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_biquad_cascade_df1_32x64_q31.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. October 2015
mbed_official 5:3762170b6d4d 5 * $Revision: V.1.4.5 a
emilmont 1:fdd22bb7aa52 6 *
emilmont 2:da51fb522205 7 * Project: CMSIS DSP Library
emilmont 2:da51fb522205 8 * Title: arm_biquad_cascade_df1_32x64_q31.c
emilmont 1:fdd22bb7aa52 9 *
emilmont 2:da51fb522205 10 * Description: High precision Q31 Biquad cascade filter processing function
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 * @defgroup BiquadCascadeDF1_32x64 High Precision Q31 Biquad Cascade Filter
emilmont 1:fdd22bb7aa52 49 *
emilmont 1:fdd22bb7aa52 50 * This function implements a high precision Biquad cascade filter which operates on
emilmont 1:fdd22bb7aa52 51 * Q31 data values. The filter coefficients are in 1.31 format and the state variables
emilmont 1:fdd22bb7aa52 52 * are in 1.63 format. The double precision state variables reduce quantization noise
emilmont 1:fdd22bb7aa52 53 * in the filter and provide a cleaner output.
emilmont 1:fdd22bb7aa52 54 * These filters are particularly useful when implementing filters in which the
emilmont 1:fdd22bb7aa52 55 * singularities are close to the unit circle. This is common for low pass or high
emilmont 1:fdd22bb7aa52 56 * pass filters with very low cutoff frequencies.
emilmont 1:fdd22bb7aa52 57 *
emilmont 1:fdd22bb7aa52 58 * The function operates on blocks of input and output data
emilmont 1:fdd22bb7aa52 59 * and each call to the function processes <code>blockSize</code> samples through
emilmont 1:fdd22bb7aa52 60 * the filter. <code>pSrc</code> and <code>pDst</code> points to input and output arrays
emilmont 1:fdd22bb7aa52 61 * containing <code>blockSize</code> Q31 values.
emilmont 1:fdd22bb7aa52 62 *
emilmont 1:fdd22bb7aa52 63 * \par Algorithm
emilmont 1:fdd22bb7aa52 64 * Each Biquad stage implements a second order filter using the difference equation:
emilmont 1:fdd22bb7aa52 65 * <pre>
emilmont 1:fdd22bb7aa52 66 * y[n] = b0 * x[n] + b1 * x[n-1] + b2 * x[n-2] + a1 * y[n-1] + a2 * y[n-2]
emilmont 1:fdd22bb7aa52 67 * </pre>
emilmont 1:fdd22bb7aa52 68 * A Direct Form I algorithm is used with 5 coefficients and 4 state variables per stage.
emilmont 1:fdd22bb7aa52 69 * \image html Biquad.gif "Single Biquad filter stage"
emilmont 1:fdd22bb7aa52 70 * Coefficients <code>b0, b1, and b2 </code> multiply the input signal <code>x[n]</code> and are referred to as the feedforward coefficients.
emilmont 1:fdd22bb7aa52 71 * Coefficients <code>a1</code> and <code>a2</code> multiply the output signal <code>y[n]</code> and are referred to as the feedback coefficients.
emilmont 1:fdd22bb7aa52 72 * Pay careful attention to the sign of the feedback coefficients.
emilmont 1:fdd22bb7aa52 73 * Some design tools use the difference equation
emilmont 1:fdd22bb7aa52 74 * <pre>
emilmont 1:fdd22bb7aa52 75 * y[n] = b0 * x[n] + b1 * x[n-1] + b2 * x[n-2] - a1 * y[n-1] - a2 * y[n-2]
emilmont 1:fdd22bb7aa52 76 * </pre>
emilmont 1:fdd22bb7aa52 77 * In this case the feedback coefficients <code>a1</code> and <code>a2</code> must be negated when used with the CMSIS DSP Library.
emilmont 1:fdd22bb7aa52 78 *
emilmont 1:fdd22bb7aa52 79 * \par
emilmont 1:fdd22bb7aa52 80 * Higher order filters are realized as a cascade of second order sections.
emilmont 1:fdd22bb7aa52 81 * <code>numStages</code> refers to the number of second order stages used.
emilmont 1:fdd22bb7aa52 82 * For example, an 8th order filter would be realized with <code>numStages=4</code> second order stages.
emilmont 1:fdd22bb7aa52 83 * \image html BiquadCascade.gif "8th order filter using a cascade of Biquad stages"
emilmont 1:fdd22bb7aa52 84 * A 9th order filter would be realized with <code>numStages=5</code> second order stages with the coefficients for one of the stages configured as a first order filter (<code>b2=0</code> and <code>a2=0</code>).
emilmont 1:fdd22bb7aa52 85 *
emilmont 1:fdd22bb7aa52 86 * \par
emilmont 1:fdd22bb7aa52 87 * The <code>pState</code> points to state variables array .
emilmont 1:fdd22bb7aa52 88 * Each Biquad stage has 4 state variables <code>x[n-1], x[n-2], y[n-1],</code> and <code>y[n-2]</code> and each state variable in 1.63 format to improve precision.
emilmont 1:fdd22bb7aa52 89 * The state variables are arranged in the array as:
emilmont 1:fdd22bb7aa52 90 * <pre>
emilmont 1:fdd22bb7aa52 91 * {x[n-1], x[n-2], y[n-1], y[n-2]}
emilmont 1:fdd22bb7aa52 92 * </pre>
emilmont 1:fdd22bb7aa52 93 *
emilmont 1:fdd22bb7aa52 94 * \par
emilmont 1:fdd22bb7aa52 95 * The 4 state variables for stage 1 are first, then the 4 state variables for stage 2, and so on.
emilmont 1:fdd22bb7aa52 96 * The state array has a total length of <code>4*numStages</code> values of data in 1.63 format.
emilmont 1:fdd22bb7aa52 97 * The state variables are updated after each block of data is processed; the coefficients are untouched.
emilmont 1:fdd22bb7aa52 98 *
emilmont 1:fdd22bb7aa52 99 * \par Instance Structure
emilmont 1:fdd22bb7aa52 100 * The coefficients and state variables for a filter are stored together in an instance data structure.
emilmont 1:fdd22bb7aa52 101 * A separate instance structure must be defined for each filter.
emilmont 1:fdd22bb7aa52 102 * Coefficient arrays may be shared among several instances while state variable arrays cannot be shared.
emilmont 1:fdd22bb7aa52 103 *
emilmont 1:fdd22bb7aa52 104 * \par Init Function
emilmont 1:fdd22bb7aa52 105 * There is also an associated initialization function which performs the following operations:
emilmont 1:fdd22bb7aa52 106 * - Sets the values of the internal structure fields.
emilmont 1:fdd22bb7aa52 107 * - Zeros out the values in the state buffer.
mbed_official 3:7a284390b0ce 108 * To do this manually without calling the init function, assign the follow subfields of the instance structure:
mbed_official 3:7a284390b0ce 109 * numStages, pCoeffs, postShift, pState. Also set all of the values in pState to zero.
mbed_official 3:7a284390b0ce 110 *
emilmont 1:fdd22bb7aa52 111 * \par
emilmont 1:fdd22bb7aa52 112 * Use of the initialization function is optional.
emilmont 1:fdd22bb7aa52 113 * However, if the initialization function is used, then the instance structure cannot be placed into a const data section.
emilmont 1:fdd22bb7aa52 114 * To place an instance structure into a const data section, the instance structure must be manually initialized.
emilmont 1:fdd22bb7aa52 115 * Set the values in the state buffer to zeros before static initialization.
emilmont 1:fdd22bb7aa52 116 * For example, to statically initialize the filter instance structure use
emilmont 1:fdd22bb7aa52 117 * <pre>
emilmont 1:fdd22bb7aa52 118 * arm_biquad_cas_df1_32x64_ins_q31 S1 = {numStages, pState, pCoeffs, postShift};
emilmont 1:fdd22bb7aa52 119 * </pre>
emilmont 1:fdd22bb7aa52 120 * where <code>numStages</code> is the number of Biquad stages in the filter; <code>pState</code> is the address of the state buffer;
emilmont 1:fdd22bb7aa52 121 * <code>pCoeffs</code> is the address of the coefficient buffer; <code>postShift</code> shift to be applied which is described in detail below.
emilmont 1:fdd22bb7aa52 122 * \par Fixed-Point Behavior
emilmont 1:fdd22bb7aa52 123 * Care must be taken while using Biquad Cascade 32x64 filter function.
emilmont 1:fdd22bb7aa52 124 * Following issues must be considered:
emilmont 1:fdd22bb7aa52 125 * - Scaling of coefficients
emilmont 1:fdd22bb7aa52 126 * - Filter gain
emilmont 1:fdd22bb7aa52 127 * - Overflow and saturation
emilmont 1:fdd22bb7aa52 128 *
emilmont 1:fdd22bb7aa52 129 * \par
emilmont 1:fdd22bb7aa52 130 * Filter coefficients are represented as fractional values and
emilmont 1:fdd22bb7aa52 131 * restricted to lie in the range <code>[-1 +1)</code>.
emilmont 1:fdd22bb7aa52 132 * The processing function has an additional scaling parameter <code>postShift</code>
emilmont 1:fdd22bb7aa52 133 * which allows the filter coefficients to exceed the range <code>[+1 -1)</code>.
emilmont 1:fdd22bb7aa52 134 * At the output of the filter's accumulator is a shift register which shifts the result by <code>postShift</code> bits.
emilmont 1:fdd22bb7aa52 135 * \image html BiquadPostshift.gif "Fixed-point Biquad with shift by postShift bits after accumulator"
emilmont 1:fdd22bb7aa52 136 * This essentially scales the filter coefficients by <code>2^postShift</code>.
emilmont 1:fdd22bb7aa52 137 * For example, to realize the coefficients
emilmont 1:fdd22bb7aa52 138 * <pre>
emilmont 1:fdd22bb7aa52 139 * {1.5, -0.8, 1.2, 1.6, -0.9}
emilmont 1:fdd22bb7aa52 140 * </pre>
emilmont 1:fdd22bb7aa52 141 * set the Coefficient array to:
emilmont 1:fdd22bb7aa52 142 * <pre>
emilmont 1:fdd22bb7aa52 143 * {0.75, -0.4, 0.6, 0.8, -0.45}
emilmont 1:fdd22bb7aa52 144 * </pre>
emilmont 1:fdd22bb7aa52 145 * and set <code>postShift=1</code>
emilmont 1:fdd22bb7aa52 146 *
emilmont 1:fdd22bb7aa52 147 * \par
emilmont 1:fdd22bb7aa52 148 * The second thing to keep in mind is the gain through the filter.
emilmont 1:fdd22bb7aa52 149 * The frequency response of a Biquad filter is a function of its coefficients.
emilmont 1:fdd22bb7aa52 150 * It is possible for the gain through the filter to exceed 1.0 meaning that the filter increases the amplitude of certain frequencies.
emilmont 1:fdd22bb7aa52 151 * This means that an input signal with amplitude < 1.0 may result in an output > 1.0 and these are saturated or overflowed based on the implementation of the filter.
emilmont 1:fdd22bb7aa52 152 * To avoid this behavior the filter needs to be scaled down such that its peak gain < 1.0 or the input signal must be scaled down so that the combination of input and filter are never overflowed.
emilmont 1:fdd22bb7aa52 153 *
emilmont 1:fdd22bb7aa52 154 * \par
emilmont 1:fdd22bb7aa52 155 * The third item to consider is the overflow and saturation behavior of the fixed-point Q31 version.
emilmont 1:fdd22bb7aa52 156 * This is described in the function specific documentation below.
emilmont 1:fdd22bb7aa52 157 */
emilmont 1:fdd22bb7aa52 158
emilmont 1:fdd22bb7aa52 159 /**
emilmont 1:fdd22bb7aa52 160 * @addtogroup BiquadCascadeDF1_32x64
emilmont 1:fdd22bb7aa52 161 * @{
emilmont 1:fdd22bb7aa52 162 */
emilmont 1:fdd22bb7aa52 163
emilmont 1:fdd22bb7aa52 164 /**
emilmont 1:fdd22bb7aa52 165 * @details
emilmont 1:fdd22bb7aa52 166
emilmont 1:fdd22bb7aa52 167 * @param[in] *S points to an instance of the high precision Q31 Biquad cascade filter.
emilmont 1:fdd22bb7aa52 168 * @param[in] *pSrc points to the block of input data.
emilmont 1:fdd22bb7aa52 169 * @param[out] *pDst points to the block of output data.
emilmont 1:fdd22bb7aa52 170 * @param[in] blockSize number of samples to process.
emilmont 1:fdd22bb7aa52 171 * @return none.
emilmont 1:fdd22bb7aa52 172 *
emilmont 1:fdd22bb7aa52 173 * \par
emilmont 1:fdd22bb7aa52 174 * The function is implemented using an internal 64-bit accumulator.
emilmont 1:fdd22bb7aa52 175 * The accumulator has a 2.62 format and maintains full precision of the intermediate multiplication results but provides only a single guard bit.
emilmont 1:fdd22bb7aa52 176 * Thus, if the accumulator result overflows it wraps around rather than clip.
emilmont 1:fdd22bb7aa52 177 * In order to avoid overflows completely the input signal must be scaled down by 2 bits and lie in the range [-0.25 +0.25).
emilmont 1:fdd22bb7aa52 178 * After all 5 multiply-accumulates are performed, the 2.62 accumulator is shifted by <code>postShift</code> bits and the result truncated to
emilmont 1:fdd22bb7aa52 179 * 1.31 format by discarding the low 32 bits.
emilmont 1:fdd22bb7aa52 180 *
emilmont 1:fdd22bb7aa52 181 * \par
emilmont 1:fdd22bb7aa52 182 * Two related functions are provided in the CMSIS DSP library.
emilmont 1:fdd22bb7aa52 183 * <code>arm_biquad_cascade_df1_q31()</code> implements a Biquad cascade with 32-bit coefficients and state variables with a Q63 accumulator.
emilmont 1:fdd22bb7aa52 184 * <code>arm_biquad_cascade_df1_fast_q31()</code> implements a Biquad cascade with 32-bit coefficients and state variables with a Q31 accumulator.
emilmont 1:fdd22bb7aa52 185 */
emilmont 1:fdd22bb7aa52 186
emilmont 1:fdd22bb7aa52 187 void arm_biquad_cas_df1_32x64_q31(
emilmont 1:fdd22bb7aa52 188 const arm_biquad_cas_df1_32x64_ins_q31 * S,
emilmont 1:fdd22bb7aa52 189 q31_t * pSrc,
emilmont 1:fdd22bb7aa52 190 q31_t * pDst,
emilmont 1:fdd22bb7aa52 191 uint32_t blockSize)
emilmont 1:fdd22bb7aa52 192 {
emilmont 1:fdd22bb7aa52 193 q31_t *pIn = pSrc; /* input pointer initialization */
emilmont 1:fdd22bb7aa52 194 q31_t *pOut = pDst; /* output pointer initialization */
emilmont 1:fdd22bb7aa52 195 q63_t *pState = S->pState; /* state pointer initialization */
emilmont 1:fdd22bb7aa52 196 q31_t *pCoeffs = S->pCoeffs; /* coeff pointer initialization */
emilmont 1:fdd22bb7aa52 197 q63_t acc; /* accumulator */
emilmont 1:fdd22bb7aa52 198 q31_t Xn1, Xn2; /* Input Filter state variables */
emilmont 1:fdd22bb7aa52 199 q63_t Yn1, Yn2; /* Output Filter state variables */
emilmont 1:fdd22bb7aa52 200 q31_t b0, b1, b2, a1, a2; /* Filter coefficients */
emilmont 1:fdd22bb7aa52 201 q31_t Xn; /* temporary input */
emilmont 1:fdd22bb7aa52 202 int32_t shift = (int32_t) S->postShift + 1; /* Shift to be applied to the output */
emilmont 1:fdd22bb7aa52 203 uint32_t sample, stage = S->numStages; /* loop counters */
emilmont 1:fdd22bb7aa52 204 q31_t acc_l, acc_h; /* temporary output */
emilmont 1:fdd22bb7aa52 205 uint32_t uShift = ((uint32_t) S->postShift + 1u);
emilmont 1:fdd22bb7aa52 206 uint32_t lShift = 32u - uShift; /* Shift to be applied to the output */
emilmont 1:fdd22bb7aa52 207
emilmont 1:fdd22bb7aa52 208
mbed_official 3:7a284390b0ce 209 #ifndef ARM_MATH_CM0_FAMILY
emilmont 1:fdd22bb7aa52 210
emilmont 1:fdd22bb7aa52 211 /* Run the below code for Cortex-M4 and Cortex-M3 */
emilmont 1:fdd22bb7aa52 212
emilmont 1:fdd22bb7aa52 213 do
emilmont 1:fdd22bb7aa52 214 {
emilmont 1:fdd22bb7aa52 215 /* Reading the coefficients */
emilmont 1:fdd22bb7aa52 216 b0 = *pCoeffs++;
emilmont 1:fdd22bb7aa52 217 b1 = *pCoeffs++;
emilmont 1:fdd22bb7aa52 218 b2 = *pCoeffs++;
emilmont 1:fdd22bb7aa52 219 a1 = *pCoeffs++;
emilmont 1:fdd22bb7aa52 220 a2 = *pCoeffs++;
emilmont 1:fdd22bb7aa52 221
emilmont 1:fdd22bb7aa52 222 /* Reading the state values */
emilmont 1:fdd22bb7aa52 223 Xn1 = (q31_t) (pState[0]);
emilmont 1:fdd22bb7aa52 224 Xn2 = (q31_t) (pState[1]);
emilmont 1:fdd22bb7aa52 225 Yn1 = pState[2];
emilmont 1:fdd22bb7aa52 226 Yn2 = pState[3];
emilmont 1:fdd22bb7aa52 227
emilmont 1:fdd22bb7aa52 228 /* Apply loop unrolling and compute 4 output values simultaneously. */
emilmont 1:fdd22bb7aa52 229 /* The variable acc hold output value that is being computed and
emilmont 1:fdd22bb7aa52 230 * stored in the destination buffer
emilmont 1:fdd22bb7aa52 231 * acc = b0 * x[n] + b1 * x[n-1] + b2 * x[n-2] + a1 * y[n-1] + a2 * y[n-2]
emilmont 1:fdd22bb7aa52 232 */
emilmont 1:fdd22bb7aa52 233
emilmont 1:fdd22bb7aa52 234 sample = blockSize >> 2u;
emilmont 1:fdd22bb7aa52 235
emilmont 1:fdd22bb7aa52 236 /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
emilmont 1:fdd22bb7aa52 237 ** a second loop below computes the remaining 1 to 3 samples. */
emilmont 1:fdd22bb7aa52 238 while(sample > 0u)
emilmont 1:fdd22bb7aa52 239 {
emilmont 1:fdd22bb7aa52 240 /* Read the input */
emilmont 1:fdd22bb7aa52 241 Xn = *pIn++;
emilmont 1:fdd22bb7aa52 242
emilmont 1:fdd22bb7aa52 243 /* acc = b0 * x[n] + b1 * x[n-1] + b2 * x[n-2] + a1 * y[n-1] + a2 * y[n-2] */
emilmont 1:fdd22bb7aa52 244
emilmont 1:fdd22bb7aa52 245 /* acc = b0 * x[n] */
emilmont 1:fdd22bb7aa52 246 acc = (q63_t) Xn *b0;
emilmont 1:fdd22bb7aa52 247
emilmont 1:fdd22bb7aa52 248 /* acc += b1 * x[n-1] */
emilmont 1:fdd22bb7aa52 249 acc += (q63_t) Xn1 *b1;
emilmont 1:fdd22bb7aa52 250
emilmont 1:fdd22bb7aa52 251 /* acc += b[2] * x[n-2] */
emilmont 1:fdd22bb7aa52 252 acc += (q63_t) Xn2 *b2;
emilmont 1:fdd22bb7aa52 253
emilmont 1:fdd22bb7aa52 254 /* acc += a1 * y[n-1] */
emilmont 1:fdd22bb7aa52 255 acc += mult32x64(Yn1, a1);
emilmont 1:fdd22bb7aa52 256
emilmont 1:fdd22bb7aa52 257 /* acc += a2 * y[n-2] */
emilmont 1:fdd22bb7aa52 258 acc += mult32x64(Yn2, a2);
emilmont 1:fdd22bb7aa52 259
emilmont 1:fdd22bb7aa52 260 /* The result is converted to 1.63 , Yn2 variable is reused */
emilmont 1:fdd22bb7aa52 261 Yn2 = acc << shift;
emilmont 1:fdd22bb7aa52 262
emilmont 1:fdd22bb7aa52 263 /* Calc lower part of acc */
emilmont 1:fdd22bb7aa52 264 acc_l = acc & 0xffffffff;
emilmont 1:fdd22bb7aa52 265
emilmont 1:fdd22bb7aa52 266 /* Calc upper part of acc */
emilmont 1:fdd22bb7aa52 267 acc_h = (acc >> 32) & 0xffffffff;
emilmont 1:fdd22bb7aa52 268
emilmont 1:fdd22bb7aa52 269 /* Apply shift for lower part of acc and upper part of acc */
emilmont 1:fdd22bb7aa52 270 acc_h = (uint32_t) acc_l >> lShift | acc_h << uShift;
emilmont 1:fdd22bb7aa52 271
emilmont 1:fdd22bb7aa52 272 /* Store the output in the destination buffer in 1.31 format. */
emilmont 1:fdd22bb7aa52 273 *pOut = acc_h;
emilmont 1:fdd22bb7aa52 274
emilmont 1:fdd22bb7aa52 275 /* Read the second input into Xn2, to reuse the value */
emilmont 1:fdd22bb7aa52 276 Xn2 = *pIn++;
emilmont 1:fdd22bb7aa52 277
emilmont 1:fdd22bb7aa52 278 /* acc = b0 * x[n] + b1 * x[n-1] + b2 * x[n-2] + a1 * y[n-1] + a2 * y[n-2] */
emilmont 1:fdd22bb7aa52 279
emilmont 1:fdd22bb7aa52 280 /* acc += b1 * x[n-1] */
emilmont 1:fdd22bb7aa52 281 acc = (q63_t) Xn *b1;
emilmont 1:fdd22bb7aa52 282
emilmont 1:fdd22bb7aa52 283 /* acc = b0 * x[n] */
emilmont 1:fdd22bb7aa52 284 acc += (q63_t) Xn2 *b0;
emilmont 1:fdd22bb7aa52 285
emilmont 1:fdd22bb7aa52 286 /* acc += b[2] * x[n-2] */
emilmont 1:fdd22bb7aa52 287 acc += (q63_t) Xn1 *b2;
emilmont 1:fdd22bb7aa52 288
emilmont 1:fdd22bb7aa52 289 /* acc += a1 * y[n-1] */
emilmont 1:fdd22bb7aa52 290 acc += mult32x64(Yn2, a1);
emilmont 1:fdd22bb7aa52 291
emilmont 1:fdd22bb7aa52 292 /* acc += a2 * y[n-2] */
emilmont 1:fdd22bb7aa52 293 acc += mult32x64(Yn1, a2);
emilmont 1:fdd22bb7aa52 294
emilmont 1:fdd22bb7aa52 295 /* The result is converted to 1.63, Yn1 variable is reused */
emilmont 1:fdd22bb7aa52 296 Yn1 = acc << shift;
emilmont 1:fdd22bb7aa52 297
emilmont 1:fdd22bb7aa52 298 /* Calc lower part of acc */
emilmont 1:fdd22bb7aa52 299 acc_l = acc & 0xffffffff;
emilmont 1:fdd22bb7aa52 300
emilmont 1:fdd22bb7aa52 301 /* Calc upper part of acc */
emilmont 1:fdd22bb7aa52 302 acc_h = (acc >> 32) & 0xffffffff;
emilmont 1:fdd22bb7aa52 303
emilmont 1:fdd22bb7aa52 304 /* Apply shift for lower part of acc and upper part of acc */
emilmont 1:fdd22bb7aa52 305 acc_h = (uint32_t) acc_l >> lShift | acc_h << uShift;
emilmont 1:fdd22bb7aa52 306
emilmont 1:fdd22bb7aa52 307 /* Read the third input into Xn1, to reuse the value */
emilmont 1:fdd22bb7aa52 308 Xn1 = *pIn++;
emilmont 1:fdd22bb7aa52 309
emilmont 1:fdd22bb7aa52 310 /* The result is converted to 1.31 */
emilmont 1:fdd22bb7aa52 311 /* Store the output in the destination buffer. */
emilmont 1:fdd22bb7aa52 312 *(pOut + 1u) = acc_h;
emilmont 1:fdd22bb7aa52 313
emilmont 1:fdd22bb7aa52 314 /* acc = b0 * x[n] + b1 * x[n-1] + b2 * x[n-2] + a1 * y[n-1] + a2 * y[n-2] */
emilmont 1:fdd22bb7aa52 315
emilmont 1:fdd22bb7aa52 316 /* acc = b0 * x[n] */
emilmont 1:fdd22bb7aa52 317 acc = (q63_t) Xn1 *b0;
emilmont 1:fdd22bb7aa52 318
emilmont 1:fdd22bb7aa52 319 /* acc += b1 * x[n-1] */
emilmont 1:fdd22bb7aa52 320 acc += (q63_t) Xn2 *b1;
emilmont 1:fdd22bb7aa52 321
emilmont 1:fdd22bb7aa52 322 /* acc += b[2] * x[n-2] */
emilmont 1:fdd22bb7aa52 323 acc += (q63_t) Xn *b2;
emilmont 1:fdd22bb7aa52 324
emilmont 1:fdd22bb7aa52 325 /* acc += a1 * y[n-1] */
emilmont 1:fdd22bb7aa52 326 acc += mult32x64(Yn1, a1);
emilmont 1:fdd22bb7aa52 327
emilmont 1:fdd22bb7aa52 328 /* acc += a2 * y[n-2] */
emilmont 1:fdd22bb7aa52 329 acc += mult32x64(Yn2, a2);
emilmont 1:fdd22bb7aa52 330
emilmont 1:fdd22bb7aa52 331 /* The result is converted to 1.63, Yn2 variable is reused */
emilmont 1:fdd22bb7aa52 332 Yn2 = acc << shift;
emilmont 1:fdd22bb7aa52 333
emilmont 1:fdd22bb7aa52 334 /* Calc lower part of acc */
emilmont 1:fdd22bb7aa52 335 acc_l = acc & 0xffffffff;
emilmont 1:fdd22bb7aa52 336
emilmont 1:fdd22bb7aa52 337 /* Calc upper part of acc */
emilmont 1:fdd22bb7aa52 338 acc_h = (acc >> 32) & 0xffffffff;
emilmont 1:fdd22bb7aa52 339
emilmont 1:fdd22bb7aa52 340 /* Apply shift for lower part of acc and upper part of acc */
emilmont 1:fdd22bb7aa52 341 acc_h = (uint32_t) acc_l >> lShift | acc_h << uShift;
emilmont 1:fdd22bb7aa52 342
emilmont 1:fdd22bb7aa52 343 /* Store the output in the destination buffer in 1.31 format. */
emilmont 1:fdd22bb7aa52 344 *(pOut + 2u) = acc_h;
emilmont 1:fdd22bb7aa52 345
emilmont 1:fdd22bb7aa52 346 /* Read the fourth input into Xn, to reuse the value */
emilmont 1:fdd22bb7aa52 347 Xn = *pIn++;
emilmont 1:fdd22bb7aa52 348
emilmont 1:fdd22bb7aa52 349 /* acc = b0 * x[n] + b1 * x[n-1] + b2 * x[n-2] + a1 * y[n-1] + a2 * y[n-2] */
emilmont 1:fdd22bb7aa52 350 /* acc = b0 * x[n] */
emilmont 1:fdd22bb7aa52 351 acc = (q63_t) Xn *b0;
emilmont 1:fdd22bb7aa52 352
emilmont 1:fdd22bb7aa52 353 /* acc += b1 * x[n-1] */
emilmont 1:fdd22bb7aa52 354 acc += (q63_t) Xn1 *b1;
emilmont 1:fdd22bb7aa52 355
emilmont 1:fdd22bb7aa52 356 /* acc += b[2] * x[n-2] */
emilmont 1:fdd22bb7aa52 357 acc += (q63_t) Xn2 *b2;
emilmont 1:fdd22bb7aa52 358
emilmont 1:fdd22bb7aa52 359 /* acc += a1 * y[n-1] */
emilmont 1:fdd22bb7aa52 360 acc += mult32x64(Yn2, a1);
emilmont 1:fdd22bb7aa52 361
emilmont 1:fdd22bb7aa52 362 /* acc += a2 * y[n-2] */
emilmont 1:fdd22bb7aa52 363 acc += mult32x64(Yn1, a2);
emilmont 1:fdd22bb7aa52 364
emilmont 1:fdd22bb7aa52 365 /* The result is converted to 1.63, Yn1 variable is reused */
emilmont 1:fdd22bb7aa52 366 Yn1 = acc << shift;
emilmont 1:fdd22bb7aa52 367
emilmont 1:fdd22bb7aa52 368 /* Calc lower part of acc */
emilmont 1:fdd22bb7aa52 369 acc_l = acc & 0xffffffff;
emilmont 1:fdd22bb7aa52 370
emilmont 1:fdd22bb7aa52 371 /* Calc upper part of acc */
emilmont 1:fdd22bb7aa52 372 acc_h = (acc >> 32) & 0xffffffff;
emilmont 1:fdd22bb7aa52 373
emilmont 1:fdd22bb7aa52 374 /* Apply shift for lower part of acc and upper part of acc */
emilmont 1:fdd22bb7aa52 375 acc_h = (uint32_t) acc_l >> lShift | acc_h << uShift;
emilmont 1:fdd22bb7aa52 376
emilmont 1:fdd22bb7aa52 377 /* Store the output in the destination buffer in 1.31 format. */
emilmont 1:fdd22bb7aa52 378 *(pOut + 3u) = acc_h;
emilmont 1:fdd22bb7aa52 379
emilmont 1:fdd22bb7aa52 380 /* Every time after the output is computed state should be updated. */
emilmont 1:fdd22bb7aa52 381 /* The states should be updated as: */
emilmont 1:fdd22bb7aa52 382 /* Xn2 = Xn1 */
emilmont 1:fdd22bb7aa52 383 /* Xn1 = Xn */
emilmont 1:fdd22bb7aa52 384 /* Yn2 = Yn1 */
emilmont 1:fdd22bb7aa52 385 /* Yn1 = acc */
emilmont 1:fdd22bb7aa52 386 Xn2 = Xn1;
emilmont 1:fdd22bb7aa52 387 Xn1 = Xn;
emilmont 1:fdd22bb7aa52 388
emilmont 1:fdd22bb7aa52 389 /* update output pointer */
emilmont 1:fdd22bb7aa52 390 pOut += 4u;
emilmont 1:fdd22bb7aa52 391
emilmont 1:fdd22bb7aa52 392 /* decrement the loop counter */
emilmont 1:fdd22bb7aa52 393 sample--;
emilmont 1:fdd22bb7aa52 394 }
emilmont 1:fdd22bb7aa52 395
emilmont 1:fdd22bb7aa52 396 /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
emilmont 1:fdd22bb7aa52 397 ** No loop unrolling is used. */
emilmont 1:fdd22bb7aa52 398 sample = (blockSize & 0x3u);
emilmont 1:fdd22bb7aa52 399
emilmont 1:fdd22bb7aa52 400 while(sample > 0u)
emilmont 1:fdd22bb7aa52 401 {
emilmont 1:fdd22bb7aa52 402 /* Read the input */
emilmont 1:fdd22bb7aa52 403 Xn = *pIn++;
emilmont 1:fdd22bb7aa52 404
emilmont 1:fdd22bb7aa52 405 /* acc = b0 * x[n] + b1 * x[n-1] + b2 * x[n-2] + a1 * y[n-1] + a2 * y[n-2] */
emilmont 1:fdd22bb7aa52 406
emilmont 1:fdd22bb7aa52 407 /* acc = b0 * x[n] */
emilmont 1:fdd22bb7aa52 408 acc = (q63_t) Xn *b0;
emilmont 1:fdd22bb7aa52 409 /* acc += b1 * x[n-1] */
emilmont 1:fdd22bb7aa52 410 acc += (q63_t) Xn1 *b1;
emilmont 1:fdd22bb7aa52 411 /* acc += b[2] * x[n-2] */
emilmont 1:fdd22bb7aa52 412 acc += (q63_t) Xn2 *b2;
emilmont 1:fdd22bb7aa52 413 /* acc += a1 * y[n-1] */
emilmont 1:fdd22bb7aa52 414 acc += mult32x64(Yn1, a1);
emilmont 1:fdd22bb7aa52 415 /* acc += a2 * y[n-2] */
emilmont 1:fdd22bb7aa52 416 acc += mult32x64(Yn2, a2);
emilmont 1:fdd22bb7aa52 417
emilmont 1:fdd22bb7aa52 418 /* Every time after the output is computed state should be updated. */
emilmont 1:fdd22bb7aa52 419 /* The states should be updated as: */
emilmont 1:fdd22bb7aa52 420 /* Xn2 = Xn1 */
emilmont 1:fdd22bb7aa52 421 /* Xn1 = Xn */
emilmont 1:fdd22bb7aa52 422 /* Yn2 = Yn1 */
emilmont 1:fdd22bb7aa52 423 /* Yn1 = acc */
emilmont 1:fdd22bb7aa52 424 Xn2 = Xn1;
emilmont 1:fdd22bb7aa52 425 Xn1 = Xn;
emilmont 1:fdd22bb7aa52 426 Yn2 = Yn1;
emilmont 1:fdd22bb7aa52 427 /* The result is converted to 1.63, Yn1 variable is reused */
emilmont 1:fdd22bb7aa52 428 Yn1 = acc << shift;
emilmont 1:fdd22bb7aa52 429
emilmont 1:fdd22bb7aa52 430 /* Calc lower part of acc */
emilmont 1:fdd22bb7aa52 431 acc_l = acc & 0xffffffff;
emilmont 1:fdd22bb7aa52 432
emilmont 1:fdd22bb7aa52 433 /* Calc upper part of acc */
emilmont 1:fdd22bb7aa52 434 acc_h = (acc >> 32) & 0xffffffff;
emilmont 1:fdd22bb7aa52 435
emilmont 1:fdd22bb7aa52 436 /* Apply shift for lower part of acc and upper part of acc */
emilmont 1:fdd22bb7aa52 437 acc_h = (uint32_t) acc_l >> lShift | acc_h << uShift;
emilmont 1:fdd22bb7aa52 438
emilmont 1:fdd22bb7aa52 439 /* Store the output in the destination buffer in 1.31 format. */
emilmont 1:fdd22bb7aa52 440 *pOut++ = acc_h;
mbed_official 5:3762170b6d4d 441 /* Yn1 = acc << shift; */
emilmont 1:fdd22bb7aa52 442
emilmont 1:fdd22bb7aa52 443 /* Store the output in the destination buffer in 1.31 format. */
mbed_official 5:3762170b6d4d 444 /* *pOut++ = (q31_t) (acc >> (32 - shift)); */
emilmont 1:fdd22bb7aa52 445
emilmont 1:fdd22bb7aa52 446 /* decrement the loop counter */
emilmont 1:fdd22bb7aa52 447 sample--;
emilmont 1:fdd22bb7aa52 448 }
emilmont 1:fdd22bb7aa52 449
emilmont 1:fdd22bb7aa52 450 /* The first stage output is given as input to the second stage. */
emilmont 1:fdd22bb7aa52 451 pIn = pDst;
emilmont 1:fdd22bb7aa52 452
emilmont 1:fdd22bb7aa52 453 /* Reset to destination buffer working pointer */
emilmont 1:fdd22bb7aa52 454 pOut = pDst;
emilmont 1:fdd22bb7aa52 455
emilmont 1:fdd22bb7aa52 456 /* Store the updated state variables back into the pState array */
emilmont 1:fdd22bb7aa52 457 /* Store the updated state variables back into the pState array */
emilmont 1:fdd22bb7aa52 458 *pState++ = (q63_t) Xn1;
emilmont 1:fdd22bb7aa52 459 *pState++ = (q63_t) Xn2;
emilmont 1:fdd22bb7aa52 460 *pState++ = Yn1;
emilmont 1:fdd22bb7aa52 461 *pState++ = Yn2;
emilmont 1:fdd22bb7aa52 462
emilmont 1:fdd22bb7aa52 463 } while(--stage);
emilmont 1:fdd22bb7aa52 464
emilmont 1:fdd22bb7aa52 465 #else
emilmont 1:fdd22bb7aa52 466
emilmont 1:fdd22bb7aa52 467 /* Run the below code for Cortex-M0 */
emilmont 1:fdd22bb7aa52 468
emilmont 1:fdd22bb7aa52 469 do
emilmont 1:fdd22bb7aa52 470 {
emilmont 1:fdd22bb7aa52 471 /* Reading the coefficients */
emilmont 1:fdd22bb7aa52 472 b0 = *pCoeffs++;
emilmont 1:fdd22bb7aa52 473 b1 = *pCoeffs++;
emilmont 1:fdd22bb7aa52 474 b2 = *pCoeffs++;
emilmont 1:fdd22bb7aa52 475 a1 = *pCoeffs++;
emilmont 1:fdd22bb7aa52 476 a2 = *pCoeffs++;
emilmont 1:fdd22bb7aa52 477
emilmont 1:fdd22bb7aa52 478 /* Reading the state values */
emilmont 1:fdd22bb7aa52 479 Xn1 = pState[0];
emilmont 1:fdd22bb7aa52 480 Xn2 = pState[1];
emilmont 1:fdd22bb7aa52 481 Yn1 = pState[2];
emilmont 1:fdd22bb7aa52 482 Yn2 = pState[3];
emilmont 1:fdd22bb7aa52 483
emilmont 1:fdd22bb7aa52 484 /* The variable acc hold output value that is being computed and
emilmont 1:fdd22bb7aa52 485 * stored in the destination buffer
emilmont 1:fdd22bb7aa52 486 * acc = b0 * x[n] + b1 * x[n-1] + b2 * x[n-2] + a1 * y[n-1] + a2 * y[n-2]
emilmont 1:fdd22bb7aa52 487 */
emilmont 1:fdd22bb7aa52 488
emilmont 1:fdd22bb7aa52 489 sample = blockSize;
emilmont 1:fdd22bb7aa52 490
emilmont 1:fdd22bb7aa52 491 while(sample > 0u)
emilmont 1:fdd22bb7aa52 492 {
emilmont 1:fdd22bb7aa52 493 /* Read the input */
emilmont 1:fdd22bb7aa52 494 Xn = *pIn++;
emilmont 1:fdd22bb7aa52 495
emilmont 1:fdd22bb7aa52 496 /* acc = b0 * x[n] + b1 * x[n-1] + b2 * x[n-2] + a1 * y[n-1] + a2 * y[n-2] */
emilmont 1:fdd22bb7aa52 497 /* acc = b0 * x[n] */
emilmont 1:fdd22bb7aa52 498 acc = (q63_t) Xn *b0;
emilmont 1:fdd22bb7aa52 499 /* acc += b1 * x[n-1] */
emilmont 1:fdd22bb7aa52 500 acc += (q63_t) Xn1 *b1;
emilmont 1:fdd22bb7aa52 501 /* acc += b[2] * x[n-2] */
emilmont 1:fdd22bb7aa52 502 acc += (q63_t) Xn2 *b2;
emilmont 1:fdd22bb7aa52 503 /* acc += a1 * y[n-1] */
emilmont 1:fdd22bb7aa52 504 acc += mult32x64(Yn1, a1);
emilmont 1:fdd22bb7aa52 505 /* acc += a2 * y[n-2] */
emilmont 1:fdd22bb7aa52 506 acc += mult32x64(Yn2, a2);
emilmont 1:fdd22bb7aa52 507
emilmont 1:fdd22bb7aa52 508 /* Every time after the output is computed state should be updated. */
emilmont 1:fdd22bb7aa52 509 /* The states should be updated as: */
emilmont 1:fdd22bb7aa52 510 /* Xn2 = Xn1 */
emilmont 1:fdd22bb7aa52 511 /* Xn1 = Xn */
emilmont 1:fdd22bb7aa52 512 /* Yn2 = Yn1 */
emilmont 1:fdd22bb7aa52 513 /* Yn1 = acc */
emilmont 1:fdd22bb7aa52 514 Xn2 = Xn1;
emilmont 1:fdd22bb7aa52 515 Xn1 = Xn;
emilmont 1:fdd22bb7aa52 516 Yn2 = Yn1;
emilmont 1:fdd22bb7aa52 517
emilmont 1:fdd22bb7aa52 518 /* The result is converted to 1.63, Yn1 variable is reused */
emilmont 1:fdd22bb7aa52 519 Yn1 = acc << shift;
emilmont 1:fdd22bb7aa52 520
emilmont 1:fdd22bb7aa52 521 /* Calc lower part of acc */
emilmont 1:fdd22bb7aa52 522 acc_l = acc & 0xffffffff;
emilmont 1:fdd22bb7aa52 523
emilmont 1:fdd22bb7aa52 524 /* Calc upper part of acc */
emilmont 1:fdd22bb7aa52 525 acc_h = (acc >> 32) & 0xffffffff;
emilmont 1:fdd22bb7aa52 526
emilmont 1:fdd22bb7aa52 527 /* Apply shift for lower part of acc and upper part of acc */
emilmont 1:fdd22bb7aa52 528 acc_h = (uint32_t) acc_l >> lShift | acc_h << uShift;
emilmont 1:fdd22bb7aa52 529
emilmont 1:fdd22bb7aa52 530 /* Store the output in the destination buffer in 1.31 format. */
emilmont 1:fdd22bb7aa52 531 *pOut++ = acc_h;
emilmont 1:fdd22bb7aa52 532
mbed_official 5:3762170b6d4d 533 /* Yn1 = acc << shift; */
emilmont 1:fdd22bb7aa52 534
emilmont 1:fdd22bb7aa52 535 /* Store the output in the destination buffer in 1.31 format. */
mbed_official 5:3762170b6d4d 536 /* *pOut++ = (q31_t) (acc >> (32 - shift)); */
emilmont 1:fdd22bb7aa52 537
emilmont 1:fdd22bb7aa52 538 /* decrement the loop counter */
emilmont 1:fdd22bb7aa52 539 sample--;
emilmont 1:fdd22bb7aa52 540 }
emilmont 1:fdd22bb7aa52 541
emilmont 1:fdd22bb7aa52 542 /* The first stage output is given as input to the second stage. */
emilmont 1:fdd22bb7aa52 543 pIn = pDst;
emilmont 1:fdd22bb7aa52 544
emilmont 1:fdd22bb7aa52 545 /* Reset to destination buffer working pointer */
emilmont 1:fdd22bb7aa52 546 pOut = pDst;
emilmont 1:fdd22bb7aa52 547
emilmont 1:fdd22bb7aa52 548 /* Store the updated state variables back into the pState array */
emilmont 1:fdd22bb7aa52 549 *pState++ = (q63_t) Xn1;
emilmont 1:fdd22bb7aa52 550 *pState++ = (q63_t) Xn2;
emilmont 1:fdd22bb7aa52 551 *pState++ = Yn1;
emilmont 1:fdd22bb7aa52 552 *pState++ = Yn2;
emilmont 1:fdd22bb7aa52 553
emilmont 1:fdd22bb7aa52 554 } while(--stage);
emilmont 1:fdd22bb7aa52 555
mbed_official 3:7a284390b0ce 556 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
emilmont 1:fdd22bb7aa52 557 }
emilmont 1:fdd22bb7aa52 558
emilmont 1:fdd22bb7aa52 559 /**
emilmont 1:fdd22bb7aa52 560 * @} end of BiquadCascadeDF1_32x64 group
emilmont 1:fdd22bb7aa52 561 */