]> git.gir.st - tmk_keyboard.git/blob - tmk_core/tool/mbed/mbed-sdk/libraries/dsp/cmsis_dsp/FilteringFunctions/arm_fir_interpolate_q31.c
Merge commit '1fe4406f374291ab2e86e95a97341fd9c475fcb8'
[tmk_keyboard.git] / tmk_core / tool / mbed / mbed-sdk / libraries / dsp / cmsis_dsp / FilteringFunctions / arm_fir_interpolate_q31.c
1 /*-----------------------------------------------------------------------------
2 * Copyright (C) 2010-2013 ARM Limited. All rights reserved.
3 *
4 * $Date: 17. January 2013
5 * $Revision: V1.4.1
6 *
7 * Project: CMSIS DSP Library
8 * Title: arm_fir_interpolate_q31.c
9 *
10 * Description: Q31 FIR interpolation.
11 *
12 * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
13 *
14 * Redistribution and use in source and binary forms, with or without
15 * modification, are permitted provided that the following conditions
16 * are met:
17 * - Redistributions of source code must retain the above copyright
18 * notice, this list of conditions and the following disclaimer.
19 * - Redistributions in binary form must reproduce the above copyright
20 * notice, this list of conditions and the following disclaimer in
21 * the documentation and/or other materials provided with the
22 * distribution.
23 * - Neither the name of ARM LIMITED nor the names of its contributors
24 * may be used to endorse or promote products derived from this
25 * software without specific prior written permission.
26 *
27 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
28 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
29 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
30 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
31 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
32 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
33 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
34 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
35 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
36 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
37 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
38 * POSSIBILITY OF SUCH DAMAGE.
39 * ---------------------------------------------------------------------------*/
40
41 #include "arm_math.h"
42
43 /**
44 * @ingroup groupFilters
45 */
46
47 /**
48 * @addtogroup FIR_Interpolate
49 * @{
50 */
51
52 /**
53 * @brief Processing function for the Q31 FIR interpolator.
54 * @param[in] *S points to an instance of the Q31 FIR interpolator structure.
55 * @param[in] *pSrc points to the block of input data.
56 * @param[out] *pDst points to the block of output data.
57 * @param[in] blockSize number of input samples to process per call.
58 * @return none.
59 *
60 * <b>Scaling and Overflow Behavior:</b>
61 * \par
62 * The function is implemented using an internal 64-bit accumulator.
63 * The accumulator has a 2.62 format and maintains full precision of the intermediate multiplication results but provides only a single guard bit.
64 * Thus, if the accumulator result overflows it wraps around rather than clip.
65 * In order to avoid overflows completely the input signal must be scaled down by <code>1/(numTaps/L)</code>.
66 * since <code>numTaps/L</code> additions occur per output sample.
67 * After all multiply-accumulates are performed, the 2.62 accumulator is truncated to 1.32 format and then saturated to 1.31 format.
68 */
69
70 #ifndef ARM_MATH_CM0_FAMILY
71
72 /* Run the below code for Cortex-M4 and Cortex-M3 */
73
74 void arm_fir_interpolate_q31(
75 const arm_fir_interpolate_instance_q31 * S,
76 q31_t * pSrc,
77 q31_t * pDst,
78 uint32_t blockSize)
79 {
80 q31_t *pState = S->pState; /* State pointer */
81 q31_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
82 q31_t *pStateCurnt; /* Points to the current sample of the state */
83 q31_t *ptr1, *ptr2; /* Temporary pointers for state and coefficient buffers */
84 q63_t sum0; /* Accumulators */
85 q31_t x0, c0; /* Temporary variables to hold state and coefficient values */
86 uint32_t i, blkCnt, j; /* Loop counters */
87 uint16_t phaseLen = S->phaseLength, tapCnt; /* Length of each polyphase filter component */
88
89 uint32_t blkCntN2;
90 q63_t acc0, acc1;
91 q31_t x1;
92
93 /* S->pState buffer contains previous frame (phaseLen - 1) samples */
94 /* pStateCurnt points to the location where the new input data should be written */
95 pStateCurnt = S->pState + ((q31_t) phaseLen - 1);
96
97 /* Initialise blkCnt */
98 blkCnt = blockSize / 2;
99 blkCntN2 = blockSize - (2 * blkCnt);
100
101 /* Samples loop unrolled by 2 */
102 while(blkCnt > 0u)
103 {
104 /* Copy new input sample into the state buffer */
105 *pStateCurnt++ = *pSrc++;
106 *pStateCurnt++ = *pSrc++;
107
108 /* Address modifier index of coefficient buffer */
109 j = 1u;
110
111 /* Loop over the Interpolation factor. */
112 i = (S->L);
113
114 while(i > 0u)
115 {
116 /* Set accumulator to zero */
117 acc0 = 0;
118 acc1 = 0;
119
120 /* Initialize state pointer */
121 ptr1 = pState;
122
123 /* Initialize coefficient pointer */
124 ptr2 = pCoeffs + (S->L - j);
125
126 /* Loop over the polyPhase length. Unroll by a factor of 4.
127 ** Repeat until we've computed numTaps-(4*S->L) coefficients. */
128 tapCnt = phaseLen >> 2u;
129
130 x0 = *(ptr1++);
131
132 while(tapCnt > 0u)
133 {
134
135 /* Read the input sample */
136 x1 = *(ptr1++);
137
138 /* Read the coefficient */
139 c0 = *(ptr2);
140
141 /* Perform the multiply-accumulate */
142 acc0 += (q63_t) x0 *c0;
143 acc1 += (q63_t) x1 *c0;
144
145
146 /* Read the coefficient */
147 c0 = *(ptr2 + S->L);
148
149 /* Read the input sample */
150 x0 = *(ptr1++);
151
152 /* Perform the multiply-accumulate */
153 acc0 += (q63_t) x1 *c0;
154 acc1 += (q63_t) x0 *c0;
155
156
157 /* Read the coefficient */
158 c0 = *(ptr2 + S->L * 2);
159
160 /* Read the input sample */
161 x1 = *(ptr1++);
162
163 /* Perform the multiply-accumulate */
164 acc0 += (q63_t) x0 *c0;
165 acc1 += (q63_t) x1 *c0;
166
167 /* Read the coefficient */
168 c0 = *(ptr2 + S->L * 3);
169
170 /* Read the input sample */
171 x0 = *(ptr1++);
172
173 /* Perform the multiply-accumulate */
174 acc0 += (q63_t) x1 *c0;
175 acc1 += (q63_t) x0 *c0;
176
177
178 /* Upsampling is done by stuffing L-1 zeros between each sample.
179 * So instead of multiplying zeros with coefficients,
180 * Increment the coefficient pointer by interpolation factor times. */
181 ptr2 += 4 * S->L;
182
183 /* Decrement the loop counter */
184 tapCnt--;
185 }
186
187 /* If the polyPhase length is not a multiple of 4, compute the remaining filter taps */
188 tapCnt = phaseLen % 0x4u;
189
190 while(tapCnt > 0u)
191 {
192
193 /* Read the input sample */
194 x1 = *(ptr1++);
195
196 /* Read the coefficient */
197 c0 = *(ptr2);
198
199 /* Perform the multiply-accumulate */
200 acc0 += (q63_t) x0 *c0;
201 acc1 += (q63_t) x1 *c0;
202
203 /* Increment the coefficient pointer by interpolation factor times. */
204 ptr2 += S->L;
205
206 /* update states for next sample processing */
207 x0 = x1;
208
209 /* Decrement the loop counter */
210 tapCnt--;
211 }
212
213 /* The result is in the accumulator, store in the destination buffer. */
214 *pDst = (q31_t) (acc0 >> 31);
215 *(pDst + S->L) = (q31_t) (acc1 >> 31);
216
217
218 pDst++;
219
220 /* Increment the address modifier index of coefficient buffer */
221 j++;
222
223 /* Decrement the loop counter */
224 i--;
225 }
226
227 /* Advance the state pointer by 1
228 * to process the next group of interpolation factor number samples */
229 pState = pState + 2;
230
231 pDst += S->L;
232
233 /* Decrement the loop counter */
234 blkCnt--;
235 }
236
237 /* If the blockSize is not a multiple of 2, compute any remaining output samples here.
238 ** No loop unrolling is used. */
239 blkCnt = blkCntN2;
240
241 /* Loop over the blockSize. */
242 while(blkCnt > 0u)
243 {
244 /* Copy new input sample into the state buffer */
245 *pStateCurnt++ = *pSrc++;
246
247 /* Address modifier index of coefficient buffer */
248 j = 1u;
249
250 /* Loop over the Interpolation factor. */
251 i = S->L;
252 while(i > 0u)
253 {
254 /* Set accumulator to zero */
255 sum0 = 0;
256
257 /* Initialize state pointer */
258 ptr1 = pState;
259
260 /* Initialize coefficient pointer */
261 ptr2 = pCoeffs + (S->L - j);
262
263 /* Loop over the polyPhase length. Unroll by a factor of 4.
264 ** Repeat until we've computed numTaps-(4*S->L) coefficients. */
265 tapCnt = phaseLen >> 2;
266 while(tapCnt > 0u)
267 {
268
269 /* Read the coefficient */
270 c0 = *(ptr2);
271
272 /* Upsampling is done by stuffing L-1 zeros between each sample.
273 * So instead of multiplying zeros with coefficients,
274 * Increment the coefficient pointer by interpolation factor times. */
275 ptr2 += S->L;
276
277 /* Read the input sample */
278 x0 = *(ptr1++);
279
280 /* Perform the multiply-accumulate */
281 sum0 += (q63_t) x0 *c0;
282
283 /* Read the coefficient */
284 c0 = *(ptr2);
285
286 /* Increment the coefficient pointer by interpolation factor times. */
287 ptr2 += S->L;
288
289 /* Read the input sample */
290 x0 = *(ptr1++);
291
292 /* Perform the multiply-accumulate */
293 sum0 += (q63_t) x0 *c0;
294
295 /* Read the coefficient */
296 c0 = *(ptr2);
297
298 /* Increment the coefficient pointer by interpolation factor times. */
299 ptr2 += S->L;
300
301 /* Read the input sample */
302 x0 = *(ptr1++);
303
304 /* Perform the multiply-accumulate */
305 sum0 += (q63_t) x0 *c0;
306
307 /* Read the coefficient */
308 c0 = *(ptr2);
309
310 /* Increment the coefficient pointer by interpolation factor times. */
311 ptr2 += S->L;
312
313 /* Read the input sample */
314 x0 = *(ptr1++);
315
316 /* Perform the multiply-accumulate */
317 sum0 += (q63_t) x0 *c0;
318
319 /* Decrement the loop counter */
320 tapCnt--;
321 }
322
323 /* If the polyPhase length is not a multiple of 4, compute the remaining filter taps */
324 tapCnt = phaseLen & 0x3u;
325
326 while(tapCnt > 0u)
327 {
328 /* Read the coefficient */
329 c0 = *(ptr2);
330
331 /* Increment the coefficient pointer by interpolation factor times. */
332 ptr2 += S->L;
333
334 /* Read the input sample */
335 x0 = *(ptr1++);
336
337 /* Perform the multiply-accumulate */
338 sum0 += (q63_t) x0 *c0;
339
340 /* Decrement the loop counter */
341 tapCnt--;
342 }
343
344 /* The result is in the accumulator, store in the destination buffer. */
345 *pDst++ = (q31_t) (sum0 >> 31);
346
347 /* Increment the address modifier index of coefficient buffer */
348 j++;
349
350 /* Decrement the loop counter */
351 i--;
352 }
353
354 /* Advance the state pointer by 1
355 * to process the next group of interpolation factor number samples */
356 pState = pState + 1;
357
358 /* Decrement the loop counter */
359 blkCnt--;
360 }
361
362 /* Processing is complete.
363 ** Now copy the last phaseLen - 1 samples to the satrt of the state buffer.
364 ** This prepares the state buffer for the next function call. */
365
366 /* Points to the start of the state buffer */
367 pStateCurnt = S->pState;
368
369 tapCnt = (phaseLen - 1u) >> 2u;
370
371 /* copy data */
372 while(tapCnt > 0u)
373 {
374 *pStateCurnt++ = *pState++;
375 *pStateCurnt++ = *pState++;
376 *pStateCurnt++ = *pState++;
377 *pStateCurnt++ = *pState++;
378
379 /* Decrement the loop counter */
380 tapCnt--;
381 }
382
383 tapCnt = (phaseLen - 1u) % 0x04u;
384
385 /* copy data */
386 while(tapCnt > 0u)
387 {
388 *pStateCurnt++ = *pState++;
389
390 /* Decrement the loop counter */
391 tapCnt--;
392 }
393
394 }
395
396
397 #else
398
399 void arm_fir_interpolate_q31(
400 const arm_fir_interpolate_instance_q31 * S,
401 q31_t * pSrc,
402 q31_t * pDst,
403 uint32_t blockSize)
404 {
405 q31_t *pState = S->pState; /* State pointer */
406 q31_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
407 q31_t *pStateCurnt; /* Points to the current sample of the state */
408 q31_t *ptr1, *ptr2; /* Temporary pointers for state and coefficient buffers */
409
410 /* Run the below code for Cortex-M0 */
411
412 q63_t sum; /* Accumulator */
413 q31_t x0, c0; /* Temporary variables to hold state and coefficient values */
414 uint32_t i, blkCnt; /* Loop counters */
415 uint16_t phaseLen = S->phaseLength, tapCnt; /* Length of each polyphase filter component */
416
417
418 /* S->pState buffer contains previous frame (phaseLen - 1) samples */
419 /* pStateCurnt points to the location where the new input data should be written */
420 pStateCurnt = S->pState + ((q31_t) phaseLen - 1);
421
422 /* Total number of intput samples */
423 blkCnt = blockSize;
424
425 /* Loop over the blockSize. */
426 while(blkCnt > 0u)
427 {
428 /* Copy new input sample into the state buffer */
429 *pStateCurnt++ = *pSrc++;
430
431 /* Loop over the Interpolation factor. */
432 i = S->L;
433
434 while(i > 0u)
435 {
436 /* Set accumulator to zero */
437 sum = 0;
438
439 /* Initialize state pointer */
440 ptr1 = pState;
441
442 /* Initialize coefficient pointer */
443 ptr2 = pCoeffs + (i - 1u);
444
445 tapCnt = phaseLen;
446
447 while(tapCnt > 0u)
448 {
449 /* Read the coefficient */
450 c0 = *(ptr2);
451
452 /* Increment the coefficient pointer by interpolation factor times. */
453 ptr2 += S->L;
454
455 /* Read the input sample */
456 x0 = *ptr1++;
457
458 /* Perform the multiply-accumulate */
459 sum += (q63_t) x0 *c0;
460
461 /* Decrement the loop counter */
462 tapCnt--;
463 }
464
465 /* The result is in the accumulator, store in the destination buffer. */
466 *pDst++ = (q31_t) (sum >> 31);
467
468 /* Decrement the loop counter */
469 i--;
470 }
471
472 /* Advance the state pointer by 1
473 * to process the next group of interpolation factor number samples */
474 pState = pState + 1;
475
476 /* Decrement the loop counter */
477 blkCnt--;
478 }
479
480 /* Processing is complete.
481 ** Now copy the last phaseLen - 1 samples to the satrt of the state buffer.
482 ** This prepares the state buffer for the next function call. */
483
484 /* Points to the start of the state buffer */
485 pStateCurnt = S->pState;
486
487 tapCnt = phaseLen - 1u;
488
489 /* copy data */
490 while(tapCnt > 0u)
491 {
492 *pStateCurnt++ = *pState++;
493
494 /* Decrement the loop counter */
495 tapCnt--;
496 }
497
498 }
499
500 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
501
502 /**
503 * @} end of FIR_Interpolate group
504 */
Imprint / Impressum