Imported Debian version 2.4.3~trusty1
[deb_ffmpeg.git] / ffmpeg / libswresample / resample.c
CommitLineData
2ba45a60
DM
1/*
2 * audio resampling
3 * Copyright (c) 2004-2012 Michael Niedermayer <michaelni@gmx.at>
4 *
5 * This file is part of FFmpeg.
6 *
7 * FFmpeg is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
11 *
12 * FFmpeg is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
16 *
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with FFmpeg; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20 */
21
22/**
23 * @file
24 * audio resampling
25 * @author Michael Niedermayer <michaelni@gmx.at>
26 */
27
28#include "libavutil/avassert.h"
29#include "resample.h"
30
31/**
32 * 0th order modified bessel function of the first kind.
33 */
34static double bessel(double x){
35 double v=1;
36 double lastv=0;
37 double t=1;
38 int i;
39 static const double inv[100]={
40 1.0/( 1* 1), 1.0/( 2* 2), 1.0/( 3* 3), 1.0/( 4* 4), 1.0/( 5* 5), 1.0/( 6* 6), 1.0/( 7* 7), 1.0/( 8* 8), 1.0/( 9* 9), 1.0/(10*10),
41 1.0/(11*11), 1.0/(12*12), 1.0/(13*13), 1.0/(14*14), 1.0/(15*15), 1.0/(16*16), 1.0/(17*17), 1.0/(18*18), 1.0/(19*19), 1.0/(20*20),
42 1.0/(21*21), 1.0/(22*22), 1.0/(23*23), 1.0/(24*24), 1.0/(25*25), 1.0/(26*26), 1.0/(27*27), 1.0/(28*28), 1.0/(29*29), 1.0/(30*30),
43 1.0/(31*31), 1.0/(32*32), 1.0/(33*33), 1.0/(34*34), 1.0/(35*35), 1.0/(36*36), 1.0/(37*37), 1.0/(38*38), 1.0/(39*39), 1.0/(40*40),
44 1.0/(41*41), 1.0/(42*42), 1.0/(43*43), 1.0/(44*44), 1.0/(45*45), 1.0/(46*46), 1.0/(47*47), 1.0/(48*48), 1.0/(49*49), 1.0/(50*50),
45 1.0/(51*51), 1.0/(52*52), 1.0/(53*53), 1.0/(54*54), 1.0/(55*55), 1.0/(56*56), 1.0/(57*57), 1.0/(58*58), 1.0/(59*59), 1.0/(60*60),
46 1.0/(61*61), 1.0/(62*62), 1.0/(63*63), 1.0/(64*64), 1.0/(65*65), 1.0/(66*66), 1.0/(67*67), 1.0/(68*68), 1.0/(69*69), 1.0/(70*70),
47 1.0/(71*71), 1.0/(72*72), 1.0/(73*73), 1.0/(74*74), 1.0/(75*75), 1.0/(76*76), 1.0/(77*77), 1.0/(78*78), 1.0/(79*79), 1.0/(80*80),
48 1.0/(81*81), 1.0/(82*82), 1.0/(83*83), 1.0/(84*84), 1.0/(85*85), 1.0/(86*86), 1.0/(87*87), 1.0/(88*88), 1.0/(89*89), 1.0/(90*90),
49 1.0/(91*91), 1.0/(92*92), 1.0/(93*93), 1.0/(94*94), 1.0/(95*95), 1.0/(96*96), 1.0/(97*97), 1.0/(98*98), 1.0/(99*99), 1.0/(10000)
50 };
51
52 x= x*x/4;
53 for(i=0; v != lastv; i++){
54 lastv=v;
55 t *= x*inv[i];
56 v += t;
57 av_assert2(i<99);
58 }
59 return v;
60}
61
62/**
63 * builds a polyphase filterbank.
64 * @param factor resampling factor
65 * @param scale wanted sum of coefficients for each filter
66 * @param filter_type filter type
67 * @param kaiser_beta kaiser window beta
68 * @return 0 on success, negative on error
69 */
70static int build_filter(ResampleContext *c, void *filter, double factor, int tap_count, int alloc, int phase_count, int scale,
71 int filter_type, int kaiser_beta){
72 int ph, i;
73 double x, y, w;
74 double *tab = av_malloc_array(tap_count, sizeof(*tab));
75 const int center= (tap_count-1)/2;
76
77 if (!tab)
78 return AVERROR(ENOMEM);
79
80 /* if upsampling, only need to interpolate, no filter */
81 if (factor > 1.0)
82 factor = 1.0;
83
84 for(ph=0;ph<phase_count;ph++) {
85 double norm = 0;
86 for(i=0;i<tap_count;i++) {
87 x = M_PI * ((double)(i - center) - (double)ph / phase_count) * factor;
88 if (x == 0) y = 1.0;
89 else y = sin(x) / x;
90 switch(filter_type){
91 case SWR_FILTER_TYPE_CUBIC:{
92 const float d= -0.5; //first order derivative = -0.5
93 x = fabs(((double)(i - center) - (double)ph / phase_count) * factor);
94 if(x<1.0) y= 1 - 3*x*x + 2*x*x*x + d*( -x*x + x*x*x);
95 else y= d*(-4 + 8*x - 5*x*x + x*x*x);
96 break;}
97 case SWR_FILTER_TYPE_BLACKMAN_NUTTALL:
98 w = 2.0*x / (factor*tap_count) + M_PI;
99 y *= 0.3635819 - 0.4891775 * cos(w) + 0.1365995 * cos(2*w) - 0.0106411 * cos(3*w);
100 break;
101 case SWR_FILTER_TYPE_KAISER:
102 w = 2.0*x / (factor*tap_count*M_PI);
103 y *= bessel(kaiser_beta*sqrt(FFMAX(1-w*w, 0)));
104 break;
105 default:
106 av_assert0(0);
107 }
108
109 tab[i] = y;
110 norm += y;
111 }
112
113 /* normalize so that an uniform color remains the same */
114 switch(c->format){
115 case AV_SAMPLE_FMT_S16P:
116 for(i=0;i<tap_count;i++)
117 ((int16_t*)filter)[ph * alloc + i] = av_clip(lrintf(tab[i] * scale / norm), INT16_MIN, INT16_MAX);
118 break;
119 case AV_SAMPLE_FMT_S32P:
120 for(i=0;i<tap_count;i++)
121 ((int32_t*)filter)[ph * alloc + i] = av_clipl_int32(llrint(tab[i] * scale / norm));
122 break;
123 case AV_SAMPLE_FMT_FLTP:
124 for(i=0;i<tap_count;i++)
125 ((float*)filter)[ph * alloc + i] = tab[i] * scale / norm;
126 break;
127 case AV_SAMPLE_FMT_DBLP:
128 for(i=0;i<tap_count;i++)
129 ((double*)filter)[ph * alloc + i] = tab[i] * scale / norm;
130 break;
131 }
132 }
133#if 0
134 {
135#define LEN 1024
136 int j,k;
137 double sine[LEN + tap_count];
138 double filtered[LEN];
139 double maxff=-2, minff=2, maxsf=-2, minsf=2;
140 for(i=0; i<LEN; i++){
141 double ss=0, sf=0, ff=0;
142 for(j=0; j<LEN+tap_count; j++)
143 sine[j]= cos(i*j*M_PI/LEN);
144 for(j=0; j<LEN; j++){
145 double sum=0;
146 ph=0;
147 for(k=0; k<tap_count; k++)
148 sum += filter[ph * tap_count + k] * sine[k+j];
149 filtered[j]= sum / (1<<FILTER_SHIFT);
150 ss+= sine[j + center] * sine[j + center];
151 ff+= filtered[j] * filtered[j];
152 sf+= sine[j + center] * filtered[j];
153 }
154 ss= sqrt(2*ss/LEN);
155 ff= sqrt(2*ff/LEN);
156 sf= 2*sf/LEN;
157 maxff= FFMAX(maxff, ff);
158 minff= FFMIN(minff, ff);
159 maxsf= FFMAX(maxsf, sf);
160 minsf= FFMIN(minsf, sf);
161 if(i%11==0){
162 av_log(NULL, AV_LOG_ERROR, "i:%4d ss:%f ff:%13.6e-%13.6e sf:%13.6e-%13.6e\n", i, ss, maxff, minff, maxsf, minsf);
163 minff=minsf= 2;
164 maxff=maxsf= -2;
165 }
166 }
167 }
168#endif
169
170 av_free(tab);
171 return 0;
172}
173
174static ResampleContext *resample_init(ResampleContext *c, int out_rate, int in_rate, int filter_size, int phase_shift, int linear,
175 double cutoff0, enum AVSampleFormat format, enum SwrFilterType filter_type, int kaiser_beta,
176 double precision, int cheby)
177{
178 double cutoff = cutoff0? cutoff0 : 0.97;
179 double factor= FFMIN(out_rate * cutoff / in_rate, 1.0);
180 int phase_count= 1<<phase_shift;
181
182 if (!c || c->phase_shift != phase_shift || c->linear!=linear || c->factor != factor
183 || c->filter_length != FFMAX((int)ceil(filter_size/factor), 1) || c->format != format
184 || c->filter_type != filter_type || c->kaiser_beta != kaiser_beta) {
185 c = av_mallocz(sizeof(*c));
186 if (!c)
187 return NULL;
188
189 c->format= format;
190
191 c->felem_size= av_get_bytes_per_sample(c->format);
192
193 switch(c->format){
194 case AV_SAMPLE_FMT_S16P:
195 c->filter_shift = 15;
196 break;
197 case AV_SAMPLE_FMT_S32P:
198 c->filter_shift = 30;
199 break;
200 case AV_SAMPLE_FMT_FLTP:
201 case AV_SAMPLE_FMT_DBLP:
202 c->filter_shift = 0;
203 break;
204 default:
205 av_log(NULL, AV_LOG_ERROR, "Unsupported sample format\n");
206 av_assert0(0);
207 }
208
209 if (filter_size/factor > INT32_MAX/256) {
210 av_log(NULL, AV_LOG_ERROR, "Filter length too large\n");
211 goto error;
212 }
213
214 c->phase_shift = phase_shift;
215 c->phase_mask = phase_count - 1;
216 c->linear = linear;
217 c->factor = factor;
218 c->filter_length = FFMAX((int)ceil(filter_size/factor), 1);
219 c->filter_alloc = FFALIGN(c->filter_length, 8);
220 c->filter_bank = av_calloc(c->filter_alloc, (phase_count+1)*c->felem_size);
221 c->filter_type = filter_type;
222 c->kaiser_beta = kaiser_beta;
223 if (!c->filter_bank)
224 goto error;
225 if (build_filter(c, (void*)c->filter_bank, factor, c->filter_length, c->filter_alloc, phase_count, 1<<c->filter_shift, filter_type, kaiser_beta))
226 goto error;
227 memcpy(c->filter_bank + (c->filter_alloc*phase_count+1)*c->felem_size, c->filter_bank, (c->filter_alloc-1)*c->felem_size);
228 memcpy(c->filter_bank + (c->filter_alloc*phase_count )*c->felem_size, c->filter_bank + (c->filter_alloc - 1)*c->felem_size, c->felem_size);
229 }
230
231 c->compensation_distance= 0;
232 if(!av_reduce(&c->src_incr, &c->dst_incr, out_rate, in_rate * (int64_t)phase_count, INT32_MAX/2))
233 goto error;
234 c->ideal_dst_incr = c->dst_incr;
235 c->dst_incr_div = c->dst_incr / c->src_incr;
236 c->dst_incr_mod = c->dst_incr % c->src_incr;
237
238 c->index= -phase_count*((c->filter_length-1)/2);
239 c->frac= 0;
240
241 swri_resample_dsp_init(c);
242
243 return c;
244error:
245 av_freep(&c->filter_bank);
246 av_free(c);
247 return NULL;
248}
249
250static void resample_free(ResampleContext **c){
251 if(!*c)
252 return;
253 av_freep(&(*c)->filter_bank);
254 av_freep(c);
255}
256
257static int set_compensation(ResampleContext *c, int sample_delta, int compensation_distance){
258 c->compensation_distance= compensation_distance;
259 if (compensation_distance)
260 c->dst_incr = c->ideal_dst_incr - c->ideal_dst_incr * (int64_t)sample_delta / compensation_distance;
261 else
262 c->dst_incr = c->ideal_dst_incr;
263
264 c->dst_incr_div = c->dst_incr / c->src_incr;
265 c->dst_incr_mod = c->dst_incr % c->src_incr;
266
267 return 0;
268}
269
270static int swri_resample(ResampleContext *c,
271 uint8_t *dst, const uint8_t *src, int *consumed,
272 int src_size, int dst_size, int update_ctx)
273{
274 if (c->filter_length == 1 && c->phase_shift == 0) {
275 int index= c->index;
276 int frac= c->frac;
277 int64_t index2= (1LL<<32)*c->frac/c->src_incr + (1LL<<32)*index;
278 int64_t incr= (1LL<<32) * c->dst_incr / c->src_incr;
279 int new_size = (src_size * (int64_t)c->src_incr - frac + c->dst_incr - 1) / c->dst_incr;
280
281 dst_size= FFMIN(dst_size, new_size);
282 c->dsp.resample_one(dst, src, dst_size, index2, incr);
283
284 index += dst_size * c->dst_incr_div;
285 index += (frac + dst_size * (int64_t)c->dst_incr_mod) / c->src_incr;
286 av_assert2(index >= 0);
287 *consumed= index;
288 if (update_ctx) {
289 c->frac = (frac + dst_size * (int64_t)c->dst_incr_mod) % c->src_incr;
290 c->index = 0;
291 }
292 } else {
293 int64_t end_index = (1LL + src_size - c->filter_length) << c->phase_shift;
294 int64_t delta_frac = (end_index - c->index) * c->src_incr - c->frac;
295 int delta_n = (delta_frac + c->dst_incr - 1) / c->dst_incr;
296
297 dst_size = FFMIN(dst_size, delta_n);
298 if (dst_size > 0) {
299 *consumed = c->dsp.resample(c, dst, src, dst_size, update_ctx);
300 } else {
301 *consumed = 0;
302 }
303 }
304
305 return dst_size;
306}
307
308static int multiple_resample(ResampleContext *c, AudioData *dst, int dst_size, AudioData *src, int src_size, int *consumed){
309 int i, ret= -1;
310 int av_unused mm_flags = av_get_cpu_flags();
311 int need_emms = c->format == AV_SAMPLE_FMT_S16P && ARCH_X86_32 &&
312 (mm_flags & (AV_CPU_FLAG_MMX2 | AV_CPU_FLAG_SSE2)) == AV_CPU_FLAG_MMX2;
313 int64_t max_src_size = (INT64_MAX >> (c->phase_shift+1)) / c->src_incr;
314
315 if (c->compensation_distance)
316 dst_size = FFMIN(dst_size, c->compensation_distance);
317 src_size = FFMIN(src_size, max_src_size);
318
319 for(i=0; i<dst->ch_count; i++){
320 ret= swri_resample(c, dst->ch[i], src->ch[i],
321 consumed, src_size, dst_size, i+1==dst->ch_count);
322 }
323 if(need_emms)
324 emms_c();
325
326 if (c->compensation_distance) {
327 c->compensation_distance -= ret;
328 if (!c->compensation_distance) {
329 c->dst_incr = c->ideal_dst_incr;
330 c->dst_incr_div = c->dst_incr / c->src_incr;
331 c->dst_incr_mod = c->dst_incr % c->src_incr;
332 }
333 }
334
335 return ret;
336}
337
338static int64_t get_delay(struct SwrContext *s, int64_t base){
339 ResampleContext *c = s->resample;
340 int64_t num = s->in_buffer_count - (c->filter_length-1)/2;
341 num <<= c->phase_shift;
342 num -= c->index;
343 num *= c->src_incr;
344 num -= c->frac;
345 return av_rescale(num, base, s->in_sample_rate*(int64_t)c->src_incr << c->phase_shift);
346}
347
348static int resample_flush(struct SwrContext *s) {
349 AudioData *a= &s->in_buffer;
350 int i, j, ret;
351 if((ret = swri_realloc_audio(a, s->in_buffer_index + 2*s->in_buffer_count)) < 0)
352 return ret;
353 av_assert0(a->planar);
354 for(i=0; i<a->ch_count; i++){
355 for(j=0; j<s->in_buffer_count; j++){
356 memcpy(a->ch[i] + (s->in_buffer_index+s->in_buffer_count+j )*a->bps,
357 a->ch[i] + (s->in_buffer_index+s->in_buffer_count-j-1)*a->bps, a->bps);
358 }
359 }
360 s->in_buffer_count += (s->in_buffer_count+1)/2;
361 return 0;
362}
363
364// in fact the whole handle multiple ridiculously small buffers might need more thinking...
365static int invert_initial_buffer(ResampleContext *c, AudioData *dst, const AudioData *src,
366 int in_count, int *out_idx, int *out_sz)
367{
368 int n, ch, num = FFMIN(in_count + *out_sz, c->filter_length + 1), res;
369
370 if (c->index >= 0)
371 return 0;
372
373 if ((res = swri_realloc_audio(dst, c->filter_length * 2 + 1)) < 0)
374 return res;
375
376 // copy
377 for (n = *out_sz; n < num; n++) {
378 for (ch = 0; ch < src->ch_count; ch++) {
379 memcpy(dst->ch[ch] + ((c->filter_length + n) * c->felem_size),
380 src->ch[ch] + ((n - *out_sz) * c->felem_size), c->felem_size);
381 }
382 }
383
384 // if not enough data is in, return and wait for more
385 if (num < c->filter_length + 1) {
386 *out_sz = num;
387 *out_idx = c->filter_length;
388 return INT_MAX;
389 }
390
391 // else invert
392 for (n = 1; n <= c->filter_length; n++) {
393 for (ch = 0; ch < src->ch_count; ch++) {
394 memcpy(dst->ch[ch] + ((c->filter_length - n) * c->felem_size),
395 dst->ch[ch] + ((c->filter_length + n) * c->felem_size),
396 c->felem_size);
397 }
398 }
399
400 res = num - *out_sz;
401 *out_idx = c->filter_length + (c->index >> c->phase_shift);
402 *out_sz = 1 + c->filter_length * 2 - *out_idx;
403 c->index &= c->phase_mask;
404 av_assert1(res > 0);
405
406 return res;
407}
408
409struct Resampler const swri_resampler={
410 resample_init,
411 resample_free,
412 multiple_resample,
413 resample_flush,
414 set_compensation,
415 get_delay,
416 invert_initial_buffer,
417};