yading@10: /*
yading@10:  * H.26L/H.264/AVC/JVT/14496-10/... decoder
yading@10:  * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
yading@10:  *
yading@10:  * This file is part of FFmpeg.
yading@10:  *
yading@10:  * FFmpeg is free software; you can redistribute it and/or
yading@10:  * modify it under the terms of the GNU Lesser General Public
yading@10:  * License as published by the Free Software Foundation; either
yading@10:  * version 2.1 of the License, or (at your option) any later version.
yading@10:  *
yading@10:  * FFmpeg is distributed in the hope that it will be useful,
yading@10:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
yading@10:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
yading@10:  * Lesser General Public License for more details.
yading@10:  *
yading@10:  * You should have received a copy of the GNU Lesser General Public
yading@10:  * License along with FFmpeg; if not, write to the Free Software
yading@10:  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
yading@10:  */
yading@10: 
yading@10: #undef MCFUNC
yading@10: 
yading@10: #if   CHROMA_IDC == 1
yading@10: #   define MCFUNC(n) FUNC(n ## _420)
yading@10: #elif CHROMA_IDC == 2
yading@10: #   define MCFUNC(n) FUNC(n ## _422)
yading@10: #elif CHROMA_IDC == 3
yading@10: #   define MCFUNC(n) FUNC(n ## _444)
yading@10: #endif
yading@10: 
yading@10: #undef  mc_part
yading@10: #define mc_part MCFUNC(mc_part)
yading@10: 
yading@10: static void mc_part(H264Context *h, int n, int square,
yading@10:                     int height, int delta,
yading@10:                     uint8_t *dest_y, uint8_t *dest_cb,
yading@10:                     uint8_t *dest_cr,
yading@10:                     int x_offset, int y_offset,
yading@10:                     qpel_mc_func *qpix_put,
yading@10:                     h264_chroma_mc_func chroma_put,
yading@10:                     qpel_mc_func *qpix_avg,
yading@10:                     h264_chroma_mc_func chroma_avg,
yading@10:                     h264_weight_func *weight_op,
yading@10:                     h264_biweight_func *weight_avg,
yading@10:                     int list0, int list1)
yading@10: {
yading@10:     if ((h->use_weight == 2 && list0 && list1 &&
yading@10:          (h->implicit_weight[h->ref_cache[0][scan8[n]]][h->ref_cache[1][scan8[n]]][h->mb_y & 1] != 32)) ||
yading@10:         h->use_weight == 1)
yading@10:         mc_part_weighted(h, n, square, height, delta, dest_y, dest_cb, dest_cr,
yading@10:                          x_offset, y_offset, qpix_put, chroma_put,
yading@10:                          weight_op[0], weight_op[1], weight_avg[0],
yading@10:                          weight_avg[1], list0, list1, PIXEL_SHIFT, CHROMA_IDC);
yading@10:     else
yading@10:         mc_part_std(h, n, square, height, delta, dest_y, dest_cb, dest_cr,
yading@10:                     x_offset, y_offset, qpix_put, chroma_put, qpix_avg,
yading@10:                     chroma_avg, list0, list1, PIXEL_SHIFT, CHROMA_IDC);
yading@10: }
yading@10: 
yading@10: static void MCFUNC(hl_motion)(H264Context *h, uint8_t *dest_y,
yading@10:                               uint8_t *dest_cb, uint8_t *dest_cr,
yading@10:                               qpel_mc_func(*qpix_put)[16],
yading@10:                               h264_chroma_mc_func(*chroma_put),
yading@10:                               qpel_mc_func(*qpix_avg)[16],
yading@10:                               h264_chroma_mc_func(*chroma_avg),
yading@10:                               h264_weight_func *weight_op,
yading@10:                               h264_biweight_func *weight_avg)
yading@10: {
yading@10:     const int mb_xy   = h->mb_xy;
yading@10:     const int mb_type = h->cur_pic.mb_type[mb_xy];
yading@10: 
yading@10:     av_assert2(IS_INTER(mb_type));
yading@10: 
yading@10:     if (HAVE_THREADS && (h->avctx->active_thread_type & FF_THREAD_FRAME))
yading@10:         await_references(h);
yading@10:     prefetch_motion(h, 0, PIXEL_SHIFT, CHROMA_IDC);
yading@10: 
yading@10:     if (IS_16X16(mb_type)) {
yading@10:         mc_part(h, 0, 1, 16, 0, dest_y, dest_cb, dest_cr, 0, 0,
yading@10:                 qpix_put[0], chroma_put[0], qpix_avg[0], chroma_avg[0],
yading@10:                 weight_op, weight_avg,
yading@10:                 IS_DIR(mb_type, 0, 0), IS_DIR(mb_type, 0, 1));
yading@10:     } else if (IS_16X8(mb_type)) {
yading@10:         mc_part(h, 0, 0, 8, 8 << PIXEL_SHIFT, dest_y, dest_cb, dest_cr, 0, 0,
yading@10:                 qpix_put[1], chroma_put[0], qpix_avg[1], chroma_avg[0],
yading@10:                 weight_op, weight_avg,
yading@10:                 IS_DIR(mb_type, 0, 0), IS_DIR(mb_type, 0, 1));
yading@10:         mc_part(h, 8, 0, 8, 8 << PIXEL_SHIFT, dest_y, dest_cb, dest_cr, 0, 4,
yading@10:                 qpix_put[1], chroma_put[0], qpix_avg[1], chroma_avg[0],
yading@10:                 weight_op, weight_avg,
yading@10:                 IS_DIR(mb_type, 1, 0), IS_DIR(mb_type, 1, 1));
yading@10:     } else if (IS_8X16(mb_type)) {
yading@10:         mc_part(h, 0, 0, 16, 8 * h->mb_linesize, dest_y, dest_cb, dest_cr, 0, 0,
yading@10:                 qpix_put[1], chroma_put[1], qpix_avg[1], chroma_avg[1],
yading@10:                 &weight_op[1], &weight_avg[1],
yading@10:                 IS_DIR(mb_type, 0, 0), IS_DIR(mb_type, 0, 1));
yading@10:         mc_part(h, 4, 0, 16, 8 * h->mb_linesize, dest_y, dest_cb, dest_cr, 4, 0,
yading@10:                 qpix_put[1], chroma_put[1], qpix_avg[1], chroma_avg[1],
yading@10:                 &weight_op[1], &weight_avg[1],
yading@10:                 IS_DIR(mb_type, 1, 0), IS_DIR(mb_type, 1, 1));
yading@10:     } else {
yading@10:         int i;
yading@10: 
yading@10:         av_assert2(IS_8X8(mb_type));
yading@10: 
yading@10:         for (i = 0; i < 4; i++) {
yading@10:             const int sub_mb_type = h->sub_mb_type[i];
yading@10:             const int n  = 4 * i;
yading@10:             int x_offset = (i & 1) << 2;
yading@10:             int y_offset = (i & 2) << 1;
yading@10: 
yading@10:             if (IS_SUB_8X8(sub_mb_type)) {
yading@10:                 mc_part(h, n, 1, 8, 0, dest_y, dest_cb, dest_cr,
yading@10:                         x_offset, y_offset,
yading@10:                         qpix_put[1], chroma_put[1], qpix_avg[1], chroma_avg[1],
yading@10:                         &weight_op[1], &weight_avg[1],
yading@10:                         IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
yading@10:             } else if (IS_SUB_8X4(sub_mb_type)) {
yading@10:                 mc_part(h, n, 0, 4, 4 << PIXEL_SHIFT, dest_y, dest_cb, dest_cr,
yading@10:                         x_offset, y_offset,
yading@10:                         qpix_put[2], chroma_put[1], qpix_avg[2], chroma_avg[1],
yading@10:                         &weight_op[1], &weight_avg[1],
yading@10:                         IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
yading@10:                 mc_part(h, n + 2, 0, 4, 4 << PIXEL_SHIFT,
yading@10:                         dest_y, dest_cb, dest_cr, x_offset, y_offset + 2,
yading@10:                         qpix_put[2], chroma_put[1], qpix_avg[2], chroma_avg[1],
yading@10:                         &weight_op[1], &weight_avg[1],
yading@10:                         IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
yading@10:             } else if (IS_SUB_4X8(sub_mb_type)) {
yading@10:                 mc_part(h, n, 0, 8, 4 * h->mb_linesize,
yading@10:                         dest_y, dest_cb, dest_cr, x_offset, y_offset,
yading@10:                         qpix_put[2], chroma_put[2], qpix_avg[2], chroma_avg[2],
yading@10:                         &weight_op[2], &weight_avg[2],
yading@10:                         IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
yading@10:                 mc_part(h, n + 1, 0, 8, 4 * h->mb_linesize,
yading@10:                         dest_y, dest_cb, dest_cr, x_offset + 2, y_offset,
yading@10:                         qpix_put[2], chroma_put[2], qpix_avg[2], chroma_avg[2],
yading@10:                         &weight_op[2], &weight_avg[2],
yading@10:                         IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
yading@10:             } else {
yading@10:                 int j;
yading@10:                 av_assert2(IS_SUB_4X4(sub_mb_type));
yading@10:                 for (j = 0; j < 4; j++) {
yading@10:                     int sub_x_offset = x_offset + 2 * (j & 1);
yading@10:                     int sub_y_offset = y_offset + (j & 2);
yading@10:                     mc_part(h, n + j, 1, 4, 0,
yading@10:                             dest_y, dest_cb, dest_cr, sub_x_offset, sub_y_offset,
yading@10:                             qpix_put[2], chroma_put[2], qpix_avg[2], chroma_avg[2],
yading@10:                             &weight_op[2], &weight_avg[2],
yading@10:                             IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
yading@10:                 }
yading@10:             }
yading@10:         }
yading@10:     }
yading@10: 
yading@10:     prefetch_motion(h, 1, PIXEL_SHIFT, CHROMA_IDC);
yading@10: }
yading@10: