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_f32.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_f32.c
emilmont 1:fdd22bb7aa52 9 *
emilmont 1:fdd22bb7aa52 10 * Description: Floating-point FIR filter processing function.
emilmont 1:fdd22bb7aa52 11 *
emilmont 1:fdd22bb7aa52 12 * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
emilmont 1:fdd22bb7aa52 13 *
emilmont 1:fdd22bb7aa52 14 * Version 1.1.0 2012/02/15
emilmont 1:fdd22bb7aa52 15 * Updated with more optimizations, bug fixes and minor API changes.
emilmont 1:fdd22bb7aa52 16 *
emilmont 1:fdd22bb7aa52 17 * Version 1.0.2 2010/11/11
emilmont 1:fdd22bb7aa52 18 * Documentation updated.
emilmont 1:fdd22bb7aa52 19 *
emilmont 1:fdd22bb7aa52 20 * Version 1.0.1 2010/10/05
emilmont 1:fdd22bb7aa52 21 * Production release and review comments incorporated.
emilmont 1:fdd22bb7aa52 22 *
emilmont 1:fdd22bb7aa52 23 * Version 1.0.0 2010/09/20
emilmont 1:fdd22bb7aa52 24 * Production release and review comments incorporated.
emilmont 1:fdd22bb7aa52 25 *
emilmont 1:fdd22bb7aa52 26 * Version 0.0.5 2010/04/26
emilmont 1:fdd22bb7aa52 27 * incorporated review comments and updated with latest CMSIS layer
emilmont 1:fdd22bb7aa52 28 *
emilmont 1:fdd22bb7aa52 29 * Version 0.0.3 2010/03/10
emilmont 1:fdd22bb7aa52 30 * Initial version
emilmont 1:fdd22bb7aa52 31 * -------------------------------------------------------------------- */
emilmont 1:fdd22bb7aa52 32
emilmont 1:fdd22bb7aa52 33 #include "arm_math.h"
emilmont 1:fdd22bb7aa52 34
emilmont 1:fdd22bb7aa52 35 /**
emilmont 1:fdd22bb7aa52 36 * @ingroup groupFilters
emilmont 1:fdd22bb7aa52 37 */
emilmont 1:fdd22bb7aa52 38
emilmont 1:fdd22bb7aa52 39 /**
emilmont 1:fdd22bb7aa52 40 * @defgroup FIR Finite Impulse Response (FIR) Filters
emilmont 1:fdd22bb7aa52 41 *
emilmont 1:fdd22bb7aa52 42 * This set of functions implements Finite Impulse Response (FIR) filters
emilmont 1:fdd22bb7aa52 43 * for Q7, Q15, Q31, and floating-point data types. Fast versions of Q15 and Q31 are also provided.
emilmont 1:fdd22bb7aa52 44 * The functions operate on blocks of input and output data and each call to the function processes
emilmont 1:fdd22bb7aa52 45 * <code>blockSize</code> samples through the filter. <code>pSrc</code> and
emilmont 1:fdd22bb7aa52 46 * <code>pDst</code> points to input and output arrays containing <code>blockSize</code> values.
emilmont 1:fdd22bb7aa52 47 *
emilmont 1:fdd22bb7aa52 48 * \par Algorithm:
emilmont 1:fdd22bb7aa52 49 * The FIR filter algorithm is based upon a sequence of multiply-accumulate (MAC) operations.
emilmont 1:fdd22bb7aa52 50 * Each filter coefficient <code>b[n]</code> is multiplied by a state variable which equals a previous input sample <code>x[n]</code>.
emilmont 1:fdd22bb7aa52 51 * <pre>
emilmont 1:fdd22bb7aa52 52 * y[n] = b[0] * x[n] + b[1] * x[n-1] + b[2] * x[n-2] + ...+ b[numTaps-1] * x[n-numTaps+1]
emilmont 1:fdd22bb7aa52 53 * </pre>
emilmont 1:fdd22bb7aa52 54 * \par
emilmont 1:fdd22bb7aa52 55 * \image html FIR.gif "Finite Impulse Response filter"
emilmont 1:fdd22bb7aa52 56 * \par
emilmont 1:fdd22bb7aa52 57 * <code>pCoeffs</code> points to a coefficient array of size <code>numTaps</code>.
emilmont 1:fdd22bb7aa52 58 * Coefficients are stored in time reversed order.
emilmont 1:fdd22bb7aa52 59 * \par
emilmont 1:fdd22bb7aa52 60 * <pre>
emilmont 1:fdd22bb7aa52 61 * {b[numTaps-1], b[numTaps-2], b[N-2], ..., b[1], b[0]}
emilmont 1:fdd22bb7aa52 62 * </pre>
emilmont 1:fdd22bb7aa52 63 * \par
emilmont 1:fdd22bb7aa52 64 * <code>pState</code> points to a state array of size <code>numTaps + blockSize - 1</code>.
emilmont 1:fdd22bb7aa52 65 * Samples in the state buffer are stored in the following order.
emilmont 1:fdd22bb7aa52 66 * \par
emilmont 1:fdd22bb7aa52 67 * <pre>
emilmont 1:fdd22bb7aa52 68 * {x[n-numTaps+1], x[n-numTaps], x[n-numTaps-1], x[n-numTaps-2]....x[0], x[1], ..., x[blockSize-1]}
emilmont 1:fdd22bb7aa52 69 * </pre>
emilmont 1:fdd22bb7aa52 70 * \par
emilmont 1:fdd22bb7aa52 71 * Note that the length of the state buffer exceeds the length of the coefficient array by <code>blockSize-1</code>.
emilmont 1:fdd22bb7aa52 72 * The increased state buffer length allows circular addressing, which is traditionally used in the FIR filters,
emilmont 1:fdd22bb7aa52 73 * to be avoided and yields a significant speed improvement.
emilmont 1:fdd22bb7aa52 74 * The state variables are updated after each block of data is processed; the coefficients are untouched.
emilmont 1:fdd22bb7aa52 75 * \par Instance Structure
emilmont 1:fdd22bb7aa52 76 * The coefficients and state variables for a filter are stored together in an instance data structure.
emilmont 1:fdd22bb7aa52 77 * A separate instance structure must be defined for each filter.
emilmont 1:fdd22bb7aa52 78 * Coefficient arrays may be shared among several instances while state variable arrays cannot be shared.
emilmont 1:fdd22bb7aa52 79 * There are separate instance structure declarations for each of the 4 supported data types.
emilmont 1:fdd22bb7aa52 80 *
emilmont 1:fdd22bb7aa52 81 * \par Initialization Functions
emilmont 1:fdd22bb7aa52 82 * There is also an associated initialization function for each data type.
emilmont 1:fdd22bb7aa52 83 * The initialization function performs the following operations:
emilmont 1:fdd22bb7aa52 84 * - Sets the values of the internal structure fields.
emilmont 1:fdd22bb7aa52 85 * - Zeros out the values in the state buffer.
emilmont 1:fdd22bb7aa52 86 *
emilmont 1:fdd22bb7aa52 87 * \par
emilmont 1:fdd22bb7aa52 88 * Use of the initialization function is optional.
emilmont 1:fdd22bb7aa52 89 * However, if the initialization function is used, then the instance structure cannot be placed into a const data section.
emilmont 1:fdd22bb7aa52 90 * To place an instance structure into a const data section, the instance structure must be manually initialized.
emilmont 1:fdd22bb7aa52 91 * Set the values in the state buffer to zeros before static initialization.
emilmont 1:fdd22bb7aa52 92 * The code below statically initializes each of the 4 different data type filter instance structures
emilmont 1:fdd22bb7aa52 93 * <pre>
emilmont 1:fdd22bb7aa52 94 *arm_fir_instance_f32 S = {numTaps, pState, pCoeffs};
emilmont 1:fdd22bb7aa52 95 *arm_fir_instance_q31 S = {numTaps, pState, pCoeffs};
emilmont 1:fdd22bb7aa52 96 *arm_fir_instance_q15 S = {numTaps, pState, pCoeffs};
emilmont 1:fdd22bb7aa52 97 *arm_fir_instance_q7 S = {numTaps, pState, pCoeffs};
emilmont 1:fdd22bb7aa52 98 * </pre>
emilmont 1:fdd22bb7aa52 99 *
emilmont 1:fdd22bb7aa52 100 * where <code>numTaps</code> is the number of filter coefficients in the filter; <code>pState</code> is the address of the state buffer;
emilmont 1:fdd22bb7aa52 101 * <code>pCoeffs</code> is the address of the coefficient buffer.
emilmont 1:fdd22bb7aa52 102 *
emilmont 1:fdd22bb7aa52 103 * \par Fixed-Point Behavior
emilmont 1:fdd22bb7aa52 104 * Care must be taken when using the fixed-point versions of the FIR filter functions.
emilmont 1:fdd22bb7aa52 105 * In particular, the overflow and saturation behavior of the accumulator used in each function must be considered.
emilmont 1:fdd22bb7aa52 106 * Refer to the function specific documentation below for usage guidelines.
emilmont 1:fdd22bb7aa52 107 */
emilmont 1:fdd22bb7aa52 108
emilmont 1:fdd22bb7aa52 109 /**
emilmont 1:fdd22bb7aa52 110 * @addtogroup FIR
emilmont 1:fdd22bb7aa52 111 * @{
emilmont 1:fdd22bb7aa52 112 */
emilmont 1:fdd22bb7aa52 113
emilmont 1:fdd22bb7aa52 114 /**
emilmont 1:fdd22bb7aa52 115 *
emilmont 1:fdd22bb7aa52 116 * @param[in] *S points to an instance of the floating-point FIR filter structure.
emilmont 1:fdd22bb7aa52 117 * @param[in] *pSrc points to the block of input data.
emilmont 1:fdd22bb7aa52 118 * @param[out] *pDst points to the block of output data.
emilmont 1:fdd22bb7aa52 119 * @param[in] blockSize number of samples to process per call.
emilmont 1:fdd22bb7aa52 120 * @return none.
emilmont 1:fdd22bb7aa52 121 *
emilmont 1:fdd22bb7aa52 122 */
emilmont 1:fdd22bb7aa52 123
emilmont 1:fdd22bb7aa52 124 #ifndef ARM_MATH_CM0
emilmont 1:fdd22bb7aa52 125
emilmont 1:fdd22bb7aa52 126 /* Run the below code for Cortex-M4 and Cortex-M3 */
emilmont 1:fdd22bb7aa52 127
emilmont 1:fdd22bb7aa52 128 void arm_fir_f32(
emilmont 1:fdd22bb7aa52 129 const arm_fir_instance_f32 * S,
emilmont 1:fdd22bb7aa52 130 float32_t * pSrc,
emilmont 1:fdd22bb7aa52 131 float32_t * pDst,
emilmont 1:fdd22bb7aa52 132 uint32_t blockSize)
emilmont 1:fdd22bb7aa52 133 {
emilmont 1:fdd22bb7aa52 134 float32_t *pState = S->pState; /* State pointer */
emilmont 1:fdd22bb7aa52 135 float32_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
emilmont 1:fdd22bb7aa52 136 float32_t *pStateCurnt; /* Points to the current sample of the state */
emilmont 1:fdd22bb7aa52 137 float32_t *px, *pb; /* Temporary pointers for state and coefficient buffers */
emilmont 1:fdd22bb7aa52 138 float32_t acc0, acc1, acc2, acc3, acc4, acc5, acc6, acc7; /* Accumulators */
emilmont 1:fdd22bb7aa52 139 float32_t x0, x1, x2, x3, x4, x5, x6, x7, c0; /* Temporary variables to hold state and coefficient values */
emilmont 1:fdd22bb7aa52 140 uint32_t numTaps = S->numTaps; /* Number of filter coefficients in the filter */
emilmont 1:fdd22bb7aa52 141 uint32_t i, tapCnt, blkCnt; /* Loop counters */
emilmont 1:fdd22bb7aa52 142
emilmont 1:fdd22bb7aa52 143 /* S->pState points to state array which contains previous frame (numTaps - 1) samples */
emilmont 1:fdd22bb7aa52 144 /* pStateCurnt points to the location where the new input data should be written */
emilmont 1:fdd22bb7aa52 145 pStateCurnt = &(S->pState[(numTaps - 1u)]);
emilmont 1:fdd22bb7aa52 146
emilmont 1:fdd22bb7aa52 147 /* Apply loop unrolling and compute 4 output values simultaneously.
emilmont 1:fdd22bb7aa52 148 * The variables acc0 ... acc3 hold output values that are being computed:
emilmont 1:fdd22bb7aa52 149 *
emilmont 1:fdd22bb7aa52 150 * 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 151 * 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 152 * 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 153 * 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 154 */
emilmont 1:fdd22bb7aa52 155 blkCnt = blockSize >> 3;
emilmont 1:fdd22bb7aa52 156
emilmont 1:fdd22bb7aa52 157 /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
emilmont 1:fdd22bb7aa52 158 ** a second loop below computes the remaining 1 to 3 samples. */
emilmont 1:fdd22bb7aa52 159 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 160 {
emilmont 1:fdd22bb7aa52 161 /* Copy four new input samples into the state buffer */
emilmont 1:fdd22bb7aa52 162 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 163 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 164 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 165 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 166 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 167 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 168 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 169 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 170
emilmont 1:fdd22bb7aa52 171 /* Set all accumulators to zero */
emilmont 1:fdd22bb7aa52 172 acc0 = 0.0f;
emilmont 1:fdd22bb7aa52 173 acc1 = 0.0f;
emilmont 1:fdd22bb7aa52 174 acc2 = 0.0f;
emilmont 1:fdd22bb7aa52 175 acc3 = 0.0f;
emilmont 1:fdd22bb7aa52 176 acc4 = 0.0f;
emilmont 1:fdd22bb7aa52 177 acc5 = 0.0f;
emilmont 1:fdd22bb7aa52 178 acc6 = 0.0f;
emilmont 1:fdd22bb7aa52 179 acc7 = 0.0f;
emilmont 1:fdd22bb7aa52 180
emilmont 1:fdd22bb7aa52 181 /* Initialize state pointer */
emilmont 1:fdd22bb7aa52 182 px = pState;
emilmont 1:fdd22bb7aa52 183
emilmont 1:fdd22bb7aa52 184 /* Initialize coeff pointer */
emilmont 1:fdd22bb7aa52 185 pb = (pCoeffs);
emilmont 1:fdd22bb7aa52 186
emilmont 1:fdd22bb7aa52 187 /* Read the first three samples from the state buffer: x[n-numTaps], x[n-numTaps-1], x[n-numTaps-2] */
emilmont 1:fdd22bb7aa52 188 x0 = *px++;
emilmont 1:fdd22bb7aa52 189 x1 = *px++;
emilmont 1:fdd22bb7aa52 190 x2 = *px++;
emilmont 1:fdd22bb7aa52 191 x3 = *px++;
emilmont 1:fdd22bb7aa52 192 x4 = *px++;
emilmont 1:fdd22bb7aa52 193 x5 = *px++;
emilmont 1:fdd22bb7aa52 194 x6 = *px++;
emilmont 1:fdd22bb7aa52 195
emilmont 1:fdd22bb7aa52 196 /* Loop unrolling. Process 4 taps at a time. */
emilmont 1:fdd22bb7aa52 197 tapCnt = numTaps >> 3u;
emilmont 1:fdd22bb7aa52 198
emilmont 1:fdd22bb7aa52 199 /* Loop over the number of taps. Unroll by a factor of 4.
emilmont 1:fdd22bb7aa52 200 ** Repeat until we've computed numTaps-4 coefficients. */
emilmont 1:fdd22bb7aa52 201 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 202 {
emilmont 1:fdd22bb7aa52 203 /* Read the b[numTaps-1] coefficient */
emilmont 1:fdd22bb7aa52 204 c0 = *(pb++);
emilmont 1:fdd22bb7aa52 205
emilmont 1:fdd22bb7aa52 206 /* Read x[n-numTaps-3] sample */
emilmont 1:fdd22bb7aa52 207 x7 = *(px++);
emilmont 1:fdd22bb7aa52 208
emilmont 1:fdd22bb7aa52 209 /* acc0 += b[numTaps-1] * x[n-numTaps] */
emilmont 1:fdd22bb7aa52 210 acc0 += x0 * c0;
emilmont 1:fdd22bb7aa52 211
emilmont 1:fdd22bb7aa52 212 /* acc1 += b[numTaps-1] * x[n-numTaps-1] */
emilmont 1:fdd22bb7aa52 213 acc1 += x1 * c0;
emilmont 1:fdd22bb7aa52 214
emilmont 1:fdd22bb7aa52 215 /* acc2 += b[numTaps-1] * x[n-numTaps-2] */
emilmont 1:fdd22bb7aa52 216 acc2 += x2 * c0;
emilmont 1:fdd22bb7aa52 217
emilmont 1:fdd22bb7aa52 218 /* acc3 += b[numTaps-1] * x[n-numTaps-3] */
emilmont 1:fdd22bb7aa52 219 acc3 += x3 * c0;
emilmont 1:fdd22bb7aa52 220
emilmont 1:fdd22bb7aa52 221 /* acc4 += b[numTaps-1] * x[n-numTaps-4] */
emilmont 1:fdd22bb7aa52 222 acc4 += x4 * c0;
emilmont 1:fdd22bb7aa52 223
emilmont 1:fdd22bb7aa52 224 /* acc1 += b[numTaps-1] * x[n-numTaps-5] */
emilmont 1:fdd22bb7aa52 225 acc5 += x5 * c0;
emilmont 1:fdd22bb7aa52 226
emilmont 1:fdd22bb7aa52 227 /* acc2 += b[numTaps-1] * x[n-numTaps-6] */
emilmont 1:fdd22bb7aa52 228 acc6 += x6 * c0;
emilmont 1:fdd22bb7aa52 229
emilmont 1:fdd22bb7aa52 230 /* acc3 += b[numTaps-1] * x[n-numTaps-7] */
emilmont 1:fdd22bb7aa52 231 acc7 += x7 * c0;
emilmont 1:fdd22bb7aa52 232
emilmont 1:fdd22bb7aa52 233 /* Read the b[numTaps-2] coefficient */
emilmont 1:fdd22bb7aa52 234 c0 = *(pb++);
emilmont 1:fdd22bb7aa52 235
emilmont 1:fdd22bb7aa52 236 /* Read x[n-numTaps-4] sample */
emilmont 1:fdd22bb7aa52 237 x0 = *(px++);
emilmont 1:fdd22bb7aa52 238
emilmont 1:fdd22bb7aa52 239 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 240 acc0 += x1 * c0;
emilmont 1:fdd22bb7aa52 241 acc1 += x2 * c0;
emilmont 1:fdd22bb7aa52 242 acc2 += x3 * c0;
emilmont 1:fdd22bb7aa52 243 acc3 += x4 * c0;
emilmont 1:fdd22bb7aa52 244 acc4 += x5 * c0;
emilmont 1:fdd22bb7aa52 245 acc5 += x6 * c0;
emilmont 1:fdd22bb7aa52 246 acc6 += x7 * c0;
emilmont 1:fdd22bb7aa52 247 acc7 += x0 * c0;
emilmont 1:fdd22bb7aa52 248
emilmont 1:fdd22bb7aa52 249 /* Read the b[numTaps-3] coefficient */
emilmont 1:fdd22bb7aa52 250 c0 = *(pb++);
emilmont 1:fdd22bb7aa52 251
emilmont 1:fdd22bb7aa52 252 /* Read x[n-numTaps-5] sample */
emilmont 1:fdd22bb7aa52 253 x1 = *(px++);
emilmont 1:fdd22bb7aa52 254
emilmont 1:fdd22bb7aa52 255 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 256 acc0 += x2 * c0;
emilmont 1:fdd22bb7aa52 257 acc1 += x3 * c0;
emilmont 1:fdd22bb7aa52 258 acc2 += x4 * c0;
emilmont 1:fdd22bb7aa52 259 acc3 += x5 * c0;
emilmont 1:fdd22bb7aa52 260 acc4 += x6 * c0;
emilmont 1:fdd22bb7aa52 261 acc5 += x7 * c0;
emilmont 1:fdd22bb7aa52 262 acc6 += x0 * c0;
emilmont 1:fdd22bb7aa52 263 acc7 += x1 * c0;
emilmont 1:fdd22bb7aa52 264
emilmont 1:fdd22bb7aa52 265 /* Read the b[numTaps-4] coefficient */
emilmont 1:fdd22bb7aa52 266 c0 = *(pb++);
emilmont 1:fdd22bb7aa52 267
emilmont 1:fdd22bb7aa52 268 /* Read x[n-numTaps-6] sample */
emilmont 1:fdd22bb7aa52 269 x2 = *(px++);
emilmont 1:fdd22bb7aa52 270
emilmont 1:fdd22bb7aa52 271 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 272 acc0 += x3 * c0;
emilmont 1:fdd22bb7aa52 273 acc1 += x4 * c0;
emilmont 1:fdd22bb7aa52 274 acc2 += x5 * c0;
emilmont 1:fdd22bb7aa52 275 acc3 += x6 * c0;
emilmont 1:fdd22bb7aa52 276 acc4 += x7 * c0;
emilmont 1:fdd22bb7aa52 277 acc5 += x0 * c0;
emilmont 1:fdd22bb7aa52 278 acc6 += x1 * c0;
emilmont 1:fdd22bb7aa52 279 acc7 += x2 * c0;
emilmont 1:fdd22bb7aa52 280
emilmont 1:fdd22bb7aa52 281 /* Read the b[numTaps-4] coefficient */
emilmont 1:fdd22bb7aa52 282 c0 = *(pb++);
emilmont 1:fdd22bb7aa52 283
emilmont 1:fdd22bb7aa52 284 /* Read x[n-numTaps-6] sample */
emilmont 1:fdd22bb7aa52 285 x3 = *(px++);
emilmont 1:fdd22bb7aa52 286
emilmont 1:fdd22bb7aa52 287 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 288 acc0 += x4 * c0;
emilmont 1:fdd22bb7aa52 289 acc1 += x5 * c0;
emilmont 1:fdd22bb7aa52 290 acc2 += x6 * c0;
emilmont 1:fdd22bb7aa52 291 acc3 += x7 * c0;
emilmont 1:fdd22bb7aa52 292 acc4 += x0 * c0;
emilmont 1:fdd22bb7aa52 293 acc5 += x1 * c0;
emilmont 1:fdd22bb7aa52 294 acc6 += x2 * c0;
emilmont 1:fdd22bb7aa52 295 acc7 += x3 * c0;
emilmont 1:fdd22bb7aa52 296
emilmont 1:fdd22bb7aa52 297 /* Read the b[numTaps-4] coefficient */
emilmont 1:fdd22bb7aa52 298 c0 = *(pb++);
emilmont 1:fdd22bb7aa52 299
emilmont 1:fdd22bb7aa52 300 /* Read x[n-numTaps-6] sample */
emilmont 1:fdd22bb7aa52 301 x4 = *(px++);
emilmont 1:fdd22bb7aa52 302
emilmont 1:fdd22bb7aa52 303 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 304 acc0 += x5 * c0;
emilmont 1:fdd22bb7aa52 305 acc1 += x6 * c0;
emilmont 1:fdd22bb7aa52 306 acc2 += x7 * c0;
emilmont 1:fdd22bb7aa52 307 acc3 += x0 * c0;
emilmont 1:fdd22bb7aa52 308 acc4 += x1 * c0;
emilmont 1:fdd22bb7aa52 309 acc5 += x2 * c0;
emilmont 1:fdd22bb7aa52 310 acc6 += x3 * c0;
emilmont 1:fdd22bb7aa52 311 acc7 += x4 * c0;
emilmont 1:fdd22bb7aa52 312
emilmont 1:fdd22bb7aa52 313 /* Read the b[numTaps-4] coefficient */
emilmont 1:fdd22bb7aa52 314 c0 = *(pb++);
emilmont 1:fdd22bb7aa52 315
emilmont 1:fdd22bb7aa52 316 /* Read x[n-numTaps-6] sample */
emilmont 1:fdd22bb7aa52 317 x5 = *(px++);
emilmont 1:fdd22bb7aa52 318
emilmont 1:fdd22bb7aa52 319 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 320 acc0 += x6 * c0;
emilmont 1:fdd22bb7aa52 321 acc1 += x7 * c0;
emilmont 1:fdd22bb7aa52 322 acc2 += x0 * c0;
emilmont 1:fdd22bb7aa52 323 acc3 += x1 * c0;
emilmont 1:fdd22bb7aa52 324 acc4 += x2 * c0;
emilmont 1:fdd22bb7aa52 325 acc5 += x3 * c0;
emilmont 1:fdd22bb7aa52 326 acc6 += x4 * c0;
emilmont 1:fdd22bb7aa52 327 acc7 += x5 * c0;
emilmont 1:fdd22bb7aa52 328
emilmont 1:fdd22bb7aa52 329 /* Read the b[numTaps-4] coefficient */
emilmont 1:fdd22bb7aa52 330 c0 = *(pb++);
emilmont 1:fdd22bb7aa52 331
emilmont 1:fdd22bb7aa52 332 /* Read x[n-numTaps-6] sample */
emilmont 1:fdd22bb7aa52 333 x6 = *(px++);
emilmont 1:fdd22bb7aa52 334
emilmont 1:fdd22bb7aa52 335 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 336 acc0 += x7 * c0;
emilmont 1:fdd22bb7aa52 337 acc1 += x0 * c0;
emilmont 1:fdd22bb7aa52 338 acc2 += x1 * c0;
emilmont 1:fdd22bb7aa52 339 acc3 += x2 * c0;
emilmont 1:fdd22bb7aa52 340 acc4 += x3 * c0;
emilmont 1:fdd22bb7aa52 341 acc5 += x4 * c0;
emilmont 1:fdd22bb7aa52 342 acc6 += x5 * c0;
emilmont 1:fdd22bb7aa52 343 acc7 += x6 * c0;
emilmont 1:fdd22bb7aa52 344
emilmont 1:fdd22bb7aa52 345 tapCnt--;
emilmont 1:fdd22bb7aa52 346 }
emilmont 1:fdd22bb7aa52 347
emilmont 1:fdd22bb7aa52 348 /* If the filter length is not a multiple of 4, compute the remaining filter taps */
emilmont 1:fdd22bb7aa52 349 tapCnt = numTaps % 0x8u;
emilmont 1:fdd22bb7aa52 350
emilmont 1:fdd22bb7aa52 351 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 352 {
emilmont 1:fdd22bb7aa52 353 /* Read coefficients */
emilmont 1:fdd22bb7aa52 354 c0 = *(pb++);
emilmont 1:fdd22bb7aa52 355
emilmont 1:fdd22bb7aa52 356 /* Fetch 1 state variable */
emilmont 1:fdd22bb7aa52 357 x7 = *(px++);
emilmont 1:fdd22bb7aa52 358
emilmont 1:fdd22bb7aa52 359 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 360 acc0 += x0 * c0;
emilmont 1:fdd22bb7aa52 361 acc1 += x1 * c0;
emilmont 1:fdd22bb7aa52 362 acc2 += x2 * c0;
emilmont 1:fdd22bb7aa52 363 acc3 += x3 * c0;
emilmont 1:fdd22bb7aa52 364 acc4 += x4 * c0;
emilmont 1:fdd22bb7aa52 365 acc5 += x5 * c0;
emilmont 1:fdd22bb7aa52 366 acc6 += x6 * c0;
emilmont 1:fdd22bb7aa52 367 acc7 += x7 * c0;
emilmont 1:fdd22bb7aa52 368
emilmont 1:fdd22bb7aa52 369 /* Reuse the present sample states for next sample */
emilmont 1:fdd22bb7aa52 370 x0 = x1;
emilmont 1:fdd22bb7aa52 371 x1 = x2;
emilmont 1:fdd22bb7aa52 372 x2 = x3;
emilmont 1:fdd22bb7aa52 373 x3 = x4;
emilmont 1:fdd22bb7aa52 374 x4 = x5;
emilmont 1:fdd22bb7aa52 375 x5 = x6;
emilmont 1:fdd22bb7aa52 376 x6 = x7;
emilmont 1:fdd22bb7aa52 377
emilmont 1:fdd22bb7aa52 378 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 379 tapCnt--;
emilmont 1:fdd22bb7aa52 380 }
emilmont 1:fdd22bb7aa52 381
emilmont 1:fdd22bb7aa52 382 /* Advance the state pointer by 4 to process the next group of 4 samples */
emilmont 1:fdd22bb7aa52 383 pState = pState + 8;
emilmont 1:fdd22bb7aa52 384
emilmont 1:fdd22bb7aa52 385 /* The results in the 4 accumulators, store in the destination buffer. */
emilmont 1:fdd22bb7aa52 386 *pDst++ = acc0;
emilmont 1:fdd22bb7aa52 387 *pDst++ = acc1;
emilmont 1:fdd22bb7aa52 388 *pDst++ = acc2;
emilmont 1:fdd22bb7aa52 389 *pDst++ = acc3;
emilmont 1:fdd22bb7aa52 390 *pDst++ = acc4;
emilmont 1:fdd22bb7aa52 391 *pDst++ = acc5;
emilmont 1:fdd22bb7aa52 392 *pDst++ = acc6;
emilmont 1:fdd22bb7aa52 393 *pDst++ = acc7;
emilmont 1:fdd22bb7aa52 394
emilmont 1:fdd22bb7aa52 395 blkCnt--;
emilmont 1:fdd22bb7aa52 396 }
emilmont 1:fdd22bb7aa52 397
emilmont 1:fdd22bb7aa52 398 /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
emilmont 1:fdd22bb7aa52 399 ** No loop unrolling is used. */
emilmont 1:fdd22bb7aa52 400 blkCnt = blockSize % 0x8u;
emilmont 1:fdd22bb7aa52 401
emilmont 1:fdd22bb7aa52 402 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 403 {
emilmont 1:fdd22bb7aa52 404 /* Copy one sample at a time into state buffer */
emilmont 1:fdd22bb7aa52 405 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 406
emilmont 1:fdd22bb7aa52 407 /* Set the accumulator to zero */
emilmont 1:fdd22bb7aa52 408 acc0 = 0.0f;
emilmont 1:fdd22bb7aa52 409
emilmont 1:fdd22bb7aa52 410 /* Initialize state pointer */
emilmont 1:fdd22bb7aa52 411 px = pState;
emilmont 1:fdd22bb7aa52 412
emilmont 1:fdd22bb7aa52 413 /* Initialize Coefficient pointer */
emilmont 1:fdd22bb7aa52 414 pb = (pCoeffs);
emilmont 1:fdd22bb7aa52 415
emilmont 1:fdd22bb7aa52 416 i = numTaps;
emilmont 1:fdd22bb7aa52 417
emilmont 1:fdd22bb7aa52 418 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 419 do
emilmont 1:fdd22bb7aa52 420 {
emilmont 1:fdd22bb7aa52 421 acc0 += *px++ * *pb++;
emilmont 1:fdd22bb7aa52 422 i--;
emilmont 1:fdd22bb7aa52 423
emilmont 1:fdd22bb7aa52 424 } while(i > 0u);
emilmont 1:fdd22bb7aa52 425
emilmont 1:fdd22bb7aa52 426 /* The result is store in the destination buffer. */
emilmont 1:fdd22bb7aa52 427 *pDst++ = acc0;
emilmont 1:fdd22bb7aa52 428
emilmont 1:fdd22bb7aa52 429 /* Advance state pointer by 1 for the next sample */
emilmont 1:fdd22bb7aa52 430 pState = pState + 1;
emilmont 1:fdd22bb7aa52 431
emilmont 1:fdd22bb7aa52 432 blkCnt--;
emilmont 1:fdd22bb7aa52 433 }
emilmont 1:fdd22bb7aa52 434
emilmont 1:fdd22bb7aa52 435 /* Processing is complete.
emilmont 1:fdd22bb7aa52 436 ** Now copy the last numTaps - 1 samples to the satrt of the state buffer.
emilmont 1:fdd22bb7aa52 437 ** This prepares the state buffer for the next function call. */
emilmont 1:fdd22bb7aa52 438
emilmont 1:fdd22bb7aa52 439 /* Points to the start of the state buffer */
emilmont 1:fdd22bb7aa52 440 pStateCurnt = S->pState;
emilmont 1:fdd22bb7aa52 441
emilmont 1:fdd22bb7aa52 442 tapCnt = (numTaps - 1u) >> 2u;
emilmont 1:fdd22bb7aa52 443
emilmont 1:fdd22bb7aa52 444 /* copy data */
emilmont 1:fdd22bb7aa52 445 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 446 {
emilmont 1:fdd22bb7aa52 447 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 448 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 449 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 450 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 451
emilmont 1:fdd22bb7aa52 452 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 453 tapCnt--;
emilmont 1:fdd22bb7aa52 454 }
emilmont 1:fdd22bb7aa52 455
emilmont 1:fdd22bb7aa52 456 /* Calculate remaining number of copies */
emilmont 1:fdd22bb7aa52 457 tapCnt = (numTaps - 1u) % 0x4u;
emilmont 1:fdd22bb7aa52 458
emilmont 1:fdd22bb7aa52 459 /* Copy the remaining q31_t data */
emilmont 1:fdd22bb7aa52 460 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 461 {
emilmont 1:fdd22bb7aa52 462 *pStateCurnt++ = *pState++;
emilmont 1:fdd22bb7aa52 463
emilmont 1:fdd22bb7aa52 464 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 465 tapCnt--;
emilmont 1:fdd22bb7aa52 466 }
emilmont 1:fdd22bb7aa52 467 }
emilmont 1:fdd22bb7aa52 468
emilmont 1:fdd22bb7aa52 469 #else
emilmont 1:fdd22bb7aa52 470
emilmont 1:fdd22bb7aa52 471 void arm_fir_f32(
emilmont 1:fdd22bb7aa52 472 const arm_fir_instance_f32 * S,
emilmont 1:fdd22bb7aa52 473 float32_t * pSrc,
emilmont 1:fdd22bb7aa52 474 float32_t * pDst,
emilmont 1:fdd22bb7aa52 475 uint32_t blockSize)
emilmont 1:fdd22bb7aa52 476 {
emilmont 1:fdd22bb7aa52 477 float32_t *pState = S->pState; /* State pointer */
emilmont 1:fdd22bb7aa52 478 float32_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
emilmont 1:fdd22bb7aa52 479 float32_t *pStateCurnt; /* Points to the current sample of the state */
emilmont 1:fdd22bb7aa52 480 float32_t *px, *pb; /* Temporary pointers for state and coefficient buffers */
emilmont 1:fdd22bb7aa52 481 uint32_t numTaps = S->numTaps; /* Number of filter coefficients in the filter */
emilmont 1:fdd22bb7aa52 482 uint32_t i, tapCnt, blkCnt; /* Loop counters */
emilmont 1:fdd22bb7aa52 483
emilmont 1:fdd22bb7aa52 484 /* Run the below code for Cortex-M0 */
emilmont 1:fdd22bb7aa52 485
emilmont 1:fdd22bb7aa52 486 float32_t acc;
emilmont 1:fdd22bb7aa52 487
emilmont 1:fdd22bb7aa52 488 /* S->pState points to state array which contains previous frame (numTaps - 1) samples */
emilmont 1:fdd22bb7aa52 489 /* pStateCurnt points to the location where the new input data should be written */
emilmont 1:fdd22bb7aa52 490 pStateCurnt = &(S->pState[(numTaps - 1u)]);
emilmont 1:fdd22bb7aa52 491
emilmont 1:fdd22bb7aa52 492 /* Initialize blkCnt with blockSize */
emilmont 1:fdd22bb7aa52 493 blkCnt = blockSize;
emilmont 1:fdd22bb7aa52 494
emilmont 1:fdd22bb7aa52 495 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 496 {
emilmont 1:fdd22bb7aa52 497 /* Copy one sample at a time into state buffer */
emilmont 1:fdd22bb7aa52 498 *pStateCurnt++ = *pSrc++;
emilmont 1:fdd22bb7aa52 499
emilmont 1:fdd22bb7aa52 500 /* Set the accumulator to zero */
emilmont 1:fdd22bb7aa52 501 acc = 0.0f;
emilmont 1:fdd22bb7aa52 502
emilmont 1:fdd22bb7aa52 503 /* Initialize state pointer */
emilmont 1:fdd22bb7aa52 504 px = pState;
emilmont 1:fdd22bb7aa52 505
emilmont 1:fdd22bb7aa52 506 /* Initialize Coefficient pointer */
emilmont 1:fdd22bb7aa52 507 pb = pCoeffs;
emilmont 1:fdd22bb7aa52 508
emilmont 1:fdd22bb7aa52 509 i = numTaps;
emilmont 1:fdd22bb7aa52 510
emilmont 1:fdd22bb7aa52 511 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 512 do
emilmont 1:fdd22bb7aa52 513 {
emilmont 1:fdd22bb7aa52 514 /* acc = 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 515 acc += *px++ * *pb++;
emilmont 1:fdd22bb7aa52 516 i--;
emilmont 1:fdd22bb7aa52 517
emilmont 1:fdd22bb7aa52 518 } while(i > 0u);
emilmont 1:fdd22bb7aa52 519
emilmont 1:fdd22bb7aa52 520 /* The result is store in the destination buffer. */
emilmont 1:fdd22bb7aa52 521 *pDst++ = acc;
emilmont 1:fdd22bb7aa52 522
emilmont 1:fdd22bb7aa52 523 /* Advance state pointer by 1 for the next sample */
emilmont 1:fdd22bb7aa52 524 pState = pState + 1;
emilmont 1:fdd22bb7aa52 525
emilmont 1:fdd22bb7aa52 526 blkCnt--;
emilmont 1:fdd22bb7aa52 527 }
emilmont 1:fdd22bb7aa52 528
emilmont 1:fdd22bb7aa52 529 /* Processing is complete.
emilmont 1:fdd22bb7aa52 530 ** Now copy the last numTaps - 1 samples to the starting of the state buffer.
emilmont 1:fdd22bb7aa52 531 ** This prepares the state buffer for the next function call. */
emilmont 1:fdd22bb7aa52 532
emilmont 1:fdd22bb7aa52 533 /* Points to the start of the state buffer */
emilmont 1:fdd22bb7aa52 534 pStateCurnt = S->pState;
emilmont 1:fdd22bb7aa52 535
emilmont 1:fdd22bb7aa52 536 /* Copy numTaps number of values */
emilmont 1:fdd22bb7aa52 537 tapCnt = numTaps - 1u;
emilmont 1:fdd22bb7aa52 538
emilmont 1:fdd22bb7aa52 539 /* Copy data */
emilmont 1:fdd22bb7aa52 540 while(tapCnt > 0u)
emilmont 1:fdd22bb7aa52 541 {
emilmont 1:fdd22bb7aa52 542 *pStateCurnt++ = *pState++;
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 }
emilmont 1:fdd22bb7aa52 549
emilmont 1:fdd22bb7aa52 550 #endif /* #ifndef ARM_MATH_CM0 */
emilmont 1:fdd22bb7aa52 551
emilmont 1:fdd22bb7aa52 552 /**
emilmont 1:fdd22bb7aa52 553 * @} end of FIR group
emilmont 1:fdd22bb7aa52 554 */