int stride = 16; // candidates are either all hpel or all qpel, so one stride is enough        uint8_t *src0, *src1, *src2, *src3;        src0 = h->mc.get_ref( m->p_fref, m->i_stride[0], pix[0], &stride, omx, omy-2, bw, bh );        src2 = h->mc.get_ref( m->p_fref, m->i_stride[0], pix[2], &stride, omx-2, omy, bw, bh );        if( (omx|omy)&1 )        {            src1 = h->mc.get_ref( m->p_fref, m->i_stride[0], pix[1], &stride, omx, omy+2, bw, bh );            src3 = h->mc.get_ref( m->p_fref, m->i_stride[0], pix[3], &stride, omx+2, omy, bw, bh );        }        else        {            src1 = src0 + stride;            src3 = src2 + 1;        }        h->pixf.sad_x4[i_pixel]( m->p_fenc[0], src0, src1, src2, src3, stride, costs );        COPY2_IF_LT( bcost, costs[0] + p_cost_mvx[omx  ] + p_cost_mvy[omy-2], bmy, omy-2 );        COPY2_IF_LT( bcost, costs[1] + p_cost_mvx[omx  ] + p_cost_mvy[omy+2], bmy, omy+2 );        COPY3_IF_LT( bcost, costs[2] + p_cost_mvx[omx-2] + p_cost_mvy[omy  ], bmx, omx-2, bmy, omy );        COPY3_IF_LT( bcost, costs[3] + p_cost_mvx[omx+2] + p_cost_mvy[omy  ], bmx, omx+2, bmy, omy );        if( bmx == omx && bmy == omy )            break;    }    if( !b_refine_qpel )    {        bcost = COST_MAX;        COST_MV_SATD( bmx, bmy, -1 );    }    /* early termination when examining multiple reference frames */    if( p_halfpel_thresh )    {        if( (bcost*7)>>3 > *p_halfpel_thresh )        {            m->cost = bcost;            m->mv[0] = bmx;            m->mv[1] = bmy;            // don't need cost_mv            return;        }        else if( bcost < *p_halfpel_thresh )            *p_halfpel_thresh = bcost;    }    /* quarterpel diamond search */    bdir = -1;    for( i = qpel_iters; i > 0; i-- )    {        odir = bdir;        omx = bmx;        omy = bmy;        COST_MV_SATD( omx, omy - 1, 0 );        COST_MV_SATD( omx, omy + 1, 1 );        COST_MV_SATD( omx - 1, omy, 2 );        COST_MV_SATD( omx + 1, omy, 3 );        if( bmx == omx && bmy == omy )            break;    }    m->cost = bcost;    m->mv[0] = bmx;    m->mv[1] = bmy;    m->cost_mv = p_cost_mvx[ bmx ] + p_cost_mvy[ bmy ];}#define BIME_CACHE( dx, dy ) \{ \    int i = 4 + 3*dx + dy; \    h->mc.mc_luma( m0->p_fref, m0->i_stride[0], pix0[i], bw, om0x+dx, om0y+dy, bw, bh ); \    h->mc.mc_luma( m1->p_fref, m1->i_stride[0], pix1[i], bw, om1x+dx, om1y+dy, bw, bh ); \}#define BIME_CACHE2(a,b) \    BIME_CACHE(a,b) \    BIME_CACHE(-(a),-(b))#define COST_BIMV_SATD( m0x, m0y, m1x, m1y ) \if( pass == 0 || !visited[(m0x)&7][(m0y)&7][(m1x)&7][(m1y)&7] ) \{ \    int cost; \    int i0 = 4 + 3*(m0x-om0x) + (m0y-om0y); \    int i1 = 4 + 3*(m1x-om1x) + (m1y-om1y); \    visited[(m0x)&7][(m0y)&7][(m1x)&7][(m1y)&7] = 1; \    memcpy( pix, pix0[i0], bs ); \    if( i_weight == 32 ) \        h->mc.avg[i_pixel]( pix, bw, pix1[i1], bw ); \    else \        h->mc.avg_weight[i_pixel]( pix, bw, pix1[i1], bw, i_weight ); \    cost = h->pixf.mbcmp[i_pixel]( m0->p_fenc[0], FENC_STRIDE, pix, bw ) \         + p_cost_m0x[ m0x ] + p_cost_m0y[ m0y ] \         + p_cost_m1x[ m1x ] + p_cost_m1y[ m1y ]; \    if( cost < bcost ) \    {                  \        bcost = cost;  \        bm0x = m0x;    \        bm0y = m0y;    \        bm1x = m1x;    \        bm1y = m1y;    \    } \}#define CHECK_BIDIR(a,b,c,d) \    COST_BIMV_SATD(om0x+a, om0y+b, om1x+c, om1y+d)#define CHECK_BIDIR2(a,b,c,d) \    CHECK_BIDIR(a,b,c,d) \    CHECK_BIDIR(-(a),-(b),-(c),-(d))#define CHECK_BIDIR8(a,b,c,d) \    CHECK_BIDIR2(a,b,c,d) \    CHECK_BIDIR2(b,c,d,a) \    CHECK_BIDIR2(c,d,a,b) \    CHECK_BIDIR2(d,a,b,c)int x264_me_refine_bidir( x264_t *h, x264_me_t *m0, x264_me_t *m1, int i_weight ){    const int i_pixel = m0->i_pixel;    const int bw = x264_pixel_size[i_pixel].w;    const int bh = x264_pixel_size[i_pixel].h;    const int bs = bw*bh;    const int16_t *p_cost_m0x = m0->p_cost_mv - x264_clip3( m0->mvp[0], h->mb.mv_min_spel[0], h->mb.mv_max_spel[0] );    const int16_t *p_cost_m0y = m0->p_cost_mv - x264_clip3( m0->mvp[1], h->mb.mv_min_spel[0], h->mb.mv_max_spel[0] );    const int16_t *p_cost_m1x = m1->p_cost_mv - x264_clip3( m1->mvp[0], h->mb.mv_min_spel[0], h->mb.mv_max_spel[0] );    const int16_t *p_cost_m1y = m1->p_cost_mv - x264_clip3( m1->mvp[1], h->mb.mv_min_spel[0], h->mb.mv_max_spel[0] );    DECLARE_ALIGNED( uint8_t, pix0[9][16*16], 16 );    DECLARE_ALIGNED( uint8_t, pix1[9][16*16], 16 );    DECLARE_ALIGNED( uint8_t, pix[16*16], 16 );    int bm0x = m0->mv[0], om0x = bm0x;    int bm0y = m0->mv[1], om0y = bm0y;    int bm1x = m1->mv[0], om1x = bm1x;    int bm1y = m1->mv[1], om1y = bm1y;    int bcost = COST_MAX;    int pass = 0;    uint8_t visited[8][8][8][8];    memset( visited, 0, sizeof(visited) );    BIME_CACHE( 0, 0 );    CHECK_BIDIR( 0, 0, 0, 0 );    for( pass = 0; pass < 8; pass++ )    {        /* check all mv pairs that differ in at most 2 components from the current mvs. */        /* doesn't do chroma ME. this probably doesn't matter, as the gains         * from bidir ME are the same with and without chroma ME. */        BIME_CACHE2( 1, 0 );        BIME_CACHE2( 0, 1 );        BIME_CACHE2( 1, 1 );        BIME_CACHE2( 1,-1 );        CHECK_BIDIR8( 0, 0, 0, 1 );        CHECK_BIDIR8( 0, 0, 1, 1 );        CHECK_BIDIR2( 0, 1, 0, 1 );        CHECK_BIDIR2( 1, 0, 1, 0 );        CHECK_BIDIR8( 0, 0,-1, 1 );        CHECK_BIDIR2( 0,-1, 0, 1 );        CHECK_BIDIR2(-1, 0, 1, 0 );        if( om0x == bm0x && om0y == bm0y && om1x == bm1x && om1y == bm1y )            break;        om0x = bm0x;        om0y = bm0y;        om1x = bm1x;        om1y = bm1y;        BIME_CACHE( 0, 0 );    }    m0->mv[0] = bm0x;    m0->mv[1] = bm0y;    m1->mv[0] = bm1x;    m1->mv[1] = bm1y;    return bcost;}#undef COST_MV_SATD#define COST_MV_SATD( mx, my, dst ) \{ \    int stride = 16; \    uint8_t *src = h->mc.get_ref( m->p_fref, m->i_stride[0], pix, &stride, mx, my, bw*4, bh*4 ); \    dst = h->pixf.mbcmp[i_pixel]( m->p_fenc[0], FENC_STRIDE, src, stride ) \        + p_cost_mvx[mx] + p_cost_mvy[my]; \    COPY1_IF_LT( bsatd, dst ); \}#define COST_MV_RD( mx, my, satd, dir ) \{ \    if( satd <= bsatd * SATD_THRESH \        && (dir^1) != odir \        && (dir<0 || !p_visited[(mx)+(my)*16]) ) \    { \        int cost; \        cache_mv[0] = cache_mv2[0] = mx; \        cache_mv[1] = cache_mv2[1] = my; \        cost = x264_rd_cost_part( h, i_lambda2, i8, m->i_pixel ); \        COPY3_IF_LT( bcost, cost, bmx, mx, bmy, my ); \        if(dir>=0) p_visited[(mx)+(my)*16] = 1; \    } \}#define SATD_THRESH 17/16void x264_me_refine_qpel_rd( x264_t *h, x264_me_t *m, int i_lambda2, int i8 ){    // don't have to fill the whole mv cache rectangle    static const int pixel_mv_offs[] = { 0, 4, 4*8, 0 };    int16_t *cache_mv = h->mb.cache.mv[0][x264_scan8[i8*4]];    int16_t *cache_mv2 = cache_mv + pixel_mv_offs[m->i_pixel];    const int16_t *p_cost_mvx, *p_cost_mvy;    const int bw = x264_pixel_size[m->i_pixel].w>>2;    const int bh = x264_pixel_size[m->i_pixel].h>>2;    const int i_pixel = m->i_pixel;    DECLARE_ALIGNED( uint8_t, pix[16*16], 16 );    int bcost = m->i_pixel == PIXEL_16x16 ? m->cost : COST_MAX;    int bmx = m->mv[0];    int bmy = m->mv[1];    int pmx, pmy, omx, omy, i;    int odir = -1, bdir;    unsigned bsatd, satds[4];    int visited[16*13] = {0}; // only need 13x13, but 16 is more convenient    int *p_visited = &visited[6+6*16];    if( m->i_pixel != PIXEL_16x16 && i8 != 0 )        x264_mb_predict_mv( h, 0, i8*4, bw, m->mvp );    pmx = m->mvp[0];    pmy = m->mvp[1];    p_cost_mvx = m->p_cost_mv - pmx;    p_cost_mvy = m->p_cost_mv - pmy;    COST_MV_SATD( bmx, bmy, bsatd );    if( m->i_pixel != PIXEL_16x16 )        COST_MV_RD( bmx, bmy, 0, -1 );    /* check the predicted mv */    if( (bmx != pmx || bmy != pmy)        && pmx >= h->mb.mv_min_spel[0] && pmx <= h->mb.mv_max_spel[0]        && pmy >= h->mb.mv_min_spel[1] && pmy <= h->mb.mv_max_spel[1] )    {        int satd;        COST_MV_SATD( pmx, pmy, satd );        COST_MV_RD( pmx, pmy, satd, -1 );    }    /* mark mv and mvp as visited */    p_visited[0] = 1;    p_visited -= bmx + bmy*16;    {        int mx = bmx ^ m->mv[0] ^ pmx;        int my = bmy ^ m->mv[1] ^ pmy;        if( abs(mx-bmx) < 7 && abs(my-bmy) < 7 )            p_visited[mx + my*16] = 1;    }    /* hpel diamond */    bdir = -1;    for( i = 0; i < 2; i++ )    {         omx = bmx;         omy = bmy;         odir = bdir;         COST_MV_SATD( omx, omy - 2, satds[0] );         COST_MV_SATD( omx, omy + 2, satds[1] );         COST_MV_SATD( omx - 2, omy, satds[2] );         COST_MV_SATD( omx + 2, omy, satds[3] );         COST_MV_RD( omx, omy - 2, satds[0], 0 );         COST_MV_RD( omx, omy + 2, satds[1], 1 );         COST_MV_RD( omx - 2, omy, satds[2], 2 );         COST_MV_RD( omx + 2, omy, satds[3], 3 );         if( bmx == omx && bmy == omy )            break;    }    /* qpel diamond */    bdir = -1;    for( i = 0; i < 2; i++ )    {         omx = bmx;         omy = bmy;         odir = bdir;         COST_MV_SATD( omx, omy - 1, satds[0] );         COST_MV_SATD( omx, omy + 1, satds[1] );         COST_MV_SATD( omx - 1, omy, satds[2] );         COST_MV_SATD( omx + 1, omy, satds[3] );         COST_MV_RD( omx, omy - 1, satds[0], 0 );         COST_MV_RD( omx, omy + 1, satds[1], 1 );         COST_MV_RD( omx - 1, omy, satds[2], 2 );         COST_MV_RD( omx + 1, omy, satds[3], 3 );         if( bmx == omx && bmy == omy )            break;    }    m->cost = bcost;    m->mv[0] = bmx;    m->mv[1] = bmy;    x264_macroblock_cache_mv ( h, 2*(i8&1), i8&2, bw, bh, 0, bmx, bmy );    x264_macroblock_cache_mvd( h, 2*(i8&1), i8&2, bw, bh, 0, bmx - pmx, bmy - pmy );}