libavutil/float_dsp.c

   1 /*
   2  * This file is part of Libav.
   3  *
   4  * Libav is free software; you can redistribute it and/or
   5  * modify it under the terms of the GNU Lesser General Public
   6  * License as published by the Free Software Foundation; either
   7  * version 2.1 of the License, or (at your option) any later version.
   8  *
   9  * Libav is distributed in the hope that it will be useful,
  10  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  11  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  12  * Lesser General Public License for more details.
  13  *
  14  * You should have received a copy of the GNU Lesser General Public
  15  * License along with Libav; if not, write to the Free Software
  16  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  17  */
  18
  19 #include "config.h"
  20
  21 #include "float_dsp.h"
  22
  23 static void vector_fmul_c(float *dst, const float *src0, const float *src1,
  24                           int len)
  25 {
  26     int i;
  27     for (i = 0; i < len; i++)
  28         dst[i] = src0[i] * src1[i];
  29 }
  30
  31 static void vector_fmac_scalar_c(float *dst, const float *src, float mul,
  32                                  int len)
  33 {
  34     int i;
  35     for (i = 0; i < len; i++)
  36         dst[i] += src[i] * mul;
  37 }
  38
  39 static void vector_fmul_scalar_c(float *dst, const float *src, float mul,
  40                                  int len)
  41 {
  42     int i;
  43     for (i = 0; i < len; i++)
  44         dst[i] = src[i] * mul;
  45 }
  46
  47 static void vector_dmul_scalar_c(double *dst, const double *src, double mul,
  48                                  int len)
  49 {
  50     int i;
  51     for (i = 0; i < len; i++)
  52         dst[i] = src[i] * mul;
  53 }
  54
  55 static void vector_fmul_window_c(float *dst, const float *src0,
  56                                  const float *src1, const float *win, int len)
  57 {
  58     int i, j;
  59
  60     dst  += len;
  61     win  += len;
  62     src0 += len;
  63
  64     for (i = -len, j = len - 1; i < 0; i++, j--) {
  65         float s0 = src0[i];
  66         float s1 = src1[j];
  67         float wi = win[i];
  68         float wj = win[j];
  69         dst[i] = s0 * wj - s1 * wi;
  70         dst[j] = s0 * wi + s1 * wj;
  71     }
  72 }
  73
  74 static void vector_fmul_add_c(float *dst, const float *src0, const float *src1,
  75                               const float *src2, int len){
  76     int i;
  77
  78     for (i = 0; i < len; i++)
  79         dst[i] = src0[i] * src1[i] + src2[i];
  80 }
  81
  82 static void vector_fmul_reverse_c(float *dst, const float *src0,
  83                                   const float *src1, int len)
  84 {
  85     int i;
  86
  87     src1 += len-1;
  88     for (i = 0; i < len; i++)
  89         dst[i] = src0[i] * src1[-i];
  90 }
  91
  92 static void butterflies_float_c(float *restrict v1, float *restrict v2,
  93                                 int len)
  94 {
  95     int i;
  96
  97     for (i = 0; i < len; i++) {
  98         float t = v1[i] - v2[i];
  99         v1[i] += v2[i];
 100         v2[i] = t;
 101     }
 102 }
 103
 104 float avpriv_scalarproduct_float_c(const float *v1, const float *v2, int len)
 105 {
 106     float p = 0.0;
 107     int i;
 108
 109     for (i = 0; i < len; i++)
 110         p += v1[i] * v2[i];
 111
 112     return p;
 113 }
 114
 115 void avpriv_float_dsp_init(AVFloatDSPContext *fdsp, int bit_exact)
 116 {
 117     fdsp->vector_fmul = vector_fmul_c;
 118     fdsp->vector_fmac_scalar = vector_fmac_scalar_c;
 119     fdsp->vector_fmul_scalar = vector_fmul_scalar_c;
 120     fdsp->vector_dmul_scalar = vector_dmul_scalar_c;
 121     fdsp->vector_fmul_window = vector_fmul_window_c;
 122     fdsp->vector_fmul_add = vector_fmul_add_c;
 123     fdsp->vector_fmul_reverse = vector_fmul_reverse_c;
 124     fdsp->butterflies_float = butterflies_float_c;
 125     fdsp->scalarproduct_float = avpriv_scalarproduct_float_c;
 126
 127 #if ARCH_ARM
 128     ff_float_dsp_init_arm(fdsp);
 129 #elif ARCH_PPC
 130     ff_float_dsp_init_ppc(fdsp, bit_exact);
 131 #elif ARCH_X86
 132     ff_float_dsp_init_x86(fdsp);
 133 #endif
 134 }