summaryrefslogtreecommitdiff
path: root/lib/rbcodec/codecs/libspeex/resample.c
diff options
context:
space:
mode:
Diffstat (limited to 'lib/rbcodec/codecs/libspeex/resample.c')
-rw-r--r--lib/rbcodec/codecs/libspeex/resample.c1140
1 files changed, 1140 insertions, 0 deletions
diff --git a/lib/rbcodec/codecs/libspeex/resample.c b/lib/rbcodec/codecs/libspeex/resample.c
new file mode 100644
index 0000000000..65dfef28a3
--- /dev/null
+++ b/lib/rbcodec/codecs/libspeex/resample.c
@@ -0,0 +1,1140 @@
1/* Copyright (C) 2007 Jean-Marc Valin
2
3 File: resample.c
4 Arbitrary resampling code
5
6 Redistribution and use in source and binary forms, with or without
7 modification, are permitted provided that the following conditions are
8 met:
9
10 1. Redistributions of source code must retain the above copyright notice,
11 this list of conditions and the following disclaimer.
12
13 2. Redistributions in binary form must reproduce the above copyright
14 notice, this list of conditions and the following disclaimer in the
15 documentation and/or other materials provided with the distribution.
16
17 3. The name of the author may not be used to endorse or promote products
18 derived from this software without specific prior written permission.
19
20 THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
21 IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
22 OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
23 DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT,
24 INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
25 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
26 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
28 STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
29 ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
30 POSSIBILITY OF SUCH DAMAGE.
31*/
32
33/*
34 The design goals of this code are:
35 - Very fast algorithm
36 - SIMD-friendly algorithm
37 - Low memory requirement
38 - Good *perceptual* quality (and not best SNR)
39
40 Warning: This resampler is relatively new. Although I think I got rid of
41 all the major bugs and I don't expect the API to change anymore, there
42 may be something I've missed. So use with caution.
43
44 This algorithm is based on this original resampling algorithm:
45 Smith, Julius O. Digital Audio Resampling Home Page
46 Center for Computer Research in Music and Acoustics (CCRMA),
47 Stanford University, 2007.
48 Web published at http://www-ccrma.stanford.edu/~jos/resample/.
49
50 There is one main difference, though. This resampler uses cubic
51 interpolation instead of linear interpolation in the above paper. This
52 makes the table much smaller and makes it possible to compute that table
53 on a per-stream basis. In turn, being able to tweak the table for each
54 stream makes it possible to both reduce complexity on simple ratios
55 (e.g. 2/3), and get rid of the rounding operations in the inner loop.
56 The latter both reduces CPU time and makes the algorithm more SIMD-friendly.
57*/
58
59#ifdef HAVE_CONFIG_H
60#include "config-speex.h"
61#endif
62
63#ifdef OUTSIDE_SPEEX
64#include <stdlib.h>
65static void *speex_alloc (int size) {return calloc(size,1);}
66static void *speex_realloc (void *ptr, int size) {return realloc(ptr, size);}
67static void speex_free (void *ptr) {free(ptr);}
68#include "speex_resampler.h"
69#include "arch.h"
70#else /* OUTSIDE_SPEEX */
71
72#include "speex/speex_resampler.h"
73#include "arch.h"
74#include "os_support.h"
75#endif /* OUTSIDE_SPEEX */
76
77#include <math.h>
78
79#ifndef M_PI
80#define M_PI 3.14159263
81#endif
82
83#ifdef FIXED_POINT
84#define WORD2INT(x) ((x) < -32767 ? -32768 : ((x) > 32766 ? 32767 : (x)))
85#else
86#define WORD2INT(x) ((x) < -32767.5f ? -32768 : ((x) > 32766.5f ? 32767 : floor(.5+(x))))
87#endif
88
89/*#define float double*/
90#define FILTER_SIZE 64
91#define OVERSAMPLE 8
92
93#define IMAX(a,b) ((a) > (b) ? (a) : (b))
94#define IMIN(a,b) ((a) < (b) ? (a) : (b))
95
96#ifndef NULL
97#define NULL 0
98#endif
99
100typedef int (*resampler_basic_func)(SpeexResamplerState *, spx_uint32_t , const spx_word16_t *, spx_uint32_t *, spx_word16_t *, spx_uint32_t *);
101
102struct SpeexResamplerState_ {
103 spx_uint32_t in_rate;
104 spx_uint32_t out_rate;
105 spx_uint32_t num_rate;
106 spx_uint32_t den_rate;
107
108 int quality;
109 spx_uint32_t nb_channels;
110 spx_uint32_t filt_len;
111 spx_uint32_t mem_alloc_size;
112 int int_advance;
113 int frac_advance;
114 float cutoff;
115 spx_uint32_t oversample;
116 int initialised;
117 int started;
118
119 /* These are per-channel */
120 spx_int32_t *last_sample;
121 spx_uint32_t *samp_frac_num;
122 spx_uint32_t *magic_samples;
123
124 spx_word16_t *mem;
125 spx_word16_t *sinc_table;
126 spx_uint32_t sinc_table_length;
127 resampler_basic_func resampler_ptr;
128
129 int in_stride;
130 int out_stride;
131} ;
132
133static double kaiser12_table[68] = {
134 0.99859849, 1.00000000, 0.99859849, 0.99440475, 0.98745105, 0.97779076,
135 0.96549770, 0.95066529, 0.93340547, 0.91384741, 0.89213598, 0.86843014,
136 0.84290116, 0.81573067, 0.78710866, 0.75723148, 0.72629970, 0.69451601,
137 0.66208321, 0.62920216, 0.59606986, 0.56287762, 0.52980938, 0.49704014,
138 0.46473455, 0.43304576, 0.40211431, 0.37206735, 0.34301800, 0.31506490,
139 0.28829195, 0.26276832, 0.23854851, 0.21567274, 0.19416736, 0.17404546,
140 0.15530766, 0.13794294, 0.12192957, 0.10723616, 0.09382272, 0.08164178,
141 0.07063950, 0.06075685, 0.05193064, 0.04409466, 0.03718069, 0.03111947,
142 0.02584161, 0.02127838, 0.01736250, 0.01402878, 0.01121463, 0.00886058,
143 0.00691064, 0.00531256, 0.00401805, 0.00298291, 0.00216702, 0.00153438,
144 0.00105297, 0.00069463, 0.00043489, 0.00025272, 0.00013031, 0.0000527734,
145 0.00001000, 0.00000000};
146/*
147static double kaiser12_table[36] = {
148 0.99440475, 1.00000000, 0.99440475, 0.97779076, 0.95066529, 0.91384741,
149 0.86843014, 0.81573067, 0.75723148, 0.69451601, 0.62920216, 0.56287762,
150 0.49704014, 0.43304576, 0.37206735, 0.31506490, 0.26276832, 0.21567274,
151 0.17404546, 0.13794294, 0.10723616, 0.08164178, 0.06075685, 0.04409466,
152 0.03111947, 0.02127838, 0.01402878, 0.00886058, 0.00531256, 0.00298291,
153 0.00153438, 0.00069463, 0.00025272, 0.0000527734, 0.00000500, 0.00000000};
154*/
155static double kaiser10_table[36] = {
156 0.99537781, 1.00000000, 0.99537781, 0.98162644, 0.95908712, 0.92831446,
157 0.89005583, 0.84522401, 0.79486424, 0.74011713, 0.68217934, 0.62226347,
158 0.56155915, 0.50119680, 0.44221549, 0.38553619, 0.33194107, 0.28205962,
159 0.23636152, 0.19515633, 0.15859932, 0.12670280, 0.09935205, 0.07632451,
160 0.05731132, 0.04193980, 0.02979584, 0.02044510, 0.01345224, 0.00839739,
161 0.00488951, 0.00257636, 0.00115101, 0.00035515, 0.00000000, 0.00000000};
162
163static double kaiser8_table[36] = {
164 0.99635258, 1.00000000, 0.99635258, 0.98548012, 0.96759014, 0.94302200,
165 0.91223751, 0.87580811, 0.83439927, 0.78875245, 0.73966538, 0.68797126,
166 0.63451750, 0.58014482, 0.52566725, 0.47185369, 0.41941150, 0.36897272,
167 0.32108304, 0.27619388, 0.23465776, 0.19672670, 0.16255380, 0.13219758,
168 0.10562887, 0.08273982, 0.06335451, 0.04724088, 0.03412321, 0.02369490,
169 0.01563093, 0.00959968, 0.00527363, 0.00233883, 0.00050000, 0.00000000};
170
171static double kaiser6_table[36] = {
172 0.99733006, 1.00000000, 0.99733006, 0.98935595, 0.97618418, 0.95799003,
173 0.93501423, 0.90755855, 0.87598009, 0.84068475, 0.80211977, 0.76076565,
174 0.71712752, 0.67172623, 0.62508937, 0.57774224, 0.53019925, 0.48295561,
175 0.43647969, 0.39120616, 0.34752997, 0.30580127, 0.26632152, 0.22934058,
176 0.19505503, 0.16360756, 0.13508755, 0.10953262, 0.08693120, 0.06722600,
177 0.05031820, 0.03607231, 0.02432151, 0.01487334, 0.00752000, 0.00000000};
178
179struct FuncDef {
180 double *table;
181 int oversample;
182};
183
184static struct FuncDef _KAISER12 = {kaiser12_table, 64};
185#define KAISER12 (&_KAISER12)
186/*static struct FuncDef _KAISER12 = {kaiser12_table, 32};
187#define KAISER12 (&_KAISER12)*/
188static struct FuncDef _KAISER10 = {kaiser10_table, 32};
189#define KAISER10 (&_KAISER10)
190static struct FuncDef _KAISER8 = {kaiser8_table, 32};
191#define KAISER8 (&_KAISER8)
192static struct FuncDef _KAISER6 = {kaiser6_table, 32};
193#define KAISER6 (&_KAISER6)
194
195struct QualityMapping {
196 int base_length;
197 int oversample;
198 float downsample_bandwidth;
199 float upsample_bandwidth;
200 struct FuncDef *window_func;
201};
202
203
204/* This table maps conversion quality to internal parameters. There are two
205 reasons that explain why the up-sampling bandwidth is larger than the
206 down-sampling bandwidth:
207 1) When up-sampling, we can assume that the spectrum is already attenuated
208 close to the Nyquist rate (from an A/D or a previous resampling filter)
209 2) Any aliasing that occurs very close to the Nyquist rate will be masked
210 by the sinusoids/noise just below the Nyquist rate (guaranteed only for
211 up-sampling).
212*/
213static const struct QualityMapping quality_map[11] = {
214 { 8, 4, 0.830f, 0.860f, KAISER6 }, /* Q0 */
215 { 16, 4, 0.850f, 0.880f, KAISER6 }, /* Q1 */
216 { 32, 4, 0.882f, 0.910f, KAISER6 }, /* Q2 */ /* 82.3% cutoff ( ~60 dB stop) 6 */
217 { 48, 8, 0.895f, 0.917f, KAISER8 }, /* Q3 */ /* 84.9% cutoff ( ~80 dB stop) 8 */
218 { 64, 8, 0.921f, 0.940f, KAISER8 }, /* Q4 */ /* 88.7% cutoff ( ~80 dB stop) 8 */
219 { 80, 16, 0.922f, 0.940f, KAISER10}, /* Q5 */ /* 89.1% cutoff (~100 dB stop) 10 */
220 { 96, 16, 0.940f, 0.945f, KAISER10}, /* Q6 */ /* 91.5% cutoff (~100 dB stop) 10 */
221 {128, 16, 0.950f, 0.950f, KAISER10}, /* Q7 */ /* 93.1% cutoff (~100 dB stop) 10 */
222 {160, 16, 0.960f, 0.960f, KAISER10}, /* Q8 */ /* 94.5% cutoff (~100 dB stop) 10 */
223 {192, 32, 0.968f, 0.968f, KAISER12}, /* Q9 */ /* 95.5% cutoff (~100 dB stop) 10 */
224 {256, 32, 0.975f, 0.975f, KAISER12}, /* Q10 */ /* 96.6% cutoff (~100 dB stop) 10 */
225};
226/*8,24,40,56,80,104,128,160,200,256,320*/
227static double compute_func(float x, struct FuncDef *func)
228{
229 float y, frac;
230 double interp[4];
231 int ind;
232 y = x*func->oversample;
233 ind = (int)floor(y);
234 frac = (y-ind);
235 /* CSE with handle the repeated powers */
236 interp[3] = -0.1666666667*frac + 0.1666666667*(frac*frac*frac);
237 interp[2] = frac + 0.5*(frac*frac) - 0.5*(frac*frac*frac);
238 /*interp[2] = 1.f - 0.5f*frac - frac*frac + 0.5f*frac*frac*frac;*/
239 interp[0] = -0.3333333333*frac + 0.5*(frac*frac) - 0.1666666667*(frac*frac*frac);
240 /* Just to make sure we don't have rounding problems */
241 interp[1] = 1.f-interp[3]-interp[2]-interp[0];
242
243 /*sum = frac*accum[1] + (1-frac)*accum[2];*/
244 return interp[0]*func->table[ind] + interp[1]*func->table[ind+1] + interp[2]*func->table[ind+2] + interp[3]*func->table[ind+3];
245}
246
247#if 0
248#include <stdio.h>
249int main(int argc, char **argv)
250{
251 int i;
252 for (i=0;i<256;i++)
253 {
254 printf ("%f\n", compute_func(i/256., KAISER12));
255 }
256 return 0;
257}
258#endif
259
260#ifdef FIXED_POINT
261/* The slow way of computing a sinc for the table. Should improve that some day */
262static spx_word16_t sinc(float cutoff, float x, int N, struct FuncDef *window_func)
263{
264 /*fprintf (stderr, "%f ", x);*/
265 float xx = x * cutoff;
266 if (fabs(x)<1e-6f)
267 return WORD2INT(32768.*cutoff);
268 else if (fabs(x) > .5f*N)
269 return 0;
270 /*FIXME: Can it really be any slower than this? */
271 return WORD2INT(32768.*cutoff*sin(M_PI*xx)/(M_PI*xx) * compute_func(fabs(2.*x/N), window_func));
272}
273#else
274/* The slow way of computing a sinc for the table. Should improve that some day */
275static spx_word16_t sinc(float cutoff, float x, int N, struct FuncDef *window_func)
276{
277 /*fprintf (stderr, "%f ", x);*/
278 float xx = x * cutoff;
279 if (fabs(x)<1e-6)
280 return cutoff;
281 else if (fabs(x) > .5*N)
282 return 0;
283 /*FIXME: Can it really be any slower than this? */
284 return cutoff*sin(M_PI*xx)/(M_PI*xx) * compute_func(fabs(2.*x/N), window_func);
285}
286#endif
287
288#ifdef FIXED_POINT
289static void cubic_coef(spx_word16_t x, spx_word16_t interp[4])
290{
291 /* Compute interpolation coefficients. I'm not sure whether this corresponds to cubic interpolation
292 but I know it's MMSE-optimal on a sinc */
293 spx_word16_t x2, x3;
294 x2 = MULT16_16_P15(x, x);
295 x3 = MULT16_16_P15(x, x2);
296 interp[0] = PSHR32(MULT16_16(QCONST16(-0.16667f, 15),x) + MULT16_16(QCONST16(0.16667f, 15),x3),15);
297 interp[1] = EXTRACT16(EXTEND32(x) + SHR32(SUB32(EXTEND32(x2),EXTEND32(x3)),1));
298 interp[3] = PSHR32(MULT16_16(QCONST16(-0.33333f, 15),x) + MULT16_16(QCONST16(.5f,15),x2) - MULT16_16(QCONST16(0.16667f, 15),x3),15);
299 /* Just to make sure we don't have rounding problems */
300 interp[2] = Q15_ONE-interp[0]-interp[1]-interp[3];
301 if (interp[2]<32767)
302 interp[2]+=1;
303}
304#else
305static void cubic_coef(spx_word16_t frac, spx_word16_t interp[4])
306{
307 /* Compute interpolation coefficients. I'm not sure whether this corresponds to cubic interpolation
308 but I know it's MMSE-optimal on a sinc */
309 interp[0] = -0.16667f*frac + 0.16667f*frac*frac*frac;
310 interp[1] = frac + 0.5f*frac*frac - 0.5f*frac*frac*frac;
311 /*interp[2] = 1.f - 0.5f*frac - frac*frac + 0.5f*frac*frac*frac;*/
312 interp[3] = -0.33333f*frac + 0.5f*frac*frac - 0.16667f*frac*frac*frac;
313 /* Just to make sure we don't have rounding problems */
314 interp[2] = 1.-interp[0]-interp[1]-interp[3];
315}
316#endif
317
318static int resampler_basic_direct_single(SpeexResamplerState *st, spx_uint32_t channel_index, const spx_word16_t *in, spx_uint32_t *in_len, spx_word16_t *out, spx_uint32_t *out_len)
319{
320 int N = st->filt_len;
321 int out_sample = 0;
322 spx_word16_t *mem;
323 int last_sample = st->last_sample[channel_index];
324 spx_uint32_t samp_frac_num = st->samp_frac_num[channel_index];
325 mem = st->mem + channel_index * st->mem_alloc_size;
326 while (!(last_sample >= (spx_int32_t)*in_len || out_sample >= (spx_int32_t)*out_len))
327 {
328 int j;
329 spx_word32_t sum=0;
330
331 /* We already have all the filter coefficients pre-computed in the table */
332 const spx_word16_t *ptr;
333 /* Do the memory part */
334 for (j=0;last_sample-N+1+j < 0;j++)
335 {
336 sum += MULT16_16(mem[last_sample+j],st->sinc_table[samp_frac_num*st->filt_len+j]);
337 }
338
339 /* Do the new part */
340 if (in != NULL)
341 {
342 ptr = in+st->in_stride*(last_sample-N+1+j);
343 for (;j<N;j++)
344 {
345 sum += MULT16_16(*ptr,st->sinc_table[samp_frac_num*st->filt_len+j]);
346 ptr += st->in_stride;
347 }
348 }
349
350 *out = PSHR32(sum,15);
351 out += st->out_stride;
352 out_sample++;
353 last_sample += st->int_advance;
354 samp_frac_num += st->frac_advance;
355 if (samp_frac_num >= st->den_rate)
356 {
357 samp_frac_num -= st->den_rate;
358 last_sample++;
359 }
360 }
361 st->last_sample[channel_index] = last_sample;
362 st->samp_frac_num[channel_index] = samp_frac_num;
363 return out_sample;
364}
365
366#ifdef FIXED_POINT
367#else
368/* This is the same as the previous function, except with a double-precision accumulator */
369static int resampler_basic_direct_double(SpeexResamplerState *st, spx_uint32_t channel_index, const spx_word16_t *in, spx_uint32_t *in_len, spx_word16_t *out, spx_uint32_t *out_len)
370{
371 int N = st->filt_len;
372 int out_sample = 0;
373 spx_word16_t *mem;
374 int last_sample = st->last_sample[channel_index];
375 spx_uint32_t samp_frac_num = st->samp_frac_num[channel_index];
376 mem = st->mem + channel_index * st->mem_alloc_size;
377 while (!(last_sample >= (spx_int32_t)*in_len || out_sample >= (spx_int32_t)*out_len))
378 {
379 int j;
380 double sum=0;
381
382 /* We already have all the filter coefficients pre-computed in the table */
383 const spx_word16_t *ptr;
384 /* Do the memory part */
385 for (j=0;last_sample-N+1+j < 0;j++)
386 {
387 sum += MULT16_16(mem[last_sample+j],(double)st->sinc_table[samp_frac_num*st->filt_len+j]);
388 }
389
390 /* Do the new part */
391 if (in != NULL)
392 {
393 ptr = in+st->in_stride*(last_sample-N+1+j);
394 for (;j<N;j++)
395 {
396 sum += MULT16_16(*ptr,(double)st->sinc_table[samp_frac_num*st->filt_len+j]);
397 ptr += st->in_stride;
398 }
399 }
400
401 *out = sum;
402 out += st->out_stride;
403 out_sample++;
404 last_sample += st->int_advance;
405 samp_frac_num += st->frac_advance;
406 if (samp_frac_num >= st->den_rate)
407 {
408 samp_frac_num -= st->den_rate;
409 last_sample++;
410 }
411 }
412 st->last_sample[channel_index] = last_sample;
413 st->samp_frac_num[channel_index] = samp_frac_num;
414 return out_sample;
415}
416#endif
417
418static int resampler_basic_interpolate_single(SpeexResamplerState *st, spx_uint32_t channel_index, const spx_word16_t *in, spx_uint32_t *in_len, spx_word16_t *out, spx_uint32_t *out_len)
419{
420 int N = st->filt_len;
421 int out_sample = 0;
422 spx_word16_t *mem;
423 int last_sample = st->last_sample[channel_index];
424 spx_uint32_t samp_frac_num = st->samp_frac_num[channel_index];
425 mem = st->mem + channel_index * st->mem_alloc_size;
426 while (!(last_sample >= (spx_int32_t)*in_len || out_sample >= (spx_int32_t)*out_len))
427 {
428 int j;
429 spx_word32_t sum=0;
430
431 /* We need to interpolate the sinc filter */
432 spx_word32_t accum[4] = {0.f,0.f, 0.f, 0.f};
433 spx_word16_t interp[4];
434 const spx_word16_t *ptr;
435 int offset;
436 spx_word16_t frac;
437 offset = samp_frac_num*st->oversample/st->den_rate;
438#ifdef FIXED_POINT
439 frac = PDIV32(SHL32((samp_frac_num*st->oversample) % st->den_rate,15),st->den_rate);
440#else
441 frac = ((float)((samp_frac_num*st->oversample) % st->den_rate))/st->den_rate;
442#endif
443 /* This code is written like this to make it easy to optimise with SIMD.
444 For most DSPs, it would be best to split the loops in two because most DSPs
445 have only two accumulators */
446 for (j=0;last_sample-N+1+j < 0;j++)
447 {
448 spx_word16_t curr_mem = mem[last_sample+j];
449 accum[0] += MULT16_16(curr_mem,st->sinc_table[4+(j+1)*st->oversample-offset-2]);
450 accum[1] += MULT16_16(curr_mem,st->sinc_table[4+(j+1)*st->oversample-offset-1]);
451 accum[2] += MULT16_16(curr_mem,st->sinc_table[4+(j+1)*st->oversample-offset]);
452 accum[3] += MULT16_16(curr_mem,st->sinc_table[4+(j+1)*st->oversample-offset+1]);
453 }
454
455 if (in != NULL)
456 {
457 ptr = in+st->in_stride*(last_sample-N+1+j);
458 /* Do the new part */
459 for (;j<N;j++)
460 {
461 spx_word16_t curr_in = *ptr;
462 ptr += st->in_stride;
463 accum[0] += MULT16_16(curr_in,st->sinc_table[4+(j+1)*st->oversample-offset-2]);
464 accum[1] += MULT16_16(curr_in,st->sinc_table[4+(j+1)*st->oversample-offset-1]);
465 accum[2] += MULT16_16(curr_in,st->sinc_table[4+(j+1)*st->oversample-offset]);
466 accum[3] += MULT16_16(curr_in,st->sinc_table[4+(j+1)*st->oversample-offset+1]);
467 }
468 }
469 cubic_coef(frac, interp);
470 sum = MULT16_32_Q15(interp[0],accum[0]) + MULT16_32_Q15(interp[1],accum[1]) + MULT16_32_Q15(interp[2],accum[2]) + MULT16_32_Q15(interp[3],accum[3]);
471
472 *out = PSHR32(sum,15);
473 out += st->out_stride;
474 out_sample++;
475 last_sample += st->int_advance;
476 samp_frac_num += st->frac_advance;
477 if (samp_frac_num >= st->den_rate)
478 {
479 samp_frac_num -= st->den_rate;
480 last_sample++;
481 }
482 }
483 st->last_sample[channel_index] = last_sample;
484 st->samp_frac_num[channel_index] = samp_frac_num;
485 return out_sample;
486}
487
488#ifdef FIXED_POINT
489#else
490/* This is the same as the previous function, except with a double-precision accumulator */
491static int resampler_basic_interpolate_double(SpeexResamplerState *st, spx_uint32_t channel_index, const spx_word16_t *in, spx_uint32_t *in_len, spx_word16_t *out, spx_uint32_t *out_len)
492{
493 int N = st->filt_len;
494 int out_sample = 0;
495 spx_word16_t *mem;
496 int last_sample = st->last_sample[channel_index];
497 spx_uint32_t samp_frac_num = st->samp_frac_num[channel_index];
498 mem = st->mem + channel_index * st->mem_alloc_size;
499 while (!(last_sample >= (spx_int32_t)*in_len || out_sample >= (spx_int32_t)*out_len))
500 {
501 int j;
502 spx_word32_t sum=0;
503
504 /* We need to interpolate the sinc filter */
505 double accum[4] = {0.f,0.f, 0.f, 0.f};
506 float interp[4];
507 const spx_word16_t *ptr;
508 float alpha = ((float)samp_frac_num)/st->den_rate;
509 int offset = samp_frac_num*st->oversample/st->den_rate;
510 float frac = alpha*st->oversample - offset;
511 /* This code is written like this to make it easy to optimise with SIMD.
512 For most DSPs, it would be best to split the loops in two because most DSPs
513 have only two accumulators */
514 for (j=0;last_sample-N+1+j < 0;j++)
515 {
516 double curr_mem = mem[last_sample+j];
517 accum[0] += MULT16_16(curr_mem,st->sinc_table[4+(j+1)*st->oversample-offset-2]);
518 accum[1] += MULT16_16(curr_mem,st->sinc_table[4+(j+1)*st->oversample-offset-1]);
519 accum[2] += MULT16_16(curr_mem,st->sinc_table[4+(j+1)*st->oversample-offset]);
520 accum[3] += MULT16_16(curr_mem,st->sinc_table[4+(j+1)*st->oversample-offset+1]);
521 }
522 if (in != NULL)
523 {
524 ptr = in+st->in_stride*(last_sample-N+1+j);
525 /* Do the new part */
526 for (;j<N;j++)
527 {
528 double curr_in = *ptr;
529 ptr += st->in_stride;
530 accum[0] += MULT16_16(curr_in,st->sinc_table[4+(j+1)*st->oversample-offset-2]);
531 accum[1] += MULT16_16(curr_in,st->sinc_table[4+(j+1)*st->oversample-offset-1]);
532 accum[2] += MULT16_16(curr_in,st->sinc_table[4+(j+1)*st->oversample-offset]);
533 accum[3] += MULT16_16(curr_in,st->sinc_table[4+(j+1)*st->oversample-offset+1]);
534 }
535 }
536 cubic_coef(frac, interp);
537 sum = interp[0]*accum[0] + interp[1]*accum[1] + interp[2]*accum[2] + interp[3]*accum[3];
538
539 *out = PSHR32(sum,15);
540 out += st->out_stride;
541 out_sample++;
542 last_sample += st->int_advance;
543 samp_frac_num += st->frac_advance;
544 if (samp_frac_num >= st->den_rate)
545 {
546 samp_frac_num -= st->den_rate;
547 last_sample++;
548 }
549 }
550 st->last_sample[channel_index] = last_sample;
551 st->samp_frac_num[channel_index] = samp_frac_num;
552 return out_sample;
553}
554#endif
555
556static void update_filter(SpeexResamplerState *st)
557{
558 spx_uint32_t old_length;
559
560 old_length = st->filt_len;
561 st->oversample = quality_map[st->quality].oversample;
562 st->filt_len = quality_map[st->quality].base_length;
563
564 if (st->num_rate > st->den_rate)
565 {
566 /* down-sampling */
567 st->cutoff = quality_map[st->quality].downsample_bandwidth * st->den_rate / st->num_rate;
568 /* FIXME: divide the numerator and denominator by a certain amount if they're too large */
569 st->filt_len = st->filt_len*st->num_rate / st->den_rate;
570 /* Round down to make sure we have a multiple of 4 */
571 st->filt_len &= (~0x3);
572 if (2*st->den_rate < st->num_rate)
573 st->oversample >>= 1;
574 if (4*st->den_rate < st->num_rate)
575 st->oversample >>= 1;
576 if (8*st->den_rate < st->num_rate)
577 st->oversample >>= 1;
578 if (16*st->den_rate < st->num_rate)
579 st->oversample >>= 1;
580 if (st->oversample < 1)
581 st->oversample = 1;
582 } else {
583 /* up-sampling */
584 st->cutoff = quality_map[st->quality].upsample_bandwidth;
585 }
586
587 /* Choose the resampling type that requires the least amount of memory */
588 if (st->den_rate <= st->oversample)
589 {
590 spx_uint32_t i;
591 if (!st->sinc_table)
592 st->sinc_table = (spx_word16_t *)speex_alloc(st->filt_len*st->den_rate*sizeof(spx_word16_t));
593 else if (st->sinc_table_length < st->filt_len*st->den_rate)
594 {
595 st->sinc_table = (spx_word16_t *)speex_realloc(st->sinc_table,st->filt_len*st->den_rate*sizeof(spx_word16_t));
596 st->sinc_table_length = st->filt_len*st->den_rate;
597 }
598 for (i=0;i<st->den_rate;i++)
599 {
600 spx_int32_t j;
601 for (j=0;j<st->filt_len;j++)
602 {
603 st->sinc_table[i*st->filt_len+j] = sinc(st->cutoff,((j-(spx_int32_t)st->filt_len/2+1)-((float)i)/st->den_rate), st->filt_len, quality_map[st->quality].window_func);
604 }
605 }
606#ifdef FIXED_POINT
607 st->resampler_ptr = resampler_basic_direct_single;
608#else
609 if (st->quality>8)
610 st->resampler_ptr = resampler_basic_direct_double;
611 else
612 st->resampler_ptr = resampler_basic_direct_single;
613#endif
614 /*fprintf (stderr, "resampler uses direct sinc table and normalised cutoff %f\n", cutoff);*/
615 } else {
616 spx_int32_t i;
617 if (!st->sinc_table)
618 st->sinc_table = (spx_word16_t *)speex_alloc((st->filt_len*st->oversample+8)*sizeof(spx_word16_t));
619 else if (st->sinc_table_length < st->filt_len*st->oversample+8)
620 {
621 st->sinc_table = (spx_word16_t *)speex_realloc(st->sinc_table,(st->filt_len*st->oversample+8)*sizeof(spx_word16_t));
622 st->sinc_table_length = st->filt_len*st->oversample+8;
623 }
624 for (i=-4;i<(spx_int32_t)(st->oversample*st->filt_len+4);i++)
625 st->sinc_table[i+4] = sinc(st->cutoff,(i/(float)st->oversample - st->filt_len/2), st->filt_len, quality_map[st->quality].window_func);
626#ifdef FIXED_POINT
627 st->resampler_ptr = resampler_basic_interpolate_single;
628#else
629 if (st->quality>8)
630 st->resampler_ptr = resampler_basic_interpolate_double;
631 else
632 st->resampler_ptr = resampler_basic_interpolate_single;
633#endif
634 /*fprintf (stderr, "resampler uses interpolated sinc table and normalised cutoff %f\n", cutoff);*/
635 }
636 st->int_advance = st->num_rate/st->den_rate;
637 st->frac_advance = st->num_rate%st->den_rate;
638
639
640 /* Here's the place where we update the filter memory to take into account
641 the change in filter length. It's probably the messiest part of the code
642 due to handling of lots of corner cases. */
643 if (!st->mem)
644 {
645 spx_uint32_t i;
646 st->mem = (spx_word16_t*)speex_alloc(st->nb_channels*(st->filt_len-1) * sizeof(spx_word16_t));
647 for (i=0;i<st->nb_channels*(st->filt_len-1);i++)
648 st->mem[i] = 0;
649 st->mem_alloc_size = st->filt_len-1;
650 /*speex_warning("init filter");*/
651 } else if (!st->started)
652 {
653 spx_uint32_t i;
654 st->mem = (spx_word16_t*)speex_realloc(st->mem, st->nb_channels*(st->filt_len-1) * sizeof(spx_word16_t));
655 for (i=0;i<st->nb_channels*(st->filt_len-1);i++)
656 st->mem[i] = 0;
657 st->mem_alloc_size = st->filt_len-1;
658 /*speex_warning("reinit filter");*/
659 } else if (st->filt_len > old_length)
660 {
661 spx_int32_t i;
662 /* Increase the filter length */
663 /*speex_warning("increase filter size");*/
664 int old_alloc_size = st->mem_alloc_size;
665 if (st->filt_len-1 > st->mem_alloc_size)
666 {
667 st->mem = (spx_word16_t*)speex_realloc(st->mem, st->nb_channels*(st->filt_len-1) * sizeof(spx_word16_t));
668 st->mem_alloc_size = st->filt_len-1;
669 }
670 for (i=st->nb_channels-1;i>=0;i--)
671 {
672 spx_int32_t j;
673 spx_uint32_t olen = old_length;
674 /*if (st->magic_samples[i])*/
675 {
676 /* Try and remove the magic samples as if nothing had happened */
677
678 /* FIXME: This is wrong but for now we need it to avoid going over the array bounds */
679 olen = old_length + 2*st->magic_samples[i];
680 for (j=old_length-2+st->magic_samples[i];j>=0;j--)
681 st->mem[i*st->mem_alloc_size+j+st->magic_samples[i]] = st->mem[i*old_alloc_size+j];
682 for (j=0;j<st->magic_samples[i];j++)
683 st->mem[i*st->mem_alloc_size+j] = 0;
684 st->magic_samples[i] = 0;
685 }
686 if (st->filt_len > olen)
687 {
688 /* If the new filter length is still bigger than the "augmented" length */
689 /* Copy data going backward */
690 for (j=0;j<olen-1;j++)
691 st->mem[i*st->mem_alloc_size+(st->filt_len-2-j)] = st->mem[i*st->mem_alloc_size+(olen-2-j)];
692 /* Then put zeros for lack of anything better */
693 for (;j<st->filt_len-1;j++)
694 st->mem[i*st->mem_alloc_size+(st->filt_len-2-j)] = 0;
695 /* Adjust last_sample */
696 st->last_sample[i] += (st->filt_len - olen)/2;
697 } else {
698 /* Put back some of the magic! */
699 st->magic_samples[i] = (olen - st->filt_len)/2;
700 for (j=0;j<st->filt_len-1+st->magic_samples[i];j++)
701 st->mem[i*st->mem_alloc_size+j] = st->mem[i*st->mem_alloc_size+j+st->magic_samples[i]];
702 }
703 }
704 } else if (st->filt_len < old_length)
705 {
706 spx_uint32_t i;
707 /* Reduce filter length, this a bit tricky. We need to store some of the memory as "magic"
708 samples so they can be used directly as input the next time(s) */
709 for (i=0;i<st->nb_channels;i++)
710 {
711 spx_uint32_t j;
712 spx_uint32_t old_magic = st->magic_samples[i];
713 st->magic_samples[i] = (old_length - st->filt_len)/2;
714 /* We must copy some of the memory that's no longer used */
715 /* Copy data going backward */
716 for (j=0;j<st->filt_len-1+st->magic_samples[i]+old_magic;j++)
717 st->mem[i*st->mem_alloc_size+j] = st->mem[i*st->mem_alloc_size+j+st->magic_samples[i]];
718 st->magic_samples[i] += old_magic;
719 }
720 }
721
722}
723
724SpeexResamplerState *speex_resampler_init(spx_uint32_t nb_channels, spx_uint32_t in_rate, spx_uint32_t out_rate, int quality, int *err)
725{
726 return speex_resampler_init_frac(nb_channels, in_rate, out_rate, in_rate, out_rate, quality, err);
727}
728
729SpeexResamplerState *speex_resampler_init_frac(spx_uint32_t nb_channels, spx_uint32_t ratio_num, spx_uint32_t ratio_den, spx_uint32_t in_rate, spx_uint32_t out_rate, int quality, int *err)
730{
731 spx_uint32_t i;
732 SpeexResamplerState *st;
733 if (quality > 10 || quality < 0)
734 {
735 if (err)
736 *err = RESAMPLER_ERR_INVALID_ARG;
737 return NULL;
738 }
739 st = (SpeexResamplerState *)speex_alloc(sizeof(SpeexResamplerState));
740 st->initialised = 0;
741 st->started = 0;
742 st->in_rate = 0;
743 st->out_rate = 0;
744 st->num_rate = 0;
745 st->den_rate = 0;
746 st->quality = -1;
747 st->sinc_table_length = 0;
748 st->mem_alloc_size = 0;
749 st->filt_len = 0;
750 st->mem = 0;
751 st->resampler_ptr = 0;
752
753 st->cutoff = 1.f;
754 st->nb_channels = nb_channels;
755 st->in_stride = 1;
756 st->out_stride = 1;
757
758 /* Per channel data */
759 st->last_sample = (spx_int32_t*)speex_alloc(nb_channels*sizeof(int));
760 st->magic_samples = (spx_uint32_t*)speex_alloc(nb_channels*sizeof(int));
761 st->samp_frac_num = (spx_uint32_t*)speex_alloc(nb_channels*sizeof(int));
762 for (i=0;i<nb_channels;i++)
763 {
764 st->last_sample[i] = 0;
765 st->magic_samples[i] = 0;
766 st->samp_frac_num[i] = 0;
767 }
768
769 speex_resampler_set_quality(st, quality);
770 speex_resampler_set_rate_frac(st, ratio_num, ratio_den, in_rate, out_rate);
771
772
773 update_filter(st);
774
775 st->initialised = 1;
776 if (err)
777 *err = RESAMPLER_ERR_SUCCESS;
778
779 return st;
780}
781
782void speex_resampler_destroy(SpeexResamplerState *st)
783{
784 speex_free(st->mem);
785 speex_free(st->sinc_table);
786 speex_free(st->last_sample);
787 speex_free(st->magic_samples);
788 speex_free(st->samp_frac_num);
789 speex_free(st);
790}
791
792
793
794static int speex_resampler_process_native(SpeexResamplerState *st, spx_uint32_t channel_index, const spx_word16_t *in, spx_uint32_t *in_len, spx_word16_t *out, spx_uint32_t *out_len)
795{
796 int j=0;
797 int N = st->filt_len;
798 int out_sample = 0;
799 spx_word16_t *mem;
800 spx_uint32_t tmp_out_len = 0;
801 mem = st->mem + channel_index * st->mem_alloc_size;
802 st->started = 1;
803
804 /* Handle the case where we have samples left from a reduction in filter length */
805 if (st->magic_samples[channel_index])
806 {
807 int istride_save;
808 spx_uint32_t tmp_in_len;
809 spx_uint32_t tmp_magic;
810
811 istride_save = st->in_stride;
812 tmp_in_len = st->magic_samples[channel_index];
813 tmp_out_len = *out_len;
814 /* magic_samples needs to be set to zero to avoid infinite recursion */
815 tmp_magic = st->magic_samples[channel_index];
816 st->magic_samples[channel_index] = 0;
817 st->in_stride = 1;
818 speex_resampler_process_native(st, channel_index, mem+N-1, &tmp_in_len, out, &tmp_out_len);
819 st->in_stride = istride_save;
820 /*speex_warning_int("extra samples:", tmp_out_len);*/
821 /* If we couldn't process all "magic" input samples, save the rest for next time */
822 if (tmp_in_len < tmp_magic)
823 {
824 spx_uint32_t i;
825 st->magic_samples[channel_index] = tmp_magic-tmp_in_len;
826 for (i=0;i<st->magic_samples[channel_index];i++)
827 mem[N-1+i]=mem[N-1+i+tmp_in_len];
828 }
829 out += tmp_out_len*st->out_stride;
830 *out_len -= tmp_out_len;
831 }
832
833 /* Call the right resampler through the function ptr */
834 out_sample = st->resampler_ptr(st, channel_index, in, in_len, out, out_len);
835
836 if (st->last_sample[channel_index] < (spx_int32_t)*in_len)
837 *in_len = st->last_sample[channel_index];
838 *out_len = out_sample+tmp_out_len;
839 st->last_sample[channel_index] -= *in_len;
840
841 for (j=0;j<N-1-(spx_int32_t)*in_len;j++)
842 mem[j] = mem[j+*in_len];
843 for (;j<N-1;j++)
844 mem[j] = in[st->in_stride*(j+*in_len-N+1)];
845
846 return RESAMPLER_ERR_SUCCESS;
847}
848
849#define FIXED_STACK_ALLOC 1024
850
851#ifdef FIXED_POINT
852int speex_resampler_process_float(SpeexResamplerState *st, spx_uint32_t channel_index, const float *in, spx_uint32_t *in_len, float *out, spx_uint32_t *out_len)
853{
854 spx_uint32_t i;
855 int istride_save, ostride_save;
856#ifdef VAR_ARRAYS
857 spx_word16_t x[*in_len];
858 spx_word16_t y[*out_len];
859 /*VARDECL(spx_word16_t *x);
860 VARDECL(spx_word16_t *y);
861 ALLOC(x, *in_len, spx_word16_t);
862 ALLOC(y, *out_len, spx_word16_t);*/
863 istride_save = st->in_stride;
864 ostride_save = st->out_stride;
865 for (i=0;i<*in_len;i++)
866 x[i] = WORD2INT(in[i*st->in_stride]);
867 st->in_stride = st->out_stride = 1;
868 speex_resampler_process_native(st, channel_index, x, in_len, y, out_len);
869 st->in_stride = istride_save;
870 st->out_stride = ostride_save;
871 for (i=0;i<*out_len;i++)
872 out[i*st->out_stride] = y[i];
873#else
874 spx_word16_t x[FIXED_STACK_ALLOC];
875 spx_word16_t y[FIXED_STACK_ALLOC];
876 spx_uint32_t ilen=*in_len, olen=*out_len;
877 istride_save = st->in_stride;
878 ostride_save = st->out_stride;
879 while (ilen && olen)
880 {
881 spx_uint32_t ichunk, ochunk;
882 ichunk = ilen;
883 ochunk = olen;
884 if (ichunk>FIXED_STACK_ALLOC)
885 ichunk=FIXED_STACK_ALLOC;
886 if (ochunk>FIXED_STACK_ALLOC)
887 ochunk=FIXED_STACK_ALLOC;
888 for (i=0;i<ichunk;i++)
889 x[i] = WORD2INT(in[i*st->in_stride]);
890 st->in_stride = st->out_stride = 1;
891 speex_resampler_process_native(st, channel_index, x, &ichunk, y, &ochunk);
892 st->in_stride = istride_save;
893 st->out_stride = ostride_save;
894 for (i=0;i<ochunk;i++)
895 out[i*st->out_stride] = y[i];
896 out += ochunk;
897 in += ichunk;
898 ilen -= ichunk;
899 olen -= ochunk;
900 }
901 *in_len -= ilen;
902 *out_len -= olen;
903#endif
904 return RESAMPLER_ERR_SUCCESS;
905}
906int speex_resampler_process_int(SpeexResamplerState *st, spx_uint32_t channel_index, const spx_int16_t *in, spx_uint32_t *in_len, spx_int16_t *out, spx_uint32_t *out_len)
907{
908 return speex_resampler_process_native(st, channel_index, in, in_len, out, out_len);
909}
910#else
911int speex_resampler_process_float(SpeexResamplerState *st, spx_uint32_t channel_index, const float *in, spx_uint32_t *in_len, float *out, spx_uint32_t *out_len)
912{
913 return speex_resampler_process_native(st, channel_index, in, in_len, out, out_len);
914}
915int speex_resampler_process_int(SpeexResamplerState *st, spx_uint32_t channel_index, const spx_int16_t *in, spx_uint32_t *in_len, spx_int16_t *out, spx_uint32_t *out_len)
916{
917 spx_uint32_t i;
918 int istride_save, ostride_save;
919#ifdef VAR_ARRAYS
920 spx_word16_t x[*in_len];
921 spx_word16_t y[*out_len];
922 /*VARDECL(spx_word16_t *x);
923 VARDECL(spx_word16_t *y);
924 ALLOC(x, *in_len, spx_word16_t);
925 ALLOC(y, *out_len, spx_word16_t);*/
926 istride_save = st->in_stride;
927 ostride_save = st->out_stride;
928 for (i=0;i<*in_len;i++)
929 x[i] = in[i*st->in_stride];
930 st->in_stride = st->out_stride = 1;
931 speex_resampler_process_native(st, channel_index, x, in_len, y, out_len);
932 st->in_stride = istride_save;
933 st->out_stride = ostride_save;
934 for (i=0;i<*out_len;i++)
935 out[i*st->out_stride] = WORD2INT(y[i]);
936#else
937 spx_word16_t x[FIXED_STACK_ALLOC];
938 spx_word16_t y[FIXED_STACK_ALLOC];
939 spx_uint32_t ilen=*in_len, olen=*out_len;
940 istride_save = st->in_stride;
941 ostride_save = st->out_stride;
942 while (ilen && olen)
943 {
944 spx_uint32_t ichunk, ochunk;
945 ichunk = ilen;
946 ochunk = olen;
947 if (ichunk>FIXED_STACK_ALLOC)
948 ichunk=FIXED_STACK_ALLOC;
949 if (ochunk>FIXED_STACK_ALLOC)
950 ochunk=FIXED_STACK_ALLOC;
951 for (i=0;i<ichunk;i++)
952 x[i] = in[i*st->in_stride];
953 st->in_stride = st->out_stride = 1;
954 speex_resampler_process_native(st, channel_index, x, &ichunk, y, &ochunk);
955 st->in_stride = istride_save;
956 st->out_stride = ostride_save;
957 for (i=0;i<ochunk;i++)
958 out[i*st->out_stride] = WORD2INT(y[i]);
959 out += ochunk;
960 in += ichunk;
961 ilen -= ichunk;
962 olen -= ochunk;
963 }
964 *in_len -= ilen;
965 *out_len -= olen;
966#endif
967 return RESAMPLER_ERR_SUCCESS;
968}
969#endif
970
971int speex_resampler_process_interleaved_float(SpeexResamplerState *st, const float *in, spx_uint32_t *in_len, float *out, spx_uint32_t *out_len)
972{
973 spx_uint32_t i;
974 int istride_save, ostride_save;
975 spx_uint32_t bak_len = *out_len;
976 istride_save = st->in_stride;
977 ostride_save = st->out_stride;
978 st->in_stride = st->out_stride = st->nb_channels;
979 for (i=0;i<st->nb_channels;i++)
980 {
981 *out_len = bak_len;
982 if (in != NULL)
983 speex_resampler_process_float(st, i, in+i, in_len, out+i, out_len);
984 else
985 speex_resampler_process_float(st, i, NULL, in_len, out+i, out_len);
986 }
987 st->in_stride = istride_save;
988 st->out_stride = ostride_save;
989 return RESAMPLER_ERR_SUCCESS;
990}
991
992
993int speex_resampler_process_interleaved_int(SpeexResamplerState *st, const spx_int16_t *in, spx_uint32_t *in_len, spx_int16_t *out, spx_uint32_t *out_len)
994{
995 spx_uint32_t i;
996 int istride_save, ostride_save;
997 spx_uint32_t bak_len = *out_len;
998 istride_save = st->in_stride;
999 ostride_save = st->out_stride;
1000 st->in_stride = st->out_stride = st->nb_channels;
1001 for (i=0;i<st->nb_channels;i++)
1002 {
1003 *out_len = bak_len;
1004 if (in != NULL)
1005 speex_resampler_process_int(st, i, in+i, in_len, out+i, out_len);
1006 else
1007 speex_resampler_process_int(st, i, NULL, in_len, out+i, out_len);
1008 }
1009 st->in_stride = istride_save;
1010 st->out_stride = ostride_save;
1011 return RESAMPLER_ERR_SUCCESS;
1012}
1013
1014int speex_resampler_set_rate(SpeexResamplerState *st, spx_uint32_t in_rate, spx_uint32_t out_rate)
1015{
1016 return speex_resampler_set_rate_frac(st, in_rate, out_rate, in_rate, out_rate);
1017}
1018
1019void speex_resampler_get_rate(SpeexResamplerState *st, spx_uint32_t *in_rate, spx_uint32_t *out_rate)
1020{
1021 *in_rate = st->in_rate;
1022 *out_rate = st->out_rate;
1023}
1024
1025int speex_resampler_set_rate_frac(SpeexResamplerState *st, spx_uint32_t ratio_num, spx_uint32_t ratio_den, spx_uint32_t in_rate, spx_uint32_t out_rate)
1026{
1027 spx_uint32_t fact;
1028 spx_uint32_t old_den;
1029 spx_uint32_t i;
1030 if (st->in_rate == in_rate && st->out_rate == out_rate && st->num_rate == ratio_num && st->den_rate == ratio_den)
1031 return RESAMPLER_ERR_SUCCESS;
1032
1033 old_den = st->den_rate;
1034 st->in_rate = in_rate;
1035 st->out_rate = out_rate;
1036 st->num_rate = ratio_num;
1037 st->den_rate = ratio_den;
1038 /* FIXME: This is terribly inefficient, but who cares (at least for now)? */
1039 for (fact=2;fact<=IMIN(st->num_rate, st->den_rate);fact++)
1040 {
1041 while ((st->num_rate % fact == 0) && (st->den_rate % fact == 0))
1042 {
1043 st->num_rate /= fact;
1044 st->den_rate /= fact;
1045 }
1046 }
1047
1048 if (old_den > 0)
1049 {
1050 for (i=0;i<st->nb_channels;i++)
1051 {
1052 st->samp_frac_num[i]=st->samp_frac_num[i]*st->den_rate/old_den;
1053 /* Safety net */
1054 if (st->samp_frac_num[i] >= st->den_rate)
1055 st->samp_frac_num[i] = st->den_rate-1;
1056 }
1057 }
1058
1059 if (st->initialised)
1060 update_filter(st);
1061 return RESAMPLER_ERR_SUCCESS;
1062}
1063
1064void speex_resampler_get_ratio(SpeexResamplerState *st, spx_uint32_t *ratio_num, spx_uint32_t *ratio_den)
1065{
1066 *ratio_num = st->num_rate;
1067 *ratio_den = st->den_rate;
1068}
1069
1070int speex_resampler_set_quality(SpeexResamplerState *st, int quality)
1071{
1072 if (quality > 10 || quality < 0)
1073 return RESAMPLER_ERR_INVALID_ARG;
1074 if (st->quality == quality)
1075 return RESAMPLER_ERR_SUCCESS;
1076 st->quality = quality;
1077 if (st->initialised)
1078 update_filter(st);
1079 return RESAMPLER_ERR_SUCCESS;
1080}
1081
1082void speex_resampler_get_quality(SpeexResamplerState *st, int *quality)
1083{
1084 *quality = st->quality;
1085}
1086
1087void speex_resampler_set_input_stride(SpeexResamplerState *st, spx_uint32_t stride)
1088{
1089 st->in_stride = stride;
1090}
1091
1092void speex_resampler_get_input_stride(SpeexResamplerState *st, spx_uint32_t *stride)
1093{
1094 *stride = st->in_stride;
1095}
1096
1097void speex_resampler_set_output_stride(SpeexResamplerState *st, spx_uint32_t stride)
1098{
1099 st->out_stride = stride;
1100}
1101
1102void speex_resampler_get_output_stride(SpeexResamplerState *st, spx_uint32_t *stride)
1103{
1104 *stride = st->out_stride;
1105}
1106
1107int speex_resampler_skip_zeros(SpeexResamplerState *st)
1108{
1109 spx_uint32_t i;
1110 for (i=0;i<st->nb_channels;i++)
1111 st->last_sample[i] = st->filt_len/2;
1112 return RESAMPLER_ERR_SUCCESS;
1113}
1114
1115int speex_resampler_reset_mem(SpeexResamplerState *st)
1116{
1117 spx_uint32_t i;
1118 for (i=0;i<st->nb_channels*(st->filt_len-1);i++)
1119 st->mem[i] = 0;
1120 return RESAMPLER_ERR_SUCCESS;
1121}
1122
1123const char *speex_resampler_strerror(int err)
1124{
1125 switch (err)
1126 {
1127 case RESAMPLER_ERR_SUCCESS:
1128 return "Success.";
1129 case RESAMPLER_ERR_ALLOC_FAILED:
1130 return "Memory allocation failed.";
1131 case RESAMPLER_ERR_BAD_STATE:
1132 return "Bad resampler state.";
1133 case RESAMPLER_ERR_INVALID_ARG:
1134 return "Invalid argument.";
1135 case RESAMPLER_ERR_PTR_OVERLAP:
1136 return "Input and output buffers overlap.";
1137 default:
1138 return "Unknown error. Bad error code or strange version mismatch.";
1139 }
1140}