Coverage Report

Created: 2025-07-11 06:54

/src/aac/libFDK/src/FDK_trigFcts.cpp
Line
Count
Source (jump to first uncovered line)
1
/* -----------------------------------------------------------------------------
2
Software License for The Fraunhofer FDK AAC Codec Library for Android
3
4
© Copyright  1995 - 2018 Fraunhofer-Gesellschaft zur Förderung der angewandten
5
Forschung e.V. All rights reserved.
6
7
 1.    INTRODUCTION
8
The Fraunhofer FDK AAC Codec Library for Android ("FDK AAC Codec") is software
9
that implements the MPEG Advanced Audio Coding ("AAC") encoding and decoding
10
scheme for digital audio. This FDK AAC Codec software is intended to be used on
11
a wide variety of Android devices.
12
13
AAC's HE-AAC and HE-AAC v2 versions are regarded as today's most efficient
14
general perceptual audio codecs. AAC-ELD is considered the best-performing
15
full-bandwidth communications codec by independent studies and is widely
16
deployed. AAC has been standardized by ISO and IEC as part of the MPEG
17
specifications.
18
19
Patent licenses for necessary patent claims for the FDK AAC Codec (including
20
those of Fraunhofer) may be obtained through Via Licensing
21
(www.vialicensing.com) or through the respective patent owners individually for
22
the purpose of encoding or decoding bit streams in products that are compliant
23
with the ISO/IEC MPEG audio standards. Please note that most manufacturers of
24
Android devices already license these patent claims through Via Licensing or
25
directly from the patent owners, and therefore FDK AAC Codec software may
26
already be covered under those patent licenses when it is used for those
27
licensed purposes only.
28
29
Commercially-licensed AAC software libraries, including floating-point versions
30
with enhanced sound quality, are also available from Fraunhofer. Users are
31
encouraged to check the Fraunhofer website for additional applications
32
information and documentation.
33
34
2.    COPYRIGHT LICENSE
35
36
Redistribution and use in source and binary forms, with or without modification,
37
are permitted without payment of copyright license fees provided that you
38
satisfy the following conditions:
39
40
You must retain the complete text of this software license in redistributions of
41
the FDK AAC Codec or your modifications thereto in source code form.
42
43
You must retain the complete text of this software license in the documentation
44
and/or other materials provided with redistributions of the FDK AAC Codec or
45
your modifications thereto in binary form. You must make available free of
46
charge copies of the complete source code of the FDK AAC Codec and your
47
modifications thereto to recipients of copies in binary form.
48
49
The name of Fraunhofer may not be used to endorse or promote products derived
50
from this library without prior written permission.
51
52
You may not charge copyright license fees for anyone to use, copy or distribute
53
the FDK AAC Codec software or your modifications thereto.
54
55
Your modified versions of the FDK AAC Codec must carry prominent notices stating
56
that you changed the software and the date of any change. For modified versions
57
of the FDK AAC Codec, the term "Fraunhofer FDK AAC Codec Library for Android"
58
must be replaced by the term "Third-Party Modified Version of the Fraunhofer FDK
59
AAC Codec Library for Android."
60
61
3.    NO PATENT LICENSE
62
63
NO EXPRESS OR IMPLIED LICENSES TO ANY PATENT CLAIMS, including without
64
limitation the patents of Fraunhofer, ARE GRANTED BY THIS SOFTWARE LICENSE.
65
Fraunhofer provides no warranty of patent non-infringement with respect to this
66
software.
67
68
You may use this FDK AAC Codec software or modifications thereto only for
69
purposes that are authorized by appropriate patent licenses.
70
71
4.    DISCLAIMER
72
73
This FDK AAC Codec software is provided by Fraunhofer on behalf of the copyright
74
holders and contributors "AS IS" and WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES,
75
including but not limited to the implied warranties of merchantability and
76
fitness for a particular purpose. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR
77
CONTRIBUTORS BE LIABLE for any direct, indirect, incidental, special, exemplary,
78
or consequential damages, including but not limited to procurement of substitute
79
goods or services; loss of use, data, or profits, or business interruption,
80
however caused and on any theory of liability, whether in contract, strict
81
liability, or tort (including negligence), arising in any way out of the use of
82
this software, even if advised of the possibility of such damage.
83
84
5.    CONTACT INFORMATION
85
86
Fraunhofer Institute for Integrated Circuits IIS
87
Attention: Audio and Multimedia Departments - FDK AAC LL
88
Am Wolfsmantel 33
89
91058 Erlangen, Germany
90
91
www.iis.fraunhofer.de/amm
92
amm-info@iis.fraunhofer.de
93
----------------------------------------------------------------------------- */
94
95
/******************* Library for basic calculation routines ********************
96
97
   Author(s):   Haricharan Lakshman, Manuel Jander
98
99
   Description: Trigonometric functions fixed point fractional implementation.
100
101
*******************************************************************************/
102
103
#include "FDK_trigFcts.h"
104
105
#include "fixpoint_math.h"
106
107
#define IMPROVE_ATAN2_ACCURACY 1 /* 0 --> 59 dB SNR     1 --> 65 dB SNR */
108
#define MINSFTAB 7
109
0
#define MAXSFTAB 25
110
111
#if IMPROVE_ATAN2_ACCURACY
112
static const FIXP_DBL f_atan_expand_range[MAXSFTAB - (MINSFTAB - 1)] = {
113
    /*****************************************************************************
114
     *
115
     *  Table holds fixp_atan() output values which are outside of input range
116
     *  of fixp_atan() to improve SNR of fixp_atan2().
117
     *
118
     *  This Table might also be used in fixp_atan() so there a wider input
119
     *  range can be covered, too.
120
     *
121
     *****************************************************************************/
122
    FL2FXCONST_DBL(7.775862990872099e-001),
123
    FL2FXCONST_DBL(7.814919928673978e-001),
124
    FL2FXCONST_DBL(7.834450483314648e-001),
125
    FL2FXCONST_DBL(7.844216021392089e-001),
126
    FL2FXCONST_DBL(7.849098823026687e-001),
127
    FL2FXCONST_DBL(7.851540227918509e-001),
128
    FL2FXCONST_DBL(7.852760930873737e-001),
129
    FL2FXCONST_DBL(7.853371282415015e-001),
130
    FL2FXCONST_DBL(7.853676458193612e-001),
131
    FL2FXCONST_DBL(7.853829046083906e-001),
132
    FL2FXCONST_DBL(7.853905340029177e-001),
133
    FL2FXCONST_DBL(7.853943487001828e-001),
134
    FL2FXCONST_DBL(7.853962560488155e-001),
135
    FL2FXCONST_DBL(7.853972097231319e-001),
136
    FL2FXCONST_DBL(7.853976865602901e-001),
137
    FL2FXCONST_DBL(7.853979249788692e-001),
138
    FL2FXCONST_DBL(7.853980441881587e-001),
139
    FL2FXCONST_DBL(7.853981037928035e-001),
140
    FL2FXCONST_DBL(7.853981335951259e-001)
141
    /* pi/4 = 0.785398163397448 = pi/2/ATO_SCALE */
142
};
143
#endif
144
145
0
FIXP_DBL fixp_atan2(FIXP_DBL y, FIXP_DBL x) {
146
0
  FIXP_DBL q;
147
0
  FIXP_DBL at;  /* atan  out */
148
0
  FIXP_DBL at2; /* atan2 out */
149
0
  FIXP_DBL ret = FL2FXCONST_DBL(-1.0f);
150
0
  INT sf, sfo, stf;
151
152
  /* --- division */
153
154
0
  if (y > FL2FXCONST_DBL(0.0f)) {
155
0
    if (x > FL2FXCONST_DBL(0.0f)) {
156
0
      q = fDivNormHighPrec(y, x, &sf); /* both pos. */
157
0
    } else if (x < FL2FXCONST_DBL(0.0f)) {
158
0
      q = -fDivNormHighPrec(y, -x, &sf); /* x neg. */
159
0
    } else {                             /* (x == FL2FXCONST_DBL(0.0f)) */
160
0
      q = FL2FXCONST_DBL(+1.0f);         /* y/x = pos/zero = +Inf */
161
0
      sf = 0;
162
0
    }
163
0
  } else if (y < FL2FXCONST_DBL(0.0f)) {
164
0
    if (x > FL2FXCONST_DBL(0.0f)) {
165
0
      q = -fDivNormHighPrec(-y, x, &sf); /* y neg. */
166
0
    } else if (x < FL2FXCONST_DBL(0.0f)) {
167
0
      q = fDivNormHighPrec(-y, -x, &sf); /* both neg. */
168
0
    } else {                             /* (x == FL2FXCONST_DBL(0.0f)) */
169
0
      q = FL2FXCONST_DBL(-1.0f);         /* y/x = neg/zero = -Inf */
170
0
      sf = 0;
171
0
    }
172
0
  } else { /* (y == FL2FXCONST_DBL(0.0f)) */
173
0
    q = FL2FXCONST_DBL(0.0f);
174
0
    sf = 0;
175
0
  }
176
0
  sfo = sf;
177
178
  /* --- atan() */
179
180
0
  if (sfo > ATI_SF) {
181
  /* --- could not calc fixp_atan() here bec of input data out of range */
182
  /*     ==> therefore give back boundary values */
183
184
0
#if IMPROVE_ATAN2_ACCURACY
185
0
    if (sfo > MAXSFTAB) sfo = MAXSFTAB;
186
0
#endif
187
188
0
    if (q > FL2FXCONST_DBL(0.0f)) {
189
0
#if IMPROVE_ATAN2_ACCURACY
190
0
      at = +f_atan_expand_range[sfo - ATI_SF - 1];
191
#else
192
      at = FL2FXCONST_DBL(+M_PI / 2 / ATO_SCALE);
193
#endif
194
0
    } else if (q < FL2FXCONST_DBL(0.0f)) {
195
0
#if IMPROVE_ATAN2_ACCURACY
196
0
      at = -f_atan_expand_range[sfo - ATI_SF - 1];
197
#else
198
      at = FL2FXCONST_DBL(-M_PI / 2 / ATO_SCALE);
199
#endif
200
0
    } else { /* q == FL2FXCONST_DBL(0.0f) */
201
0
      at = FL2FXCONST_DBL(0.0f);
202
0
    }
203
0
  } else {
204
    /* --- calc of fixp_atan() is possible; input data within range */
205
    /*     ==> set q on fixed scale level as desired from fixp_atan() */
206
0
    stf = sfo - ATI_SF;
207
0
    if (stf > 0)
208
0
      q = q << (INT)fMin(stf, DFRACT_BITS - 1);
209
0
    else
210
0
      q = q >> (INT)fMin(-stf, DFRACT_BITS - 1);
211
0
    at = fixp_atan(q); /* ATO_SF */
212
0
  }
213
214
  // --- atan2()
215
216
0
  at2 = at >> (AT2O_SF - ATO_SF);  // now AT2O_SF for atan2
217
0
  if (x > FL2FXCONST_DBL(0.0f)) {
218
0
    ret = at2;
219
0
  } else if (x < FL2FXCONST_DBL(0.0f)) {
220
0
    if (y >= FL2FXCONST_DBL(0.0f)) {
221
0
      ret = at2 + FL2FXCONST_DBL(M_PI / AT2O_SCALE);
222
0
    } else {
223
0
      ret = at2 - FL2FXCONST_DBL(M_PI / AT2O_SCALE);
224
0
    }
225
0
  } else {
226
    // x == 0
227
0
    if (y > FL2FXCONST_DBL(0.0f)) {
228
0
      ret = FL2FXCONST_DBL(+M_PI / 2 / AT2O_SCALE);
229
0
    } else if (y < FL2FXCONST_DBL(0.0f)) {
230
0
      ret = FL2FXCONST_DBL(-M_PI / 2 / AT2O_SCALE);
231
0
    } else if (y == FL2FXCONST_DBL(0.0f)) {
232
0
      ret = FL2FXCONST_DBL(0.0f);
233
0
    }
234
0
  }
235
0
  return ret;
236
0
}
237
238
0
FIXP_DBL fixp_atan(FIXP_DBL x) {
239
0
  INT sign;
240
0
  FIXP_DBL result, temp;
241
242
  /* SNR of fixp_atan() = 56 dB */
243
0
  FIXP_DBL P281 = (FIXP_DBL)0x00013000;     // 0.281 in q18
244
0
  FIXP_DBL ONEP571 = (FIXP_DBL)0x6487ef00;  // 1.571 in q30
245
246
0
  if (x < FIXP_DBL(0)) {
247
0
    sign = 1;
248
0
    x = -x;
249
0
  } else {
250
0
    sign = 0;
251
0
  }
252
0
  FDK_ASSERT(FL2FXCONST_DBL(1.0 / 64.0) == Q(Q_ATANINP));
253
  /* calc of arctan */
254
0
  if (x < FL2FXCONST_DBL(1.0 / 64.0))
255
  /*
256
    Chebyshev polynomial approximation of atan(x)
257
    5th-order approximation: atan(x) = a1*x + a2*x^3 + a3*x^5 = x(a1 + x^2*(a2 +
258
    a3*x^2)); a1 = 0.9949493661166540f, a2 = 0.2870606355326520f, a3 =
259
    0.0780371764464410f; 7th-order approximation: atan(x) = a1*x + a2*x^3 +
260
    a3*x^5 + a3*x^7 = x(a1 + x^2*(a2 + x^2*(a3 + a4*x^2))); a1 =
261
    0.9991334482227801, a2 = -0.3205332923816640, a3 = 0.1449824901444650, a4 =
262
    -0.0382544649702990; 7th-order approximation in use (the most accurate
263
    solution)
264
  */
265
0
  {
266
0
    x <<= ATI_SF;
267
0
    FIXP_DBL x2 = fPow2(x);
268
0
    temp = fMultAddDiv2((FL2FXCONST_DBL(0.1449824901444650f) >> 1), x2,
269
0
                        FL2FXCONST_DBL(-0.0382544649702990));
270
0
    temp = fMultAddDiv2((FL2FXCONST_DBL(-0.3205332923816640f) >> 2), x2, temp);
271
0
    temp = fMultAddDiv2((FL2FXCONST_DBL(0.9991334482227801f) >> 3), x2, temp);
272
0
    result = fMult(x, (temp << 2));
273
0
  } else if (x < FL2FXCONST_DBL(1.28 / 64.0)) {
274
0
    FIXP_DBL delta_fix;
275
0
    FIXP_DBL PI_BY_4 = FL2FXCONST_DBL(3.1415926 / 4.0) >> 1; /* pi/4 in q30 */
276
277
0
    delta_fix = (x - FL2FXCONST_DBL(1.0 / 64.0)) << 5; /* q30 */
278
0
    result = PI_BY_4 + (delta_fix >> 1) - (fPow2Div2(delta_fix));
279
0
  } else {
280
    /* Other approximation for |x| > 1.28 */
281
0
    INT res_e;
282
283
0
    temp = fPow2Div2(x); /* q25 * q25 - (DFRACT_BITS-1) - 1 = q18 */
284
0
    temp = temp + P281;  /* q18 + q18 = q18 */
285
0
    result = fDivNorm(x, temp, &res_e);
286
0
    result = scaleValue(result,
287
0
                        (Q_ATANOUT - Q_ATANINP + 18 - DFRACT_BITS + 1) + res_e);
288
0
    result = ONEP571 - result; /* q30 + q30 = q30 */
289
0
  }
290
0
  if (sign) {
291
0
    result = -result;
292
0
  }
293
294
0
  return (result);
295
0
}
296
297
#include "FDK_tools_rom.h"
298
299
4.33M
FIXP_DBL fixp_cos(FIXP_DBL x, int scale) {
300
4.33M
  FIXP_DBL residual, error, sine, cosine;
301
302
4.33M
  residual = fixp_sin_cos_residual_inline(x, scale, &sine, &cosine);
303
4.33M
  error = fMult(sine, residual);
304
305
4.33M
#ifdef SINETABLE_16BIT
306
4.33M
  return cosine - error;
307
#else
308
  /* Undo downscaling by 1 which was done at fixp_sin_cos_residual_inline */
309
  return SATURATE_LEFT_SHIFT(cosine - error, 1, DFRACT_BITS);
310
#endif
311
4.33M
}
312
313
0
FIXP_DBL fixp_sin(FIXP_DBL x, int scale) {
314
0
  FIXP_DBL residual, error, sine, cosine;
315
316
0
  residual = fixp_sin_cos_residual_inline(x, scale, &sine, &cosine);
317
0
  error = fMult(cosine, residual);
318
319
0
#ifdef SINETABLE_16BIT
320
0
  return sine + error;
321
#else
322
  return SATURATE_LEFT_SHIFT(sine + error, 1, DFRACT_BITS);
323
#endif
324
0
}
325
326
0
void fixp_cos_sin(FIXP_DBL x, int scale, FIXP_DBL *cos, FIXP_DBL *sin) {
327
0
  FIXP_DBL residual, error0, error1, sine, cosine;
328
329
0
  residual = fixp_sin_cos_residual_inline(x, scale, &sine, &cosine);
330
0
  error0 = fMult(sine, residual);
331
0
  error1 = fMult(cosine, residual);
332
333
0
#ifdef SINETABLE_16BIT
334
0
  *cos = cosine - error0;
335
0
  *sin = sine + error1;
336
#else
337
  *cos = SATURATE_LEFT_SHIFT(cosine - error0, 1, DFRACT_BITS);
338
  *sin = SATURATE_LEFT_SHIFT(sine + error1, 1, DFRACT_BITS);
339
#endif
340
0
}