/*****************************************************************************
* mc.c: motion compensation
*****************************************************************************
- * Copyright (C) 2003-2010 x264 project
+ * Copyright (C) 2003-2014 x264 project
*
* Authors: Laurent Aimar <fenrir@via.ecp.fr>
* Loren Merritt <lorenm@u.washington.edu>
#endif
-static inline void pixel_avg( pixel *dst, int i_dst_stride,
- pixel *src1, int i_src1_stride,
- pixel *src2, int i_src2_stride,
- int i_width, int i_height )
+static inline void pixel_avg( pixel *dst, intptr_t i_dst_stride,
+ pixel *src1, intptr_t i_src1_stride,
+ pixel *src2, intptr_t i_src2_stride, int i_width, int i_height )
{
for( int y = 0; y < i_height; y++ )
{
}
}
-static inline void pixel_avg_wxh( pixel *dst, int i_dst, pixel *src1, int i_src1, pixel *src2, int i_src2, int width, int height )
+static inline void pixel_avg_wxh( pixel *dst, intptr_t i_dst,
+ pixel *src1, intptr_t i_src1,
+ pixel *src2, intptr_t i_src2, int width, int height )
{
for( int y = 0; y < height; y++ )
{
/* Implicit weighted bipred only:
* assumes log2_denom = 5, offset = 0, weight1 + weight2 = 64 */
-static inline void pixel_avg_weight_wxh( pixel *dst, int i_dst, pixel *src1, int i_src1, pixel *src2, int i_src2, int width, int height, int i_weight1 )
+static inline void pixel_avg_weight_wxh( pixel *dst, intptr_t i_dst,
+ pixel *src1, intptr_t i_src1,
+ pixel *src2, intptr_t i_src2, int width, int height, int i_weight1 )
{
- const int i_weight2 = 64 - i_weight1;
+ int i_weight2 = 64 - i_weight1;
for( int y = 0; y<height; y++, dst += i_dst, src1 += i_src1, src2 += i_src2 )
for( int x = 0; x<width; x++ )
dst[x] = x264_clip_pixel( (src1[x]*i_weight1 + src2[x]*i_weight2 + (1<<5)) >> 6 );
#undef op_scale2
#define PIXEL_AVG_C( name, width, height ) \
-static void name( pixel *pix1, int i_stride_pix1, \
- pixel *pix2, int i_stride_pix2, \
- pixel *pix3, int i_stride_pix3, int weight ) \
+static void name( pixel *pix1, intptr_t i_stride_pix1, \
+ pixel *pix2, intptr_t i_stride_pix2, \
+ pixel *pix3, intptr_t i_stride_pix3, int weight ) \
{ \
if( weight == 32 ) \
pixel_avg_wxh( pix1, i_stride_pix1, pix2, i_stride_pix2, pix3, i_stride_pix3, width, height ); \
PIXEL_AVG_C( pixel_avg_8x16, 8, 16 )
PIXEL_AVG_C( pixel_avg_8x8, 8, 8 )
PIXEL_AVG_C( pixel_avg_8x4, 8, 4 )
+PIXEL_AVG_C( pixel_avg_4x16, 4, 16 )
PIXEL_AVG_C( pixel_avg_4x8, 4, 8 )
PIXEL_AVG_C( pixel_avg_4x4, 4, 4 )
PIXEL_AVG_C( pixel_avg_4x2, 4, 2 )
+PIXEL_AVG_C( pixel_avg_2x8, 2, 8 )
PIXEL_AVG_C( pixel_avg_2x4, 2, 4 )
PIXEL_AVG_C( pixel_avg_2x2, 2, 2 )
}
#define opscale(x) dst[x] = x264_clip_pixel( ((src[x] * scale + (1<<(denom - 1))) >> denom) + offset )
#define opscale_noden(x) dst[x] = x264_clip_pixel( src[x] * scale + offset )
-static void mc_weight( pixel *dst, int i_dst_stride, pixel *src, int i_src_stride, const x264_weight_t *weight, int i_width, int i_height )
+static void mc_weight( pixel *dst, intptr_t i_dst_stride, pixel *src, intptr_t i_src_stride,
+ const x264_weight_t *weight, int i_width, int i_height )
{
int offset = weight->i_offset << (BIT_DEPTH-8);
int scale = weight->i_scale;
}
#define MC_WEIGHT_C( name, width ) \
- static void name( pixel *dst, int i_dst_stride, pixel *src, int i_src_stride, const x264_weight_t *weight, int height ) \
+ static void name( pixel *dst, intptr_t i_dst_stride, pixel *src, intptr_t i_src_stride, const x264_weight_t *weight, int height ) \
{ \
mc_weight( dst, i_dst_stride, src, i_src_stride, weight, width, height );\
}
mc_weight_w16,
mc_weight_w20,
};
-const x264_weight_t weight_none[3] = { {{0}} };
-static void mc_copy( pixel *src, int i_src_stride, pixel *dst, int i_dst_stride, int i_width, int i_height )
+const x264_weight_t x264_weight_none[3] = { {{0}} };
+static void mc_copy( pixel *src, intptr_t i_src_stride, pixel *dst, intptr_t i_dst_stride, int i_width, int i_height )
{
for( int y = 0; y < i_height; y++ )
{
#define TAPFILTER(pix, d) ((pix)[x-2*d] + (pix)[x+3*d] - 5*((pix)[x-d] + (pix)[x+2*d]) + 20*((pix)[x] + (pix)[x+d]))
static void hpel_filter( pixel *dsth, pixel *dstv, pixel *dstc, pixel *src,
- int stride, int width, int height, int16_t *buf )
+ intptr_t stride, int width, int height, int16_t *buf )
{
const int pad = (BIT_DEPTH > 9) ? (-10 * PIXEL_MAX) : 0;
for( int y = 0; y < height; y++ )
static const uint8_t hpel_ref0[16] = {0,1,1,1,0,1,1,1,2,3,3,3,0,1,1,1};
static const uint8_t hpel_ref1[16] = {0,0,0,0,2,2,3,2,2,2,3,2,2,2,3,2};
-static void mc_luma( pixel *dst, int i_dst_stride,
- pixel *src[4], int i_src_stride,
+static void mc_luma( pixel *dst, intptr_t i_dst_stride,
+ pixel *src[4], intptr_t i_src_stride,
int mvx, int mvy,
int i_width, int i_height, const x264_weight_t *weight )
{
mc_copy( src1, i_src_stride, dst, i_dst_stride, i_width, i_height );
}
-static pixel *get_ref( pixel *dst, int *i_dst_stride,
- pixel *src[4], int i_src_stride,
+static pixel *get_ref( pixel *dst, intptr_t *i_dst_stride,
+ pixel *src[4], intptr_t i_src_stride,
int mvx, int mvy,
int i_width, int i_height, const x264_weight_t *weight )
{
}
/* full chroma mc (ie until 1/8 pixel)*/
-static void mc_chroma( pixel *dstu, pixel *dstv, int i_dst_stride,
- pixel *src, int i_src_stride,
+static void mc_chroma( pixel *dstu, pixel *dstv, intptr_t i_dst_stride,
+ pixel *src, intptr_t i_src_stride,
int mvx, int mvy,
int i_width, int i_height )
{
}
#define MC_COPY(W) \
-static void mc_copy_w##W( pixel *dst, int i_dst, pixel *src, int i_src, int i_height ) \
+static void mc_copy_w##W( pixel *dst, intptr_t i_dst, pixel *src, intptr_t i_src, int i_height ) \
{ \
mc_copy( src, i_src, dst, i_dst, W, i_height ); \
}
MC_COPY( 8 )
MC_COPY( 4 )
-void x264_plane_copy_c( pixel *dst, int i_dst,
- pixel *src, int i_src, int w, int h )
+void x264_plane_copy_c( pixel *dst, intptr_t i_dst,
+ pixel *src, intptr_t i_src, int w, int h )
{
while( h-- )
{
}
}
-void x264_plane_copy_interleave_c( pixel *dst, int i_dst,
- pixel *srcu, int i_srcu,
- pixel *srcv, int i_srcv, int w, int h )
+void x264_plane_copy_interleave_c( pixel *dst, intptr_t i_dst,
+ pixel *srcu, intptr_t i_srcu,
+ pixel *srcv, intptr_t i_srcv, int w, int h )
{
for( int y=0; y<h; y++, dst+=i_dst, srcu+=i_srcu, srcv+=i_srcv )
for( int x=0; x<w; x++ )
}
}
-void x264_plane_copy_deinterleave_c( pixel *dstu, int i_dstu,
- pixel *dstv, int i_dstv,
- pixel *src, int i_src, int w, int h )
+static void x264_plane_copy_deinterleave_c( pixel *dstu, intptr_t i_dstu,
+ pixel *dstv, intptr_t i_dstv,
+ pixel *src, intptr_t i_src, int w, int h )
{
for( int y=0; y<h; y++, dstu+=i_dstu, dstv+=i_dstv, src+=i_src )
for( int x=0; x<w; x++ )
}
}
-static void store_interleave_8x8x2( pixel *dst, int i_dst, pixel *srcu, pixel *srcv )
+static void x264_plane_copy_deinterleave_rgb_c( pixel *dsta, intptr_t i_dsta,
+ pixel *dstb, intptr_t i_dstb,
+ pixel *dstc, intptr_t i_dstc,
+ pixel *src, intptr_t i_src, int pw, int w, int h )
{
- for( int y=0; y<8; y++, dst+=i_dst, srcu+=FDEC_STRIDE, srcv+=FDEC_STRIDE )
+ for( int y=0; y<h; y++, dsta+=i_dsta, dstb+=i_dstb, dstc+=i_dstc, src+=i_src )
+ {
+ for( int x=0; x<w; x++ )
+ {
+ dsta[x] = src[x*pw];
+ dstb[x] = src[x*pw+1];
+ dstc[x] = src[x*pw+2];
+ }
+ }
+}
+
+void x264_plane_copy_deinterleave_v210_c( pixel *dsty, intptr_t i_dsty,
+ pixel *dstc, intptr_t i_dstc,
+ uint32_t *src, intptr_t i_src, int w, int h )
+{
+ for( int l = 0; l < h; l++ )
+ {
+ pixel *dsty0 = dsty;
+ pixel *dstc0 = dstc;
+ uint32_t *src0 = src;
+
+ for( int n = 0; n < w; n += 3 )
+ {
+ *(dstc0++) = *src0 & 0x03FF;
+ *(dsty0++) = ( *src0 >> 10 ) & 0x03FF;
+ *(dstc0++) = ( *src0 >> 20 ) & 0x03FF;
+ src0++;
+ *(dsty0++) = *src0 & 0x03FF;
+ *(dstc0++) = ( *src0 >> 10 ) & 0x03FF;
+ *(dsty0++) = ( *src0 >> 20 ) & 0x03FF;
+ src0++;
+ }
+
+ dsty += i_dsty;
+ dstc += i_dstc;
+ src += i_src;
+ }
+}
+
+static void store_interleave_chroma( pixel *dst, intptr_t i_dst, pixel *srcu, pixel *srcv, int height )
+{
+ for( int y=0; y<height; y++, dst+=i_dst, srcu+=FDEC_STRIDE, srcv+=FDEC_STRIDE )
for( int x=0; x<8; x++ )
{
dst[2*x] = srcu[x];
}
}
-static void load_deinterleave_8x8x2_fenc( pixel *dst, pixel *src, int i_src )
+static void load_deinterleave_chroma_fenc( pixel *dst, pixel *src, intptr_t i_src, int height )
{
- x264_plane_copy_deinterleave_c( dst, FENC_STRIDE, dst+FENC_STRIDE/2, FENC_STRIDE, src, i_src, 8, 8 );
+ x264_plane_copy_deinterleave_c( dst, FENC_STRIDE, dst+FENC_STRIDE/2, FENC_STRIDE, src, i_src, 8, height );
}
-static void load_deinterleave_8x8x2_fdec( pixel *dst, pixel *src, int i_src )
+static void load_deinterleave_chroma_fdec( pixel *dst, pixel *src, intptr_t i_src, int height )
{
- x264_plane_copy_deinterleave_c( dst, FDEC_STRIDE, dst+FDEC_STRIDE/2, FDEC_STRIDE, src, i_src, 8, 8 );
+ x264_plane_copy_deinterleave_c( dst, FDEC_STRIDE, dst+FDEC_STRIDE/2, FDEC_STRIDE, src, i_src, 8, height );
}
-static void prefetch_fenc_null( pixel *pix_y, int stride_y,
- pixel *pix_uv, int stride_uv, int mb_x )
+static void prefetch_fenc_null( pixel *pix_y, intptr_t stride_y,
+ pixel *pix_uv, intptr_t stride_uv, int mb_x )
{}
-static void prefetch_ref_null( pixel *pix, int stride, int parity )
+static void prefetch_ref_null( pixel *pix, intptr_t stride, int parity )
{}
-static void memzero_aligned( void * dst, int n )
+static void memzero_aligned( void * dst, size_t n )
{
memset( dst, 0, n );
}
-static void integral_init4h( uint16_t *sum, pixel *pix, int stride )
+static void integral_init4h( uint16_t *sum, pixel *pix, intptr_t stride )
{
int v = pix[0]+pix[1]+pix[2]+pix[3];
for( int x = 0; x < stride-4; x++ )
}
}
-static void integral_init8h( uint16_t *sum, pixel *pix, int stride )
+static void integral_init8h( uint16_t *sum, pixel *pix, intptr_t stride )
{
int v = pix[0]+pix[1]+pix[2]+pix[3]+pix[4]+pix[5]+pix[6]+pix[7];
for( int x = 0; x < stride-8; x++ )
}
}
-static void integral_init4v( uint16_t *sum8, uint16_t *sum4, int stride )
+static void integral_init4v( uint16_t *sum8, uint16_t *sum4, intptr_t stride )
{
for( int x = 0; x < stride-8; x++ )
sum4[x] = sum8[x+4*stride] - sum8[x];
sum8[x] = sum8[x+8*stride] + sum8[x+8*stride+4] - sum8[x] - sum8[x+4];
}
-static void integral_init8v( uint16_t *sum8, int stride )
+static void integral_init8v( uint16_t *sum8, intptr_t stride )
{
for( int x = 0; x < stride-8; x++ )
sum8[x] = sum8[x+8*stride] - sum8[x];
}
static void frame_init_lowres_core( pixel *src0, pixel *dst0, pixel *dsth, pixel *dstv, pixel *dstc,
- int src_stride, int dst_stride, int width, int height )
+ intptr_t src_stride, intptr_t dst_stride, int width, int height )
{
for( int y = 0; y < height; y++ )
{
}
}
-void x264_mc_init( int cpu, x264_mc_functions_t *pf )
+void x264_mc_init( int cpu, x264_mc_functions_t *pf, int cpu_independent )
{
pf->mc_luma = mc_luma;
pf->get_ref = get_ref;
+
pf->mc_chroma = mc_chroma;
pf->avg[PIXEL_16x16]= pixel_avg_16x16;
pf->avg[PIXEL_8x16] = pixel_avg_8x16;
pf->avg[PIXEL_8x8] = pixel_avg_8x8;
pf->avg[PIXEL_8x4] = pixel_avg_8x4;
+ pf->avg[PIXEL_4x16] = pixel_avg_4x16;
pf->avg[PIXEL_4x8] = pixel_avg_4x8;
pf->avg[PIXEL_4x4] = pixel_avg_4x4;
pf->avg[PIXEL_4x2] = pixel_avg_4x2;
+ pf->avg[PIXEL_2x8] = pixel_avg_2x8;
pf->avg[PIXEL_2x4] = pixel_avg_2x4;
pf->avg[PIXEL_2x2] = pixel_avg_2x2;
pf->copy[PIXEL_8x8] = mc_copy_w8;
pf->copy[PIXEL_4x4] = mc_copy_w4;
- pf->store_interleave_8x8x2 = store_interleave_8x8x2;
- pf->load_deinterleave_8x8x2_fenc = load_deinterleave_8x8x2_fenc;
- pf->load_deinterleave_8x8x2_fdec = load_deinterleave_8x8x2_fdec;
+ pf->store_interleave_chroma = store_interleave_chroma;
+ pf->load_deinterleave_chroma_fenc = load_deinterleave_chroma_fenc;
+ pf->load_deinterleave_chroma_fdec = load_deinterleave_chroma_fdec;
pf->plane_copy = x264_plane_copy_c;
pf->plane_copy_interleave = x264_plane_copy_interleave_c;
pf->plane_copy_deinterleave = x264_plane_copy_deinterleave_c;
+ pf->plane_copy_deinterleave_rgb = x264_plane_copy_deinterleave_rgb_c;
+ pf->plane_copy_deinterleave_v210 = x264_plane_copy_deinterleave_v210_c;
pf->hpel_filter = hpel_filter;
- pf->prefetch_fenc = prefetch_fenc_null;
+ pf->prefetch_fenc_420 = prefetch_fenc_null;
+ pf->prefetch_fenc_422 = prefetch_fenc_null;
pf->prefetch_ref = prefetch_ref_null;
pf->memcpy_aligned = memcpy;
pf->memzero_aligned = memzero_aligned;
#if HAVE_ARMV6
x264_mc_init_arm( cpu, pf );
#endif
+
+ if( cpu_independent )
+ pf->mbtree_propagate_cost = mbtree_propagate_cost;
}
void x264_frame_filter( x264_t *h, x264_frame_t *frame, int mb_y, int b_end )
{
- const int b_interlaced = h->sh.b_mbaff;
- const int stride = frame->i_stride[0] << b_interlaced;
- const int width = frame->i_width[0];
- int start = (mb_y*16 >> b_interlaced) - 8; // buffer = 4 for deblock + 3 for 6tap, rounded to 8
- int height = ((b_end ? frame->i_lines[0] : mb_y*16) >> b_interlaced) + 8;
- int offs = start*stride - 8; // buffer = 3 for 6tap, aligned to 8 for simd
+ const int b_interlaced = PARAM_INTERLACED;
+ int start = mb_y*16 - 8; // buffer = 4 for deblock + 3 for 6tap, rounded to 8
+ int height = (b_end ? frame->i_lines[0] + 16*PARAM_INTERLACED : (mb_y+b_interlaced)*16) + 8;
if( mb_y & b_interlaced )
return;
- for( int y = 0; y <= b_interlaced; y++, offs += frame->i_stride[0] )
+ for( int p = 0; p < (CHROMA444 ? 3 : 1); p++ )
{
- h->mc.hpel_filter(
- frame->filtered[1] + offs,
- frame->filtered[2] + offs,
- frame->filtered[3] + offs,
- frame->plane[0] + offs,
- stride, width + 16, height - start,
- h->scratch_buffer );
+ int stride = frame->i_stride[p];
+ const int width = frame->i_width[p];
+ int offs = start*stride - 8; // buffer = 3 for 6tap, aligned to 8 for simd
+
+ if( !b_interlaced || h->mb.b_adaptive_mbaff )
+ h->mc.hpel_filter(
+ frame->filtered[p][1] + offs,
+ frame->filtered[p][2] + offs,
+ frame->filtered[p][3] + offs,
+ frame->plane[p] + offs,
+ stride, width + 16, height - start,
+ h->scratch_buffer );
+
+ if( b_interlaced )
+ {
+ /* MC must happen between pixels in the same field. */
+ stride = frame->i_stride[p] << 1;
+ start = (mb_y*16 >> 1) - 8;
+ int height_fld = ((b_end ? frame->i_lines[p] : mb_y*16) >> 1) + 8;
+ offs = start*stride - 8;
+ for( int i = 0; i < 2; i++, offs += frame->i_stride[p] )
+ {
+ h->mc.hpel_filter(
+ frame->filtered_fld[p][1] + offs,
+ frame->filtered_fld[p][2] + offs,
+ frame->filtered_fld[p][3] + offs,
+ frame->plane_fld[p] + offs,
+ stride, width + 16, height_fld - start,
+ h->scratch_buffer );
+ }
+ }
}
/* generate integral image:
if( frame->integral )
{
+ int stride = frame->i_stride[0];
if( start < 0 )
{
memset( frame->integral - PADV * stride - PADH, 0, stride * sizeof(uint16_t) );
projects / x264 / blobdiff