont 1:fdd22bb7aa52 158 {
emilmont 1:fdd22bb7aa52 159 /* Read g1(n-1), g3(n-1) .... from state */
emilmont 1:fdd22bb7aa52 160 gcurnt1 = *px;
emilmont 1:fdd22bb7aa52 161
emilmont 1:fdd22bb7aa52 162 /* save g1(n) in state buffer */
emilmont 1:fdd22bb7aa52 163 *px++ = (q15_t) gnext4;
emilmont 1:fdd22bb7aa52 164
emilmont 1:fdd22bb7aa52 165 /* Process first sample for 2nd, 6th .. tap */
emilmont 1:fdd22bb7aa52 166 /* Sample processing for K2, K6.... */
emilmont 1:fdd22bb7aa52 167 /* f1(n) = f0(n) + K1 * g0(n-1) */
emilmont 1:fdd22bb7aa52 168 fnext1 = (q31_t) ((gcurnt1 * (*pk)) >> 15u) + fcurnt1;
emilmont 1:fdd22bb7aa52 169 fnext1 = __SSAT(fnext1, 16);
emilmont 1:fdd22bb7aa52 170
emilmont 1:fdd22bb7aa52 171
emilmont 1:fdd22bb7aa52 172 /* Process second sample for 2nd, 6th .. tap */
emilmont 1:fdd22bb7aa52 173 /* for sample 2 processing */
emilmont 1:fdd22bb7aa52 174 fnext2 = (q31_t) ((gnext1 * (*pk)) >> 15u) + fcurnt2;
emilmont 1:fdd22bb7aa52 175 fnext2 = __SSAT(fnext2, 16);
emilmont 1:fdd22bb7aa52 176 /* Process third sample for 2nd, 6th .. tap */
emilmont 1:fdd22bb7aa52 177 fnext3 = (q31_t) ((gnext2 * (*pk)) >> 15u) + fcurnt3;
emilmont 1:fdd22bb7aa52 178 fnext3 = __SSAT(fnext3, 16);
emilmont 1:fdd22bb7aa52 179 /* Process fourth sample for 2nd, 6th .. tap */
emilmont 1:fdd22bb7aa52 180 /* fnext4 = fcurnt4 + (*pk) * gnext3; */
emilmont 1:fdd22bb7aa52 181 fnext4 = (q31_t) ((gnext3 * (*pk)) >> 15u) + fcurnt4;
emilmont 1:fdd22bb7aa52 182 fnext4 = __SSAT(fnext4, 16);
emilmont 1:fdd22bb7aa52 183
emilmont 1:fdd22bb7aa52 184 /* g1(n) = f0(n) * K1 + g0(n-1) */
emilmont 1:fdd22bb7aa52 185 /* Calculation of state values for next stage */
emilmont 1:fdd22bb7aa52 186 gnext4 = (q31_t) ((fcurnt4 * (*pk)) >> 15u) + gnext3;
emilmont 1:fdd22bb7aa52 187 gnext4 = __SSAT(gnext4, 16);
emilmont 1:fdd22bb7aa52 188 gnext3 = (q31_t) ((fcurnt3 * (*pk)) >> 15u) + gnext2;
emilmont 1:fdd22bb7aa52 189 gnext3 = __SSAT(gnext3, 16);
emilmont 1:fdd22bb7aa52 190
emilmont 1:fdd22bb7aa52 191 gnext2 = (q31_t) ((fcurnt2 * (*pk)) >> 15u) + gnext1;
emilmont 1:fdd22bb7aa52 192 gnext2 = __SSAT(gnext2, 16);
emilmont 1:fdd22bb7aa52 193
emilmont 1:fdd22bb7aa52 194 gnext1 = (q31_t) ((fcurnt1 * (*pk++)) >> 15u) + gcurnt1;
emilmont 1:fdd22bb7aa52 195 gnext1 = __SSAT(gnext1, 16);
emilmont 1:fdd22bb7aa52 196
emilmont 1:fdd22bb7aa52 197
emilmont 1:fdd22bb7aa52 198 /* Read g2(n-1), g4(n-1) .... from state */
emilmont 1:fdd22bb7aa52 199 gcurnt1 = *px;
emilmont 1:fdd22bb7aa52 200
emilmont 1:fdd22bb7aa52 201 /* save g1(n) in state buffer */
emilmont 1:fdd22bb7aa52 202 *px++ = (q15_t) gnext4;
emilmont 1:fdd22bb7aa52 203
emilmont 1:fdd22bb7aa52 204 /* Sample processing for K3, K7.... */
emilmont 1:fdd22bb7aa52 205 /* Process first sample for 3rd, 7th .. tap */
emilmont 1:fdd22bb7aa52 206 /* f3(n) = f2(n) + K3 * g2(n-1) */
emilmont 1:fdd22bb7aa52 207 fcurnt1 = (q31_t) ((gcurnt1 * (*pk)) >> 15u) + fnext1;
emilmont 1:fdd22bb7aa52 208 fcurnt1 = __SSAT(fcurnt1, 16);
emilmont 1:fdd22bb7aa52 209
emilmont 1:fdd22bb7aa52 210 /* Process second sample for 3rd, 7th .. tap */
emilmont 1:fdd22bb7aa52 211 fcurnt2 = (q31_t) ((gnext1 * (*pk)) >> 15u) + fnext2;
emilmont 1:fdd22bb7aa52 212 fcurnt2 = __SSAT(fcurnt2, 16);
emilmont 1:fdd22bb7aa52 213
emilmont 1:fdd22bb7aa52 214 /* Process third sample for 3rd, 7th .. tap */
emilmont 1:fdd22bb7aa52 215 fcurnt3 = (q31_t) ((gnext2 * (*pk)) >> 15u) + fnext3;
emilmont 1:fdd22bb7aa52 216 fcurnt3 = __SSAT(fcurnt3, 16);
emilmont 1:fdd22bb7aa52 217
emilmont 1:fdd22bb7aa52 218 /* Process fourth sample for 3rd, 7th .. tap */
emilmont 1:fdd22bb7aa52 219 fcurnt4 = (q31_t) ((gnext3 * (*pk)) >> 15u) + fnext4;
emilmont 1:fdd22bb7aa52 220 fcurnt4 = __SSAT(fcurnt4, 16);
emilmont 1:fdd22bb7aa52 221
emilmont 1:fdd22bb7aa52 222 /* Calculation of state values for next stage */
emilmont 1:fdd22bb7aa52 223 /* g3(n) = f2(n) * K3 + g2(n-1) */
emilmont 1:fdd22bb7aa52 224 gnext4 = (q31_t) ((fnext4 * (*pk)) >> 15u) + gnext3;
emilmont 1:fdd22bb7aa52 225 gnext4 = __SSAT(gnext4, 16);
emilmont 1:fdd22bb7aa52 226
emilmont 1:fdd22bb7aa52 227 gnext3 = (q31_t) ((fnext3 * (*pk)) >> 15u) + gnext2;
emilmont 1:fdd22bb7aa52 228 gnext3 = __SSAT(gnext3, 16);
emilmont 1:fdd22bb7aa52 229
emilmont 1:fdd22bb7aa52 230 gnext2 = (q31_t) ((fnext2 * (*pk)) >> 15u) + gnext1;
emilmont 1:fdd22bb7aa52 231 gnext2 = __SSAT(gnext2, 16);
emilmont 1:fdd22bb7aa52 232
emilmont 1:fdd22bb7aa52 233 gnext1 = (q31_t) ((fnext1 * (*pk++)) >> 15u) + gcurnt1;
emilmont 1:fdd22bb7aa52 234 gnext1 = __SSAT(gnext1, 16);
emilmont 1:fdd22bb7aa52 235
emilmont 1:fdd22bb7aa52 236 /* Read g1(n-1), g3(n-1) .... from state */
emilmont 1:fdd22bb7aa52 237 gcurnt1 = *px;
emilmont 1:fdd22bb7aa52 238
emilmont 1:fdd22bb7aa52 239 /* save g1(n) in state buffer */
emilmont 1:fdd22bb7aa52 240 *px++ = (q15_t) gnext4;
emilmont 1:fdd22bb7aa52 241
emilmont 1:fdd22bb7aa52 242 /* Sample processing for K4, K8.... */
emilmont 1:fdd22bb7aa52 243 /* Process first sample for 4th, 8th .. tap */
emilmont 1:fdd22bb7aa52 244 /* f4(n) = f3(n) + K4 * g3(n-1) */
emilmont 1:fdd22bb7aa52 245 fnext1 = (q31_t) ((gcurnt1 * (*pk)) >> 15u) + fcurnt1;
emilmont 1:fdd22bb7aa52 246 fnext1 = __SSAT(fnext1, 16);
emilmont 1:fdd22bb7aa52 247
emilmont 1:fdd22bb7aa52 248 /* Process second sample for 4th, 8th .. tap */
emilmont 1:fdd22bb7aa52 249 /* for sample 2 processing */
emilmont 1:fdd22bb7aa52 250 fnext2 = (q31_t) ((gnext1 * (*pk)) >> 15u) + fcurnt2;
emilmont 1:fdd22bb7aa52 251 fnext2 = __SSAT(fnext2, 16);
emilmont 1:fdd22bb7aa52 252
emilmont 1:fdd22bb7aa52 253 /* Process third sample for 4th, 8th .. tap */
emilmont 1:fdd22bb7aa52 254 fnext3 = (q31_t) ((gnext2 * (*pk)) >> 15u) + fcurnt3;
emilmont 1:fdd22bb7aa52 255 fnext3 = __SSAT(fnext3, 16);
emilmont 1:fdd22bb7aa52 256
emilmont 1:fdd22bb7aa52 257 /* Process fourth sample for 4th, 8th .. tap */
emilmont 1:fdd22bb7aa52 258 fnext4 = (q31_t) ((gnext3 * (*pk)) >> 15u) + fcurnt4;
emilmont 1:fdd22bb7aa52 259 fnext4 = __SSAT(fnext4, 16);
emilmont 1:fdd22bb7aa52 260
emilmont 1:fdd22bb7aa52 261 /* g4(n) = f3(n) * K4 + g3(n-1) */
emilmont 1:fdd22bb7aa52 262 /* Calculation of state values for next stage */
emilmont 1:fdd22bb7aa52 263 gnext4 = (q31_t) ((fcurnt4 * (*pk)) >> 15u) + gnext3;
emilmont 1:fdd22bb7aa52 264 gnext4 = __SSAT(gnext4, 16);
emilmont 1:fdd22bb7aa52 265
emilmont 1:fdd22bb7aa52 266 gnext3 = (q31_t) ((fcurnt3 * (*pk)) >> 15u) + gnext2;
emilmont 1:fdd22bb7aa52 267 gnext3 = __SSAT(gnext3, 16);
emilmont 1:fdd22bb7aa52 268
emilmont 1:fdd22bb7aa52 269 gnext2 = (q31_t) ((fcurnt2 * (*pk)) >> 15u) + gnext1;
emilmont 1:fdd22bb7aa52 270 gnext2 = __SSAT(gnext2, 16);
emilmont 1:fdd22bb7aa52 271 gnext1 = (q31_t) ((fcurnt1 * (*pk++)) >> 15u) + gcurnt1;
emilmont 1:fdd22bb7aa52 272 gnext1 = __SSAT(gnext1, 16);
emilmont 1:fdd22bb7aa52 273
emilmont 1:fdd22bb7aa52 274
emilmont 1:fdd22bb7aa52 275 /* Read g2(n-1), g4(n-1) .... from state */
emilmont 1:fdd22bb7aa52 276 gcurnt1 = *px;
emilmont 1:fdd22bb7aa52 277
emilmont 1:fdd22bb7aa52 278 /* save g4(n) in state buffer */
emilmont 1:fdd22bb7aa52 279 *px++ = (q15_t) gnext4;
emilmont 1:fdd22bb7aa52 280
emilmont 1:fdd22bb7aa52 281 /* Sample processing for K5, K9.... */
emilmont 1:fdd22bb7aa52 282 /* Process first sample for 5th, 9th .. tap */
emilmont 1:fdd22bb7aa52 283 /* f5(n) = f4(n) + K5 * g4(n-1) */
emilmont 1:fdd22bb7aa52 284 fcurnt1 = (q31_t) ((gcurnt1 * (*pk)) >> 15u) + fnext1;
emilmont 1:fdd22bb7aa52 285 fcurnt1 = __SSAT(fcurnt1, 16);
emilmont 1:fdd22bb7aa52 286
emilmont 1:fdd22bb7aa52 287 /* Process second sample for 5th, 9th .. tap */
emilmont 1:fdd22bb7aa52 288 fcurnt2 = (q31_t) ((gnext1 * (*pk)) >> 15u) + fnext2;
emilmont 1:fdd22bb7aa52 289 fcurnt2 = __SSAT(fcurnt2, 16);
emilmont 1:fdd22bb7aa52 290
emilmont 1:fdd22bb7aa52 291 /* Process third sample for 5th, 9th .. tap */
emilmont 1:fdd22bb7aa52 292 fcurnt3 = (q31_t) ((gnext2 * (*pk)) >> 15u) + fnext3;
emilmont 1:fdd22bb7aa52 293 fcurnt3 = __SSAT(fcurnt3, 16);
emilmont 1:fdd22bb7aa52 294
emilmont 1:fdd22bb7aa52 295 /* Process fourth sample for 5th, 9th .. tap */
emilmont 1:fdd22bb7aa52 296 fcurnt4 = (q31_t) ((gnext3 * (*pk)) >> 15u) + fnext4;
emilmont 1:fdd22bb7aa52 297 fcurnt4 = __SSAT(fcurnt4, 16);
emilmont 1:fdd22bb7aa52 298
emilmont 1:fdd22bb7aa52 299 /* Calculation of state values for next stage */
emilmont 1:fdd22bb7aa52 300 /* g5(n) = f4(n) * K5 + g4(n-1) */
emilmont 1:fdd22bb7aa52 301 gnext4 = (q31_t) ((fnext4 * (*pk)) >> 15u) + gnext3;
emilmont 1:fdd22bb7aa52 302 gnext4 = __SSAT(gnext4, 16);
emilmont 1:fdd22bb7aa52 303 gnext3 = (q31_t) ((fnext3 * (*pk)) >> 15u) + gnext2;
emilmont 1:fdd22bb7aa52 304 gnext3 = __SSAT(gnext3, 16);
emilmont 1:fdd22bb7aa52 305 gnext2 = (q31_t) ((fnext2 * (*pk)) >> 15u) + gnext1;
emilmont 1:fdd22bb7aa52 306 gnext2 = __SSAT(gnext2, 16);
emilmont 1:fdd22bb7aa52 307 gnext1 = (q31_t) ((fnext1 * (*pk++)) >> 15u) + gcurnt1;
emilmont 1:fdd22bb7aa52 308 gnext1 = __SSAT(gnext1, 16);
emilmont 1:fdd22bb7aa52 309
emilmont 1:fdd22bb7aa52 310 stageCnt--;
emilmont 1:fdd22bb7aa52 311 }
emilmont 1:fdd22bb7aa52 312
emilmont 1:fdd22bb7aa52 313 /* If the (filter length -1) is not a multiple of 4, compute the remaining filter taps */
emilmont 1:fdd22bb7aa52 314 stageCnt = (numStages - 1u) % 0x4u;
emilmont 1:fdd22bb7aa52 315
emilmont 1:fdd22bb7aa52 316 while(stageCnt > 0u)
emilmont 1:fdd22bb7aa52 317 {
emilmont 1:fdd22bb7aa52 318 gcurnt1 = *px;
emilmont 1:fdd22bb7aa52 319
emilmont 1:fdd22bb7aa52 320 /* save g value in state buffer */
emilmont 1:fdd22bb7aa52 321 *px++ = (q15_t) gnext4;
emilmont 1:fdd22bb7aa52 322
emilmont 1:fdd22bb7aa52 323 /* Process four samples for last three taps here */
emilmont 1:fdd22bb7aa52 324 fnext1 = (q31_t) ((gcurnt1 * (*pk)) >> 15u) + fcurnt1;
emilmont 1:fdd22bb7aa52 325 fnext1 = __SSAT(fnext1, 16);
emilmont 1:fdd22bb7aa52 326 fnext2 = (q31_t) ((gnext1 * (*pk)) >> 15u) + fcurnt2;
emilmont 1:fdd22bb7aa52 327 fnext2 = __SSAT(fnext2, 16);
emilmont 1:fdd22bb7aa52 328
emilmont 1:fdd22bb7aa52 329 fnext3 = (q31_t) ((gnext2 * (*pk)) >> 15u) + fcurnt3;
emilmont 1:fdd22bb7aa52 330 fnext3 = __SSAT(fnext3, 16);
emilmont 1:fdd22bb7aa52 331
emilmont 1:fdd22bb7aa52 332 fnext4 = (q31_t) ((gnext3 * (*pk)) >> 15u) + fcurnt4;
emilmont 1:fdd22bb7aa52 333 fnext4 = __SSAT(fnext4, 16);
emilmont 1:fdd22bb7aa52 334
emilmont 1:fdd22bb7aa52 335 /* g1(n) = f0(n) * K1 + g0(n-1) */
emilmont 1:fdd22bb7aa52 336 gnext4 = (q31_t) ((fcurnt4 * (*pk)) >> 15u) + gnext3;
emilmont 1:fdd22bb7aa52 337 gnext4 = __SSAT(gnext4, 16);
emilmont 1:fdd22bb7aa52 338 gnext3 = (q31_t) ((fcurnt3 * (*pk)) >> 15u) + gnext2;
emilmont 1:fdd22bb7aa52 339 gnext3 = __SSAT(gnext3, 16);
emilmont 1:fdd22bb7aa52 340 gnext2 = (q31_t) ((fcurnt2 * (*pk)) >> 15u) + gnext1;
emilmont 1:fdd22bb7aa52 341 gnext2 = __SSAT(gnext2, 16);
emilmont 1:fdd22bb7aa52 342 gnext1 = (q31_t) ((fcurnt1 * (*pk++)) >> 15u) + gcurnt1;
emilmont 1:fdd22bb7aa52 343 gnext1 = __SSAT(gnext1, 16);
emilmont 1:fdd22bb7aa52 344
emilmont 1:fdd22bb7aa52 345 /* Update of f values for next coefficient set processing */
emilmont 1:fdd22bb7aa52 346 fcurnt1 = fnext1;
emilmont 1:fdd22bb7aa52 347 fcurnt2 = fnext2;
emilmont 1:fdd22bb7aa52 348 fcurnt3 = fnext3;
emilmont 1:fdd22bb7aa52 349 fcurnt4 = fnext4;
emilmont 1:fdd22bb7aa52 350
emilmont 1:fdd22bb7aa52 351 stageCnt--;
emilmont 1:fdd22bb7aa52 352
emilmont 1:fdd22bb7aa52 353 }
emilmont 1:fdd22bb7aa52 354
emilmont 1:fdd22bb7aa52 355 /* The results in the 4 accumulators, store in the destination buffer. */
emilmont 1:fdd22bb7aa52 356 /* y(n) = fN(n) */
emilmont 1:fdd22bb7aa52 357
emilmont 1:fdd22bb7aa52 358 #ifndef ARM_MATH_BIG_ENDIAN
emilmont 1:fdd22bb7aa52 359
emilmont 1:fdd22bb7aa52 360 *__SIMD32(pDst)++ = __PKHBT(fcurnt1, fcurnt2, 16);
emilmont 1:fdd22bb7aa52 361 *__SIMD32(pDst)++ = __PKHBT(fcurnt3, fcurnt4, 16);
emilmont 1:fdd22bb7aa52 362
emilmont 1:fdd22bb7aa52 363 #else
emilmont 1:fdd22bb7aa52 364
emilmont 1:fdd22bb7aa52 365 *__SIMD32(pDst)++ = __PKHBT(fcurnt2, fcurnt1, 16);
emilmont 1:fdd22bb7aa52 366 *__SIMD32(pDst)++ = __PKHBT(fcurnt4, fcurnt3, 16);
emilmont 1:fdd22bb7aa52 367
emilmont 1:fdd22bb7aa52 368 #endif /* #ifndef ARM_MATH_BIG_ENDIAN */
emilmont 1:fdd22bb7aa52 369
emilmont 1:fdd22bb7aa52 370 blkCnt--;
emilmont 1:fdd22bb7aa52 371 }
emilmont 1:fdd22bb7aa52 372
emilmont 1:fdd22bb7aa52 373 /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
emilmont 1:fdd22bb7aa52 374 ** No loop unrolling is used. */
emilmont 1:fdd22bb7aa52 375 blkCnt = blockSize % 0x4u;
emilmont 1:fdd22bb7aa52 376
emilmont 1:fdd22bb7aa52 377 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 378 {
emilmont 1:fdd22bb7aa52 379 /* f0(n) = x(n) */
emilmont 1:fdd22bb7aa52 380 fcurnt1 = *pSrc++;
emilmont 1:fdd22bb7aa52 381
emilmont 1:fdd22bb7aa52 382 /* Initialize coeff pointer */
emilmont 1:fdd22bb7aa52 383 pk = (pCoeffs);
emilmont 1:fdd22bb7aa52 384
emilmont 1:fdd22bb7aa52 385 /* Initialize state pointer */
emilmont 1:fdd22bb7aa52 386 px = pState;
emilmont 1:fdd22bb7aa52 387
emilmont 1:fdd22bb7aa52 388 /* read g2(n) from state buffer */
emilmont 1:fdd22bb7aa52 389 gcurnt1 = *px;
emilmont 1:fdd22bb7aa52 390
emilmont 1:fdd22bb7aa52 391 /* for sample 1 processing */
emilmont 1:fdd22bb7aa52 392 /* f1(n) = f0(n) + K1 * g0(n-1) */
emilmont 1:fdd22bb7aa52 393 fnext1 = (((q31_t) gcurnt1 * (*pk)) >> 15u) + fcurnt1;
emilmont 1:fdd22bb7aa52 394 fnext1 = __SSAT(fnext1, 16);
emilmont 1:fdd22bb7aa52 395
emilmont 1:fdd22bb7aa52 396
emilmont 1:fdd22bb7aa52 397 /* g1(n) = f0(n) * K1 + g0(n-1) */
emilmont 1:fdd22bb7aa52 398 gnext1 = (((q31_t) fcurnt1 * (*pk++)) >> 15u) + gcurnt1;
emilmont 1:fdd22bb7aa52 399 gnext1 = __SSAT(gnext1, 16);
emilmont 1:fdd22bb7aa52 400
emilmont 1:fdd22bb7aa52 401 /* save g1(n) in state buffer */
emilmont 1:fdd22bb7aa52 402 *px++ = (q15_t) fcurnt1;
emilmont 1:fdd22bb7aa52 403
emilmont 1:fdd22bb7aa52 404 /* f1(n) is saved in fcurnt1
emilmont 1:fdd22bb7aa52 405 for next stage processing */
emilmont 1:fdd22bb7aa52 406 fcurnt1 = fnext1;
emilmont 1:fdd22bb7aa52 407
emilmont 1:fdd22bb7aa52 408 stageCnt = (numStages - 1u);
emilmont 1:fdd22bb7aa52 409
emilmont 1:fdd22bb7aa52 410 /* stage loop */
emilmont 1:fdd22bb7aa52 411 while(stageCnt > 0u)
emilmont 1:fdd22bb7aa52 412 {
emilmont 1:fdd22bb7aa52 413 /* read g2(n) from state buffer */
emilmont 1:fdd22bb7aa52 414 gcurnt1 = *px;
emilmont 1:fdd22bb7aa52 415
emilmont 1:fdd22bb7aa52 416 /* save g1(n) in state buffer */
emilmont 1:fdd22bb7aa52 417 *px++ = (q15_t) gnext1;
emilmont 1:fdd22bb7aa52 418
emilmont 1:fdd22bb7aa52 419 /* Sample processing for K2, K3.... */
emilmont 1:fdd22bb7aa52 420 /* f2(n) = f1(n) + K2 * g1(n-1) */
emilmont 1:fdd22bb7aa52 421 fnext1 = (((q31_t) gcurnt1 * (*pk)) >> 15u) + fcurnt1;
emilmont 1:fdd22bb7aa52 422 fnext1 = __SSAT(fnext1, 16);
emilmont 1:fdd22bb7aa52 423
emilmont 1:fdd22bb7aa52 424 /* g2(n) = f1(n) * K2 + g1(n-1) */
emilmont 1:fdd22bb7aa52 425 gnext1 = (((q31_t) fcurnt1 * (*pk++)) >> 15u) + gcurnt1;
emilmont 1:fdd22bb7aa52 426 gnext1 = __SSAT(gnext1, 16);
emilmont 1:fdd22bb7aa52 427
emilmont 1:fdd22bb7aa52 428
emilmont 1:fdd22bb7aa52 429 /* f1(n) is saved in fcurnt1
emilmont 1:fdd22bb7aa52 430 for next stage processing */
emilmont 1:fdd22bb7aa52 431 fcurnt1 = fnext1;
emilmont 1:fdd22bb7aa52 432
emilmont 1:fdd22bb7aa52 433 stageCnt--;
emilmont 1:fdd22bb7aa52 434
emilmont 1:fdd22bb7aa52 435 }
emilmont 1:fdd22bb7aa52 436
emilmont 1:fdd22bb7aa52 437 /* y(n) = fN(n) */
emilmont 1:fdd22bb7aa52 438 *pDst++ = __SSAT(fcurnt1, 16);
emilmont 1:fdd22bb7aa52 439
emilmont 1:fdd22bb7aa52 440
emilmont 1:fdd22bb7aa52 441 blkCnt--;
emilmont 1:fdd22bb7aa52 442
emilmont 1:fdd22bb7aa52 443 }
emilmont 1:fdd22bb7aa52 444
emilmont 1:fdd22bb7aa52 445 #else
emilmont 1:fdd22bb7aa52 446
emilmont 1:fdd22bb7aa52 447 /* Run the below code for Cortex-M0 */
emilmont 1:fdd22bb7aa52 448
emilmont 1:fdd22bb7aa52 449 q31_t fcurnt, fnext, gcurnt, gnext; /* temporary variables */
emilmont 1:fdd22bb7aa52 450 uint32_t numStages = S->numStages; /* Length of the filter */
emilmont 1:fdd22bb7aa52 451 uint32_t blkCnt, stageCnt; /* temporary variables for counts */
emilmont 1:fdd22bb7aa52 452
emilmont 1:fdd22bb7aa52 453 pState = &S->pState[0];
emilmont 1:fdd22bb7aa52 454
emilmont 1:fdd22bb7aa52 455 blkCnt = blockSize;
emilmont 1:fdd22bb7aa52 456
emilmont 1:fdd22bb7aa52 457 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 458 {
emilmont 1:fdd22bb7aa52 459 /* f0(n) = x(n) */
emilmont 1:fdd22bb7aa52 460 fcurnt = *pSrc++;
emilmont 1:fdd22bb7aa52 461
emilmont 1:fdd22bb7aa52 462 /* Initialize coeff pointer */
emilmont 1:fdd22bb7aa52 463 pk = (pCoeffs);
emilmont 1:fdd22bb7aa52 464
emilmont 1:fdd22bb7aa52 465 /* Initialize state pointer */
emilmont 1:fdd22bb7aa52 466 px = pState;
emilmont 1:fdd22bb7aa52 467
emilmont 1:fdd22bb7aa52 468 /* read g0(n-1) from state buffer */
emilmont 1:fdd22bb7aa52 469 gcurnt = *px;
emilmont 1:fdd22bb7aa52 470
emilmont 1:fdd22bb7aa52 471 /* for sample 1 processing */
emilmont 1:fdd22bb7aa52 472 /* f1(n) = f0(n) + K1 * g0(n-1) */
emilmont 1:fdd22bb7aa52 473 fnext = ((gcurnt * (*pk)) >> 15u) + fcurnt;
emilmont 1:fdd22bb7aa52 474 fnext = __SSAT(fnext, 16);
emilmont 1:fdd22bb7aa52 475
emilmont 1:fdd22bb7aa52 476
emilmont 1:fdd22bb7aa52 477 /* g1(n) = f0(n) * K1 + g0(n-1) */
emilmont 1:fdd22bb7aa52 478 gnext = ((fcurnt * (*pk++)) >> 15u) + gcurnt;
emilmont 1:fdd22bb7aa52 479 gnext = __SSAT(gnext, 16);
emilmont 1:fdd22bb7aa52 480
emilmont 1:fdd22bb7aa52 481 /* save f0(n) in state buffer */
emilmont 1:fdd22bb7aa52 482 *px++ = (q15_t) fcurnt;
emilmont 1:fdd22bb7aa52 483
emilmont 1:fdd22bb7aa52 484 /* f1(n) is saved in fcurnt
emilmont 1:fdd22bb7aa52 485 for next stage processing */
emilmont 1:fdd22bb7aa52 486 fcurnt = fnext;
emilmont 1:fdd22bb7aa52 487
emilmont 1:fdd22bb7aa52 488 stageCnt = (numStages - 1u);
emilmont 1:fdd22bb7aa52 489
emilmont 1:fdd22bb7aa52 490 /* stage loop */
emilmont 1:fdd22bb7aa52 491 while(stageCnt > 0u)
emilmont 1:fdd22bb7aa52 492 {
emilmont 1:fdd22bb7aa52 493 /* read g1(n-1) from state buffer */
emilmont 1:fdd22bb7aa52 494 gcurnt = *px;
emilmont 1:fdd22bb7aa52 495
emilmont 1:fdd22bb7aa52 496 /* save g0(n-1) in state buffer */
emilmont 1:fdd22bb7aa52 497 *px++ = (q15_t) gnext;
emilmont 1:fdd22bb7aa52 498
emilmont 1:fdd22bb7aa52 499 /* Sample processing for K2, K3.... */
emilmont 1:fdd22bb7aa52 500 /* f2(n) = f1(n) + K2 * g1(n-1) */
emilmont 1:fdd22bb7aa52 501 fnext = ((gcurnt * (*pk)) >> 15u) + fcurnt;
emilmont 1:fdd22bb7aa52 502 fnext = __SSAT(fnext, 16);
emilmont 1:fdd22bb7aa52 503
emilmont 1:fdd22bb7aa52 504 /* g2(n) = f1(n) * K2 + g1(n-1) */
emilmont 1:fdd22bb7aa52 505 gnext = ((fcurnt * (*pk++)) >> 15u) + gcurnt;
emilmont 1:fdd22bb7aa52 506 gnext = __SSAT(gnext, 16);
emilmont 1:fdd22bb7aa52 507
emilmont 1:fdd22bb7aa52 508
emilmont 1:fdd22bb7aa52 509 /* f1(n) is saved in fcurnt
emilmont 1:fdd22bb7aa52 510 for next stage processing */
emilmont 1:fdd22bb7aa52 511 fcurnt = fnext;
emilmont 1:fdd22bb7aa52 512
emilmont 1:fdd22bb7aa52 513 stageCnt--;
emilmont 1:fdd22bb7aa52 514
emilmont 1:fdd22bb7aa52 515 }
emilmont 1:fdd22bb7aa52 516
emilmont 1:fdd22bb7aa52 517 /* y(n) = fN(n) */
emilmont 1:fdd22bb7aa52 518 *pDst++ = __SSAT(fcurnt, 16);
emilmont 1:fdd22bb7aa52 519
emilmont 1:fdd22bb7aa52 520
emilmont 1:fdd22bb7aa52 521 blkCnt--;
emilmont 1:fdd22bb7aa52 522
emilmont 1:fdd22bb7aa52 523 }
emilmont 1:fdd22bb7aa52 524
emilmont 1:fdd22bb7aa52 525 #endif /* #ifndef ARM_MATH_CM0 */
emilmont 1:fdd22bb7aa52 526
emilmont 1:fdd22bb7aa52 527 }
emilmont 1:fdd22bb7aa52 528
emilmont 1:fdd22bb7aa52 529 /**
emilmont 1:fdd22bb7aa52 530 * @} end of FIR_Lattice group
emilmont 1:fdd22bb7aa52 531 */