x264 - x264_mb_analyse_inter_p16x16
2014-06-14 13:39
501 查看
static void x264_mb_analyse_inter_p16x16( x264_t *h, x264_mb_analysis_t *a )
{
x264_me_t m;
int i_mvc;
ALIGNED_4( int16_t mvc[8][2] );
int i_halfpel_thresh = INT_MAX;
int *p_halfpel_thresh = (a->b_early_terminate && h->mb.pic.i_fref[0]>1) ? &i_halfpel_thresh : NULL;
/* 16x16 Search on all ref frame */
m.i_pixel = PIXEL_16x16;
// 将m的YUV 3个平面的编码指针指向h->mb.pic.p_fenc的YUV 3 个平面
LOAD_FENC( &m, h->mb.pic.p_fenc, 0, 0 );
// 初始化l0 最大估计值
a->l0.me16x16.cost = INT_MAX;
// 遍历l0 参考帧
for( int i_ref = 0; i_ref < h->mb.pic.i_fref[0]; i_ref++ )
{
// 得到 l0 参考帧 i_ref的代价
m.i_ref_cost = REF_COST( 0, i_ref );
i_halfpel_thresh -= m.i_ref_cost;
/* search with ref */
// i_ref 为参考帧序号
// 装载半像素平面, 见LOAD_HPELS
LOAD_HPELS( &m, h->mb.pic.p_fref[0][i_ref], 0, i_ref, 0, 0 );
LOAD_WPELS( &m, h->mb.pic.p_fref_w[i_ref], 0, i_ref, 0, 0 );
// 用当前宏块的邻块的运动矢量作为当前宏块的预测运动矢量
// left, top, topright, or topleft
// or median of left, top and (topright or topleft)
x264_mb_predict_mv_16x16( h, 0, i_ref, m.mvp );
if( h->mb.ref_blind_dupe == i_ref )
{
CP32( m.mv, a->l0.mvc[0][0] );
x264_me_refine_qpel_refdupe( h, &m, p_halfpel_thresh );
}
else
{
// 寻找其他候选运动矢量作为候选的预测运动矢量
// 参考帧 i_ref 对应宏块的left, top, topleft, topright mb 的运动矢量
// 次采样
// 时间预测
x264_mb_predict_mv_ref16x16( h, 0, i_ref, mvc, &i_mvc );
// 搜索最佳预测运动向量
x264_me_search_ref( h, &m, mvc, i_mvc, p_halfpel_thresh );
}
// 保存最佳运动向量
/* save mv for predicting neighbors */
CP32( h->mb.mvr[0][i_ref][h->mb.i_mb_xy], m.mv );
CP32( a->l0.mvc[i_ref][0], m.mv );
/* early termination
* SSD threshold would probably be better than SATD */
if( i_ref == 0
&& a->b_try_skip
&& m.cost-m.cost_mv < 300*a->i_lambda
&& abs(m.mv[0]-h->mb.cache.pskip_mv[0])
+ abs(m.mv[1]-h->mb.cache.pskip_mv[1]) <= 1
&& x264_macroblock_probe_pskip( h ) )
{
h->mb.i_type = P_SKIP;
x264_analyse_update_cache( h, a );
assert( h->mb.cache.pskip_mv[1] <= h->mb.mv_max_spel[1] || h->i_thread_frames == 1 );
return;
}
m.cost += m.i_ref_cost;
i_halfpel_thresh += m.i_ref_cost;
if( m.cost < a->l0.me16x16.cost )
h->mc.memcpy_aligned( &a->l0.me16x16, &m, sizeof(x264_me_t) );
}
x264_macroblock_cache_ref( h, 0, 0, 4, 4, 0, a->l0.me16x16.i_ref );
assert( a->l0.me16x16.mv[1] <= h->mb.mv_max_spel[1] || h->i_thread_frames == 1 );
h->mb.i_type = P_L0;
if( a->i_mbrd )
{
x264_mb_init_fenc_cache( h, a->i_mbrd >= 2 || h->param.analyse.inter & X264_ANALYSE_PSUB8x8 );
if( a->l0.me16x16.i_ref == 0 && M32( a->l0.me16x16.mv ) == M32( h->mb.cache.pskip_mv ) && !a->b_force_intra )
{
h->mb.i_partition = D_16x16;
x264_macroblock_cache_mv_ptr( h, 0, 0, 4, 4, 0, a->l0.me16x16.mv );
a->l0.i_rd16x16 = x264_rd_cost_mb( h, a->i_lambda2 );
if( !(h->mb.i_cbp_luma|h->mb.i_cbp_chroma) )
h->mb.i_type = P_SKIP;
}
}
}
#define LOAD_HPELS(m, src, list, ref, xoff, yoff) \
{ \
// 指向整像素平面Yn
(m)->p_fref_w = (m)->p_fref[0] = &(src)[0][(xoff)+(yoff)*(m)->i_stride[0]]; \
// 指向水平半像素平面 Yh
(m)->p_fref[1] = &(src)[1][(xoff)+(yoff)*(m)->i_stride[0]]; \
// 指向垂直半像素平面 Yv
(m)->p_fref[2] = &(src)[2][(xoff)+(yoff)*(m)->i_stride[0]]; \
// 指向斜对角半像素平面 Yhv
(m)->p_fref[3] = &(src)[3][(xoff)+(yoff)*(m)->i_stride[0]]; \
if( CHROMA444 ) \
{ \
// 指向整像素平面 Un
(m)->p_fref[ 4] = &(src)[ 4][(xoff)+(yoff)*(m)->i_stride[1]]; \
// 指向水平半像素平面 Uh
(m)->p_fref[ 5] = &(src)[ 5][(xoff)+(yoff)*(m)->i_stride[1]]; \
// 指向垂直半像素平面 Uv
(m)->p_fref[ 6] = &(src)[ 6][(xoff)+(yoff)*(m)->i_stride[1]]; \
// 指向斜对角半像素平面 Uhv
(m)->p_fref[ 7] = &(src)[ 7][(xoff)+(yoff)*(m)->i_stride[1]]; \
// 指向整像素平面 Vn
(m)->p_fref[ 8] = &(src)[ 8][(xoff)+(yoff)*(m)->i_stride[2]]; \
// 指向水平半像素平面 Vh
(m)->p_fref[ 9] = &(src)[ 9][(xoff)+(yoff)*(m)->i_stride[2]]; \
// 指向垂直半像素平面 Vv
(m)->p_fref[10] = &(src)[10][(xoff)+(yoff)*(m)->i_stride[2]]; \
// 指向斜对角半像素平面 Vhv
(m)->p_fref[11] = &(src)[11][(xoff)+(yoff)*(m)->i_stride[2]]; \
} \
else \
// 指向整像素平面 UV
(m)->p_fref[4] = &(src)[4][(xoff)+((yoff)>>CHROMA_V_SHIFT)*(m)->i_stride[1]]; \
(m)->integral = &h->mb.pic.p_integral[list][ref][(xoff)+(yoff)*(m)->i_stride[0]]; \
(m)->weight = x264_weight_none; \
(m)->i_ref = ref; \
}
{
x264_me_t m;
int i_mvc;
ALIGNED_4( int16_t mvc[8][2] );
int i_halfpel_thresh = INT_MAX;
int *p_halfpel_thresh = (a->b_early_terminate && h->mb.pic.i_fref[0]>1) ? &i_halfpel_thresh : NULL;
/* 16x16 Search on all ref frame */
m.i_pixel = PIXEL_16x16;
// 将m的YUV 3个平面的编码指针指向h->mb.pic.p_fenc的YUV 3 个平面
LOAD_FENC( &m, h->mb.pic.p_fenc, 0, 0 );
// 初始化l0 最大估计值
a->l0.me16x16.cost = INT_MAX;
// 遍历l0 参考帧
for( int i_ref = 0; i_ref < h->mb.pic.i_fref[0]; i_ref++ )
{
// 得到 l0 参考帧 i_ref的代价
m.i_ref_cost = REF_COST( 0, i_ref );
i_halfpel_thresh -= m.i_ref_cost;
/* search with ref */
// i_ref 为参考帧序号
// 装载半像素平面, 见LOAD_HPELS
LOAD_HPELS( &m, h->mb.pic.p_fref[0][i_ref], 0, i_ref, 0, 0 );
LOAD_WPELS( &m, h->mb.pic.p_fref_w[i_ref], 0, i_ref, 0, 0 );
// 用当前宏块的邻块的运动矢量作为当前宏块的预测运动矢量
// left, top, topright, or topleft
// or median of left, top and (topright or topleft)
x264_mb_predict_mv_16x16( h, 0, i_ref, m.mvp );
if( h->mb.ref_blind_dupe == i_ref )
{
CP32( m.mv, a->l0.mvc[0][0] );
x264_me_refine_qpel_refdupe( h, &m, p_halfpel_thresh );
}
else
{
// 寻找其他候选运动矢量作为候选的预测运动矢量
// 参考帧 i_ref 对应宏块的left, top, topleft, topright mb 的运动矢量
// 次采样
// 时间预测
x264_mb_predict_mv_ref16x16( h, 0, i_ref, mvc, &i_mvc );
// 搜索最佳预测运动向量
x264_me_search_ref( h, &m, mvc, i_mvc, p_halfpel_thresh );
}
// 保存最佳运动向量
/* save mv for predicting neighbors */
CP32( h->mb.mvr[0][i_ref][h->mb.i_mb_xy], m.mv );
CP32( a->l0.mvc[i_ref][0], m.mv );
/* early termination
* SSD threshold would probably be better than SATD */
if( i_ref == 0
&& a->b_try_skip
&& m.cost-m.cost_mv < 300*a->i_lambda
&& abs(m.mv[0]-h->mb.cache.pskip_mv[0])
+ abs(m.mv[1]-h->mb.cache.pskip_mv[1]) <= 1
&& x264_macroblock_probe_pskip( h ) )
{
h->mb.i_type = P_SKIP;
x264_analyse_update_cache( h, a );
assert( h->mb.cache.pskip_mv[1] <= h->mb.mv_max_spel[1] || h->i_thread_frames == 1 );
return;
}
m.cost += m.i_ref_cost;
i_halfpel_thresh += m.i_ref_cost;
if( m.cost < a->l0.me16x16.cost )
h->mc.memcpy_aligned( &a->l0.me16x16, &m, sizeof(x264_me_t) );
}
x264_macroblock_cache_ref( h, 0, 0, 4, 4, 0, a->l0.me16x16.i_ref );
assert( a->l0.me16x16.mv[1] <= h->mb.mv_max_spel[1] || h->i_thread_frames == 1 );
h->mb.i_type = P_L0;
if( a->i_mbrd )
{
x264_mb_init_fenc_cache( h, a->i_mbrd >= 2 || h->param.analyse.inter & X264_ANALYSE_PSUB8x8 );
if( a->l0.me16x16.i_ref == 0 && M32( a->l0.me16x16.mv ) == M32( h->mb.cache.pskip_mv ) && !a->b_force_intra )
{
h->mb.i_partition = D_16x16;
x264_macroblock_cache_mv_ptr( h, 0, 0, 4, 4, 0, a->l0.me16x16.mv );
a->l0.i_rd16x16 = x264_rd_cost_mb( h, a->i_lambda2 );
if( !(h->mb.i_cbp_luma|h->mb.i_cbp_chroma) )
h->mb.i_type = P_SKIP;
}
}
}
#define LOAD_HPELS(m, src, list, ref, xoff, yoff) \
{ \
// 指向整像素平面Yn
(m)->p_fref_w = (m)->p_fref[0] = &(src)[0][(xoff)+(yoff)*(m)->i_stride[0]]; \
// 指向水平半像素平面 Yh
(m)->p_fref[1] = &(src)[1][(xoff)+(yoff)*(m)->i_stride[0]]; \
// 指向垂直半像素平面 Yv
(m)->p_fref[2] = &(src)[2][(xoff)+(yoff)*(m)->i_stride[0]]; \
// 指向斜对角半像素平面 Yhv
(m)->p_fref[3] = &(src)[3][(xoff)+(yoff)*(m)->i_stride[0]]; \
if( CHROMA444 ) \
{ \
// 指向整像素平面 Un
(m)->p_fref[ 4] = &(src)[ 4][(xoff)+(yoff)*(m)->i_stride[1]]; \
// 指向水平半像素平面 Uh
(m)->p_fref[ 5] = &(src)[ 5][(xoff)+(yoff)*(m)->i_stride[1]]; \
// 指向垂直半像素平面 Uv
(m)->p_fref[ 6] = &(src)[ 6][(xoff)+(yoff)*(m)->i_stride[1]]; \
// 指向斜对角半像素平面 Uhv
(m)->p_fref[ 7] = &(src)[ 7][(xoff)+(yoff)*(m)->i_stride[1]]; \
// 指向整像素平面 Vn
(m)->p_fref[ 8] = &(src)[ 8][(xoff)+(yoff)*(m)->i_stride[2]]; \
// 指向水平半像素平面 Vh
(m)->p_fref[ 9] = &(src)[ 9][(xoff)+(yoff)*(m)->i_stride[2]]; \
// 指向垂直半像素平面 Vv
(m)->p_fref[10] = &(src)[10][(xoff)+(yoff)*(m)->i_stride[2]]; \
// 指向斜对角半像素平面 Vhv
(m)->p_fref[11] = &(src)[11][(xoff)+(yoff)*(m)->i_stride[2]]; \
} \
else \
// 指向整像素平面 UV
(m)->p_fref[4] = &(src)[4][(xoff)+((yoff)>>CHROMA_V_SHIFT)*(m)->i_stride[1]]; \
(m)->integral = &h->mb.pic.p_integral[list][ref][(xoff)+(yoff)*(m)->i_stride[0]]; \
(m)->weight = x264_weight_none; \
(m)->i_ref = ref; \
}
相关文章推荐
- x264代码剖析(十三):核心算法之帧间预测函数x264_mb_analyse_inter_*()
- x264代码剖析(十三):核心算法之帧间预测函数x264_mb_analyse_inter_*()
- x264代码剖析(十三):核心算法之帧间预測函数x264_mb_analyse_inter_*()
- x264 - x264_mb_analyse_inter_b16x16
- x264代码剖析(十二):核心算法之帧内预测函数x264_mb_analyse_intra()
- x264代码剖析(十二):核心算法之帧内预测函数x264_mb_analyse_intra()
- X264 rev1198 MB-Tree Ratecontrol
- x264代码追踪之x264_macroblock_analyse模块分析
- x264 - x264.mb.i_neighbour8
- x264源码分析 -- x264_slicetype_mb_cost
- X264代码跟踪之x264_ratecontrol_mb
- x264_macroblock_analyse(x264_t *h ) 分析
- x264_macroblock_analyse(x264_t *h ) 分析
- x264 - x264_mb_predict_intra4x4_mode
- x264 - x264_ratecontrol_mb
- x264_analyse_update_cache 内部各种函数注释
- x264源代码简单分析:宏块分析(Analysis)部分-帧间宏块(Inter)
- x264_mb_predict_mv_16x16 函数分析
- X264中x264_macroblock_analyse模块分析
- x264_mb_predict_mv_16x16