libavcodec/hq_hqadsp.c

   1 /*
   2  * Canopus HQ/HQA decoder
   3  *
   4  * This file is part of FFmpeg.
   5  *
   6  * FFmpeg is free software; you can redistribute it and/or
   7  * modify it under the terms of the GNU Lesser General Public
   8  * License as published by the Free Software Foundation; either
   9  * version 2.1 of the License, or (at your option) any later version.
  10  *
  11  * FFmpeg is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14  * Lesser General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU Lesser General Public
  17  * License along with FFmpeg; if not, write to the Free Software
  18  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  19  */
  20
  21 #include <stdint.h>
  22
  23 #include "libavutil/attributes.h"
  24 #include "libavutil/common.h"
  25
  26 #include "hq_hqadsp.h"
  27
  28 #define FIX_1_082 17734
  29 #define FIX_1_847 30274
  30 #define FIX_1_414 23170
  31 #define FIX_2_613 21407 // divided by two to fit the range
  32
  33 #define IDCTMUL(a, b) ((int)((a) * (unsigned)(b)) >> 16)
  34
  35 static inline void idct_row(int16_t *blk)
  36 {
  37     int tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7, tmp8, tmp9, tmpA;
  38     int tmpB, tmpC, tmpD, tmpE, tmpF, tmp10, tmp11, tmp12, tmp13, tmp14;
  39
  40     tmp0 = blk[5] - blk[3];
  41     tmp1 = blk[5] + blk[3];
  42     tmp2 = blk[1] - blk[7];
  43     tmp3 = blk[1] + blk[7];
  44     tmp4 = tmp3 - tmp1;
  45     tmp5 = IDCTMUL(tmp0 + tmp2, FIX_1_847);
  46     tmp6 = IDCTMUL(tmp2,        FIX_1_082) - tmp5;
  47     tmp7 = tmp5 - IDCTMUL(tmp0, FIX_2_613) * 2;
  48     tmp8 = tmp3 + tmp1;
  49     tmp9 = tmp7 * 4 - tmp8;
  50     tmpA = IDCTMUL(tmp4, FIX_1_414) * 4 - tmp9;
  51     tmpB = tmp6 * 4 + tmpA;
  52     tmpC = blk[2] + blk[6];
  53     tmpD = blk[2] - blk[6];
  54     tmpE = blk[0] - blk[4];
  55     tmpF = blk[0] + blk[4];
  56
  57     tmp10 = IDCTMUL(tmpD, FIX_1_414) * 4 - tmpC;
  58     tmp11 = tmpE - tmp10;
  59     tmp12 = tmpF - tmpC;
  60     tmp13 = tmpE + tmp10;
  61     tmp14 = tmpF + tmpC;
  62
  63     blk[0] = tmp14 + tmp8;
  64     blk[1] = tmp13 + tmp9;
  65     blk[2] = tmp11 + tmpA;
  66     blk[3] = tmp12 - tmpB;
  67     blk[4] = tmp12 + tmpB;
  68     blk[5] = tmp11 - tmpA;
  69     blk[6] = tmp13 - tmp9;
  70     blk[7] = tmp14 - tmp8;
  71 }
  72
  73 static inline void idct_col(int16_t *blk)
  74 {
  75     int tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7, tmp8, tmp9, tmpA;
  76     int tmpB, tmpC, tmpD, tmpE, tmpF, tmp10, tmp11, tmp12, tmp13, tmp14;
  77
  78     tmp0 = blk[5 * 8] - blk[3 * 8];
  79     tmp1 = blk[5 * 8] + blk[3 * 8];
  80     tmp2 = blk[1 * 8] * 2 - (blk[7 * 8] >> 2);
  81     tmp3 = blk[1 * 8] * 2 + (blk[7 * 8] >> 2);
  82     tmp4 = tmp3 - tmp1;
  83     tmp5 = IDCTMUL(tmp0 + tmp2, FIX_1_847);
  84     tmp6 = IDCTMUL(tmp2,        FIX_1_082) - tmp5;
  85     tmp7 = tmp5 - IDCTMUL(tmp0, FIX_2_613) * 2;
  86     tmp8 = (tmp3 + tmp1) >> 1;
  87     tmp9 = tmp7 * 2 - tmp8;
  88     tmpA = IDCTMUL(tmp4, FIX_1_414) * 2 - tmp9;
  89     tmpB = tmp6 * 2 + tmpA;
  90     tmpC =  blk[2 * 8] + (blk[6 * 8] >> 1) >> 1;
  91     tmpD =  blk[2 * 8] - (blk[6 * 8] >> 1);
  92     tmpE = (blk[0 * 8] >> 1) - (blk[4 * 8] >> 1) + 0x2020;
  93     tmpF = (blk[0 * 8] >> 1) + (blk[4 * 8] >> 1) + 0x2020;
  94
  95     tmp10 = IDCTMUL(tmpD, FIX_1_414) * 2 - tmpC;
  96     tmp11 = tmpE - tmp10;
  97     tmp12 = tmpF - tmpC;
  98     tmp13 = tmpE + tmp10;
  99     tmp14 = tmpF + tmpC;
 100
 101     blk[0 * 8] = (tmp14 + tmp8) >> 6;
 102     blk[1 * 8] = (tmp13 + tmp9) >> 6;
 103     blk[2 * 8] = (tmp11 + tmpA) >> 6;
 104     blk[3 * 8] = (tmp12 - tmpB) >> 6;
 105     blk[4 * 8] = (tmp12 + tmpB) >> 6;
 106     blk[5 * 8] = (tmp11 - tmpA) >> 6;
 107     blk[6 * 8] = (tmp13 - tmp9) >> 6;
 108     blk[7 * 8] = (tmp14 - tmp8) >> 6;
 109 }
 110
 111 static void hq_idct_put(uint8_t *dst, int stride, int16_t *block)
 112 {
 113     int i, j;
 114
 115     for (i = 0; i < 8; i++)
 116         idct_row(block + i * 8);
 117     for (i = 0; i < 8; i++)
 118         idct_col(block + i);
 119
 120     for (i = 0; i < 8; i++) {
 121         for (j = 0; j < 8; j++)
 122             dst[j] = av_clip_uint8(block[j + i * 8]);
 123         dst += stride;
 124     }
 125 }
 126
 127 av_cold void ff_hqdsp_init(HQDSPContext *c)
 128 {
 129     c->idct_put = hq_idct_put;
 130 }