Coverage Report

Created: 2026-08-13 07:04

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
/src/opus/celt/mdct.c
Line
Count
Source
1
/* Copyright (c) 2007-2008 CSIRO
2
   Copyright (c) 2007-2008 Xiph.Org Foundation
3
   Written by Jean-Marc Valin */
4
/*
5
   Redistribution and use in source and binary forms, with or without
6
   modification, are permitted provided that the following conditions
7
   are met:
8
9
   - Redistributions of source code must retain the above copyright
10
   notice, this list of conditions and the following disclaimer.
11
12
   - Redistributions in binary form must reproduce the above copyright
13
   notice, this list of conditions and the following disclaimer in the
14
   documentation and/or other materials provided with the distribution.
15
16
   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
17
   ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
18
   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
19
   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
20
   OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
21
   EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
22
   PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
23
   PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
24
   LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
25
   NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
26
   SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
27
*/
28
29
/* This is a simple MDCT implementation that uses a N/4 complex FFT
30
   to do most of the work. It should be relatively straightforward to
31
   plug in pretty much and FFT here.
32
33
   This replaces the Vorbis FFT (and uses the exact same API), which
34
   was a bit too messy and that was ending up duplicating code
35
   (might as well use the same FFT everywhere).
36
37
   The algorithm is similar to (and inspired from) Fabrice Bellard's
38
   MDCT implementation in FFMPEG, but has differences in signs, ordering
39
   and scaling in many places.
40
*/
41
42
#ifndef SKIP_CONFIG_H
43
#ifdef HAVE_CONFIG_H
44
#include "config.h"
45
#endif
46
#endif
47
48
#include "mdct.h"
49
#include "kiss_fft.h"
50
#include "_kiss_fft_guts.h"
51
#include <math.h>
52
#include "os_support.h"
53
#include "mathops.h"
54
#include "stack_alloc.h"
55
56
#if defined(FIXED_POINT) && defined(__mips) && __mips == 32
57
#include "mips/mdct_mipsr1.h"
58
#endif
59
60
#ifndef M_PI
61
#define M_PI 3.141592653
62
#endif
63
64
#ifdef CUSTOM_MODES
65
66
int clt_mdct_init(mdct_lookup *l,int N, int maxshift, int arch)
67
{
68
   int i;
69
   kiss_twiddle_scalar *trig;
70
   int shift;
71
   int N2=N>>1;
72
   l->n = N;
73
   l->maxshift = maxshift;
74
   for (i=0;i<=maxshift;i++)
75
   {
76
      if (i==0)
77
         l->kfft[i] = opus_fft_alloc(N>>2>>i, 0, 0, arch);
78
      else
79
         l->kfft[i] = opus_fft_alloc_twiddles(N>>2>>i, 0, 0, l->kfft[0], arch);
80
#ifndef ENABLE_TI_DSPLIB55
81
      if (l->kfft[i]==NULL)
82
         return 0;
83
#endif
84
   }
85
   l->trig = trig = (kiss_twiddle_scalar*)opus_alloc((N-(N2>>maxshift))*sizeof(kiss_twiddle_scalar));
86
   if (l->trig==NULL)
87
     return 0;
88
   for (shift=0;shift<=maxshift;shift++)
89
   {
90
      /* We have enough points that sine isn't necessary */
91
#if defined(FIXED_POINT)
92
#ifndef ENABLE_QEXT
93
      for (i=0;i<N2;i++)
94
         trig[i] = TRIG_UPSCALE*celt_cos_norm(DIV32(ADD32(SHL32(EXTEND32(i),17),N2+16384),N));
95
#else
96
      for (i=0;i<N2;i++)
97
         trig[i] = (kiss_twiddle_scalar)MAX32(-2147483647,MIN32(2147483647,floor(.5+2147483648*cos(2*M_PI*(i+.125)/N))));
98
#endif
99
#else
100
      for (i=0;i<N2;i++)
101
         trig[i] = (kiss_twiddle_scalar)cos(2*PI*(i+.125)/N);
102
#endif
103
      trig += N2;
104
      N2 >>= 1;
105
      N >>= 1;
106
   }
107
   return 1;
108
}
109
110
void clt_mdct_clear(mdct_lookup *l, int arch)
111
{
112
   int i;
113
   for (i=0;i<=l->maxshift;i++)
114
      opus_fft_free(l->kfft[i], arch);
115
   opus_free((kiss_twiddle_scalar*)l->trig);
116
}
117
118
#endif /* CUSTOM_MODES */
119
120
/* Forward MDCT trashes the input array */
121
#ifndef OVERRIDE_clt_mdct_forward
122
void clt_mdct_forward_c(const mdct_lookup *l, kiss_fft_scalar *in, kiss_fft_scalar * OPUS_RESTRICT out,
123
      const celt_coef *window, int overlap, int shift, int stride, int arch)
124
70.7M
{
125
70.7M
   int i;
126
70.7M
   int N, N2, N4;
127
70.7M
   VARDECL(kiss_fft_scalar, f);
128
70.7M
   VARDECL(kiss_fft_cpx, f2);
129
70.7M
   const kiss_fft_state *st = l->kfft[shift];
130
70.7M
   const kiss_twiddle_scalar *trig;
131
70.7M
   celt_coef scale;
132
#ifdef FIXED_POINT
133
   /* Allows us to scale with MULT16_32_Q16(), which is faster than
134
      MULT16_32_Q15() on ARM. */
135
   int scale_shift = st->scale_shift-1;
136
   int headroom;
137
#endif
138
70.7M
   SAVE_STACK;
139
70.7M
   (void)arch;
140
70.7M
   scale = st->scale;
141
142
70.7M
   N = l->n;
143
70.7M
   trig = l->trig;
144
221M
   for (i=0;i<shift;i++)
145
150M
   {
146
150M
      N >>= 1;
147
150M
      trig += N;
148
150M
   }
149
70.7M
   N2 = N>>1;
150
70.7M
   N4 = N>>2;
151
152
70.7M
   ALLOC(f, N2, kiss_fft_scalar);
153
70.7M
   ALLOC(f2, N4, kiss_fft_cpx);
154
155
   /* Consider the input to be composed of four blocks: [a, b, c, d] */
156
   /* Window, shuffle, fold */
157
70.7M
   {
158
      /* Temp pointers to make it really clear to the compiler what we're doing */
159
70.7M
      const kiss_fft_scalar * OPUS_RESTRICT xp1 = in+(overlap>>1);
160
70.7M
      const kiss_fft_scalar * OPUS_RESTRICT xp2 = in+N2-1+(overlap>>1);
161
70.7M
      kiss_fft_scalar * OPUS_RESTRICT yp = f;
162
70.7M
      const celt_coef * OPUS_RESTRICT wp1 = window+(overlap>>1);
163
70.7M
      const celt_coef * OPUS_RESTRICT wp2 = window+(overlap>>1)-1;
164
2.19G
      for(i=0;i<((overlap+3)>>2);i++)
165
2.12G
      {
166
         /* Real part arranged as -d-cR, Imag part arranged as -b+aR*/
167
2.12G
         *yp++ = S_MUL(xp1[N2], *wp2) + S_MUL(*xp2, *wp1);
168
2.12G
         *yp++ = S_MUL(*xp1, *wp1)    - S_MUL(xp2[-N2], *wp2);
169
2.12G
         xp1+=2;
170
2.12G
         xp2-=2;
171
2.12G
         wp1+=2;
172
2.12G
         wp2-=2;
173
2.12G
      }
174
70.7M
      wp1 = window;
175
70.7M
      wp2 = window+overlap-1;
176
6.89G
      for(;i<N4-((overlap+3)>>2);i++)
177
6.82G
      {
178
         /* Real part arranged as a-bR, Imag part arranged as -c-dR */
179
6.82G
         *yp++ = *xp2;
180
6.82G
         *yp++ = *xp1;
181
6.82G
         xp1+=2;
182
6.82G
         xp2-=2;
183
6.82G
      }
184
2.19G
      for(;i<N4;i++)
185
2.12G
      {
186
         /* Real part arranged as a-bR, Imag part arranged as -c-dR */
187
2.12G
         *yp++ =  -S_MUL(xp1[-N2], *wp1) + S_MUL(*xp2, *wp2);
188
2.12G
         *yp++ = S_MUL(*xp1, *wp2)     + S_MUL(xp2[N2], *wp1);
189
2.12G
         xp1+=2;
190
2.12G
         xp2-=2;
191
2.12G
         wp1+=2;
192
2.12G
         wp2-=2;
193
2.12G
      }
194
70.7M
   }
195
   /* Pre-rotation */
196
70.7M
   {
197
70.7M
      kiss_fft_scalar * OPUS_RESTRICT yp = f;
198
70.7M
      const kiss_twiddle_scalar *t = &trig[0];
199
#ifdef FIXED_POINT
200
      opus_val32 maxval=1;
201
#endif
202
11.1G
      for(i=0;i<N4;i++)
203
11.0G
      {
204
11.0G
         kiss_fft_cpx yc;
205
11.0G
         kiss_twiddle_scalar t0, t1;
206
11.0G
         kiss_fft_scalar re, im, yr, yi;
207
11.0G
         t0 = t[i];
208
11.0G
         t1 = t[N4+i];
209
11.0G
         re = *yp++;
210
11.0G
         im = *yp++;
211
11.0G
         yr = S_MUL(re,t0)  -  S_MUL(im,t1);
212
11.0G
         yi = S_MUL(im,t0)  +  S_MUL(re,t1);
213
         /* For QEXT, it's best to scale before the FFT, but otherwise it's best to scale after.
214
            For floating-point it doesn't matter. */
215
#ifdef ENABLE_QEXT
216
         yc.r = yr;
217
         yc.i = yi;
218
#else
219
11.0G
         yc.r = S_MUL2(yr, scale);
220
11.0G
         yc.i = S_MUL2(yi, scale);
221
11.0G
#endif
222
#ifdef FIXED_POINT
223
         maxval = MAX32(maxval, MAX32(ABS32(yc.r), ABS32(yc.i)));
224
#endif
225
11.0G
         f2[st->bitrev[i]] = yc;
226
11.0G
      }
227
#ifdef FIXED_POINT
228
      headroom = IMAX(0, IMIN(scale_shift, 28-celt_ilog2(maxval)));
229
#endif
230
70.7M
   }
231
232
   /* N/4 complex FFT, does not downscale anymore */
233
70.7M
   opus_fft_impl(st, f2 ARG_FIXED(scale_shift-headroom));
234
235
   /* Post-rotate */
236
70.7M
   {
237
      /* Temp pointers to make it really clear to the compiler what we're doing */
238
70.7M
      const kiss_fft_cpx * OPUS_RESTRICT fp = f2;
239
70.7M
      kiss_fft_scalar * OPUS_RESTRICT yp1 = out;
240
70.7M
      kiss_fft_scalar * OPUS_RESTRICT yp2 = out+stride*(N2-1);
241
70.7M
      const kiss_twiddle_scalar *t = &trig[0];
242
      /* Temp pointers to make it really clear to the compiler what we're doing */
243
11.1G
      for(i=0;i<N4;i++)
244
11.0G
      {
245
11.0G
         kiss_fft_scalar yr, yi;
246
11.0G
         kiss_fft_scalar t0, t1;
247
#ifdef ENABLE_QEXT
248
         t0 = S_MUL2(t[i], scale);
249
         t1 = S_MUL2(t[N4+i], scale);
250
#else
251
11.0G
         t0 = t[i];
252
11.0G
         t1 = t[N4+i];
253
11.0G
#endif
254
11.0G
         yr = PSHR32(S_MUL(fp->i,t1) - S_MUL(fp->r,t0), headroom);
255
11.0G
         yi = PSHR32(S_MUL(fp->r,t1) + S_MUL(fp->i,t0), headroom);
256
11.0G
         *yp1 = yr;
257
11.0G
         *yp2 = yi;
258
11.0G
         fp++;
259
11.0G
         yp1 += 2*stride;
260
11.0G
         yp2 -= 2*stride;
261
11.0G
      }
262
70.7M
   }
263
70.7M
   RESTORE_STACK;
264
70.7M
}
265
#endif /* OVERRIDE_clt_mdct_forward */
266
267
#ifndef OVERRIDE_clt_mdct_backward
268
void clt_mdct_backward_c(const mdct_lookup *l, kiss_fft_scalar *in, kiss_fft_scalar * OPUS_RESTRICT out,
269
      const celt_coef * OPUS_RESTRICT window, int overlap, int shift, int stride, int arch)
270
0
{
271
0
   int i;
272
0
   int N, N2, N4;
273
0
   const kiss_twiddle_scalar *trig;
274
#ifdef FIXED_POINT
275
   int pre_shift, post_shift, fft_shift;
276
#endif
277
0
   (void) arch;
278
279
0
   N = l->n;
280
0
   trig = l->trig;
281
0
   for (i=0;i<shift;i++)
282
0
   {
283
0
      N >>= 1;
284
0
      trig += N;
285
0
   }
286
0
   N2 = N>>1;
287
0
   N4 = N>>2;
288
289
#ifdef FIXED_POINT
290
   {
291
      opus_val32 sumval=N2;
292
      opus_val32 maxval=0;
293
      for (i=0;i<N2;i++) {
294
         maxval = MAX32(maxval, ABS32(in[i*stride]));
295
         sumval = ADD32_ovflw(sumval, ABS32(SHR32(in[i*stride],11)));
296
      }
297
      pre_shift = IMAX(0, 29-celt_zlog2(1+maxval));
298
      /* Worst-case where all the energy goes to a single sample. */
299
      post_shift = IMAX(0, 19-celt_ilog2(ABS32(sumval)));
300
      post_shift = IMIN(post_shift, pre_shift);
301
      fft_shift = pre_shift - post_shift;
302
   }
303
#endif
304
   /* Pre-rotate */
305
0
   {
306
      /* Temp pointers to make it really clear to the compiler what we're doing */
307
0
      const kiss_fft_scalar * OPUS_RESTRICT xp1 = in;
308
0
      const kiss_fft_scalar * OPUS_RESTRICT xp2 = in+stride*(N2-1);
309
0
      kiss_fft_scalar * OPUS_RESTRICT yp = out+(overlap>>1);
310
0
      const kiss_twiddle_scalar * OPUS_RESTRICT t = &trig[0];
311
0
      const opus_int16 * OPUS_RESTRICT bitrev = l->kfft[shift]->bitrev;
312
0
      for(i=0;i<N4;i++)
313
0
      {
314
0
         int rev;
315
0
         kiss_fft_scalar yr, yi;
316
0
         opus_val32 x1, x2;
317
0
         rev = *bitrev++;
318
0
         x1 = SHL32_ovflw(*xp1, pre_shift);
319
0
         x2 = SHL32_ovflw(*xp2, pre_shift);
320
0
         yr = ADD32_ovflw(S_MUL(x2, t[i]), S_MUL(x1, t[N4+i]));
321
0
         yi = SUB32_ovflw(S_MUL(x1, t[i]), S_MUL(x2, t[N4+i]));
322
         /* We swap real and imag because we use an FFT instead of an IFFT. */
323
0
         yp[2*rev+1] = yr;
324
0
         yp[2*rev] = yi;
325
         /* Storing the pre-rotation directly in the bitrev order. */
326
0
         xp1+=2*stride;
327
0
         xp2-=2*stride;
328
0
      }
329
0
   }
330
331
0
   opus_fft_impl(l->kfft[shift], (kiss_fft_cpx*)(out+(overlap>>1)) ARG_FIXED(fft_shift));
332
333
   /* Post-rotate and de-shuffle from both ends of the buffer at once to make
334
      it in-place. */
335
0
   {
336
0
      kiss_fft_scalar * yp0 = out+(overlap>>1);
337
0
      kiss_fft_scalar * yp1 = out+(overlap>>1)+N2-2;
338
0
      const kiss_twiddle_scalar *t = &trig[0];
339
      /* Loop to (N4+1)>>1 to handle odd N4. When N4 is odd, the
340
         middle pair will be computed twice. */
341
0
      for(i=0;i<(N4+1)>>1;i++)
342
0
      {
343
0
         kiss_fft_scalar re, im, yr, yi;
344
0
         kiss_twiddle_scalar t0, t1;
345
         /* We swap real and imag because we're using an FFT instead of an IFFT. */
346
0
         re = yp0[1];
347
0
         im = yp0[0];
348
0
         t0 = t[i];
349
0
         t1 = t[N4+i];
350
         /* We'd scale up by 2 here, but instead it's done when mixing the windows */
351
0
         yr = PSHR32_ovflw(ADD32_ovflw(S_MUL(re,t0), S_MUL(im,t1)), post_shift);
352
0
         yi = PSHR32_ovflw(SUB32_ovflw(S_MUL(re,t1), S_MUL(im,t0)), post_shift);
353
         /* We swap real and imag because we're using an FFT instead of an IFFT. */
354
0
         re = yp1[1];
355
0
         im = yp1[0];
356
0
         yp0[0] = yr;
357
0
         yp1[1] = yi;
358
359
0
         t0 = t[(N4-i-1)];
360
0
         t1 = t[(N2-i-1)];
361
         /* We'd scale up by 2 here, but instead it's done when mixing the windows */
362
0
         yr = PSHR32_ovflw(ADD32_ovflw(S_MUL(re,t0), S_MUL(im,t1)), post_shift);
363
0
         yi = PSHR32_ovflw(SUB32_ovflw(S_MUL(re,t1), S_MUL(im,t0)), post_shift);
364
0
         yp1[0] = yr;
365
0
         yp0[1] = yi;
366
0
         yp0 += 2;
367
0
         yp1 -= 2;
368
0
      }
369
0
   }
370
371
   /* Mirror on both sides for TDAC */
372
0
   {
373
0
      kiss_fft_scalar * OPUS_RESTRICT xp1 = out+overlap-1;
374
0
      kiss_fft_scalar * OPUS_RESTRICT yp1 = out;
375
0
      const celt_coef * OPUS_RESTRICT wp1 = window;
376
0
      const celt_coef * OPUS_RESTRICT wp2 = window+overlap-1;
377
378
0
      for(i = 0; i < overlap/2; i++)
379
0
      {
380
0
         kiss_fft_scalar x1, x2;
381
0
         x1 = *xp1;
382
0
         x2 = *yp1;
383
0
         *yp1++ = SUB32_ovflw(S_MUL(x2, *wp2), S_MUL(x1, *wp1));
384
0
         *xp1-- = ADD32_ovflw(S_MUL(x2, *wp1), S_MUL(x1, *wp2));
385
0
         wp1++;
386
0
         wp2--;
387
0
      }
388
0
   }
389
0
}
390
#endif /* OVERRIDE_clt_mdct_backward */