Statistics
| Branch: | Revision:

ffmpeg / libavcodec / fft.c @ c009df3f

History | View | Annotate | Download (6.1 KB)

1 bb6f5690 Fabrice Bellard
/*
2
 * FFT/IFFT transforms
3
 * Copyright (c) 2002 Fabrice Bellard.
4
 *
5
 * This library is free software; you can redistribute it and/or
6
 * modify it under the terms of the GNU Lesser General Public
7
 * License as published by the Free Software Foundation; either
8
 * version 2 of the License, or (at your option) any later version.
9
 *
10
 * This library is distributed in the hope that it will be useful,
11
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
13
 * Lesser General Public License for more details.
14
 *
15
 * You should have received a copy of the GNU Lesser General Public
16
 * License along with this library; if not, write to the Free Software
17
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
18
 */
19 983e3246 Michael Niedermayer
20
/**
21
 * @file fft.c
22
 * FFT/IFFT transforms.
23
 */
24
25 bb6f5690 Fabrice Bellard
#include "dsputil.h"
26
27
/**
28
 * The size of the FFT is 2^nbits. If inverse is TRUE, inverse FFT is
29
 * done 
30
 */
31
int fft_init(FFTContext *s, int nbits, int inverse)
32
{
33
    int i, j, m, n;
34
    float alpha, c1, s1, s2;
35
    
36
    s->nbits = nbits;
37
    n = 1 << nbits;
38
39
    s->exptab = av_malloc((n / 2) * sizeof(FFTComplex));
40
    if (!s->exptab)
41
        goto fail;
42
    s->revtab = av_malloc(n * sizeof(uint16_t));
43
    if (!s->revtab)
44
        goto fail;
45
    s->inverse = inverse;
46
47
    s2 = inverse ? 1.0 : -1.0;
48
        
49
    for(i=0;i<(n/2);i++) {
50
        alpha = 2 * M_PI * (float)i / (float)n;
51
        c1 = cos(alpha);
52
        s1 = sin(alpha) * s2;
53
        s->exptab[i].re = c1;
54
        s->exptab[i].im = s1;
55
    }
56
    s->fft_calc = fft_calc_c;
57
    s->exptab1 = NULL;
58
59
    /* compute constant table for HAVE_SSE version */
60 8d268a7d Fabrice Bellard
#if (defined(HAVE_MMX) && defined(HAVE_BUILTIN_VECTOR)) || defined(HAVE_ALTIVEC)
61
    {
62 db40a39a Michael Niedermayer
        int has_vectors = 0;
63 8d268a7d Fabrice Bellard
64
#if defined(HAVE_MMX)
65
        has_vectors = mm_support() & MM_SSE;
66 e629ab68 Romain Dolbeau
#endif
67 db40a39a Michael Niedermayer
#if defined(HAVE_ALTIVEC) && !defined(ALTIVEC_USE_REFERENCE_C_CODE)
68 e629ab68 Romain Dolbeau
        has_vectors = mm_support() & MM_ALTIVEC;
69 8d268a7d Fabrice Bellard
#endif
70
        if (has_vectors) {
71
            int np, nblocks, np2, l;
72
            FFTComplex *q;
73
            
74
            np = 1 << nbits;
75
            nblocks = np >> 3;
76
            np2 = np >> 1;
77
            s->exptab1 = av_malloc(np * 2 * sizeof(FFTComplex));
78
            if (!s->exptab1)
79
                goto fail;
80
            q = s->exptab1;
81
            do {
82
                for(l = 0; l < np2; l += 2 * nblocks) {
83
                    *q++ = s->exptab[l];
84
                    *q++ = s->exptab[l + nblocks];
85 bb6f5690 Fabrice Bellard
86 8d268a7d Fabrice Bellard
                    q->re = -s->exptab[l].im;
87
                    q->im = s->exptab[l].re;
88
                    q++;
89
                    q->re = -s->exptab[l + nblocks].im;
90
                    q->im = s->exptab[l + nblocks].re;
91
                    q++;
92
                }
93
                nblocks = nblocks >> 1;
94
            } while (nblocks != 0);
95
            av_freep(&s->exptab);
96
#if defined(HAVE_MMX)
97
            s->fft_calc = fft_calc_sse;
98
#else
99
            s->fft_calc = fft_calc_altivec;
100
#endif
101
        }
102 bb6f5690 Fabrice Bellard
    }
103
#endif
104
105
    /* compute bit reverse table */
106
107
    for(i=0;i<n;i++) {
108
        m=0;
109
        for(j=0;j<nbits;j++) {
110
            m |= ((i >> j) & 1) << (nbits-j-1);
111
        }
112
        s->revtab[i]=m;
113
    }
114
    return 0;
115
 fail:
116
    av_freep(&s->revtab);
117
    av_freep(&s->exptab);
118
    av_freep(&s->exptab1);
119
    return -1;
120
}
121
122
/* butter fly op */
123
#define BF(pre, pim, qre, qim, pre1, pim1, qre1, qim1) \
124
{\
125
  FFTSample ax, ay, bx, by;\
126
  bx=pre1;\
127
  by=pim1;\
128
  ax=qre1;\
129
  ay=qim1;\
130
  pre = (bx + ax);\
131
  pim = (by + ay);\
132
  qre = (bx - ax);\
133
  qim = (by - ay);\
134
}
135
136
#define MUL16(a,b) ((a) * (b))
137
138
#define CMUL(pre, pim, are, aim, bre, bim) \
139
{\
140
   pre = (MUL16(are, bre) - MUL16(aim, bim));\
141
   pim = (MUL16(are, bim) + MUL16(bre, aim));\
142
}
143
144
/**
145
 * Do a complex FFT with the parameters defined in fft_init(). The
146
 * input data must be permuted before with s->revtab table. No
147
 * 1.0/sqrt(n) normalization is done.  
148
 */
149
void fft_calc_c(FFTContext *s, FFTComplex *z)
150
{
151
    int ln = s->nbits;
152
    int        j, np, np2;
153
    int        nblocks, nloops;
154
    register FFTComplex *p, *q;
155
    FFTComplex *exptab = s->exptab;
156
    int l;
157
    FFTSample tmp_re, tmp_im;
158
159
    np = 1 << ln;
160
161
    /* pass 0 */
162
163
    p=&z[0];
164
    j=(np >> 1);
165
    do {
166
        BF(p[0].re, p[0].im, p[1].re, p[1].im, 
167
           p[0].re, p[0].im, p[1].re, p[1].im);
168
        p+=2;
169
    } while (--j != 0);
170
171
    /* pass 1 */
172
173
    
174
    p=&z[0];
175
    j=np >> 2;
176
    if (s->inverse) {
177
        do {
178
            BF(p[0].re, p[0].im, p[2].re, p[2].im, 
179
               p[0].re, p[0].im, p[2].re, p[2].im);
180
            BF(p[1].re, p[1].im, p[3].re, p[3].im, 
181
               p[1].re, p[1].im, -p[3].im, p[3].re);
182
            p+=4;
183
        } while (--j != 0);
184
    } else {
185
        do {
186
            BF(p[0].re, p[0].im, p[2].re, p[2].im, 
187
               p[0].re, p[0].im, p[2].re, p[2].im);
188
            BF(p[1].re, p[1].im, p[3].re, p[3].im, 
189
               p[1].re, p[1].im, p[3].im, -p[3].re);
190
            p+=4;
191
        } while (--j != 0);
192
    }
193
    /* pass 2 .. ln-1 */
194
195
    nblocks = np >> 3;
196
    nloops = 1 << 2;
197
    np2 = np >> 1;
198
    do {
199
        p = z;
200
        q = z + nloops;
201
        for (j = 0; j < nblocks; ++j) {
202
            BF(p->re, p->im, q->re, q->im,
203
               p->re, p->im, q->re, q->im);
204
            
205
            p++;
206
            q++;
207
            for(l = nblocks; l < np2; l += nblocks) {
208
                CMUL(tmp_re, tmp_im, exptab[l].re, exptab[l].im, q->re, q->im);
209
                BF(p->re, p->im, q->re, q->im,
210
                   p->re, p->im, tmp_re, tmp_im);
211
                p++;
212
                q++;
213
            }
214
215
            p += nloops;
216
            q += nloops;
217
        }
218
        nblocks = nblocks >> 1;
219
        nloops = nloops << 1;
220
    } while (nblocks != 0);
221
}
222
223
/**
224
 * Do the permutation needed BEFORE calling fft_calc()
225
 */
226
void fft_permute(FFTContext *s, FFTComplex *z)
227
{
228
    int j, k, np;
229
    FFTComplex tmp;
230
    const uint16_t *revtab = s->revtab;
231
    
232
    /* reverse */
233
    np = 1 << s->nbits;
234
    for(j=0;j<np;j++) {
235
        k = revtab[j];
236
        if (k < j) {
237
            tmp = z[k];
238
            z[k] = z[j];
239
            z[j] = tmp;
240
        }
241
    }
242
}
243
244
void fft_end(FFTContext *s)
245
{
246
    av_freep(&s->revtab);
247
    av_freep(&s->exptab);
248
    av_freep(&s->exptab1);
249
}