Coverage Report

Created: 2026-02-14 06:59

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
/src/opus/celt/bands.c
Line
Count
Source
1
/* Copyright (c) 2007-2008 CSIRO
2
   Copyright (c) 2007-2009 Xiph.Org Foundation
3
   Copyright (c) 2008-2009 Gregory Maxwell
4
   Written by Jean-Marc Valin and Gregory Maxwell */
5
/*
6
   Redistribution and use in source and binary forms, with or without
7
   modification, are permitted provided that the following conditions
8
   are met:
9
10
   - Redistributions of source code must retain the above copyright
11
   notice, this list of conditions and the following disclaimer.
12
13
   - 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
   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
18
   ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
19
   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
20
   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
21
   OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
22
   EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
23
   PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
24
   PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
25
   LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
26
   NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
27
   SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
28
*/
29
30
#ifdef HAVE_CONFIG_H
31
#include "config.h"
32
#endif
33
34
#include <math.h>
35
#include "bands.h"
36
#include "modes.h"
37
#include "vq.h"
38
#include "cwrs.h"
39
#include "stack_alloc.h"
40
#include "os_support.h"
41
#include "mathops.h"
42
#include "rate.h"
43
#include "quant_bands.h"
44
#include "pitch.h"
45
46
int hysteresis_decision(opus_val16 val, const opus_val16 *thresholds, const opus_val16 *hysteresis, int N, int prev)
47
0
{
48
0
   int i;
49
0
   for (i=0;i<N;i++)
50
0
   {
51
0
      if (val < thresholds[i])
52
0
         break;
53
0
   }
54
0
   if (i>prev && val < thresholds[prev]+hysteresis[prev])
55
0
      i=prev;
56
0
   if (i<prev && val > thresholds[prev-1]-hysteresis[prev-1])
57
0
      i=prev;
58
0
   return i;
59
0
}
60
61
opus_uint32 celt_lcg_rand(opus_uint32 seed)
62
62.5M
{
63
62.5M
   return 1664525 * seed + 1013904223;
64
62.5M
}
65
66
/* This is a cos() approximation designed to be bit-exact on any platform. Bit exactness
67
   with this approximation is important because it has an impact on the bit allocation */
68
opus_int16 bitexact_cos(opus_int16 x)
69
499k
{
70
499k
   opus_int32 tmp;
71
499k
   opus_int16 x2;
72
499k
   tmp = (4096+((opus_int32)(x)*(x)))>>13;
73
499k
   celt_sig_assert(tmp<=32767);
74
499k
   x2 = tmp;
75
499k
   x2 = (32767-x2) + FRAC_MUL16(x2, (-7651 + FRAC_MUL16(x2, (8277 + FRAC_MUL16(-626, x2)))));
76
499k
   celt_sig_assert(x2<=32766);
77
499k
   return 1+x2;
78
499k
}
79
80
int bitexact_log2tan(int isin,int icos)
81
249k
{
82
249k
   int lc;
83
249k
   int ls;
84
249k
   lc=EC_ILOG(icos);
85
249k
   ls=EC_ILOG(isin);
86
249k
   icos<<=15-lc;
87
249k
   isin<<=15-ls;
88
249k
   return (ls-lc)*(1<<11)
89
249k
         +FRAC_MUL16(isin, FRAC_MUL16(isin, -2597) + 7932)
90
249k
         -FRAC_MUL16(icos, FRAC_MUL16(icos, -2597) + 7932);
91
249k
}
92
93
#ifdef FIXED_POINT
94
/* Compute the amplitude (sqrt energy) in each of the bands */
95
void compute_band_energies(const CELTMode *m, const celt_sig *X, celt_ener *bandE, int end, int C, int LM, int arch)
96
{
97
   int i, c, N;
98
   const opus_int16 *eBands = m->eBands;
99
   (void)arch;
100
   N = m->shortMdctSize<<LM;
101
   c=0; do {
102
      for (i=0;i<end;i++)
103
      {
104
         int j;
105
         opus_val32 maxval=0;
106
         opus_val32 sum = 0;
107
108
         maxval = celt_maxabs32(&X[c*N+(eBands[i]<<LM)], (eBands[i+1]-eBands[i])<<LM);
109
         if (maxval > 0)
110
         {
111
            int shift = IMAX(0, 30 - celt_ilog2(maxval+(maxval>>14)+1) - ((((m->logN[i]+7)>>BITRES)+LM+1)>>1));
112
            j=eBands[i]<<LM; do {
113
               opus_val32 x = SHL32(X[j+c*N],shift);
114
               sum = ADD32(sum, MULT32_32_Q31(x, x));
115
            } while (++j<eBands[i+1]<<LM);
116
            bandE[i+c*m->nbEBands] = MAX32(maxval, PSHR32(celt_sqrt32(SHR32(sum,1)), shift));
117
         } else {
118
            bandE[i+c*m->nbEBands] = EPSILON;
119
         }
120
      }
121
   } while (++c<C);
122
}
123
124
/* Normalise each band such that the energy is one. */
125
void normalise_bands(const CELTMode *m, const celt_sig * OPUS_RESTRICT freq, celt_norm * OPUS_RESTRICT X, const celt_ener *bandE, int end, int C, int M)
126
{
127
   int i, c, N;
128
   const opus_int16 *eBands = m->eBands;
129
   N = M*m->shortMdctSize;
130
   c=0; do {
131
      i=0; do {
132
         int j,shift;
133
         opus_val32 E;
134
         opus_val32 g;
135
         E = bandE[i+c*m->nbEBands];
136
         /* For very low energies, we need this to make sure not to prevent energy rounding from
137
            blowing up the normalized signal. */
138
         if (E < 10) E += EPSILON;
139
         shift = 30-celt_zlog2(E);
140
         E = SHL32(E, shift);
141
         g = celt_rcp_norm32(E);
142
         j=M*eBands[i]; do {
143
            X[j+c*N] = PSHR32(MULT32_32_Q31(g, SHL32(freq[j+c*N], shift)), 30-NORM_SHIFT);
144
         } while (++j<M*eBands[i+1]);
145
      } while (++i<end);
146
   } while (++c<C);
147
}
148
149
#else /* FIXED_POINT */
150
/* Compute the amplitude (sqrt energy) in each of the bands */
151
void compute_band_energies(const CELTMode *m, const celt_sig *X, celt_ener *bandE, int end, int C, int LM, int arch)
152
0
{
153
0
   int i, c, N;
154
0
   const opus_int16 *eBands = m->eBands;
155
0
   N = m->shortMdctSize<<LM;
156
0
   c=0; do {
157
0
      for (i=0;i<end;i++)
158
0
      {
159
0
         opus_val32 sum;
160
0
         sum = 1e-27f + celt_inner_prod(&X[c*N+(eBands[i]<<LM)], &X[c*N+(eBands[i]<<LM)], (eBands[i+1]-eBands[i])<<LM, arch);
161
0
         bandE[i+c*m->nbEBands] = celt_sqrt(sum);
162
         /*printf ("%f ", bandE[i+c*m->nbEBands]);*/
163
0
      }
164
0
   } while (++c<C);
165
   /*printf ("\n");*/
166
0
}
167
168
/* Normalise each band such that the energy is one. */
169
void normalise_bands(const CELTMode *m, const celt_sig * OPUS_RESTRICT freq, celt_norm * OPUS_RESTRICT X, const celt_ener *bandE, int end, int C, int M)
170
0
{
171
0
   int i, c, N;
172
0
   const opus_int16 *eBands = m->eBands;
173
0
   N = M*m->shortMdctSize;
174
0
   c=0; do {
175
0
      for (i=0;i<end;i++)
176
0
      {
177
0
         int j;
178
0
         opus_val16 g = 1.f/(1e-27f+bandE[i+c*m->nbEBands]);
179
0
         for (j=M*eBands[i];j<M*eBands[i+1];j++)
180
0
            X[j+c*N] = freq[j+c*N]*g;
181
0
      }
182
0
   } while (++c<C);
183
0
}
184
185
#endif /* FIXED_POINT */
186
187
/* De-normalise the energy to produce the synthesis from the unit-energy bands */
188
void denormalise_bands(const CELTMode *m, const celt_norm * OPUS_RESTRICT X,
189
      celt_sig * OPUS_RESTRICT freq, const celt_glog *bandLogE, int start,
190
      int end, int M, int downsample, int silence)
191
285k
{
192
285k
   int i, N;
193
285k
   int bound;
194
285k
   celt_sig * OPUS_RESTRICT f;
195
285k
   const celt_norm * OPUS_RESTRICT x;
196
285k
   const opus_int16 *eBands = m->eBands;
197
285k
   N = M*m->shortMdctSize;
198
285k
   bound = M*eBands[end];
199
285k
   if (downsample!=1)
200
0
      bound = IMIN(bound, N/downsample);
201
285k
   if (silence)
202
33.5k
   {
203
33.5k
      bound = 0;
204
33.5k
      start = end = 0;
205
33.5k
   }
206
285k
   f = freq;
207
285k
   x = X+M*eBands[start];
208
285k
   if (start != 0)
209
55.9k
   {
210
13.0M
      for (i=0;i<M*eBands[start];i++)
211
12.9M
         *f++ = 0;
212
229k
   } else {
213
229k
      f += M*eBands[start];
214
229k
   }
215
4.00M
   for (i=start;i<end;i++)
216
3.71M
   {
217
3.71M
      int j, band_end;
218
3.71M
      opus_val32 g;
219
3.71M
      celt_glog lg;
220
#ifdef FIXED_POINT
221
      int shift;
222
#endif
223
3.71M
      j=M*eBands[i];
224
3.71M
      band_end = M*eBands[i+1];
225
3.71M
      lg = ADD32(bandLogE[i], SHL32((opus_val32)eMeans[i],DB_SHIFT-4));
226
3.71M
#ifndef FIXED_POINT
227
3.71M
      g = celt_exp2_db(MIN32(32.f, lg));
228
#else
229
      /* Handle the integer part of the log energy */
230
      shift = 17-(lg>>DB_SHIFT);
231
      if (shift>=31)
232
      {
233
         shift=0;
234
         g=0;
235
      } else {
236
         /* Handle the fractional part. */
237
         g = SHL32(celt_exp2_db_frac((lg&((1<<DB_SHIFT)-1))), 2);
238
      }
239
      /* Handle extreme gains with negative shift. */
240
      if (shift<0)
241
      {
242
         /* To avoid overflow, we're
243
            capping the gain here, which is equivalent to a cap of 18 on lg.
244
            This shouldn't trigger unless the bitstream is already corrupted. */
245
         g = 2147483647;
246
         shift = 0;
247
      }
248
#endif
249
65.3M
      do {
250
65.3M
         *f++ = PSHR32(MULT32_32_Q31(SHL32(*x, 30-NORM_SHIFT), g), shift);
251
65.3M
         x++;
252
65.3M
      } while (++j<band_end);
253
3.71M
   }
254
285k
   celt_assert(start <= end);
255
285k
   OPUS_CLEAR(&freq[bound], N-bound);
256
285k
}
257
258
/* This prevents energy collapse for transients with multiple short MDCTs */
259
void anti_collapse(const CELTMode *m, celt_norm *X_, unsigned char *collapse_masks, int LM, int C, int size,
260
      int start, int end, const celt_glog *logE, const celt_glog *prev1logE,
261
      const celt_glog *prev2logE, const int *pulses, opus_uint32 seed, int encode, int arch)
262
2.65k
{
263
2.65k
   int c, i, j, k;
264
46.3k
   for (i=start;i<end;i++)
265
43.6k
   {
266
43.6k
      int N0;
267
43.6k
      opus_val16 thresh, sqrt_1;
268
43.6k
      int depth;
269
#ifdef FIXED_POINT
270
      int shift;
271
      opus_val32 thresh32;
272
#endif
273
274
43.6k
      N0 = m->eBands[i+1]-m->eBands[i];
275
      /* depth in 1/8 bits */
276
43.6k
      celt_sig_assert(pulses[i]>=0);
277
43.6k
      depth = celt_udiv(1+pulses[i], (m->eBands[i+1]-m->eBands[i]))>>LM;
278
279
#ifdef FIXED_POINT
280
      thresh32 = SHR32(celt_exp2(-SHL16(depth, 10-BITRES)),1);
281
      thresh = MULT16_32_Q15(QCONST16(0.5f, 15), MIN32(32767,thresh32));
282
      {
283
         opus_val32 t;
284
         t = N0<<LM;
285
         shift = celt_ilog2(t)>>1;
286
         t = SHL32(t, (7-shift)<<1);
287
         sqrt_1 = celt_rsqrt_norm(t);
288
      }
289
#else
290
43.6k
      thresh = .5f*celt_exp2(-.125f*depth);
291
43.6k
      sqrt_1 = celt_rsqrt(N0<<LM);
292
43.6k
#endif
293
294
43.6k
      c=0; do
295
53.5k
      {
296
53.5k
         celt_norm *X;
297
53.5k
         celt_glog prev1;
298
53.5k
         celt_glog prev2;
299
53.5k
         opus_val32 Ediff;
300
53.5k
         celt_norm r;
301
53.5k
         int renormalize=0;
302
53.5k
         prev1 = prev1logE[c*m->nbEBands+i];
303
53.5k
         prev2 = prev2logE[c*m->nbEBands+i];
304
53.5k
         if (!encode && C==1)
305
33.8k
         {
306
33.8k
            prev1 = MAXG(prev1,prev1logE[m->nbEBands+i]);
307
33.8k
            prev2 = MAXG(prev2,prev2logE[m->nbEBands+i]);
308
33.8k
         }
309
53.5k
         Ediff = logE[c*m->nbEBands+i]-MING(prev1,prev2);
310
53.5k
         Ediff = MAX32(0, Ediff);
311
312
#ifdef FIXED_POINT
313
         if (Ediff < GCONST(16.f))
314
         {
315
            opus_val32 r32 = SHR32(celt_exp2_db(-Ediff),1);
316
            r = 2*MIN16(16383,r32);
317
         } else {
318
            r = 0;
319
         }
320
         if (LM==3)
321
            r = MULT16_16_Q14(23170, MIN32(23169, r));
322
         r = SHR16(MIN16(thresh, r),1);
323
         r = VSHR32(MULT16_16_Q15(sqrt_1, r),shift+14-NORM_SHIFT);
324
#else
325
         /* r needs to be multiplied by 2 or 2*sqrt(2) depending on LM because
326
            short blocks don't have the same energy as long */
327
53.5k
         r = 2.f*celt_exp2_db(-Ediff);
328
53.5k
         if (LM==3)
329
18.1k
            r *= 1.41421356f;
330
53.5k
         r = MIN16(thresh, r);
331
53.5k
         r = r*sqrt_1;
332
53.5k
#endif
333
53.5k
         X = X_+c*size+(m->eBands[i]<<LM);
334
340k
         for (k=0;k<1<<LM;k++)
335
286k
         {
336
            /* Detect collapse */
337
286k
            if (!(collapse_masks[i*C+c]&1<<k))
338
40.1k
            {
339
               /* Fill with noise */
340
181k
               for (j=0;j<N0;j++)
341
141k
               {
342
141k
                  seed = celt_lcg_rand(seed);
343
141k
                  X[(j<<LM)+k] = (seed&0x8000 ? r : -r);
344
141k
               }
345
40.1k
               renormalize = 1;
346
40.1k
            }
347
286k
         }
348
         /* We just added some energy, so we need to renormalise */
349
53.5k
         if (renormalize)
350
10.3k
            renormalise_vector(X, N0<<LM, Q31ONE, arch);
351
53.5k
      } while (++c<C);
352
43.6k
   }
353
2.65k
}
354
355
/* Compute the weights to use for optimizing normalized distortion across
356
   channels. We use the amplitude to weight square distortion, which means
357
   that we use the square root of the value we would have been using if we
358
   wanted to minimize the MSE in the non-normalized domain. This roughly
359
   corresponds to some quick-and-dirty perceptual experiments I ran to
360
   measure inter-aural masking (there doesn't seem to be any published data
361
   on the topic). */
362
static void compute_channel_weights(celt_ener Ex, celt_ener Ey, opus_val16 w[2])
363
0
{
364
0
   celt_ener minE;
365
#ifdef FIXED_POINT
366
   int shift;
367
#endif
368
0
   minE = MIN32(Ex, Ey);
369
   /* Adjustment to make the weights a bit more conservative. */
370
0
   Ex = ADD32(Ex, minE/3);
371
0
   Ey = ADD32(Ey, minE/3);
372
#ifdef FIXED_POINT
373
   shift = celt_ilog2(EPSILON+MAX32(Ex, Ey))-14;
374
#endif
375
0
   w[0] = VSHR32(Ex, shift);
376
0
   w[1] = VSHR32(Ey, shift);
377
0
}
378
379
static void intensity_stereo(const CELTMode *m, celt_norm * OPUS_RESTRICT X, const celt_norm * OPUS_RESTRICT Y, const celt_ener *bandE, int bandID, int N)
380
0
{
381
0
   int i = bandID;
382
0
   int j;
383
0
   opus_val16 a1, a2;
384
0
   opus_val16 left, right;
385
0
   opus_val16 norm;
386
#ifdef FIXED_POINT
387
   int shift = celt_zlog2(MAX32(bandE[i], bandE[i+m->nbEBands]))-13;
388
#endif
389
0
   left = VSHR32(bandE[i],shift);
390
0
   right = VSHR32(bandE[i+m->nbEBands],shift);
391
0
   norm = EPSILON + celt_sqrt(EPSILON+MULT16_16(left,left)+MULT16_16(right,right));
392
#ifdef FIXED_POINT
393
   left = MIN32(left, norm-1);
394
   right = MIN32(right, norm-1);
395
#endif
396
0
   a1 = DIV32_16(SHL32(EXTEND32(left),15),norm);
397
0
   a2 = DIV32_16(SHL32(EXTEND32(right),15),norm);
398
0
   for (j=0;j<N;j++)
399
0
   {
400
0
      X[j] = ADD32(MULT16_32_Q15(a1, X[j]), MULT16_32_Q15(a2, Y[j]));
401
      /* Side is not encoded, no need to calculate */
402
0
   }
403
0
}
404
405
static void stereo_split(celt_norm * OPUS_RESTRICT X, celt_norm * OPUS_RESTRICT Y, int N)
406
0
{
407
0
   int j;
408
0
   for (j=0;j<N;j++)
409
0
   {
410
0
      opus_val32 r, l;
411
0
      l = MULT32_32_Q31(QCONST32(.70710678f,31), X[j]);
412
0
      r = MULT32_32_Q31(QCONST32(.70710678f,31), Y[j]);
413
0
      X[j] = ADD32(l, r);
414
0
      Y[j] = SUB32(r, l);
415
0
   }
416
0
}
417
418
static void stereo_merge(celt_norm * OPUS_RESTRICT X, celt_norm * OPUS_RESTRICT Y, opus_val32 mid, int N, int arch)
419
200k
{
420
200k
   int j;
421
200k
   opus_val32 xp=0, side=0;
422
200k
   opus_val32 El, Er;
423
#ifdef FIXED_POINT
424
   int kl, kr;
425
#endif
426
200k
   opus_val32 t, lgain, rgain;
427
428
   /* Compute the norm of X+Y and X-Y as |X|^2 + |Y|^2 +/- sum(xy) */
429
200k
   xp = celt_inner_prod_norm_shift(Y, X, N, arch);
430
200k
   side = celt_inner_prod_norm_shift(Y, Y, N, arch);
431
   /* Compensating for the mid normalization */
432
200k
   xp = MULT32_32_Q31(mid, xp);
433
   /* mid and side are in Q15, not Q14 like X and Y */
434
200k
   El = SHR32(MULT32_32_Q31(mid, mid),3) + side - 2*xp;
435
200k
   Er = SHR32(MULT32_32_Q31(mid, mid),3) + side + 2*xp;
436
200k
   if (Er < QCONST32(6e-4f, 28) || El < QCONST32(6e-4f, 28))
437
712
   {
438
712
      OPUS_COPY(Y, X, N);
439
712
      return;
440
712
   }
441
442
#ifdef FIXED_POINT
443
   kl = celt_ilog2(El)>>1;
444
   kr = celt_ilog2(Er)>>1;
445
#endif
446
199k
   t = VSHR32(El, (kl<<1)-29);
447
199k
   lgain = celt_rsqrt_norm32(t);
448
199k
   t = VSHR32(Er, (kr<<1)-29);
449
199k
   rgain = celt_rsqrt_norm32(t);
450
451
#ifdef FIXED_POINT
452
   if (kl < 7)
453
      kl = 7;
454
   if (kr < 7)
455
      kr = 7;
456
#endif
457
458
5.56M
   for (j=0;j<N;j++)
459
5.36M
   {
460
5.36M
      celt_norm r, l;
461
      /* Apply mid scaling (side is already scaled) */
462
5.36M
      l = MULT32_32_Q31(mid, X[j]);
463
5.36M
      r = Y[j];
464
5.36M
      X[j] = VSHR32(MULT32_32_Q31(lgain, SUB32(l,r)), kl-15);
465
5.36M
      Y[j] = VSHR32(MULT32_32_Q31(rgain, ADD32(l,r)), kr-15);
466
5.36M
   }
467
199k
}
468
469
/* Decide whether we should spread the pulses in the current frame */
470
int spreading_decision(const CELTMode *m, const celt_norm *X, int *average,
471
      int last_decision, int *hf_average, int *tapset_decision, int update_hf,
472
      int end, int C, int M, const int *spread_weight)
473
0
{
474
0
   int i, c, N0;
475
0
   int sum = 0, nbBands=0;
476
0
   const opus_int16 * OPUS_RESTRICT eBands = m->eBands;
477
0
   int decision;
478
0
   int hf_sum=0;
479
480
0
   celt_assert(end>0);
481
482
0
   N0 = M*m->shortMdctSize;
483
484
0
   if (M*(eBands[end]-eBands[end-1]) <= 8)
485
0
      return SPREAD_NONE;
486
0
   c=0; do {
487
0
      for (i=0;i<end;i++)
488
0
      {
489
0
         int j, N, tmp=0;
490
0
         int tcount[3] = {0,0,0};
491
0
         const celt_norm * OPUS_RESTRICT x = X+M*eBands[i]+c*N0;
492
0
         N = M*(eBands[i+1]-eBands[i]);
493
0
         if (N<=8)
494
0
            continue;
495
         /* Compute rough CDF of |x[j]| */
496
0
         for (j=0;j<N;j++)
497
0
         {
498
0
            opus_val32 x2N; /* Q13 */
499
500
0
            x2N = MULT16_16(MULT16_16_Q15(SHR32(x[j], NORM_SHIFT-14), SHR32(x[j], NORM_SHIFT-14)), N);
501
0
            if (x2N < QCONST16(0.25f,13))
502
0
               tcount[0]++;
503
0
            if (x2N < QCONST16(0.0625f,13))
504
0
               tcount[1]++;
505
0
            if (x2N < QCONST16(0.015625f,13))
506
0
               tcount[2]++;
507
0
         }
508
509
         /* Only include four last bands (8 kHz and up) */
510
0
         if (i>m->nbEBands-4)
511
0
            hf_sum += celt_udiv(32*(tcount[1]+tcount[0]), N);
512
0
         tmp = (2*tcount[2] >= N) + (2*tcount[1] >= N) + (2*tcount[0] >= N);
513
0
         sum += tmp*spread_weight[i];
514
0
         nbBands+=spread_weight[i];
515
0
      }
516
0
   } while (++c<C);
517
518
0
   if (update_hf)
519
0
   {
520
0
      if (hf_sum)
521
0
         hf_sum = celt_udiv(hf_sum, C*(4-m->nbEBands+end));
522
0
      *hf_average = (*hf_average+hf_sum)>>1;
523
0
      hf_sum = *hf_average;
524
0
      if (*tapset_decision==2)
525
0
         hf_sum += 4;
526
0
      else if (*tapset_decision==0)
527
0
         hf_sum -= 4;
528
0
      if (hf_sum > 22)
529
0
         *tapset_decision=2;
530
0
      else if (hf_sum > 18)
531
0
         *tapset_decision=1;
532
0
      else
533
0
         *tapset_decision=0;
534
0
   }
535
   /*printf("%d %d %d\n", hf_sum, *hf_average, *tapset_decision);*/
536
0
   celt_assert(nbBands>0); /* end has to be non-zero */
537
0
   celt_assert(sum>=0);
538
0
   sum = celt_udiv((opus_int32)sum<<8, nbBands);
539
   /* Recursive averaging */
540
0
   sum = (sum+*average)>>1;
541
0
   *average = sum;
542
   /* Hysteresis */
543
0
   sum = (3*sum + (((3-last_decision)<<7) + 64) + 2)>>2;
544
0
   if (sum < 80)
545
0
   {
546
0
      decision = SPREAD_AGGRESSIVE;
547
0
   } else if (sum < 256)
548
0
   {
549
0
      decision = SPREAD_NORMAL;
550
0
   } else if (sum < 384)
551
0
   {
552
0
      decision = SPREAD_LIGHT;
553
0
   } else {
554
0
      decision = SPREAD_NONE;
555
0
   }
556
#ifdef FUZZING
557
   decision = rand()&0x3;
558
   *tapset_decision=rand()%3;
559
#endif
560
0
   return decision;
561
0
}
562
563
/* Indexing table for converting from natural Hadamard to ordery Hadamard
564
   This is essentially a bit-reversed Gray, on top of which we've added
565
   an inversion of the order because we want the DC at the end rather than
566
   the beginning. The lines are for N=2, 4, 8, 16 */
567
static const int ordery_table[] = {
568
       1,  0,
569
       3,  0,  2,  1,
570
       7,  0,  4,  3,  6,  1,  5,  2,
571
      15,  0,  8,  7, 12,  3, 11,  4, 14,  1,  9,  6, 13,  2, 10,  5,
572
};
573
574
static void deinterleave_hadamard(celt_norm *X, int N0, int stride, int hadamard)
575
177k
{
576
177k
   int i,j;
577
177k
   VARDECL(celt_norm, tmp);
578
177k
   int N;
579
177k
   SAVE_STACK;
580
177k
   N = N0*stride;
581
177k
   ALLOC(tmp, N, celt_norm);
582
177k
   celt_assert(stride>0);
583
177k
   if (hadamard)
584
111k
   {
585
111k
      const int *ordery = ordery_table+stride-2;
586
633k
      for (i=0;i<stride;i++)
587
521k
      {
588
2.29M
         for (j=0;j<N0;j++)
589
1.76M
            tmp[ordery[i]*N0+j] = X[j*stride+i];
590
521k
      }
591
111k
   } else {
592
425k
      for (i=0;i<stride;i++)
593
1.38M
         for (j=0;j<N0;j++)
594
1.02M
            tmp[i*N0+j] = X[j*stride+i];
595
66.0k
   }
596
177k
   OPUS_COPY(X, tmp, N);
597
177k
   RESTORE_STACK;
598
177k
}
599
600
static void interleave_hadamard(celt_norm *X, int N0, int stride, int hadamard)
601
232k
{
602
232k
   int i,j;
603
232k
   VARDECL(celt_norm, tmp);
604
232k
   int N;
605
232k
   SAVE_STACK;
606
232k
   N = N0*stride;
607
232k
   ALLOC(tmp, N, celt_norm);
608
232k
   if (hadamard)
609
153k
   {
610
153k
      const int *ordery = ordery_table+stride-2;
611
902k
      for (i=0;i<stride;i++)
612
3.57M
         for (j=0;j<N0;j++)
613
2.82M
            tmp[j*stride+i] = X[ordery[i]*N0+j];
614
153k
   } else {
615
544k
      for (i=0;i<stride;i++)
616
1.83M
         for (j=0;j<N0;j++)
617
1.36M
            tmp[j*stride+i] = X[i*N0+j];
618
78.5k
   }
619
232k
   OPUS_COPY(X, tmp, N);
620
232k
   RESTORE_STACK;
621
232k
}
622
623
void haar1(celt_norm *X, int N0, int stride)
624
860k
{
625
860k
   int i, j;
626
860k
   N0 >>= 1;
627
2.45M
   for (i=0;i<stride;i++)
628
9.57M
      for (j=0;j<N0;j++)
629
7.98M
      {
630
7.98M
         opus_val32 tmp1, tmp2;
631
7.98M
         tmp1 = MULT32_32_Q31(QCONST32(.70710678f,31), X[stride*2*j+i]);
632
7.98M
         tmp2 = MULT32_32_Q31(QCONST32(.70710678f,31), X[stride*(2*j+1)+i]);
633
7.98M
         X[stride*2*j+i] = ADD32(tmp1, tmp2);
634
7.98M
         X[stride*(2*j+1)+i] = SUB32(tmp1, tmp2);
635
7.98M
      }
636
860k
}
637
638
static int compute_qn(int N, int b, int offset, int pulse_cap, int stereo)
639
503k
{
640
503k
   static const opus_int16 exp2_table8[8] =
641
503k
      {16384, 17866, 19483, 21247, 23170, 25267, 27554, 30048};
642
503k
   int qn, qb;
643
503k
   int N2 = 2*N-1;
644
503k
   if (stereo && N==2)
645
74.0k
      N2--;
646
   /* The upper limit ensures that in a stereo split with itheta==16384, we'll
647
       always have enough bits left over to code at least one pulse in the
648
       side; otherwise it would collapse, since it doesn't get folded. */
649
503k
   qb = celt_sudiv(b+N2*offset, N2);
650
503k
   qb = IMIN(b-pulse_cap-(4<<BITRES), qb);
651
652
503k
   qb = IMIN(8<<BITRES, qb);
653
654
503k
   if (qb<(1<<BITRES>>1)) {
655
180k
      qn = 1;
656
323k
   } else {
657
323k
      qn = exp2_table8[qb&0x7]>>(14-(qb>>BITRES));
658
323k
      qn = (qn+1)>>1<<1;
659
323k
   }
660
503k
   celt_assert(qn <= 256);
661
503k
   return qn;
662
503k
}
663
664
struct band_ctx {
665
   int encode;
666
   int resynth;
667
   const CELTMode *m;
668
   int i;
669
   int intensity;
670
   int spread;
671
   int tf_change;
672
   ec_ctx *ec;
673
   opus_int32 remaining_bits;
674
   const celt_ener *bandE;
675
   opus_uint32 seed;
676
   int arch;
677
   int theta_round;
678
   int disable_inv;
679
   int avoid_split_noise;
680
#ifdef ENABLE_QEXT
681
   ec_ctx *ext_ec;
682
   int extra_bits;
683
   opus_int32 ext_total_bits;
684
   int extra_bands;
685
#endif
686
};
687
688
struct split_ctx {
689
   int inv;
690
   int imid;
691
   int iside;
692
   int delta;
693
   int itheta;
694
#ifdef ENABLE_QEXT
695
   int itheta_q30;
696
#endif
697
   int qalloc;
698
};
699
700
static void compute_theta(struct band_ctx *ctx, struct split_ctx *sctx,
701
      celt_norm *X, celt_norm *Y, int N, int *b, int B, int B0,
702
      int LM,
703
      int stereo, int *fill ARG_QEXT(int *ext_b))
704
503k
{
705
503k
   int qn;
706
503k
   int itheta=0;
707
503k
   int itheta_q30=0;
708
503k
   int delta;
709
503k
   int imid, iside;
710
503k
   int qalloc;
711
503k
   int pulse_cap;
712
503k
   int offset;
713
503k
   opus_int32 tell;
714
503k
   int inv=0;
715
503k
   int encode;
716
503k
   const CELTMode *m;
717
503k
   int i;
718
503k
   int intensity;
719
503k
   ec_ctx *ec;
720
503k
   const celt_ener *bandE;
721
722
503k
   encode = ctx->encode;
723
503k
   m = ctx->m;
724
503k
   i = ctx->i;
725
503k
   intensity = ctx->intensity;
726
503k
   ec = ctx->ec;
727
503k
   bandE = ctx->bandE;
728
729
   /* Decide on the resolution to give to the split parameter theta */
730
503k
   pulse_cap = m->logN[i]+LM*(1<<BITRES);
731
503k
   offset = (pulse_cap>>1) - (stereo&&N==2 ? QTHETA_OFFSET_TWOPHASE : QTHETA_OFFSET);
732
503k
   qn = compute_qn(N, *b, offset, pulse_cap, stereo);
733
503k
   if (stereo && i>=intensity)
734
234k
      qn = 1;
735
503k
   if (encode)
736
0
   {
737
      /* theta is the atan() of the ratio between the (normalized)
738
         side and mid. With just that parameter, we can re-scale both
739
         mid and side because we know that 1) they have unit norm and
740
         2) they are orthogonal. */
741
0
      itheta_q30 = stereo_itheta(X, Y, stereo, N, ctx->arch);
742
0
      itheta = itheta_q30>>16;
743
0
   }
744
503k
   tell = ec_tell_frac(ec);
745
503k
   if (qn!=1)
746
263k
   {
747
263k
      if (encode)
748
0
      {
749
0
         if (!stereo || ctx->theta_round == 0)
750
0
         {
751
0
            itheta = (itheta*(opus_int32)qn+8192)>>14;
752
0
            if (!stereo && ctx->avoid_split_noise && itheta > 0 && itheta < qn)
753
0
            {
754
               /* Check if the selected value of theta will cause the bit allocation
755
                  to inject noise on one side. If so, make sure the energy of that side
756
                  is zero. */
757
0
               int unquantized = celt_udiv((opus_int32)itheta*16384, qn);
758
0
               imid = bitexact_cos((opus_int16)unquantized);
759
0
               iside = bitexact_cos((opus_int16)(16384-unquantized));
760
0
               delta = FRAC_MUL16((N-1)<<7,bitexact_log2tan(iside,imid));
761
0
               if (delta > *b)
762
0
                  itheta = qn;
763
0
               else if (delta < -*b)
764
0
                  itheta = 0;
765
0
            }
766
0
         } else {
767
0
            int down;
768
            /* Bias quantization towards itheta=0 and itheta=16384. */
769
0
            int bias = itheta > 8192 ? 32767/qn : -32767/qn;
770
0
            down = IMIN(qn-1, IMAX(0, (itheta*(opus_int32)qn + bias)>>14));
771
0
            if (ctx->theta_round < 0)
772
0
               itheta = down;
773
0
            else
774
0
               itheta = down+1;
775
0
         }
776
0
      }
777
      /* Entropy coding of the angle. We use a uniform pdf for the
778
         time split, a step for stereo, and a triangular one for the rest. */
779
263k
      if (stereo && N>2)
780
17.2k
      {
781
17.2k
         int p0 = 3;
782
17.2k
         int x = itheta;
783
17.2k
         int x0 = qn/2;
784
17.2k
         int ft = p0*(x0+1) + x0;
785
         /* Use a probability of p0 up to itheta=8192 and then use 1 after */
786
17.2k
         if (encode)
787
0
         {
788
0
            ec_encode(ec,x<=x0?p0*x:(x-1-x0)+(x0+1)*p0,x<=x0?p0*(x+1):(x-x0)+(x0+1)*p0,ft);
789
17.2k
         } else {
790
17.2k
            int fs;
791
17.2k
            fs=ec_decode(ec,ft);
792
17.2k
            if (fs<(x0+1)*p0)
793
13.1k
               x=fs/p0;
794
4.11k
            else
795
4.11k
               x=x0+1+(fs-(x0+1)*p0);
796
17.2k
            ec_dec_update(ec,x<=x0?p0*x:(x-1-x0)+(x0+1)*p0,x<=x0?p0*(x+1):(x-x0)+(x0+1)*p0,ft);
797
17.2k
            itheta = x;
798
17.2k
         }
799
245k
      } else if (B0>1 || stereo) {
800
         /* Uniform pdf */
801
95.2k
         if (encode)
802
0
            ec_enc_uint(ec, itheta, qn+1);
803
95.2k
         else
804
95.2k
            itheta = ec_dec_uint(ec, qn+1);
805
150k
      } else {
806
150k
         int fs=1, ft;
807
150k
         ft = ((qn>>1)+1)*((qn>>1)+1);
808
150k
         if (encode)
809
0
         {
810
0
            int fl;
811
812
0
            fs = itheta <= (qn>>1) ? itheta + 1 : qn + 1 - itheta;
813
0
            fl = itheta <= (qn>>1) ? itheta*(itheta + 1)>>1 :
814
0
             ft - ((qn + 1 - itheta)*(qn + 2 - itheta)>>1);
815
816
0
            ec_encode(ec, fl, fl+fs, ft);
817
150k
         } else {
818
            /* Triangular pdf */
819
150k
            int fl=0;
820
150k
            int fm;
821
150k
            fm = ec_decode(ec, ft);
822
823
150k
            if (fm < ((qn>>1)*((qn>>1) + 1)>>1))
824
73.0k
            {
825
73.0k
               itheta = (isqrt32(8*(opus_uint32)fm + 1) - 1)>>1;
826
73.0k
               fs = itheta + 1;
827
73.0k
               fl = itheta*(itheta + 1)>>1;
828
73.0k
            }
829
77.6k
            else
830
77.6k
            {
831
77.6k
               itheta = (2*(qn + 1)
832
77.6k
                - isqrt32(8*(opus_uint32)(ft - fm - 1) + 1))>>1;
833
77.6k
               fs = qn + 1 - itheta;
834
77.6k
               fl = ft - ((qn + 1 - itheta)*(qn + 2 - itheta)>>1);
835
77.6k
            }
836
837
150k
            ec_dec_update(ec, fl, fl+fs, ft);
838
150k
         }
839
150k
      }
840
263k
      celt_assert(itheta>=0);
841
263k
      itheta = celt_udiv((opus_int32)itheta*16384, qn);
842
#ifdef ENABLE_QEXT
843
      *ext_b = IMIN(*ext_b, ctx->ext_total_bits - (opus_int32)ec_tell_frac(ctx->ext_ec));
844
      if (*ext_b >= 2*N<<BITRES && ctx->ext_total_bits-ec_tell_frac(ctx->ext_ec)-1 > 2<<BITRES) {
845
         int extra_bits;
846
         int ext_tell = ec_tell_frac(ctx->ext_ec);
847
         extra_bits = IMIN(12, IMAX(2, celt_sudiv(*ext_b, (2*N-1)<<BITRES)));
848
         if (encode) {
849
            itheta_q30 = itheta_q30 - (itheta<<16);
850
            itheta_q30 = (itheta_q30*(opus_int64)qn*((1<<extra_bits)-1)+(1<<29))>>30;
851
            itheta_q30 += (1<<(extra_bits-1))-1;
852
            itheta_q30 = IMAX(0, IMIN((1<<extra_bits)-2, itheta_q30));
853
            ec_enc_uint(ctx->ext_ec, itheta_q30, (1<<extra_bits)-1);
854
         } else {
855
            itheta_q30 = ec_dec_uint(ctx->ext_ec, (1<<extra_bits)-1);
856
         }
857
         itheta_q30 -= (1<<(extra_bits-1))-1;
858
         itheta_q30 = (itheta<<16) + itheta_q30*(opus_int64)(1<<30)/(qn*((1<<extra_bits)-1));
859
         /* Hard bounds on itheta (can only trigger on corrupted bitstreams). */
860
         itheta_q30 = IMAX(0, IMIN(itheta_q30, 1073741824));
861
         *ext_b -= ec_tell_frac(ctx->ext_ec) - ext_tell;
862
      } else {
863
         itheta_q30 = (opus_int32)itheta<<16;
864
      }
865
#endif
866
263k
      if (encode && stereo)
867
0
      {
868
0
         if (itheta==0)
869
0
            intensity_stereo(m, X, Y, bandE, i, N);
870
0
         else
871
0
            stereo_split(X, Y, N);
872
0
      }
873
      /* NOTE: Renormalising X and Y *may* help fixed-point a bit at very high rate.
874
               Let's do that at higher complexity */
875
263k
   } else if (stereo) {
876
240k
      if (encode)
877
0
      {
878
0
         inv = itheta > 8192 && !ctx->disable_inv;
879
0
         if (inv)
880
0
         {
881
0
            int j;
882
0
            for (j=0;j<N;j++)
883
0
               Y[j] = -Y[j];
884
0
         }
885
0
         intensity_stereo(m, X, Y, bandE, i, N);
886
0
      }
887
240k
      if (*b>2<<BITRES && ctx->remaining_bits > 2<<BITRES)
888
75.0k
      {
889
75.0k
         if (encode)
890
0
            ec_enc_bit_logp(ec, inv, 2);
891
75.0k
         else
892
75.0k
            inv = ec_dec_bit_logp(ec, 2);
893
75.0k
      } else
894
165k
         inv = 0;
895
      /* inv flag override to avoid problems with downmixing. */
896
240k
      if (ctx->disable_inv)
897
0
         inv = 0;
898
240k
      itheta = 0;
899
240k
      itheta_q30 = 0;
900
240k
   }
901
503k
   qalloc = ec_tell_frac(ec) - tell;
902
503k
   *b -= qalloc;
903
904
503k
   if (itheta == 0)
905
247k
   {
906
247k
      imid = 32767;
907
247k
      iside = 0;
908
247k
      *fill &= (1<<B)-1;
909
247k
      delta = -16384;
910
255k
   } else if (itheta == 16384)
911
5.74k
   {
912
5.74k
      imid = 0;
913
5.74k
      iside = 32767;
914
5.74k
      *fill &= ((1<<B)-1)<<B;
915
5.74k
      delta = 16384;
916
249k
   } else {
917
249k
      imid = bitexact_cos((opus_int16)itheta);
918
249k
      iside = bitexact_cos((opus_int16)(16384-itheta));
919
      /* This is the mid vs side allocation that minimizes squared error
920
         in that band. */
921
249k
      delta = FRAC_MUL16((N-1)<<7,bitexact_log2tan(iside,imid));
922
249k
   }
923
924
503k
   sctx->inv = inv;
925
503k
   sctx->imid = imid;
926
503k
   sctx->iside = iside;
927
503k
   sctx->delta = delta;
928
503k
   sctx->itheta = itheta;
929
#ifdef ENABLE_QEXT
930
   sctx->itheta_q30 = itheta_q30;
931
#endif
932
503k
   sctx->qalloc = qalloc;
933
503k
}
934
static unsigned quant_band_n1(struct band_ctx *ctx, celt_norm *X, celt_norm *Y,
935
      celt_norm *lowband_out)
936
154k
{
937
154k
   int c;
938
154k
   int stereo;
939
154k
   celt_norm *x = X;
940
154k
   int encode;
941
154k
   ec_ctx *ec;
942
943
154k
   encode = ctx->encode;
944
154k
   ec = ctx->ec;
945
946
154k
   stereo = Y != NULL;
947
218k
   c=0; do {
948
218k
      int sign=0;
949
218k
      if (ctx->remaining_bits>=1<<BITRES)
950
82.7k
      {
951
82.7k
         if (encode)
952
0
         {
953
0
            sign = x[0]<0;
954
0
            ec_enc_bits(ec, sign, 1);
955
82.7k
         } else {
956
82.7k
            sign = ec_dec_bits(ec, 1);
957
82.7k
         }
958
82.7k
         ctx->remaining_bits -= 1<<BITRES;
959
82.7k
      }
960
218k
      if (ctx->resynth)
961
218k
         x[0] = sign ? -NORM_SCALING : NORM_SCALING;
962
218k
      x = Y;
963
218k
   } while (++c<1+stereo);
964
154k
   if (lowband_out)
965
154k
      lowband_out[0] = SHR32(X[0],4);
966
154k
   return 1;
967
154k
}
968
969
/* This function is responsible for encoding and decoding a mono partition.
970
   It can split the band in two and transmit the energy difference with
971
   the two half-bands. It can be called recursively so bands can end up being
972
   split in 8 parts. */
973
static unsigned quant_partition(struct band_ctx *ctx, celt_norm *X,
974
      int N, int b, int B, celt_norm *lowband,
975
      int LM,
976
      opus_val32 gain, int fill
977
      ARG_QEXT(int ext_b))
978
1.66M
{
979
1.66M
   const unsigned char *cache;
980
1.66M
   int q;
981
1.66M
   int curr_bits;
982
1.66M
   int imid=0, iside=0;
983
1.66M
   int B0=B;
984
1.66M
   opus_val32 mid=0, side=0;
985
1.66M
   unsigned cm=0;
986
1.66M
   celt_norm *Y=NULL;
987
1.66M
   int encode;
988
1.66M
   const CELTMode *m;
989
1.66M
   int i;
990
1.66M
   int spread;
991
1.66M
   ec_ctx *ec;
992
993
1.66M
   encode = ctx->encode;
994
1.66M
   m = ctx->m;
995
1.66M
   i = ctx->i;
996
1.66M
   spread = ctx->spread;
997
1.66M
   ec = ctx->ec;
998
999
   /* If we need 1.5 more bit than we can produce, split the band in two. */
1000
1.66M
   cache = m->cache.bits + m->cache.index[(LM+1)*m->nbEBands+i];
1001
1.66M
   if (LM != -1 && b > cache[cache[0]]+12 && N>2)
1002
229k
   {
1003
229k
      int mbits, sbits, delta;
1004
229k
      int itheta;
1005
229k
      int qalloc;
1006
229k
      struct split_ctx sctx;
1007
229k
      celt_norm *next_lowband2=NULL;
1008
229k
      opus_int32 rebalance;
1009
1010
229k
      N >>= 1;
1011
229k
      Y = X+N;
1012
229k
      LM -= 1;
1013
229k
      if (B==1)
1014
150k
         fill = (fill&1)|(fill<<1);
1015
229k
      B = (B+1)>>1;
1016
1017
229k
      compute_theta(ctx, &sctx, X, Y, N, &b, B, B0, LM, 0, &fill ARG_QEXT(&ext_b));
1018
229k
      imid = sctx.imid;
1019
229k
      iside = sctx.iside;
1020
229k
      delta = sctx.delta;
1021
229k
      itheta = sctx.itheta;
1022
229k
      qalloc = sctx.qalloc;
1023
#ifdef FIXED_POINT
1024
# ifdef ENABLE_QEXT
1025
      (void)imid;
1026
      (void)iside;
1027
      mid = celt_cos_norm32(sctx.itheta_q30);
1028
      side = celt_cos_norm32((1<<30)-sctx.itheta_q30);
1029
# else
1030
      mid = SHL32(EXTEND32(imid), 16);
1031
      side = SHL32(EXTEND32(iside), 16);
1032
# endif
1033
#else
1034
# ifdef ENABLE_QEXT
1035
      (void)imid;
1036
      (void)iside;
1037
      mid = celt_cos_norm2(sctx.itheta_q30*(1.f/(1<<30)));
1038
      side = celt_cos_norm2(1.f-sctx.itheta_q30*(1.f/(1<<30)));
1039
# else
1040
229k
      mid = (1.f/32768)*imid;
1041
229k
      side = (1.f/32768)*iside;
1042
229k
# endif
1043
229k
#endif
1044
1045
      /* Give more bits to low-energy MDCTs than they would otherwise deserve */
1046
229k
      if (B0>1 && (itheta&0x3fff))
1047
73.8k
      {
1048
73.8k
         if (itheta > 8192)
1049
            /* Rough approximation for pre-echo masking */
1050
32.0k
            delta -= delta>>(4-LM);
1051
41.7k
         else
1052
            /* Corresponds to a forward-masking slope of 1.5 dB per 10 ms */
1053
41.7k
            delta = IMIN(0, delta + (N<<BITRES>>(5-LM)));
1054
73.8k
      }
1055
229k
      mbits = IMAX(0, IMIN(b, (b-delta)/2));
1056
229k
      sbits = b-mbits;
1057
229k
      ctx->remaining_bits -= qalloc;
1058
1059
229k
      if (lowband)
1060
170k
         next_lowband2 = lowband+N; /* >32-bit split case */
1061
1062
229k
      rebalance = ctx->remaining_bits;
1063
229k
      if (mbits >= sbits)
1064
117k
      {
1065
117k
         cm = quant_partition(ctx, X, N, mbits, B, lowband, LM,
1066
117k
               MULT32_32_Q31(gain,mid), fill ARG_QEXT(ext_b/2));
1067
117k
         rebalance = mbits - (rebalance-ctx->remaining_bits);
1068
117k
         if (rebalance > 3<<BITRES && itheta!=0)
1069
54.2k
            sbits += rebalance - (3<<BITRES);
1070
117k
         cm |= quant_partition(ctx, Y, N, sbits, B, next_lowband2, LM,
1071
117k
               MULT32_32_Q31(gain,side), fill>>B ARG_QEXT(ext_b/2))<<(B0>>1);
1072
117k
      } else {
1073
111k
         cm = quant_partition(ctx, Y, N, sbits, B, next_lowband2, LM,
1074
111k
               MULT32_32_Q31(gain,side), fill>>B ARG_QEXT(ext_b/2))<<(B0>>1);
1075
111k
         rebalance = sbits - (rebalance-ctx->remaining_bits);
1076
111k
         if (rebalance > 3<<BITRES && itheta!=16384)
1077
54.8k
            mbits += rebalance - (3<<BITRES);
1078
111k
         cm |= quant_partition(ctx, X, N, mbits, B, lowband, LM,
1079
111k
               MULT32_32_Q31(gain,mid), fill ARG_QEXT(ext_b/2));
1080
111k
      }
1081
1.43M
   } else {
1082
#ifdef ENABLE_QEXT
1083
      int extra_bits;
1084
      int ext_remaining_bits;
1085
      extra_bits = ext_b/(N-1)>>BITRES;
1086
      ext_remaining_bits = ctx->ext_total_bits-(opus_int32)ec_tell_frac(ctx->ext_ec);
1087
      if (ext_remaining_bits < ((extra_bits+1)*(N-1)+N)<<BITRES) {
1088
         extra_bits = (ext_remaining_bits-(N<<BITRES))/(N-1)>>BITRES;
1089
         extra_bits = IMAX(extra_bits-1, 0);
1090
      }
1091
      extra_bits = IMIN(12, extra_bits);
1092
#endif
1093
      /* This is the basic no-split case */
1094
1.43M
      q = bits2pulses(m, i, LM, b);
1095
1.43M
      curr_bits = pulses2bits(m, i, LM, q);
1096
1.43M
      ctx->remaining_bits -= curr_bits;
1097
1098
      /* Ensures we can never bust the budget */
1099
1.45M
      while (ctx->remaining_bits < 0 && q > 0)
1100
16.0k
      {
1101
16.0k
         ctx->remaining_bits += curr_bits;
1102
16.0k
         q--;
1103
16.0k
         curr_bits = pulses2bits(m, i, LM, q);
1104
16.0k
         ctx->remaining_bits -= curr_bits;
1105
16.0k
      }
1106
1107
1.43M
      if (q!=0)
1108
649k
      {
1109
649k
         int K = get_pulses(q);
1110
1111
         /* Finally do the actual quantization */
1112
649k
         if (encode)
1113
0
         {
1114
0
            cm = alg_quant(X, N, K, spread, B, ec, gain, ctx->resynth
1115
0
                           ARG_QEXT(ctx->ext_ec) ARG_QEXT(extra_bits),
1116
0
                           ctx->arch);
1117
649k
         } else {
1118
649k
            cm = alg_unquant(X, N, K, spread, B, ec, gain
1119
649k
                             ARG_QEXT(ctx->ext_ec) ARG_QEXT(extra_bits));
1120
649k
         }
1121
#ifdef ENABLE_QEXT
1122
      } else if (ext_b > 2*N<<BITRES)
1123
      {
1124
         extra_bits = ext_b/(N-1)>>BITRES;
1125
         ext_remaining_bits = ctx->ext_total_bits-ec_tell_frac(ctx->ext_ec);
1126
         if (ext_remaining_bits < ((extra_bits+1)*(N-1)+N)<<BITRES) {
1127
            extra_bits = (ext_remaining_bits-(N<<BITRES))/(N-1)>>BITRES;
1128
            extra_bits = IMAX(extra_bits-1, 0);
1129
         }
1130
         extra_bits = IMIN(14, extra_bits);
1131
         if (encode) cm = cubic_quant(X, N, extra_bits, B, ctx->ext_ec, gain, ctx->resynth);
1132
         else cm = cubic_unquant(X, N, extra_bits, B, ctx->ext_ec, gain);
1133
#endif
1134
789k
      } else {
1135
         /* If there's no pulse, fill the band anyway */
1136
789k
         int j;
1137
789k
         if (ctx->resynth)
1138
789k
         {
1139
789k
            unsigned cm_mask;
1140
            /* B can be as large as 16, so this shift might overflow an int on a
1141
               16-bit platform; use a long to get defined behavior.*/
1142
789k
            cm_mask = (unsigned)(1UL<<B)-1;
1143
789k
            fill &= cm_mask;
1144
789k
            if (!fill)
1145
194k
            {
1146
194k
               OPUS_CLEAR(X, N);
1147
594k
            } else {
1148
594k
               if (lowband == NULL)
1149
36.8k
               {
1150
                  /* Noise */
1151
1.03M
                  for (j=0;j<N;j++)
1152
993k
                  {
1153
993k
                     ctx->seed = celt_lcg_rand(ctx->seed);
1154
993k
                     X[j] = SHL32((celt_norm)((opus_int32)ctx->seed>>20), NORM_SHIFT-14);
1155
993k
                  }
1156
36.8k
                  cm = cm_mask;
1157
557k
               } else {
1158
                  /* Folded spectrum */
1159
10.8M
                  for (j=0;j<N;j++)
1160
10.2M
                  {
1161
10.2M
                     opus_val16 tmp;
1162
10.2M
                     ctx->seed = celt_lcg_rand(ctx->seed);
1163
                     /* About 48 dB below the "normal" folding level */
1164
10.2M
                     tmp = QCONST16(1.0f/256, NORM_SHIFT-4);
1165
10.2M
                     tmp = (ctx->seed)&0x8000 ? tmp : -tmp;
1166
10.2M
                     X[j] = lowband[j]+tmp;
1167
10.2M
                  }
1168
557k
                  cm = fill;
1169
557k
               }
1170
594k
               renormalise_vector(X, N, gain, ctx->arch);
1171
594k
            }
1172
789k
         }
1173
789k
      }
1174
1.43M
   }
1175
1176
1.66M
   return cm;
1177
1.66M
}
1178
1179
#ifdef ENABLE_QEXT
1180
static unsigned cubic_quant_partition(struct band_ctx *ctx, celt_norm *X, int N, int b, int B, ec_ctx *ec, int LM, opus_val32 gain, int resynth, int encode)
1181
{
1182
   celt_assert(LM>=0);
1183
   ctx->remaining_bits = ctx->ec->storage*8*8 - ec_tell_frac(ctx->ec);
1184
   b = IMIN(b, ctx->remaining_bits);
1185
   /* As long as we have at least two bits of depth, split all the way to LM=0 (not -1 like PVQ). */
1186
   if (LM==0 || b<=2*N<<BITRES) {
1187
      int res, ret;
1188
      b = IMIN(b + ((N-1)<<BITRES)/2, ctx->remaining_bits);
1189
      /* Resolution left after taking into account coding the cube face. */
1190
      res = (b-(1<<BITRES)-ctx->m->logN[ctx->i]-(LM<<BITRES)-1)/(N-1)>>BITRES;
1191
      res = IMIN(14, IMAX(0, res));
1192
      if (encode) ret = cubic_quant(X, N, res, B, ec, gain, resynth);
1193
      else ret = cubic_unquant(X, N, res, B, ec, gain);
1194
      ctx->remaining_bits = ctx->ec->storage*8*8 - ec_tell_frac(ctx->ec);
1195
      return ret;
1196
   } else {
1197
      celt_norm *Y;
1198
      opus_int32 itheta_q30;
1199
      opus_val32 g1, g2;
1200
      opus_int32 theta_res;
1201
      opus_int32 qtheta;
1202
      int delta;
1203
      int b1, b2;
1204
      int cm;
1205
      int N0;
1206
      N0 = N;
1207
      N >>= 1;
1208
      Y = X+N;
1209
      LM -= 1;
1210
      B = (B+1)>>1;
1211
      theta_res = IMIN(16, (b>>BITRES)/(N0-1) + 1);
1212
      if (encode) {
1213
         itheta_q30 = stereo_itheta(X, Y, 0, N, ctx->arch);
1214
         qtheta = (itheta_q30+(1<<(29-theta_res)))>>(30-theta_res);
1215
         ec_enc_uint(ec, qtheta, (1<<theta_res)+1);
1216
      } else {
1217
         qtheta = ec_dec_uint(ec, (1<<theta_res)+1);
1218
      }
1219
      itheta_q30 = qtheta<<(30-theta_res);
1220
      b -= theta_res<<BITRES;
1221
      delta = (N0-1) * 23 * ((itheta_q30>>16)-8192) >> (17-BITRES);
1222
1223
#ifdef FIXED_POINT
1224
      g1 = celt_cos_norm32(itheta_q30);
1225
      g2 = celt_cos_norm32((1<<30)-itheta_q30);
1226
#else
1227
      g1 = celt_cos_norm2(itheta_q30*(1.f/(1<<30)));
1228
      g2 = celt_cos_norm2(1.f-itheta_q30*(1.f/(1<<30)));
1229
#endif
1230
      if (itheta_q30 == 0) {
1231
         b1=b;
1232
         b2=0;
1233
      } else if (itheta_q30==1073741824) {
1234
         b1=0;
1235
         b2=b;
1236
      } else {
1237
         b1 = IMIN(b, IMAX(0, (b-delta)/2));
1238
         b2 = b-b1;
1239
      }
1240
      cm  = cubic_quant_partition(ctx, X, N, b1, B, ec, LM, MULT32_32_Q31(gain, g1), resynth, encode);
1241
      cm |= cubic_quant_partition(ctx, Y, N, b2, B, ec, LM, MULT32_32_Q31(gain, g2), resynth, encode);
1242
      return cm;
1243
   }
1244
}
1245
#endif
1246
1247
/* This function is responsible for encoding and decoding a band for the mono case. */
1248
static unsigned quant_band(struct band_ctx *ctx, celt_norm *X,
1249
      int N, int b, int B, celt_norm *lowband,
1250
      int LM, celt_norm *lowband_out,
1251
      opus_val32 gain, celt_norm *lowband_scratch, int fill
1252
      ARG_QEXT(int ext_b))
1253
1.30M
{
1254
1.30M
   int N0=N;
1255
1.30M
   int N_B=N;
1256
1.30M
   int N_B0;
1257
1.30M
   int B0=B;
1258
1.30M
   int time_divide=0;
1259
1.30M
   int recombine=0;
1260
1.30M
   int longBlocks;
1261
1.30M
   unsigned cm=0;
1262
1.30M
   int k;
1263
1.30M
   int encode;
1264
1.30M
   int tf_change;
1265
1266
1.30M
   encode = ctx->encode;
1267
1.30M
   tf_change = ctx->tf_change;
1268
1269
1.30M
   longBlocks = B0==1;
1270
1271
1.30M
   N_B = celt_udiv(N_B, B);
1272
1273
   /* Special case for one sample */
1274
1.30M
   if (N==1)
1275
91.6k
   {
1276
91.6k
      return quant_band_n1(ctx, X, NULL, lowband_out);
1277
91.6k
   }
1278
1279
1.20M
   if (tf_change>0)
1280
92.8k
      recombine = tf_change;
1281
   /* Band recombining to increase frequency resolution */
1282
1283
1.20M
   if (lowband_scratch && lowband && (recombine || ((N_B&1) == 0 && tf_change<0) || B0>1))
1284
218k
   {
1285
218k
      OPUS_COPY(lowband_scratch, lowband, N);
1286
218k
      lowband = lowband_scratch;
1287
218k
   }
1288
1289
1.36M
   for (k=0;k<recombine;k++)
1290
156k
   {
1291
156k
      static const unsigned char bit_interleave_table[16]={
1292
156k
            0,1,1,1,2,3,3,3,2,3,3,3,2,3,3,3
1293
156k
      };
1294
156k
      if (encode)
1295
0
         haar1(X, N>>k, 1<<k);
1296
156k
      if (lowband)
1297
129k
         haar1(lowband, N>>k, 1<<k);
1298
156k
      fill = bit_interleave_table[fill&0xF]|bit_interleave_table[fill>>4]<<2;
1299
156k
   }
1300
1.20M
   B>>=recombine;
1301
1.20M
   N_B<<=recombine;
1302
1303
   /* Increasing the time resolution */
1304
1.54M
   while ((N_B&1) == 0 && tf_change<0)
1305
336k
   {
1306
336k
      if (encode)
1307
0
         haar1(X, N_B, B);
1308
336k
      if (lowband)
1309
239k
         haar1(lowband, N_B, B);
1310
336k
      fill |= fill<<B;
1311
336k
      B <<= 1;
1312
336k
      N_B >>= 1;
1313
336k
      time_divide++;
1314
336k
      tf_change++;
1315
336k
   }
1316
1.20M
   B0=B;
1317
1.20M
   N_B0 = N_B;
1318
1319
   /* Reorganize the samples in time order instead of frequency order */
1320
1.20M
   if (B0>1)
1321
232k
   {
1322
232k
      if (encode)
1323
0
         deinterleave_hadamard(X, N_B>>recombine, B0<<recombine, longBlocks);
1324
232k
      if (lowband)
1325
177k
         deinterleave_hadamard(lowband, N_B>>recombine, B0<<recombine, longBlocks);
1326
232k
   }
1327
1328
#ifdef ENABLE_QEXT
1329
   if (ctx->extra_bands && b > (3*N<<BITRES)+(ctx->m->logN[ctx->i]+8+8*LM)) {
1330
      cm = cubic_quant_partition(ctx, X, N, b, B, ctx->ec, LM, gain, ctx->resynth, encode);
1331
   } else
1332
#endif
1333
1.20M
   {
1334
1.20M
      cm = quant_partition(ctx, X, N, b, B, lowband, LM, gain, fill ARG_QEXT(ext_b));
1335
1.20M
   }
1336
1337
   /* This code is used by the decoder and by the resynthesis-enabled encoder */
1338
1.20M
   if (ctx->resynth)
1339
1.20M
   {
1340
      /* Undo the sample reorganization going from time order to frequency order */
1341
1.20M
      if (B0>1)
1342
232k
         interleave_hadamard(X, N_B>>recombine, B0<<recombine, longBlocks);
1343
1344
      /* Undo time-freq changes that we did earlier */
1345
1.20M
      N_B = N_B0;
1346
1.20M
      B = B0;
1347
1.54M
      for (k=0;k<time_divide;k++)
1348
336k
      {
1349
336k
         B >>= 1;
1350
336k
         N_B <<= 1;
1351
336k
         cm |= cm>>B;
1352
336k
         haar1(X, N_B, B);
1353
336k
      }
1354
1355
1.36M
      for (k=0;k<recombine;k++)
1356
156k
      {
1357
156k
         static const unsigned char bit_deinterleave_table[16]={
1358
156k
               0x00,0x03,0x0C,0x0F,0x30,0x33,0x3C,0x3F,
1359
156k
               0xC0,0xC3,0xCC,0xCF,0xF0,0xF3,0xFC,0xFF
1360
156k
         };
1361
156k
         cm = bit_deinterleave_table[cm];
1362
156k
         haar1(X, N0>>k, 1<<k);
1363
156k
      }
1364
1.20M
      B<<=recombine;
1365
1366
      /* Scale output for later folding */
1367
1.20M
      if (lowband_out)
1368
916k
      {
1369
916k
         int j;
1370
916k
         opus_val16 n;
1371
916k
         n = celt_sqrt(SHL32(EXTEND32(N0),22));
1372
11.1M
         for (j=0;j<N0;j++)
1373
10.2M
            lowband_out[j] = MULT16_32_Q15(n,X[j]);
1374
916k
      }
1375
1.20M
      cm &= (1<<B)-1;
1376
1.20M
   }
1377
1.20M
   return cm;
1378
1.30M
}
1379
1380
#ifdef FIXED_POINT
1381
#define MIN_STEREO_ENERGY 2
1382
#else
1383
0
#define MIN_STEREO_ENERGY 1e-10f
1384
#endif
1385
1386
/* This function is responsible for encoding and decoding a band for the stereo case. */
1387
static unsigned quant_band_stereo(struct band_ctx *ctx, celt_norm *X, celt_norm *Y,
1388
      int N, int b, int B, celt_norm *lowband,
1389
      int LM, celt_norm *lowband_out,
1390
      celt_norm *lowband_scratch, int fill
1391
      ARG_QEXT(int ext_b) ARG_QEXT(const int *cap))
1392
337k
{
1393
337k
   int imid=0, iside=0;
1394
337k
   int inv = 0;
1395
337k
   opus_val32 mid=0, side=0;
1396
337k
   unsigned cm=0;
1397
337k
   int mbits, sbits, delta;
1398
337k
   int itheta;
1399
337k
   int qalloc;
1400
337k
   struct split_ctx sctx;
1401
337k
   int orig_fill;
1402
337k
   int encode;
1403
337k
   ec_ctx *ec;
1404
1405
337k
   encode = ctx->encode;
1406
337k
   ec = ctx->ec;
1407
1408
   /* Special case for one sample */
1409
337k
   if (N==1)
1410
63.3k
   {
1411
63.3k
      return quant_band_n1(ctx, X, Y, lowband_out);
1412
63.3k
   }
1413
1414
274k
   orig_fill = fill;
1415
1416
274k
   if (encode) {
1417
0
      if (ctx->bandE[ctx->i] < MIN_STEREO_ENERGY || ctx->bandE[ctx->m->nbEBands+ctx->i] < MIN_STEREO_ENERGY) {
1418
0
         if (ctx->bandE[ctx->i] > ctx->bandE[ctx->m->nbEBands+ctx->i]) OPUS_COPY(Y, X, N);
1419
0
         else OPUS_COPY(X, Y, N);
1420
0
      }
1421
0
   }
1422
274k
   compute_theta(ctx, &sctx, X, Y, N, &b, B, B, LM, 1, &fill ARG_QEXT(&ext_b));
1423
274k
   inv = sctx.inv;
1424
274k
   imid = sctx.imid;
1425
274k
   iside = sctx.iside;
1426
274k
   delta = sctx.delta;
1427
274k
   itheta = sctx.itheta;
1428
274k
   qalloc = sctx.qalloc;
1429
#ifdef FIXED_POINT
1430
# ifdef ENABLE_QEXT
1431
   (void)imid;
1432
   (void)iside;
1433
   mid = celt_cos_norm32(sctx.itheta_q30);
1434
   side = celt_cos_norm32((1<<30)-sctx.itheta_q30);
1435
# else
1436
   mid = SHL32(EXTEND32(imid), 16);
1437
   side = SHL32(EXTEND32(iside), 16);
1438
# endif
1439
#else
1440
# ifdef ENABLE_QEXT
1441
   (void)imid;
1442
   (void)iside;
1443
   mid = celt_cos_norm2(sctx.itheta_q30*(1.f/(1<<30)));
1444
   side = celt_cos_norm2(1.f-sctx.itheta_q30*(1.f/(1<<30)));
1445
# else
1446
274k
   mid = (1.f/32768)*imid;
1447
274k
   side = (1.f/32768)*iside;
1448
274k
# endif
1449
274k
#endif
1450
1451
   /* This is a special case for N=2 that only works for stereo and takes
1452
      advantage of the fact that mid and side are orthogonal to encode
1453
      the side with just one bit. */
1454
274k
   if (N==2)
1455
74.0k
   {
1456
74.0k
      int c;
1457
74.0k
      int sign=0;
1458
74.0k
      celt_norm *x2, *y2;
1459
74.0k
      mbits = b;
1460
74.0k
      sbits = 0;
1461
      /* Only need one bit for the side. */
1462
74.0k
      if (itheta != 0 && itheta != 16384)
1463
14.6k
         sbits = 1<<BITRES;
1464
74.0k
      mbits -= sbits;
1465
74.0k
      c = itheta > 8192;
1466
74.0k
      ctx->remaining_bits -= qalloc+sbits;
1467
1468
74.0k
      x2 = c ? Y : X;
1469
74.0k
      y2 = c ? X : Y;
1470
74.0k
      if (sbits)
1471
14.6k
      {
1472
14.6k
         if (encode)
1473
0
         {
1474
            /* Here we only need to encode a sign for the side. */
1475
            /* FIXME: Need to increase fixed-point precision? */
1476
0
            sign = MULT32_32_Q31(x2[0],y2[1]) - MULT32_32_Q31(x2[1],y2[0]) < 0;
1477
0
            ec_enc_bits(ec, sign, 1);
1478
14.6k
         } else {
1479
14.6k
            sign = ec_dec_bits(ec, 1);
1480
14.6k
         }
1481
14.6k
      }
1482
74.0k
      sign = 1-2*sign;
1483
      /* We use orig_fill here because we want to fold the side, but if
1484
         itheta==16384, we'll have cleared the low bits of fill. */
1485
74.0k
      cm = quant_band(ctx, x2, N, mbits, B, lowband, LM, lowband_out, Q31ONE,
1486
74.0k
            lowband_scratch, orig_fill ARG_QEXT(ext_b));
1487
      /* We don't split N=2 bands, so cm is either 1 or 0 (for a fold-collapse),
1488
         and there's no need to worry about mixing with the other channel. */
1489
74.0k
      y2[0] = -sign*x2[1];
1490
74.0k
      y2[1] = sign*x2[0];
1491
74.0k
      if (ctx->resynth)
1492
74.0k
      {
1493
74.0k
         celt_norm tmp;
1494
74.0k
         X[0] = MULT32_32_Q31(mid, X[0]);
1495
74.0k
         X[1] = MULT32_32_Q31(mid, X[1]);
1496
74.0k
         Y[0] = MULT32_32_Q31(side, Y[0]);
1497
74.0k
         Y[1] = MULT32_32_Q31(side, Y[1]);
1498
74.0k
         tmp = X[0];
1499
74.0k
         X[0] = SUB32(tmp,Y[0]);
1500
74.0k
         Y[0] = ADD32(tmp,Y[0]);
1501
74.0k
         tmp = X[1];
1502
74.0k
         X[1] = SUB32(tmp,Y[1]);
1503
74.0k
         Y[1] = ADD32(tmp,Y[1]);
1504
74.0k
      }
1505
200k
   } else {
1506
      /* "Normal" split code */
1507
200k
      opus_int32 rebalance;
1508
1509
200k
      mbits = IMAX(0, IMIN(b, (b-delta)/2));
1510
200k
      sbits = b-mbits;
1511
200k
      ctx->remaining_bits -= qalloc;
1512
1513
200k
      rebalance = ctx->remaining_bits;
1514
200k
      if (mbits >= sbits)
1515
194k
      {
1516
#ifdef ENABLE_QEXT
1517
         int qext_extra = 0;
1518
         /* Reallocate any mid bits that cannot be used to extra mid bits. */
1519
         if (cap != NULL && ext_b != 0) qext_extra = IMAX(0, IMIN(ext_b/2, mbits - cap[ctx->i]/2));
1520
#endif
1521
         /* In stereo mode, we do not apply a scaling to the mid because we need the normalized
1522
            mid for folding later. */
1523
194k
         cm = quant_band(ctx, X, N, mbits, B, lowband, LM, lowband_out, Q31ONE,
1524
194k
               lowband_scratch, fill ARG_QEXT(ext_b/2+qext_extra));
1525
194k
         rebalance = mbits - (rebalance-ctx->remaining_bits);
1526
194k
         if (rebalance > 3<<BITRES && itheta!=0)
1527
1.23k
            sbits += rebalance - (3<<BITRES);
1528
#ifdef ENABLE_QEXT
1529
         /* Guard against overflowing the EC with the angle if the cubic quant used too many bits for the mid. */
1530
         if (ctx->extra_bands) sbits = IMIN(sbits, ctx->remaining_bits);
1531
#endif
1532
         /* For a stereo split, the high bits of fill are always zero, so no
1533
            folding will be done to the side. */
1534
194k
         cm |= quant_band(ctx, Y, N, sbits, B, NULL, LM, NULL, side, NULL, fill>>B ARG_QEXT(ext_b/2-qext_extra));
1535
194k
      } else {
1536
#ifdef ENABLE_QEXT
1537
         int qext_extra = 0;
1538
         /* Reallocate any side bits that cannot be used to extra side bits. */
1539
         if (cap != NULL && ext_b != 0) qext_extra = IMAX(0, IMIN(ext_b/2, sbits - cap[ctx->i]/2));
1540
#endif
1541
         /* For a stereo split, the high bits of fill are always zero, so no
1542
            folding will be done to the side. */
1543
6.24k
         cm = quant_band(ctx, Y, N, sbits, B, NULL, LM, NULL, side, NULL, fill>>B ARG_QEXT(ext_b/2+qext_extra));
1544
6.24k
         rebalance = sbits - (rebalance-ctx->remaining_bits);
1545
6.24k
         if (rebalance > 3<<BITRES && itheta!=16384)
1546
544
            mbits += rebalance - (3<<BITRES);
1547
#ifdef ENABLE_QEXT
1548
         /* Guard against overflowing the EC with the angle if the cubic quant used too many bits for the side. */
1549
         if (ctx->extra_bands) mbits = IMIN(mbits, ctx->remaining_bits);
1550
#endif
1551
         /* In stereo mode, we do not apply a scaling to the mid because we need the normalized
1552
            mid for folding later. */
1553
6.24k
         cm |= quant_band(ctx, X, N, mbits, B, lowband, LM, lowband_out, Q31ONE,
1554
6.24k
               lowband_scratch, fill ARG_QEXT(ext_b/2-qext_extra));
1555
6.24k
      }
1556
200k
   }
1557
1558
1559
   /* This code is used by the decoder and by the resynthesis-enabled encoder */
1560
274k
   if (ctx->resynth)
1561
274k
   {
1562
274k
      if (N!=2)
1563
200k
         stereo_merge(X, Y, mid, N, ctx->arch);
1564
274k
      if (inv)
1565
16.9k
      {
1566
16.9k
         int j;
1567
206k
         for (j=0;j<N;j++)
1568
189k
            Y[j] = -Y[j];
1569
16.9k
      }
1570
274k
   }
1571
274k
   return cm;
1572
337k
}
1573
1574
#ifndef DISABLE_UPDATE_DRAFT
1575
static void special_hybrid_folding(const CELTMode *m, celt_norm *norm, celt_norm *norm2, int start, int M, int dual_stereo)
1576
93.2k
{
1577
93.2k
   int n1, n2;
1578
93.2k
   const opus_int16 * OPUS_RESTRICT eBands = m->eBands;
1579
93.2k
   n1 = M*(eBands[start+1]-eBands[start]);
1580
93.2k
   n2 = M*(eBands[start+2]-eBands[start+1]);
1581
   /* Duplicate enough of the first band folding data to be able to fold the second band.
1582
      Copies no data for CELT-only mode. */
1583
93.2k
   OPUS_COPY(&norm[n1], &norm[2*n1 - n2], n2-n1);
1584
93.2k
   if (dual_stereo)
1585
5.17k
      OPUS_COPY(&norm2[n1], &norm2[2*n1 - n2], n2-n1);
1586
93.2k
}
1587
#endif
1588
1589
void quant_all_bands(int encode, const CELTMode *m, int start, int end,
1590
      celt_norm *X_, celt_norm *Y_, unsigned char *collapse_masks,
1591
      const celt_ener *bandE, int *pulses, int shortBlocks, int spread,
1592
      int dual_stereo, int intensity, int *tf_res, opus_int32 total_bits,
1593
      opus_int32 balance, ec_ctx *ec, int LM, int codedBands,
1594
      opus_uint32 *seed, int complexity, int arch, int disable_inv
1595
      ARG_QEXT(ec_ctx *ext_ec) ARG_QEXT(int *extra_pulses)
1596
      ARG_QEXT(opus_int32 ext_total_bits) ARG_QEXT(const int *cap))
1597
93.2k
{
1598
93.2k
   int i;
1599
93.2k
   opus_int32 remaining_bits;
1600
93.2k
   const opus_int16 * OPUS_RESTRICT eBands = m->eBands;
1601
93.2k
   celt_norm * OPUS_RESTRICT norm, * OPUS_RESTRICT norm2;
1602
93.2k
   VARDECL(celt_norm, _norm);
1603
93.2k
   VARDECL(celt_norm, _lowband_scratch);
1604
93.2k
   VARDECL(celt_norm, X_save);
1605
93.2k
   VARDECL(celt_norm, Y_save);
1606
93.2k
   VARDECL(celt_norm, X_save2);
1607
93.2k
   VARDECL(celt_norm, Y_save2);
1608
93.2k
   VARDECL(celt_norm, norm_save2);
1609
93.2k
   VARDECL(unsigned char, bytes_save);
1610
93.2k
   int resynth_alloc;
1611
93.2k
   celt_norm *lowband_scratch;
1612
93.2k
   int B;
1613
93.2k
   int M;
1614
93.2k
   int lowband_offset;
1615
93.2k
   int update_lowband = 1;
1616
93.2k
   int C = Y_ != NULL ? 2 : 1;
1617
93.2k
   int norm_offset;
1618
93.2k
   int theta_rdo = encode && Y_!=NULL && !dual_stereo && complexity>=8;
1619
#ifdef RESYNTH
1620
   int resynth = 1;
1621
#else
1622
93.2k
   int resynth = !encode || theta_rdo;
1623
93.2k
#endif
1624
93.2k
   struct band_ctx ctx;
1625
#ifdef ENABLE_QEXT
1626
   int ext_b;
1627
   opus_int32 ext_balance=0;
1628
   opus_int32 ext_tell=0;
1629
   VARDECL(unsigned char, ext_bytes_save);
1630
#endif
1631
93.2k
   SAVE_STACK;
1632
1633
93.2k
   M = 1<<LM;
1634
93.2k
   B = shortBlocks ? M : 1;
1635
93.2k
   norm_offset = M*eBands[start];
1636
   /* No need to allocate norm for the last band because we don't need an
1637
      output in that band. */
1638
93.2k
   ALLOC(_norm, C*(M*eBands[m->nbEBands-1]-norm_offset), celt_norm);
1639
93.2k
   norm = _norm;
1640
93.2k
   norm2 = norm + M*eBands[m->nbEBands-1]-norm_offset;
1641
1642
   /* For decoding, we can use the last band as scratch space because we don't need that
1643
      scratch space for the last band and we don't care about the data there until we're
1644
      decoding the last band. */
1645
93.2k
   if (encode && resynth)
1646
0
      resynth_alloc = M*(eBands[m->nbEBands]-eBands[m->nbEBands-1]);
1647
93.2k
   else
1648
93.2k
      resynth_alloc = ALLOC_NONE;
1649
93.2k
   ALLOC(_lowband_scratch, resynth_alloc, celt_norm);
1650
93.2k
   if (encode && resynth)
1651
0
      lowband_scratch = _lowband_scratch;
1652
93.2k
   else
1653
93.2k
      lowband_scratch = X_+M*eBands[m->effEBands-1];
1654
93.2k
   ALLOC(X_save, resynth_alloc, celt_norm);
1655
93.2k
   ALLOC(Y_save, resynth_alloc, celt_norm);
1656
93.2k
   ALLOC(X_save2, resynth_alloc, celt_norm);
1657
93.2k
   ALLOC(Y_save2, resynth_alloc, celt_norm);
1658
93.2k
   ALLOC(norm_save2, resynth_alloc, celt_norm);
1659
1660
93.2k
   lowband_offset = 0;
1661
93.2k
   ctx.bandE = bandE;
1662
93.2k
   ctx.ec = ec;
1663
93.2k
   ctx.encode = encode;
1664
93.2k
   ctx.intensity = intensity;
1665
93.2k
   ctx.m = m;
1666
93.2k
   ctx.seed = *seed;
1667
93.2k
   ctx.spread = spread;
1668
93.2k
   ctx.arch = arch;
1669
93.2k
   ctx.disable_inv = disable_inv;
1670
93.2k
   ctx.resynth = resynth;
1671
93.2k
   ctx.theta_round = 0;
1672
#ifdef ENABLE_QEXT
1673
   ctx.ext_ec = ext_ec;
1674
   ctx.ext_total_bits = ext_total_bits;
1675
   ctx.extra_bands = end == NB_QEXT_BANDS || end == 2;
1676
   if (ctx.extra_bands) theta_rdo = 0;
1677
   ALLOC(ext_bytes_save, theta_rdo ? QEXT_PACKET_SIZE_CAP : ALLOC_NONE, unsigned char);
1678
#endif
1679
93.2k
   ALLOC(bytes_save, theta_rdo ? 1275 : ALLOC_NONE, unsigned char);
1680
1681
   /* Avoid injecting noise in the first band on transients. */
1682
93.2k
   ctx.avoid_split_noise = B > 1;
1683
1.21M
   for (i=start;i<end;i++)
1684
1.12M
   {
1685
1.12M
      opus_int32 tell;
1686
1.12M
      int b;
1687
1.12M
      int N;
1688
1.12M
      opus_int32 curr_balance;
1689
1.12M
      int effective_lowband=-1;
1690
1.12M
      celt_norm * OPUS_RESTRICT X, * OPUS_RESTRICT Y;
1691
1.12M
      int tf_change=0;
1692
1.12M
      unsigned x_cm;
1693
1.12M
      unsigned y_cm;
1694
1.12M
      int last;
1695
1696
1.12M
      ctx.i = i;
1697
1.12M
      last = (i==end-1);
1698
1699
1.12M
      X = X_+M*eBands[i];
1700
1.12M
      if (Y_!=NULL)
1701
376k
         Y = Y_+M*eBands[i];
1702
749k
      else
1703
749k
         Y = NULL;
1704
1.12M
      N = M*eBands[i+1]-M*eBands[i];
1705
1.12M
      celt_assert(N > 0);
1706
1.12M
      tell = ec_tell_frac(ec);
1707
1708
      /* Compute how many bits we want to allocate to this band */
1709
1.12M
      if (i != start)
1710
1.03M
         balance -= tell;
1711
1.12M
      remaining_bits = total_bits-tell-1;
1712
1.12M
      ctx.remaining_bits = remaining_bits;
1713
#ifdef ENABLE_QEXT
1714
      if (i != start) {
1715
         ext_balance += extra_pulses[i-1] + ext_tell;
1716
      }
1717
      ext_tell = ec_tell_frac(ext_ec);
1718
      ctx.extra_bits = extra_pulses[i];
1719
      if (i != start)
1720
         ext_balance -= ext_tell;
1721
      if (i <= codedBands-1)
1722
      {
1723
         opus_int32 ext_curr_balance = celt_sudiv(ext_balance, IMIN(3, codedBands-i));
1724
         ext_b = IMAX(0, IMIN(16383, IMIN(ext_total_bits-ext_tell,extra_pulses[i]+ext_curr_balance)));
1725
      } else {
1726
         ext_b = 0;
1727
      }
1728
#endif
1729
1.12M
      if (i <= codedBands-1)
1730
512k
      {
1731
512k
         curr_balance = celt_sudiv(balance, IMIN(3, codedBands-i));
1732
512k
         b = IMAX(0, IMIN(16383, IMIN(remaining_bits+1,pulses[i]+curr_balance)));
1733
613k
      } else {
1734
613k
         b = 0;
1735
613k
      }
1736
1737
1.12M
#ifndef DISABLE_UPDATE_DRAFT
1738
1.12M
      if (resynth && (M*eBands[i]-N >= M*eBands[start] || i==start+1) && (update_lowband || lowband_offset==0))
1739
396k
            lowband_offset = i;
1740
1.12M
      if (i == start+1)
1741
93.2k
         special_hybrid_folding(m, norm, norm2, start, M, dual_stereo);
1742
#else
1743
      if (resynth && M*eBands[i]-N >= M*eBands[start] && (update_lowband || lowband_offset==0))
1744
            lowband_offset = i;
1745
#endif
1746
1747
1.12M
      tf_change = tf_res[i];
1748
1.12M
      ctx.tf_change = tf_change;
1749
1.12M
      if (i>=m->effEBands)
1750
0
      {
1751
0
         X=norm;
1752
0
         if (Y_!=NULL)
1753
0
            Y = norm;
1754
0
         lowband_scratch = NULL;
1755
0
      }
1756
1.12M
      if (last && !theta_rdo)
1757
93.2k
         lowband_scratch = NULL;
1758
1759
      /* Get a conservative estimate of the collapse_mask's for the bands we're
1760
         going to be folding from. */
1761
1.12M
      if (lowband_offset != 0 && (spread!=SPREAD_AGGRESSIVE || B>1 || tf_change<0))
1762
1.00M
      {
1763
1.00M
         int fold_start;
1764
1.00M
         int fold_end;
1765
1.00M
         int fold_i;
1766
         /* This ensures we never repeat spectral content within one band */
1767
1.00M
         effective_lowband = IMAX(0, M*eBands[lowband_offset]-norm_offset-N);
1768
1.00M
         fold_start = lowband_offset;
1769
1.26M
         while(M*eBands[--fold_start] > effective_lowband+norm_offset);
1770
1.00M
         fold_end = lowband_offset-1;
1771
1.00M
#ifndef DISABLE_UPDATE_DRAFT
1772
1.81M
         while(++fold_end < i && M*eBands[fold_end] < effective_lowband+norm_offset+N);
1773
#else
1774
         while(M*eBands[++fold_end] < effective_lowband+norm_offset+N);
1775
#endif
1776
1.00M
         x_cm = y_cm = 0;
1777
2.07M
         fold_i = fold_start; do {
1778
2.07M
           x_cm |= collapse_masks[fold_i*C+0];
1779
2.07M
           y_cm |= collapse_masks[fold_i*C+C-1];
1780
2.07M
         } while (++fold_i<fold_end);
1781
1.00M
      }
1782
      /* Otherwise, we'll be using the LCG to fold, so all blocks will (almost
1783
         always) be non-zero. */
1784
122k
      else
1785
122k
         x_cm = y_cm = (1<<B)-1;
1786
1787
1.12M
      if (dual_stereo && i==intensity)
1788
4.80k
      {
1789
4.80k
         int j;
1790
1791
         /* Switch off dual stereo to do intensity. */
1792
4.80k
         dual_stereo = 0;
1793
4.80k
         if (resynth)
1794
132k
            for (j=0;j<M*eBands[i]-norm_offset;j++)
1795
127k
               norm[j] = HALF32(norm[j]+norm2[j]);
1796
4.80k
      }
1797
1.12M
      if (dual_stereo)
1798
38.7k
      {
1799
38.7k
         x_cm = quant_band(&ctx, X, N, b/2, B,
1800
38.7k
               effective_lowband != -1 ? norm+effective_lowband : NULL, LM,
1801
38.7k
               last?NULL:norm+M*eBands[i]-norm_offset, Q31ONE, lowband_scratch, x_cm ARG_QEXT(ext_b/2));
1802
38.7k
         y_cm = quant_band(&ctx, Y, N, b/2, B,
1803
38.7k
               effective_lowband != -1 ? norm2+effective_lowband : NULL, LM,
1804
38.7k
               last?NULL:norm2+M*eBands[i]-norm_offset, Q31ONE, lowband_scratch, y_cm ARG_QEXT(ext_b/2));
1805
1.08M
      } else {
1806
1.08M
         if (Y!=NULL)
1807
337k
         {
1808
337k
            if (theta_rdo && i < intensity)
1809
0
            {
1810
0
               ec_ctx ec_save, ec_save2;
1811
0
               struct band_ctx ctx_save, ctx_save2;
1812
0
               opus_val32 dist0, dist1;
1813
0
               unsigned cm, cm2;
1814
0
               int nstart_bytes, nend_bytes, save_bytes;
1815
0
               unsigned char *bytes_buf;
1816
#ifdef ENABLE_QEXT
1817
               ec_ctx ext_ec_save, ext_ec_save2;
1818
               unsigned char *ext_bytes_buf;
1819
               int ext_nstart_bytes, ext_nend_bytes, ext_save_bytes;
1820
#endif
1821
0
               opus_val16 w[2];
1822
0
               compute_channel_weights(bandE[i], bandE[i+m->nbEBands], w);
1823
               /* Make a copy. */
1824
0
               cm = x_cm|y_cm;
1825
0
               ec_save = *ec;
1826
#ifdef ENABLE_QEXT
1827
               ext_ec_save = *ext_ec;
1828
#endif
1829
0
               ctx_save = ctx;
1830
0
               OPUS_COPY(X_save, X, N);
1831
0
               OPUS_COPY(Y_save, Y, N);
1832
               /* Encode and round down. */
1833
0
               ctx.theta_round = -1;
1834
0
               x_cm = quant_band_stereo(&ctx, X, Y, N, b, B,
1835
0
                     effective_lowband != -1 ? norm+effective_lowband : NULL, LM,
1836
0
                     last?NULL:norm+M*eBands[i]-norm_offset, lowband_scratch, cm ARG_QEXT(ext_b) ARG_QEXT(cap));
1837
0
               dist0 = MULT16_32_Q15(w[0], celt_inner_prod_norm_shift(X_save, X, N, arch)) + MULT16_32_Q15(w[1], celt_inner_prod_norm_shift(Y_save, Y, N, arch));
1838
1839
               /* Save first result. */
1840
0
               cm2 = x_cm;
1841
0
               ec_save2 = *ec;
1842
#ifdef ENABLE_QEXT
1843
               ext_ec_save2 = *ext_ec;
1844
#endif
1845
0
               ctx_save2 = ctx;
1846
0
               OPUS_COPY(X_save2, X, N);
1847
0
               OPUS_COPY(Y_save2, Y, N);
1848
0
               if (!last)
1849
0
                  OPUS_COPY(norm_save2, norm+M*eBands[i]-norm_offset, N);
1850
0
               nstart_bytes = ec_save.offs;
1851
0
               nend_bytes = ec_save.storage;
1852
0
               bytes_buf = ec_save.buf+nstart_bytes;
1853
0
               save_bytes = nend_bytes-nstart_bytes;
1854
0
               OPUS_COPY(bytes_save, bytes_buf, save_bytes);
1855
#ifdef ENABLE_QEXT
1856
               ext_nstart_bytes = ext_ec_save.offs;
1857
               ext_nend_bytes = ext_ec_save.storage;
1858
               ext_bytes_buf = ext_ec_save.buf!=NULL ? ext_ec_save.buf+ext_nstart_bytes : NULL;
1859
               ext_save_bytes = ext_nend_bytes-ext_nstart_bytes;
1860
               if (ext_save_bytes) OPUS_COPY(ext_bytes_save, ext_bytes_buf, ext_save_bytes);
1861
#endif
1862
               /* Restore */
1863
0
               *ec = ec_save;
1864
#ifdef ENABLE_QEXT
1865
               *ext_ec = ext_ec_save;
1866
#endif
1867
0
               ctx = ctx_save;
1868
0
               OPUS_COPY(X, X_save, N);
1869
0
               OPUS_COPY(Y, Y_save, N);
1870
0
#ifndef DISABLE_UPDATE_DRAFT
1871
0
               if (i == start+1)
1872
0
                  special_hybrid_folding(m, norm, norm2, start, M, dual_stereo);
1873
0
#endif
1874
               /* Encode and round up. */
1875
0
               ctx.theta_round = 1;
1876
0
               x_cm = quant_band_stereo(&ctx, X, Y, N, b, B,
1877
0
                     effective_lowband != -1 ? norm+effective_lowband : NULL, LM,
1878
0
                     last?NULL:norm+M*eBands[i]-norm_offset, lowband_scratch, cm ARG_QEXT(ext_b) ARG_QEXT(cap));
1879
0
               dist1 = MULT16_32_Q15(w[0], celt_inner_prod_norm_shift(X_save, X, N, arch)) + MULT16_32_Q15(w[1], celt_inner_prod_norm_shift(Y_save, Y, N, arch));
1880
0
               if (dist0 >= dist1) {
1881
0
                  x_cm = cm2;
1882
0
                  *ec = ec_save2;
1883
#ifdef ENABLE_QEXT
1884
                  *ext_ec = ext_ec_save2;
1885
#endif
1886
0
                  ctx = ctx_save2;
1887
0
                  OPUS_COPY(X, X_save2, N);
1888
0
                  OPUS_COPY(Y, Y_save2, N);
1889
0
                  if (!last)
1890
0
                     OPUS_COPY(norm+M*eBands[i]-norm_offset, norm_save2, N);
1891
0
                  OPUS_COPY(bytes_buf, bytes_save, save_bytes);
1892
#ifdef ENABLE_QEXT
1893
                  if (ext_save_bytes) OPUS_COPY(ext_bytes_buf, ext_bytes_save, ext_save_bytes);
1894
#endif
1895
0
               }
1896
337k
            } else {
1897
337k
               ctx.theta_round = 0;
1898
337k
               x_cm = quant_band_stereo(&ctx, X, Y, N, b, B,
1899
337k
                     effective_lowband != -1 ? norm+effective_lowband : NULL, LM,
1900
337k
                     last?NULL:norm+M*eBands[i]-norm_offset, lowband_scratch, x_cm|y_cm ARG_QEXT(ext_b) ARG_QEXT(cap));
1901
337k
            }
1902
749k
         } else {
1903
749k
            x_cm = quant_band(&ctx, X, N, b, B,
1904
749k
                  effective_lowband != -1 ? norm+effective_lowband : NULL, LM,
1905
749k
                  last?NULL:norm+M*eBands[i]-norm_offset, Q31ONE, lowband_scratch, x_cm|y_cm ARG_QEXT(ext_b));
1906
749k
         }
1907
1.08M
         y_cm = x_cm;
1908
1.08M
      }
1909
1.12M
      collapse_masks[i*C+0] = (unsigned char)x_cm;
1910
1.12M
      collapse_masks[i*C+C-1] = (unsigned char)y_cm;
1911
1.12M
      balance += pulses[i] + tell;
1912
1913
      /* Update the folding position only as long as we have 1 bit/sample depth. */
1914
1.12M
      update_lowband = b>(N<<BITRES);
1915
      /* We only need to avoid noise on a split for the first band. After that, we
1916
         have folding. */
1917
1.12M
      ctx.avoid_split_noise = 0;
1918
1.12M
   }
1919
93.2k
   *seed = ctx.seed;
1920
1921
93.2k
   RESTORE_STACK;
1922
93.2k
}