libavcodec/mdct.c

   1 /*
   2  * MDCT/IMDCT transforms
   3  * Copyright (c) 2002 Fabrice Bellard
   4  *
   5  * This file is part of Libav.
   6  *
   7  * Libav is free software; you can redistribute it and/or
   8  * modify it under the terms of the GNU Lesser General Public
   9  * License as published by the Free Software Foundation; either
  10  * version 2.1 of the License, or (at your option) any later version.
  11  *
  12  * Libav is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  15  * Lesser General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU Lesser General Public
  18  * License along with Libav; if not, write to the Free Software
  19  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  20  */
  21
  22 #include <stdlib.h>
  23 #include <string.h>
  24 #include "libavutil/common.h"
  25 #include "libavutil/mathematics.h"
  26 #include "fft.h"
  27 #include "fft-internal.h"
  28
  29 /**
  30  * @file
  31  * MDCT/IMDCT transforms.
  32  */
  33
  34 #if CONFIG_FFT_FLOAT
  35 #   define RSCALE(x) (x)
  36 #else
  37 #   define RSCALE(x) ((x) >> 1)
  38 #endif
  39
  40 /**
  41  * init MDCT or IMDCT computation.
  42  */
  43 av_cold int ff_mdct_init(FFTContext *s, int nbits, int inverse, double scale)
  44 {
  45     int n, n4, i;
  46     double alpha, theta;
  47     int tstep;
  48
  49     memset(s, 0, sizeof(*s));
  50     n = 1 << nbits;
  51     s->mdct_bits = nbits;
  52     s->mdct_size = n;
  53     n4 = n >> 2;
  54     s->mdct_permutation = FF_MDCT_PERM_NONE;
  55
  56     if (ff_fft_init(s, s->mdct_bits - 2, inverse) < 0)
  57         goto fail;
  58
  59     s->tcos = av_malloc(n/2 * sizeof(FFTSample));
  60     if (!s->tcos)
  61         goto fail;
  62
  63     switch (s->mdct_permutation) {
  64     case FF_MDCT_PERM_NONE:
  65         s->tsin = s->tcos + n4;
  66         tstep = 1;
  67         break;
  68     case FF_MDCT_PERM_INTERLEAVE:
  69         s->tsin = s->tcos + 1;
  70         tstep = 2;
  71         break;
  72     default:
  73         goto fail;
  74     }
  75
  76     theta = 1.0 / 8.0 + (scale < 0 ? n4 : 0);
  77     scale = sqrt(fabs(scale));
  78     for(i=0;i<n4;i++) {
  79         alpha = 2 * M_PI * (i + theta) / n;
  80         s->tcos[i*tstep] = FIX15(-cos(alpha) * scale);
  81         s->tsin[i*tstep] = FIX15(-sin(alpha) * scale);
  82     }
  83     return 0;
  84  fail:
  85     ff_mdct_end(s);
  86     return -1;
  87 }
  88
  89 /**
  90  * Compute the middle half of the inverse MDCT of size N = 2^nbits,
  91  * thus excluding the parts that can be derived by symmetry
  92  * @param output N/2 samples
  93  * @param input N/2 samples
  94  */
  95 void ff_imdct_half_c(FFTContext *s, FFTSample *output, const FFTSample *input)
  96 {
  97     int k, n8, n4, n2, n, j;
  98     const uint16_t *revtab = s->revtab;
  99     const FFTSample *tcos = s->tcos;
 100     const FFTSample *tsin = s->tsin;
 101     const FFTSample *in1, *in2;
 102     FFTComplex *z = (FFTComplex *)output;
 103
 104     n = 1 << s->mdct_bits;
 105     n2 = n >> 1;
 106     n4 = n >> 2;
 107     n8 = n >> 3;
 108
 109     /* pre rotation */
 110     in1 = input;
 111     in2 = input + n2 - 1;
 112     for(k = 0; k < n4; k++) {
 113         j=revtab[k];
 114         CMUL(z[j].re, z[j].im, *in2, *in1, tcos[k], tsin[k]);
 115         in1 += 2;
 116         in2 -= 2;
 117     }
 118     s->fft_calc(s, z);
 119
 120     /* post rotation + reordering */
 121     for(k = 0; k < n8; k++) {
 122         FFTSample r0, i0, r1, i1;
 123         CMUL(r0, i1, z[n8-k-1].im, z[n8-k-1].re, tsin[n8-k-1], tcos[n8-k-1]);
 124         CMUL(r1, i0, z[n8+k  ].im, z[n8+k  ].re, tsin[n8+k  ], tcos[n8+k  ]);
 125         z[n8-k-1].re = r0;
 126         z[n8-k-1].im = i0;
 127         z[n8+k  ].re = r1;
 128         z[n8+k  ].im = i1;
 129     }
 130 }
 131
 132 /**
 133  * Compute inverse MDCT of size N = 2^nbits
 134  * @param output N samples
 135  * @param input N/2 samples
 136  */
 137 void ff_imdct_calc_c(FFTContext *s, FFTSample *output, const FFTSample *input)
 138 {
 139     int k;
 140     int n = 1 << s->mdct_bits;
 141     int n2 = n >> 1;
 142     int n4 = n >> 2;
 143
 144     ff_imdct_half_c(s, output+n4, input);
 145
 146     for(k = 0; k < n4; k++) {
 147         output[k] = -output[n2-k-1];
 148         output[n-k-1] = output[n2+k];
 149     }
 150 }
 151
 152 /**
 153  * Compute MDCT of size N = 2^nbits
 154  * @param input N samples
 155  * @param out N/2 samples
 156  */
 157 void ff_mdct_calc_c(FFTContext *s, FFTSample *out, const FFTSample *input)
 158 {
 159     int i, j, n, n8, n4, n2, n3;
 160     FFTDouble re, im;
 161     const uint16_t *revtab = s->revtab;
 162     const FFTSample *tcos = s->tcos;
 163     const FFTSample *tsin = s->tsin;
 164     FFTComplex *x = (FFTComplex *)out;
 165
 166     n = 1 << s->mdct_bits;
 167     n2 = n >> 1;
 168     n4 = n >> 2;
 169     n8 = n >> 3;
 170     n3 = 3 * n4;
 171
 172     /* pre rotation */
 173     for(i=0;i<n8;i++) {
 174         re = RSCALE(-input[2*i+n3] - input[n3-1-2*i]);
 175         im = RSCALE(-input[n4+2*i] + input[n4-1-2*i]);
 176         j = revtab[i];
 177         CMUL(x[j].re, x[j].im, re, im, -tcos[i], tsin[i]);
 178
 179         re = RSCALE( input[2*i]    - input[n2-1-2*i]);
 180         im = RSCALE(-input[n2+2*i] - input[ n-1-2*i]);
 181         j = revtab[n8 + i];
 182         CMUL(x[j].re, x[j].im, re, im, -tcos[n8 + i], tsin[n8 + i]);
 183     }
 184
 185     s->fft_calc(s, x);
 186
 187     /* post rotation */
 188     for(i=0;i<n8;i++) {
 189         FFTSample r0, i0, r1, i1;
 190         CMUL(i1, r0, x[n8-i-1].re, x[n8-i-1].im, -tsin[n8-i-1], -tcos[n8-i-1]);
 191         CMUL(i0, r1, x[n8+i  ].re, x[n8+i  ].im, -tsin[n8+i  ], -tcos[n8+i  ]);
 192         x[n8-i-1].re = r0;
 193         x[n8-i-1].im = i0;
 194         x[n8+i  ].re = r1;
 195         x[n8+i  ].im = i1;
 196     }
 197 }
 198
 199 av_cold void ff_mdct_end(FFTContext *s)
 200 {
 201     av_freep(&s->tcos);
 202     ff_fft_end(s);
 203 }