Line | Count | Source |
1 | | /* Copyright (c) 2007-2008 CSIRO |
2 | | Copyright (c) 2007-2009 Xiph.Org Foundation |
3 | | Written by Jean-Marc Valin */ |
4 | | /** |
5 | | @file pitch.c |
6 | | @brief Pitch analysis |
7 | | */ |
8 | | |
9 | | /* |
10 | | Redistribution and use in source and binary forms, with or without |
11 | | modification, are permitted provided that the following conditions |
12 | | are met: |
13 | | |
14 | | - Redistributions of source code must retain the above copyright |
15 | | notice, this list of conditions and the following disclaimer. |
16 | | |
17 | | - Redistributions in binary form must reproduce the above copyright |
18 | | notice, this list of conditions and the following disclaimer in the |
19 | | documentation and/or other materials provided with the distribution. |
20 | | |
21 | | THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS |
22 | | ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT |
23 | | LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR |
24 | | A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER |
25 | | OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, |
26 | | EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, |
27 | | PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR |
28 | | PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF |
29 | | LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING |
30 | | NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS |
31 | | SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
32 | | */ |
33 | | |
34 | | #ifdef HAVE_CONFIG_H |
35 | | #include "config.h" |
36 | | #endif |
37 | | |
38 | | #include "pitch.h" |
39 | | #include "os_support.h" |
40 | | #include "modes.h" |
41 | | #include "stack_alloc.h" |
42 | | #include "mathops.h" |
43 | | #include "celt_lpc.h" |
44 | | |
45 | | static void find_best_pitch(opus_val32 *xcorr, opus_val16 *y, int len, |
46 | | int max_pitch, int *best_pitch |
47 | | #ifdef FIXED_POINT |
48 | | , int yshift, opus_val32 maxcorr |
49 | | #endif |
50 | | ) |
51 | 36.2k | { |
52 | 36.2k | int i, j; |
53 | 36.2k | opus_val32 Syy=1; |
54 | 36.2k | opus_val16 best_num[2]; |
55 | 36.2k | opus_val32 best_den[2]; |
56 | | #ifdef FIXED_POINT |
57 | | int xshift; |
58 | | |
59 | | xshift = celt_ilog2(maxcorr)-14; |
60 | | #endif |
61 | | |
62 | 36.2k | best_num[0] = -1; |
63 | 36.2k | best_num[1] = -1; |
64 | 36.2k | best_den[0] = 0; |
65 | 36.2k | best_den[1] = 0; |
66 | 36.2k | best_pitch[0] = 0; |
67 | 36.2k | best_pitch[1] = 1; |
68 | 18.0M | for (j=0;j<len;j++) |
69 | 18.0M | Syy = ADD32(Syy, SHR32(MULT16_16(y[j],y[j]), yshift)); |
70 | 8.45M | for (i=0;i<max_pitch;i++) |
71 | 8.42M | { |
72 | 8.42M | if (xcorr[i]>0) |
73 | 1.19M | { |
74 | 1.19M | opus_val16 num; |
75 | 1.19M | opus_val32 xcorr16; |
76 | 1.19M | xcorr16 = EXTRACT16(VSHR32(xcorr[i], xshift)); |
77 | 1.19M | #ifndef FIXED_POINT |
78 | | /* Considering the range of xcorr16, this should avoid both underflows |
79 | | and overflows (inf) when squaring xcorr16 */ |
80 | 1.19M | xcorr16 *= 1e-12f; |
81 | 1.19M | #endif |
82 | 1.19M | num = MULT16_16_Q15(xcorr16,xcorr16); |
83 | 1.19M | if (MULT16_32_Q15(num,best_den[1]) > MULT16_32_Q15(best_num[1],Syy)) |
84 | 303k | { |
85 | 303k | if (MULT16_32_Q15(num,best_den[0]) > MULT16_32_Q15(best_num[0],Syy)) |
86 | 215k | { |
87 | 215k | best_num[1] = best_num[0]; |
88 | 215k | best_den[1] = best_den[0]; |
89 | 215k | best_pitch[1] = best_pitch[0]; |
90 | 215k | best_num[0] = num; |
91 | 215k | best_den[0] = Syy; |
92 | 215k | best_pitch[0] = i; |
93 | 215k | } else { |
94 | 87.6k | best_num[1] = num; |
95 | 87.6k | best_den[1] = Syy; |
96 | 87.6k | best_pitch[1] = i; |
97 | 87.6k | } |
98 | 303k | } |
99 | 1.19M | } |
100 | 8.42M | Syy += SHR32(MULT16_16(y[i+len],y[i+len]),yshift) - SHR32(MULT16_16(y[i],y[i]),yshift); |
101 | 8.42M | Syy = MAX32(1, Syy); |
102 | 8.42M | } |
103 | 36.2k | } |
104 | | |
105 | | static void celt_fir5(opus_val16 *x, |
106 | | const opus_val16 *num, |
107 | | int N) |
108 | 18.1k | { |
109 | 18.1k | int i; |
110 | 18.1k | opus_val16 num0, num1, num2, num3, num4; |
111 | 18.1k | opus_val32 mem0, mem1, mem2, mem3, mem4; |
112 | 18.1k | num0=num[0]; |
113 | 18.1k | num1=num[1]; |
114 | 18.1k | num2=num[2]; |
115 | 18.1k | num3=num[3]; |
116 | 18.1k | num4=num[4]; |
117 | 18.1k | mem0=0; |
118 | 18.1k | mem1=0; |
119 | 18.1k | mem2=0; |
120 | 18.1k | mem3=0; |
121 | 18.1k | mem4=0; |
122 | 18.5M | for (i=0;i<N;i++) |
123 | 18.5M | { |
124 | 18.5M | opus_val32 sum = SHL32(EXTEND32(x[i]), SIG_SHIFT); |
125 | 18.5M | sum = MAC16_16(sum,num0,mem0); |
126 | 18.5M | sum = MAC16_16(sum,num1,mem1); |
127 | 18.5M | sum = MAC16_16(sum,num2,mem2); |
128 | 18.5M | sum = MAC16_16(sum,num3,mem3); |
129 | 18.5M | sum = MAC16_16(sum,num4,mem4); |
130 | 18.5M | mem4 = mem3; |
131 | 18.5M | mem3 = mem2; |
132 | 18.5M | mem2 = mem1; |
133 | 18.5M | mem1 = mem0; |
134 | 18.5M | mem0 = x[i]; |
135 | 18.5M | x[i] = ROUND16(sum, SIG_SHIFT); |
136 | 18.5M | } |
137 | 18.1k | } |
138 | | |
139 | | |
140 | | void pitch_downsample(celt_sig * OPUS_RESTRICT x[], opus_val16 * OPUS_RESTRICT x_lp, |
141 | | int len, int C, int factor, int arch) |
142 | 18.1k | { |
143 | 18.1k | int i; |
144 | 18.1k | opus_val32 ac[5]; |
145 | 18.1k | opus_val16 tmp=Q15ONE; |
146 | 18.1k | opus_val16 lpc[4]; |
147 | 18.1k | opus_val16 lpc2[5]; |
148 | 18.1k | opus_val16 c1 = QCONST16(.8f,15); |
149 | 18.1k | int offset; |
150 | | #ifdef FIXED_POINT |
151 | | int shift; |
152 | | opus_val32 maxabs; |
153 | | #endif |
154 | 18.1k | offset = factor/2; |
155 | | #ifdef FIXED_POINT |
156 | | maxabs = celt_maxabs32(x[0], len*factor); |
157 | | if (C==2) |
158 | | { |
159 | | opus_val32 maxabs_1 = celt_maxabs32(x[1], len*factor); |
160 | | maxabs = MAX32(maxabs, maxabs_1); |
161 | | } |
162 | | if (maxabs<1) |
163 | | maxabs=1; |
164 | | shift = celt_ilog2(maxabs)-10; |
165 | | if (shift<0) |
166 | | shift=0; |
167 | | if (C==2) |
168 | | shift++; |
169 | | for (i=1;i<len;i++) |
170 | | x_lp[i] = SHR32(x[0][(factor*i-offset)], shift+2) + SHR32(x[0][(factor*i+offset)], shift+2) + SHR32(x[0][factor*i], shift+1); |
171 | | x_lp[0] = SHR32(x[0][offset], shift+2) + SHR32(x[0][0], shift+1); |
172 | | if (C==2) |
173 | | { |
174 | | for (i=1;i<len;i++) |
175 | | x_lp[i] += SHR32(x[1][(factor*i-offset)], shift+2) + SHR32(x[1][(factor*i+offset)], shift+2) + SHR32(x[1][factor*i], shift+1); |
176 | | x_lp[0] += SHR32(x[1][offset], shift+2) + SHR32(x[1][0], shift+1); |
177 | | } |
178 | | #else |
179 | 18.5M | for (i=1;i<len;i++) |
180 | 18.5M | x_lp[i] = .25f*x[0][(factor*i-offset)] + .25f*x[0][(factor*i+offset)] + .5f*x[0][factor*i]; |
181 | 18.1k | x_lp[0] = .25f*x[0][offset] + .5f*x[0][0]; |
182 | 18.1k | if (C==2) |
183 | 14.9k | { |
184 | 15.2M | for (i=1;i<len;i++) |
185 | 15.2M | x_lp[i] += .25f*x[1][(factor*i-offset)] + .25f*x[1][(factor*i+offset)] + .5f*x[1][factor*i]; |
186 | 14.9k | x_lp[0] += .25f*x[1][offset] + .5f*x[1][0]; |
187 | 14.9k | } |
188 | 18.1k | #endif |
189 | 18.1k | _celt_autocorr(x_lp, ac, NULL, 0, |
190 | 18.1k | 4, len, arch); |
191 | | |
192 | | /* Noise floor -40 dB */ |
193 | | #ifdef FIXED_POINT |
194 | | ac[0] += SHR32(ac[0],13); |
195 | | #else |
196 | 18.1k | ac[0] *= 1.0001f; |
197 | 18.1k | #endif |
198 | | /* Lag windowing */ |
199 | 90.5k | for (i=1;i<=4;i++) |
200 | 72.4k | { |
201 | | /*ac[i] *= exp(-.5*(2*M_PI*.002*i)*(2*M_PI*.002*i));*/ |
202 | | #ifdef FIXED_POINT |
203 | | ac[i] -= MULT16_32_Q15(2*i*i, ac[i]); |
204 | | #else |
205 | 72.4k | ac[i] -= ac[i]*(.008f*i)*(.008f*i); |
206 | 72.4k | #endif |
207 | 72.4k | } |
208 | | |
209 | 18.1k | _celt_lpc(lpc, ac, 4); |
210 | 90.5k | for (i=0;i<4;i++) |
211 | 72.4k | { |
212 | 72.4k | tmp = MULT16_16_Q15(QCONST16(.9f,15), tmp); |
213 | 72.4k | lpc[i] = MULT16_16_Q15(lpc[i], tmp); |
214 | 72.4k | } |
215 | | /* Add a zero */ |
216 | 18.1k | lpc2[0] = lpc[0] + QCONST16(.8f,SIG_SHIFT); |
217 | 18.1k | lpc2[1] = lpc[1] + MULT16_16_Q15(c1,lpc[0]); |
218 | 18.1k | lpc2[2] = lpc[2] + MULT16_16_Q15(c1,lpc[1]); |
219 | 18.1k | lpc2[3] = lpc[3] + MULT16_16_Q15(c1,lpc[2]); |
220 | 18.1k | lpc2[4] = MULT16_16_Q15(c1,lpc[3]); |
221 | 18.1k | celt_fir5(x_lp, lpc2, len); |
222 | 18.1k | } |
223 | | |
224 | | /* Pure C implementation. */ |
225 | | #ifdef FIXED_POINT |
226 | | opus_val32 |
227 | | #else |
228 | | void |
229 | | #endif |
230 | | celt_pitch_xcorr_c(const opus_val16 *_x, const opus_val16 *_y, |
231 | | opus_val32 *xcorr, int len, int max_pitch, int arch) |
232 | 0 | { |
233 | |
|
234 | | #if 0 /* This is a simple version of the pitch correlation that should work |
235 | | well on DSPs like Blackfin and TI C5x/C6x */ |
236 | | int i, j; |
237 | | #ifdef FIXED_POINT |
238 | | opus_val32 maxcorr=1; |
239 | | #endif |
240 | | #if !defined(OVERRIDE_PITCH_XCORR) |
241 | | (void)arch; |
242 | | #endif |
243 | | for (i=0;i<max_pitch;i++) |
244 | | { |
245 | | opus_val32 sum = 0; |
246 | | for (j=0;j<len;j++) |
247 | | sum = MAC16_16(sum, _x[j], _y[i+j]); |
248 | | xcorr[i] = sum; |
249 | | #ifdef FIXED_POINT |
250 | | maxcorr = MAX32(maxcorr, sum); |
251 | | #endif |
252 | | } |
253 | | #ifdef FIXED_POINT |
254 | | return maxcorr; |
255 | | #endif |
256 | | |
257 | | #else /* Unrolled version of the pitch correlation -- runs faster on x86 and ARM */ |
258 | 0 | int i; |
259 | | /*The EDSP version requires that max_pitch is at least 1, and that _x is |
260 | | 32-bit aligned. |
261 | | Since it's hard to put asserts in assembly, put them here.*/ |
262 | | #ifdef FIXED_POINT |
263 | | opus_val32 maxcorr=1; |
264 | | #endif |
265 | 0 | celt_assert(max_pitch>0); |
266 | 0 | celt_sig_assert(((size_t)_x&3)==0); |
267 | 0 | for (i=0;i<max_pitch-3;i+=4) |
268 | 0 | { |
269 | 0 | opus_val32 sum[4]={0,0,0,0}; |
270 | | #if defined(OPUS_CHECK_ASM) && defined(FIXED_POINT) |
271 | | { |
272 | | opus_val32 sum_c[4]={0,0,0,0}; |
273 | | xcorr_kernel_c(_x, _y+i, sum_c, len); |
274 | | #endif |
275 | 0 | xcorr_kernel(_x, _y+i, sum, len, arch); |
276 | | #if defined(OPUS_CHECK_ASM) && defined(FIXED_POINT) |
277 | | celt_assert(memcmp(sum, sum_c, sizeof(sum)) == 0); |
278 | | } |
279 | | #endif |
280 | 0 | xcorr[i]=sum[0]; |
281 | 0 | xcorr[i+1]=sum[1]; |
282 | 0 | xcorr[i+2]=sum[2]; |
283 | 0 | xcorr[i+3]=sum[3]; |
284 | | #ifdef FIXED_POINT |
285 | | sum[0] = MAX32(sum[0], sum[1]); |
286 | | sum[2] = MAX32(sum[2], sum[3]); |
287 | | sum[0] = MAX32(sum[0], sum[2]); |
288 | | maxcorr = MAX32(maxcorr, sum[0]); |
289 | | #endif |
290 | 0 | } |
291 | | /* In case max_pitch isn't a multiple of 4, do non-unrolled version. */ |
292 | 0 | for (;i<max_pitch;i++) |
293 | 0 | { |
294 | 0 | opus_val32 sum; |
295 | 0 | sum = celt_inner_prod(_x, _y+i, len, arch); |
296 | 0 | xcorr[i] = sum; |
297 | | #ifdef FIXED_POINT |
298 | | maxcorr = MAX32(maxcorr, sum); |
299 | | #endif |
300 | 0 | } |
301 | | #ifdef FIXED_POINT |
302 | | return maxcorr; |
303 | | #endif |
304 | 0 | #endif |
305 | 0 | } |
306 | | |
307 | | void pitch_search(const opus_val16 * OPUS_RESTRICT x_lp, opus_val16 * OPUS_RESTRICT y, |
308 | | int len, int max_pitch, int *pitch, int arch) |
309 | 18.1k | { |
310 | 18.1k | int i, j; |
311 | 18.1k | int lag; |
312 | 18.1k | int best_pitch[2]={0,0}; |
313 | 18.1k | VARDECL(opus_val16, x_lp4); |
314 | 18.1k | VARDECL(opus_val16, y_lp4); |
315 | 18.1k | VARDECL(opus_val32, xcorr); |
316 | | #ifdef FIXED_POINT |
317 | | opus_val32 maxcorr; |
318 | | opus_val32 xmax, ymax; |
319 | | int shift=0; |
320 | | #endif |
321 | 18.1k | int offset; |
322 | | |
323 | 18.1k | SAVE_STACK; |
324 | | |
325 | 18.1k | celt_assert(len>0); |
326 | 18.1k | celt_assert(max_pitch>0); |
327 | 18.1k | lag = len+max_pitch; |
328 | | |
329 | 18.1k | ALLOC(x_lp4, len>>2, opus_val16); |
330 | 18.1k | ALLOC(y_lp4, lag>>2, opus_val16); |
331 | 18.1k | ALLOC(xcorr, max_pitch>>1, opus_val32); |
332 | | |
333 | | /* Downsample by 2 again */ |
334 | 6.03M | for (j=0;j<len>>2;j++) |
335 | 6.01M | x_lp4[j] = x_lp[2*j]; |
336 | 8.83M | for (j=0;j<lag>>2;j++) |
337 | 8.82M | y_lp4[j] = y[2*j]; |
338 | | |
339 | | #ifdef FIXED_POINT |
340 | | xmax = celt_maxabs16(x_lp4, len>>2); |
341 | | ymax = celt_maxabs16(y_lp4, lag>>2); |
342 | | shift = celt_ilog2(MAX32(1, MAX32(xmax, ymax))) - 14 + celt_ilog2(len)/2; |
343 | | if (shift>0) |
344 | | { |
345 | | for (j=0;j<len>>2;j++) |
346 | | x_lp4[j] = SHR16(x_lp4[j], shift); |
347 | | for (j=0;j<lag>>2;j++) |
348 | | y_lp4[j] = SHR16(y_lp4[j], shift); |
349 | | /* Use double the shift for a MAC */ |
350 | | shift *= 2; |
351 | | } else { |
352 | | shift = 0; |
353 | | } |
354 | | #endif |
355 | | |
356 | | /* Coarse search with 4x decimation */ |
357 | | |
358 | | #ifdef FIXED_POINT |
359 | | maxcorr = |
360 | | #endif |
361 | 18.1k | celt_pitch_xcorr(x_lp4, y_lp4, xcorr, len>>2, max_pitch>>2, arch); |
362 | | |
363 | 18.1k | find_best_pitch(xcorr, y_lp4, len>>2, max_pitch>>2, best_pitch |
364 | | #ifdef FIXED_POINT |
365 | | , 0, maxcorr |
366 | | #endif |
367 | 18.1k | ); |
368 | | |
369 | | /* Finer search with 2x decimation */ |
370 | | #ifdef FIXED_POINT |
371 | | maxcorr=1; |
372 | | #endif |
373 | 5.63M | for (i=0;i<max_pitch>>1;i++) |
374 | 5.61M | { |
375 | 5.61M | opus_val32 sum; |
376 | 5.61M | xcorr[i] = 0; |
377 | 5.61M | if (abs(i-2*best_pitch[0])>2 && abs(i-2*best_pitch[1])>2) |
378 | 5.45M | continue; |
379 | | #ifdef FIXED_POINT |
380 | | sum = 0; |
381 | | for (j=0;j<len>>1;j++) |
382 | | sum += SHR32(MULT16_16(x_lp[j],y[i+j]), shift); |
383 | | #else |
384 | 161k | sum = celt_inner_prod(x_lp, y+i, len>>1, arch); |
385 | 161k | #endif |
386 | 161k | xcorr[i] = MAX32(-1, sum); |
387 | | #ifdef FIXED_POINT |
388 | | maxcorr = MAX32(maxcorr, sum); |
389 | | #endif |
390 | 161k | } |
391 | 18.1k | find_best_pitch(xcorr, y, len>>1, max_pitch>>1, best_pitch |
392 | | #ifdef FIXED_POINT |
393 | | , shift+1, maxcorr |
394 | | #endif |
395 | 18.1k | ); |
396 | | |
397 | | /* Refine by pseudo-interpolation */ |
398 | 18.1k | if (best_pitch[0]>0 && best_pitch[0]<(max_pitch>>1)-1) |
399 | 16.3k | { |
400 | 16.3k | opus_val32 a, b, c; |
401 | 16.3k | a = xcorr[best_pitch[0]-1]; |
402 | 16.3k | b = xcorr[best_pitch[0]]; |
403 | 16.3k | c = xcorr[best_pitch[0]+1]; |
404 | 16.3k | if ((c-a) > MULT16_32_Q15(QCONST16(.7f,15),b-a)) |
405 | 2.60k | offset = 1; |
406 | 13.7k | else if ((a-c) > MULT16_32_Q15(QCONST16(.7f,15),b-c)) |
407 | 3.70k | offset = -1; |
408 | 10.0k | else |
409 | 10.0k | offset = 0; |
410 | 16.3k | } else { |
411 | 1.72k | offset = 0; |
412 | 1.72k | } |
413 | 18.1k | *pitch = 2*best_pitch[0]-offset; |
414 | | |
415 | 18.1k | RESTORE_STACK; |
416 | 18.1k | } |
417 | | |
418 | | #ifdef FIXED_POINT |
419 | | static opus_val16 compute_pitch_gain(opus_val32 xy, opus_val32 xx, opus_val32 yy) |
420 | | { |
421 | | opus_val32 x2y2; |
422 | | int sx, sy, shift; |
423 | | opus_val32 g; |
424 | | opus_val16 den; |
425 | | if (xy == 0 || xx == 0 || yy == 0) |
426 | | return 0; |
427 | | sx = celt_ilog2(xx)-14; |
428 | | sy = celt_ilog2(yy)-14; |
429 | | shift = sx + sy; |
430 | | x2y2 = SHR32(MULT16_16(VSHR32(xx, sx), VSHR32(yy, sy)), 14); |
431 | | if (shift & 1) { |
432 | | if (x2y2 < 32768) |
433 | | { |
434 | | x2y2 <<= 1; |
435 | | shift--; |
436 | | } else { |
437 | | x2y2 >>= 1; |
438 | | shift++; |
439 | | } |
440 | | } |
441 | | den = celt_rsqrt_norm(x2y2); |
442 | | g = MULT16_32_Q15(den, xy); |
443 | | g = VSHR32(g, (shift>>1)-1); |
444 | | return EXTRACT16(MAX32(-Q15ONE, MIN32(g, Q15ONE))); |
445 | | } |
446 | | #else |
447 | | static opus_val16 compute_pitch_gain(opus_val32 xy, opus_val32 xx, opus_val32 yy) |
448 | 0 | { |
449 | 0 | return xy/celt_sqrt(1+xx*yy); |
450 | 0 | } |
451 | | #endif |
452 | | |
453 | | static const int second_check[16] = {0, 0, 3, 2, 3, 2, 5, 2, 3, 2, 3, 2, 5, 2, 3, 2}; |
454 | | opus_val16 remove_doubling(opus_val16 *x, int maxperiod, int minperiod, |
455 | | int N, int *T0_, int prev_period, opus_val16 prev_gain, int arch) |
456 | 0 | { |
457 | 0 | int k, i, T, T0; |
458 | 0 | opus_val16 g, g0; |
459 | 0 | opus_val16 pg; |
460 | 0 | opus_val32 xy,xx,yy,xy2; |
461 | 0 | opus_val32 xcorr[3]; |
462 | 0 | opus_val32 best_xy, best_yy; |
463 | 0 | int offset; |
464 | 0 | int minperiod0; |
465 | 0 | VARDECL(opus_val32, yy_lookup); |
466 | 0 | SAVE_STACK; |
467 | |
|
468 | 0 | minperiod0 = minperiod; |
469 | 0 | maxperiod /= 2; |
470 | 0 | minperiod /= 2; |
471 | 0 | *T0_ /= 2; |
472 | 0 | prev_period /= 2; |
473 | 0 | N /= 2; |
474 | 0 | x += maxperiod; |
475 | 0 | if (*T0_>=maxperiod) |
476 | 0 | *T0_=maxperiod-1; |
477 | |
|
478 | 0 | T = T0 = *T0_; |
479 | 0 | ALLOC(yy_lookup, maxperiod+1, opus_val32); |
480 | 0 | dual_inner_prod(x, x, x-T0, N, &xx, &xy, arch); |
481 | 0 | yy_lookup[0] = xx; |
482 | 0 | yy=xx; |
483 | 0 | for (i=1;i<=maxperiod;i++) |
484 | 0 | { |
485 | 0 | yy = yy+MULT16_16(x[-i],x[-i])-MULT16_16(x[N-i],x[N-i]); |
486 | 0 | yy_lookup[i] = MAX32(0, yy); |
487 | 0 | } |
488 | 0 | yy = yy_lookup[T0]; |
489 | 0 | best_xy = xy; |
490 | 0 | best_yy = yy; |
491 | 0 | g = g0 = compute_pitch_gain(xy, xx, yy); |
492 | | /* Look for any pitch at T/k */ |
493 | 0 | for (k=2;k<=15;k++) |
494 | 0 | { |
495 | 0 | int T1, T1b; |
496 | 0 | opus_val16 g1; |
497 | 0 | opus_val16 cont=0; |
498 | 0 | opus_val16 thresh; |
499 | 0 | T1 = celt_udiv(2*T0+k, 2*k); |
500 | 0 | if (T1 < minperiod) |
501 | 0 | break; |
502 | | /* Look for another strong correlation at T1b */ |
503 | 0 | if (k==2) |
504 | 0 | { |
505 | 0 | if (T1+T0>maxperiod) |
506 | 0 | T1b = T0; |
507 | 0 | else |
508 | 0 | T1b = T0+T1; |
509 | 0 | } else |
510 | 0 | { |
511 | 0 | T1b = celt_udiv(2*second_check[k]*T0+k, 2*k); |
512 | 0 | } |
513 | 0 | dual_inner_prod(x, &x[-T1], &x[-T1b], N, &xy, &xy2, arch); |
514 | 0 | xy = HALF32(xy + xy2); |
515 | 0 | yy = HALF32(yy_lookup[T1] + yy_lookup[T1b]); |
516 | 0 | g1 = compute_pitch_gain(xy, xx, yy); |
517 | 0 | if (abs(T1-prev_period)<=1) |
518 | 0 | cont = prev_gain; |
519 | 0 | else if (abs(T1-prev_period)<=2 && 5*k*k < T0) |
520 | 0 | cont = HALF16(prev_gain); |
521 | 0 | else |
522 | 0 | cont = 0; |
523 | 0 | thresh = MAX16(QCONST16(.3f,15), MULT16_16_Q15(QCONST16(.7f,15),g0)-cont); |
524 | | /* Bias against very high pitch (very short period) to avoid false-positives |
525 | | due to short-term correlation */ |
526 | 0 | if (T1<3*minperiod) |
527 | 0 | thresh = MAX16(QCONST16(.4f,15), MULT16_16_Q15(QCONST16(.85f,15),g0)-cont); |
528 | 0 | else if (T1<2*minperiod) |
529 | 0 | thresh = MAX16(QCONST16(.5f,15), MULT16_16_Q15(QCONST16(.9f,15),g0)-cont); |
530 | 0 | if (g1 > thresh) |
531 | 0 | { |
532 | 0 | best_xy = xy; |
533 | 0 | best_yy = yy; |
534 | 0 | T = T1; |
535 | 0 | g = g1; |
536 | 0 | } |
537 | 0 | } |
538 | 0 | best_xy = MAX32(0, best_xy); |
539 | 0 | if (best_yy <= best_xy) |
540 | 0 | pg = Q15ONE; |
541 | 0 | else |
542 | 0 | pg = SHR32(frac_div32(best_xy,best_yy+1),16); |
543 | |
|
544 | 0 | for (k=0;k<3;k++) |
545 | 0 | xcorr[k] = celt_inner_prod(x, x-(T+k-1), N, arch); |
546 | 0 | if ((xcorr[2]-xcorr[0]) > MULT16_32_Q15(QCONST16(.7f,15),xcorr[1]-xcorr[0])) |
547 | 0 | offset = 1; |
548 | 0 | else if ((xcorr[0]-xcorr[2]) > MULT16_32_Q15(QCONST16(.7f,15),xcorr[1]-xcorr[2])) |
549 | 0 | offset = -1; |
550 | 0 | else |
551 | 0 | offset = 0; |
552 | 0 | if (pg > g) |
553 | 0 | pg = g; |
554 | 0 | *T0_ = 2*T+offset; |
555 | |
|
556 | 0 | if (*T0_<minperiod0) |
557 | 0 | *T0_=minperiod0; |
558 | 0 | RESTORE_STACK; |
559 | 0 | return pg; |
560 | 0 | } |