libavcodec/h264_mc_template.c

   1 /*
   2  * H.26L/H.264/AVC/JVT/14496-10/... decoder
   3  * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
   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 #undef MCFUNC
  23
  24 #if   CHROMA_IDC == 1
  25 #   define MCFUNC(n) FUNC(n ## _420)
  26 #elif CHROMA_IDC == 2
  27 #   define MCFUNC(n) FUNC(n ## _422)
  28 #elif CHROMA_IDC == 3
  29 #   define MCFUNC(n) FUNC(n ## _444)
  30 #endif
  31
  32 #undef  mc_part
  33 #define mc_part MCFUNC(mc_part)
  34
  35 static void mc_part(H264Context *h, int n, int square,
  36                     int height, int delta,
  37                     uint8_t *dest_y, uint8_t *dest_cb,
  38                     uint8_t *dest_cr,
  39                     int x_offset, int y_offset,
  40                     qpel_mc_func *qpix_put,
  41                     h264_chroma_mc_func chroma_put,
  42                     qpel_mc_func *qpix_avg,
  43                     h264_chroma_mc_func chroma_avg,
  44                     h264_weight_func *weight_op,
  45                     h264_biweight_func *weight_avg,
  46                     int list0, int list1)
  47 {
  48     if ((h->use_weight == 2 && list0 && list1 &&
  49          (h->implicit_weight[h->ref_cache[0][scan8[n]]][h->ref_cache[1][scan8[n]]][h->mb_y & 1] != 32)) ||
  50         h->use_weight == 1)
  51         mc_part_weighted(h, n, square, height, delta, dest_y, dest_cb, dest_cr,
  52                          x_offset, y_offset, qpix_put, chroma_put,
  53                          weight_op[0], weight_op[1], weight_avg[0],
  54                          weight_avg[1], list0, list1, PIXEL_SHIFT, CHROMA_IDC);
  55     else
  56         mc_part_std(h, n, square, height, delta, dest_y, dest_cb, dest_cr,
  57                     x_offset, y_offset, qpix_put, chroma_put, qpix_avg,
  58                     chroma_avg, list0, list1, PIXEL_SHIFT, CHROMA_IDC);
  59 }
  60
  61 static void MCFUNC(hl_motion)(H264Context *h, uint8_t *dest_y,
  62                               uint8_t *dest_cb, uint8_t *dest_cr,
  63                               qpel_mc_func(*qpix_put)[16],
  64                               h264_chroma_mc_func(*chroma_put),
  65                               qpel_mc_func(*qpix_avg)[16],
  66                               h264_chroma_mc_func(*chroma_avg),
  67                               h264_weight_func *weight_op,
  68                               h264_biweight_func *weight_avg)
  69 {
  70     const int mb_xy   = h->mb_xy;
  71     const int mb_type = h->cur_pic.mb_type[mb_xy];
  72
  73     assert(IS_INTER(mb_type));
  74
  75     if (HAVE_THREADS && (h->avctx->active_thread_type & FF_THREAD_FRAME))
  76         await_references(h);
  77     prefetch_motion(h, 0, PIXEL_SHIFT, CHROMA_IDC);
  78
  79     if (IS_16X16(mb_type)) {
  80         mc_part(h, 0, 1, 16, 0, dest_y, dest_cb, dest_cr, 0, 0,
  81                 qpix_put[0], chroma_put[0], qpix_avg[0], chroma_avg[0],
  82                 weight_op, weight_avg,
  83                 IS_DIR(mb_type, 0, 0), IS_DIR(mb_type, 0, 1));
  84     } else if (IS_16X8(mb_type)) {
  85         mc_part(h, 0, 0, 8, 8 << PIXEL_SHIFT, dest_y, dest_cb, dest_cr, 0, 0,
  86                 qpix_put[1], chroma_put[0], qpix_avg[1], chroma_avg[0],
  87                 weight_op, weight_avg,
  88                 IS_DIR(mb_type, 0, 0), IS_DIR(mb_type, 0, 1));
  89         mc_part(h, 8, 0, 8, 8 << PIXEL_SHIFT, dest_y, dest_cb, dest_cr, 0, 4,
  90                 qpix_put[1], chroma_put[0], qpix_avg[1], chroma_avg[0],
  91                 weight_op, weight_avg,
  92                 IS_DIR(mb_type, 1, 0), IS_DIR(mb_type, 1, 1));
  93     } else if (IS_8X16(mb_type)) {
  94         mc_part(h, 0, 0, 16, 8 * h->mb_linesize, dest_y, dest_cb, dest_cr, 0, 0,
  95                 qpix_put[1], chroma_put[1], qpix_avg[1], chroma_avg[1],
  96                 &weight_op[1], &weight_avg[1],
  97                 IS_DIR(mb_type, 0, 0), IS_DIR(mb_type, 0, 1));
  98         mc_part(h, 4, 0, 16, 8 * h->mb_linesize, dest_y, dest_cb, dest_cr, 4, 0,
  99                 qpix_put[1], chroma_put[1], qpix_avg[1], chroma_avg[1],
 100                 &weight_op[1], &weight_avg[1],
 101                 IS_DIR(mb_type, 1, 0), IS_DIR(mb_type, 1, 1));
 102     } else {
 103         int i;
 104
 105         assert(IS_8X8(mb_type));
 106
 107         for (i = 0; i < 4; i++) {
 108             const int sub_mb_type = h->sub_mb_type[i];
 109             const int n  = 4 * i;
 110             int x_offset = (i & 1) << 2;
 111             int y_offset = (i & 2) << 1;
 112
 113             if (IS_SUB_8X8(sub_mb_type)) {
 114                 mc_part(h, n, 1, 8, 0, dest_y, dest_cb, dest_cr,
 115                         x_offset, y_offset,
 116                         qpix_put[1], chroma_put[1], qpix_avg[1], chroma_avg[1],
 117                         &weight_op[1], &weight_avg[1],
 118                         IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
 119             } else if (IS_SUB_8X4(sub_mb_type)) {
 120                 mc_part(h, n, 0, 4, 4 << PIXEL_SHIFT, dest_y, dest_cb, dest_cr,
 121                         x_offset, y_offset,
 122                         qpix_put[2], chroma_put[1], qpix_avg[2], chroma_avg[1],
 123                         &weight_op[1], &weight_avg[1],
 124                         IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
 125                 mc_part(h, n + 2, 0, 4, 4 << PIXEL_SHIFT,
 126                         dest_y, dest_cb, dest_cr, x_offset, y_offset + 2,
 127                         qpix_put[2], chroma_put[1], qpix_avg[2], chroma_avg[1],
 128                         &weight_op[1], &weight_avg[1],
 129                         IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
 130             } else if (IS_SUB_4X8(sub_mb_type)) {
 131                 mc_part(h, n, 0, 8, 4 * h->mb_linesize,
 132                         dest_y, dest_cb, dest_cr, x_offset, y_offset,
 133                         qpix_put[2], chroma_put[2], qpix_avg[2], chroma_avg[2],
 134                         &weight_op[2], &weight_avg[2],
 135                         IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
 136                 mc_part(h, n + 1, 0, 8, 4 * h->mb_linesize,
 137                         dest_y, dest_cb, dest_cr, x_offset + 2, y_offset,
 138                         qpix_put[2], chroma_put[2], qpix_avg[2], chroma_avg[2],
 139                         &weight_op[2], &weight_avg[2],
 140                         IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
 141             } else {
 142                 int j;
 143                 assert(IS_SUB_4X4(sub_mb_type));
 144                 for (j = 0; j < 4; j++) {
 145                     int sub_x_offset = x_offset + 2 * (j & 1);
 146                     int sub_y_offset = y_offset + (j & 2);
 147                     mc_part(h, n + j, 1, 4, 0,
 148                             dest_y, dest_cb, dest_cr, sub_x_offset, sub_y_offset,
 149                             qpix_put[2], chroma_put[2], qpix_avg[2], chroma_avg[2],
 150                             &weight_op[2], &weight_avg[2],
 151                             IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
 152                 }
 153             }
 154         }
 155     }
 156
 157     prefetch_motion(h, 1, PIXEL_SHIFT, CHROMA_IDC);
 158 }
 159