Important changes to repositories hosted on mbed.com
Mbed hosted mercurial repositories are deprecated and are due to be permanently deleted in July 2026.
To keep a copy of this software download the repository Zip archive or clone locally using Mercurial.
It is also possible to export all your personal repositories from the account settings page.
Fork of mbed-dsp by
cmsis_dsp/FilteringFunctions/arm_fir_decimate_q15.c
- Committer:
- emilmont
- Date:
- 2013-05-30
- Revision:
- 2:da51fb522205
- Parent:
- 1:fdd22bb7aa52
- Child:
- 3:7a284390b0ce
File content as of revision 2:da51fb522205:
/* ----------------------------------------------------------------------
* Copyright (C) 2010 ARM Limited. All rights reserved.
*
* $Date: 15. February 2012
* $Revision: V1.1.0
*
* Project: CMSIS DSP Library
* Title: arm_fir_decimate_q15.c
*
* Description: Q15 FIR Decimator.
*
* Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
*
* Version 1.1.0 2012/02/15
* Updated with more optimizations, bug fixes and minor API changes.
*
* Version 1.0.10 2011/7/15
* Big Endian support added and Merged M0 and M3/M4 Source code.
*
* Version 1.0.3 2010/11/29
* Re-organized the CMSIS folders and updated documentation.
*
* Version 1.0.2 2010/11/11
* Documentation updated.
*
* Version 1.0.1 2010/10/05
* Production release and review comments incorporated.
*
* Version 1.0.0 2010/09/20
* Production release and review comments incorporated
*
* Version 0.0.7 2010/06/10
* Misra-C changes done
* -------------------------------------------------------------------- */
#include "arm_math.h"
/**
* @ingroup groupFilters
*/
/**
* @addtogroup FIR_decimate
* @{
*/
/**
* @brief Processing function for the Q15 FIR decimator.
* @param[in] *S points to an instance of the Q15 FIR decimator structure.
* @param[in] *pSrc points to the block of input data.
* @param[out] *pDst points to the location where the output result is written.
* @param[in] blockSize number of input samples to process per call.
* @return none.
*
* <b>Scaling and Overflow Behavior:</b>
* \par
* The function is implemented using a 64-bit internal accumulator.
* Both coefficients and state variables are represented in 1.15 format and multiplications yield a 2.30 result.
* The 2.30 intermediate results are accumulated in a 64-bit accumulator in 34.30 format.
* There is no risk of internal overflow with this approach and the full precision of intermediate multiplications is preserved.
* After all additions have been performed, the accumulator is truncated to 34.15 format by discarding low 15 bits.
* Lastly, the accumulator is saturated to yield a result in 1.15 format.
*
* \par
* Refer to the function <code>arm_fir_decimate_fast_q15()</code> for a faster but less precise implementation of this function for Cortex-M3 and Cortex-M4.
*/
#ifndef ARM_MATH_CM0
#ifndef UNALIGNED_SUPPORT_DISABLE
void arm_fir_decimate_q15(
const arm_fir_decimate_instance_q15 * S,
q15_t * pSrc,
q15_t * pDst,
uint32_t blockSize)
{
q15_t *pState = S->pState; /* State pointer */
q15_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
q15_t *pStateCurnt; /* Points to the current sample of the state */
q15_t *px; /* Temporary pointer for state buffer */
q15_t *pb; /* Temporary pointer coefficient buffer */
q31_t x0, x1, c0, c1; /* Temporary variables to hold state and coefficient values */
q63_t sum0; /* Accumulators */
q63_t acc0, acc1;
q15_t *px0, *px1;
uint32_t blkCntN3;
uint32_t numTaps = S->numTaps; /* Number of taps */
uint32_t i, blkCnt, tapCnt, outBlockSize = blockSize / S->M; /* Loop counters */
/* S->pState buffer contains previous frame (numTaps - 1) samples */
/* pStateCurnt points to the location where the new input data should be written */
pStateCurnt = S->pState + (numTaps - 1u);
/* Total number of output samples to be computed */
blkCnt = outBlockSize / 2;
blkCntN3 = outBlockSize - (2 * blkCnt);
while(blkCnt > 0u)
{
/* Copy decimation factor number of new input samples into the state buffer */
i = 2 * S->M;
do
{
*pStateCurnt++ = *pSrc++;
} while(--i);
/* Set accumulator to zero */
acc0 = 0;
acc1 = 0;
/* Initialize state pointer */
px0 = pState;
px1 = pState + S->M;
/* Initialize coeff pointer */
pb = pCoeffs;
/* Loop unrolling. Process 4 taps at a time. */
tapCnt = numTaps >> 2;
/* Loop over the number of taps. Unroll by a factor of 4.
** Repeat until we've computed numTaps-4 coefficients. */
while(tapCnt > 0u)
{
/* Read the Read b[numTaps-1] and b[numTaps-2] coefficients */
c0 = *__SIMD32(pb)++;
/* Read x[n-numTaps-1] and x[n-numTaps-2]sample */
x0 = *__SIMD32(px0)++;
x1 = *__SIMD32(px1)++;
/* Perform the multiply-accumulate */
acc0 = __SMLALD(x0, c0, acc0);
acc1 = __SMLALD(x1, c0, acc1);
/* Read the b[numTaps-3] and b[numTaps-4] coefficient */
c0 = *__SIMD32(pb)++;
/* Read x[n-numTaps-2] and x[n-numTaps-3] sample */
x0 = *__SIMD32(px0)++;
x1 = *__SIMD32(px1)++;
/* Perform the multiply-accumulate */
acc0 = __SMLALD(x0, c0, acc0);
acc1 = __SMLALD(x1, c0, acc1);
/* Decrement the loop counter */
tapCnt--;
}
/* If the filter length is not a multiple of 4, compute the remaining filter taps */
tapCnt = numTaps % 0x4u;
while(tapCnt > 0u)
{
/* Read coefficients */
c0 = *pb++;
/* Fetch 1 state variable */
x0 = *px0++;
x1 = *px1++;
/* Perform the multiply-accumulate */
acc0 = __SMLALD(x0, c0, acc0);
acc1 = __SMLALD(x1, c0, acc1);
/* Decrement the loop counter */
tapCnt--;
}
/* Advance the state pointer by the decimation factor
* to process the next group of decimation factor number samples */
pState = pState + S->M * 2;
/* Store filter output, smlad returns the values in 2.14 format */
/* so downsacle by 15 to get output in 1.15 */
*pDst++ = (q15_t) (__SSAT((acc0 >> 15), 16));
*pDst++ = (q15_t) (__SSAT((acc1 >> 15), 16));
/* Decrement the loop counter */
blkCnt--;
}
while(blkCntN3 > 0u)
{
/* Copy decimation factor number of new input samples into the state buffer */
i = S->M;
do
{
*pStateCurnt++ = *pSrc++;
} while(--i);
/*Set sum to zero */
sum0 = 0;
/* Initialize state pointer */
px = pState;
/* Initialize coeff pointer */
pb = pCoeffs;
/* Loop unrolling. Process 4 taps at a time. */
tapCnt = numTaps >> 2;
/* Loop over the number of taps. Unroll by a factor of 4.
** Repeat until we've computed numTaps-4 coefficients. */
while(tapCnt > 0u)
{
/* Read the Read b[numTaps-1] and b[numTaps-2] coefficients */
c0 = *__SIMD32(pb)++;
/* Read x[n-numTaps-1] and x[n-numTaps-2]sample */
x0 = *__SIMD32(px)++;
/* Read the b[numTaps-3] and b[numTaps-4] coefficient */
c1 = *__SIMD32(pb)++;
/* Perform the multiply-accumulate */
sum0 = __SMLALD(x0, c0, sum0);
/* Read x[n-numTaps-2] and x[n-numTaps-3] sample */
x0 = *__SIMD32(px)++;
/* Perform the multiply-accumulate */
sum0 = __SMLALD(x0, c1, sum0);
/* Decrement the loop counter */
tapCnt--;
}
/* If the filter length is not a multiple of 4, compute the remaining filter taps */
tapCnt = numTaps % 0x4u;
while(tapCnt > 0u)
{
/* Read coefficients */
c0 = *pb++;
/* Fetch 1 state variable */
x0 = *px++;
/* Perform the multiply-accumulate */
sum0 = __SMLALD(x0, c0, sum0);
/* Decrement the loop counter */
tapCnt--;
}
/* Advance the state pointer by the decimation factor
* to process the next group of decimation factor number samples */
pState = pState + S->M;
/* Store filter output, smlad returns the values in 2.14 format */
/* so downsacle by 15 to get output in 1.15 */
*pDst++ = (q15_t) (__SSAT((sum0 >> 15), 16));
/* Decrement the loop counter */
blkCntN3--;
}
/* Processing is complete.
** Now copy the last numTaps - 1 samples to the satrt of the state buffer.
** This prepares the state buffer for the next function call. */
/* Points to the start of the state buffer */
pStateCurnt = S->pState;
i = (numTaps - 1u) >> 2u;
/* copy data */
while(i > 0u)
{
*__SIMD32(pStateCurnt)++ = *__SIMD32(pState)++;
*__SIMD32(pStateCurnt)++ = *__SIMD32(pState)++;
/* Decrement the loop counter */
i--;
}
i = (numTaps - 1u) % 0x04u;
/* copy data */
while(i > 0u)
{
*pStateCurnt++ = *pState++;
/* Decrement the loop counter */
i--;
}
}
#else
void arm_fir_decimate_q15(
const arm_fir_decimate_instance_q15 * S,
q15_t * pSrc,
q15_t * pDst,
uint32_t blockSize)
{
q15_t *pState = S->pState; /* State pointer */
q15_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
q15_t *pStateCurnt; /* Points to the current sample of the state */
q15_t *px; /* Temporary pointer for state buffer */
q15_t *pb; /* Temporary pointer coefficient buffer */
q15_t x0, x1, c0; /* Temporary variables to hold state and coefficient values */
q63_t sum0; /* Accumulators */
q63_t acc0, acc1;
q15_t *px0, *px1;
uint32_t blkCntN3;
uint32_t numTaps = S->numTaps; /* Number of taps */
uint32_t i, blkCnt, tapCnt, outBlockSize = blockSize / S->M; /* Loop counters */
/* S->pState buffer contains previous frame (numTaps - 1) samples */
/* pStateCurnt points to the location where the new input data should be written */
pStateCurnt = S->pState + (numTaps - 1u);
/* Total number of output samples to be computed */
blkCnt = outBlockSize / 2;
blkCntN3 = outBlockSize - (2 * blkCnt);
while(blkCnt > 0u)
{
/* Copy decimation factor number of new input samples into the state buffer */
i = 2 * S->M;
do
{
*pStateCurnt++ = *pSrc++;
} while(--i);
/* Set accumulator to zero */
acc0 = 0;
acc1 = 0;
/* Initialize state pointer */
px0 = pState;
px1 = pState + S->M;
/* Initialize coeff pointer */
pb = pCoeffs;
/* Loop unrolling. Process 4 taps at a time. */
tapCnt = numTaps >> 2;
/* Loop over the number of taps. Unroll by a factor of 4.
** Repeat until we've computed numTaps-4 coefficients. */
while(tapCnt > 0u)
{
/* Read the Read b[numTaps-1] coefficients */
c0 = *pb++;
/* Read x[n-numTaps-1] for sample 0 and for sample 1 */
x0 = *px0++;
x1 = *px1++;
/* Perform the multiply-accumulate */
acc0 += x0 * c0;
acc1 += x1 * c0;
/* Read the b[numTaps-2] coefficient */
c0 = *pb++;
/* Read x[n-numTaps-2] for sample 0 and sample 1 */
x0 = *px0++;
x1 = *px1++;
/* Perform the multiply-accumulate */
acc0 += x0 * c0;
acc1 += x1 * c0;
/* Read the b[numTaps-3] coefficients */
c0 = *pb++;
/* Read x[n-numTaps-3] for sample 0 and sample 1 */
x0 = *px0++;
x1 = *px1++;
/* Perform the multiply-accumulate */
acc0 += x0 * c0;
acc1 += x1 * c0;
/* Read the b[numTaps-4] coefficient */
c0 = *pb++;
/* Read x[n-numTaps-4] for sample 0 and sample 1 */
x0 = *px0++;
x1 = *px1++;
/* Perform the multiply-accumulate */
acc0 += x0 * c0;
acc1 += x1 * c0;
/* Decrement the loop counter */
tapCnt--;
}
/* If the filter length is not a multiple of 4, compute the remaining filter taps */
tapCnt = numTaps % 0x4u;
while(tapCnt > 0u)
{
/* Read coefficients */
c0 = *pb++;
/* Fetch 1 state variable */
x0 = *px0++;
x1 = *px1++;
/* Perform the multiply-accumulate */
acc0 += x0 * c0;
acc1 += x1 * c0;
/* Decrement the loop counter */
tapCnt--;
}
/* Advance the state pointer by the decimation factor
* to process the next group of decimation factor number samples */
pState = pState + S->M * 2;
/* Store filter output, smlad returns the values in 2.14 format */
/* so downsacle by 15 to get output in 1.15 */
*pDst++ = (q15_t) (__SSAT((acc0 >> 15), 16));
*pDst++ = (q15_t) (__SSAT((acc1 >> 15), 16));
/* Decrement the loop counter */
blkCnt--;
}
while(blkCntN3 > 0u)
{
/* Copy decimation factor number of new input samples into the state buffer */
i = S->M;
do
{
*pStateCurnt++ = *pSrc++;
} while(--i);
/*Set sum to zero */
sum0 = 0;
/* Initialize state pointer */
px = pState;
/* Initialize coeff pointer */
pb = pCoeffs;
/* Loop unrolling. Process 4 taps at a time. */
tapCnt = numTaps >> 2;
/* Loop over the number of taps. Unroll by a factor of 4.
** Repeat until we've computed numTaps-4 coefficients. */
while(tapCnt > 0u)
{
/* Read the Read b[numTaps-1] coefficients */
c0 = *pb++;
/* Read x[n-numTaps-1] and sample */
x0 = *px++;
/* Perform the multiply-accumulate */
sum0 += x0 * c0;
/* Read the b[numTaps-2] coefficient */
c0 = *pb++;
/* Read x[n-numTaps-2] and sample */
x0 = *px++;
/* Perform the multiply-accumulate */
sum0 += x0 * c0;
/* Read the b[numTaps-3] coefficients */
c0 = *pb++;
/* Read x[n-numTaps-3] sample */
x0 = *px++;
/* Perform the multiply-accumulate */
sum0 += x0 * c0;
/* Read the b[numTaps-4] coefficient */
c0 = *pb++;
/* Read x[n-numTaps-4] sample */
x0 = *px++;
/* Perform the multiply-accumulate */
sum0 += x0 * c0;
/* Decrement the loop counter */
tapCnt--;
}
/* If the filter length is not a multiple of 4, compute the remaining filter taps */
tapCnt = numTaps % 0x4u;
while(tapCnt > 0u)
{
/* Read coefficients */
c0 = *pb++;
/* Fetch 1 state variable */
x0 = *px++;
/* Perform the multiply-accumulate */
sum0 += x0 * c0;
/* Decrement the loop counter */
tapCnt--;
}
/* Advance the state pointer by the decimation factor
* to process the next group of decimation factor number samples */
pState = pState + S->M;
/* Store filter output, smlad returns the values in 2.14 format */
/* so downsacle by 15 to get output in 1.15 */
*pDst++ = (q15_t) (__SSAT((sum0 >> 15), 16));
/* Decrement the loop counter */
blkCntN3--;
}
/* Processing is complete.
** Now copy the last numTaps - 1 samples to the satrt of the state buffer.
** This prepares the state buffer for the next function call. */
/* Points to the start of the state buffer */
pStateCurnt = S->pState;
i = (numTaps - 1u) >> 2u;
/* copy data */
while(i > 0u)
{
*pStateCurnt++ = *pState++;
*pStateCurnt++ = *pState++;
*pStateCurnt++ = *pState++;
*pStateCurnt++ = *pState++;
/* Decrement the loop counter */
i--;
}
i = (numTaps - 1u) % 0x04u;
/* copy data */
while(i > 0u)
{
*pStateCurnt++ = *pState++;
/* Decrement the loop counter */
i--;
}
}
#endif /* #ifndef UNALIGNED_SUPPORT_DISABLE */
#else
void arm_fir_decimate_q15(
const arm_fir_decimate_instance_q15 * S,
q15_t * pSrc,
q15_t * pDst,
uint32_t blockSize)
{
q15_t *pState = S->pState; /* State pointer */
q15_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
q15_t *pStateCurnt; /* Points to the current sample of the state */
q15_t *px; /* Temporary pointer for state buffer */
q15_t *pb; /* Temporary pointer coefficient buffer */
q31_t x0, c0; /* Temporary variables to hold state and coefficient values */
q63_t sum0; /* Accumulators */
uint32_t numTaps = S->numTaps; /* Number of taps */
uint32_t i, blkCnt, tapCnt, outBlockSize = blockSize / S->M; /* Loop counters */
/* Run the below code for Cortex-M0 */
/* S->pState buffer contains previous frame (numTaps - 1) samples */
/* pStateCurnt points to the location where the new input data should be written */
pStateCurnt = S->pState + (numTaps - 1u);
/* Total number of output samples to be computed */
blkCnt = outBlockSize;
while(blkCnt > 0u)
{
/* Copy decimation factor number of new input samples into the state buffer */
i = S->M;
do
{
*pStateCurnt++ = *pSrc++;
} while(--i);
/*Set sum to zero */
sum0 = 0;
/* Initialize state pointer */
px = pState;
/* Initialize coeff pointer */
pb = pCoeffs;
tapCnt = numTaps;
while(tapCnt > 0u)
{
/* Read coefficients */
c0 = *pb++;
/* Fetch 1 state variable */
x0 = *px++;
/* Perform the multiply-accumulate */
sum0 += (q31_t) x0 *c0;
/* Decrement the loop counter */
tapCnt--;
}
/* Advance the state pointer by the decimation factor
* to process the next group of decimation factor number samples */
pState = pState + S->M;
/*Store filter output , smlad will return the values in 2.14 format */
/* so downsacle by 15 to get output in 1.15 */
*pDst++ = (q15_t) (__SSAT((sum0 >> 15), 16));
/* Decrement the loop counter */
blkCnt--;
}
/* Processing is complete.
** Now copy the last numTaps - 1 samples to the start of the state buffer.
** This prepares the state buffer for the next function call. */
/* Points to the start of the state buffer */
pStateCurnt = S->pState;
i = numTaps - 1u;
/* copy data */
while(i > 0u)
{
*pStateCurnt++ = *pState++;
/* Decrement the loop counter */
i--;
}
}
#endif /* #ifndef ARM_MATH_CM0 */
/**
* @} end of FIR_decimate group
*/
