[FFmpeg-devel] [RFC] RV30 1/3pel MC functions

Kostya kostya.shishkov
Thu Jan 3 13:09:15 CET 2008


On Wed, Jan 02, 2008 at 11:12:55PM +0100, Michael Niedermayer wrote:
> On Mon, Dec 31, 2007 at 09:17:59AM +0200, Kostya wrote:
> > On Wed, Dec 26, 2007 at 03:38:35AM +0100, Michael Niedermayer wrote:
> > > On Mon, Dec 24, 2007 at 01:04:00PM +0200, Kostya wrote:
> > > [...]
> > > > +        mx = (s->current_picture_ptr->motion_val[dir][mv_pos][0] + (3 << 24)) / 3 - (1 << 24);
> > > > +        my = (s->current_picture_ptr->motion_val[dir][mv_pos][1] + (3 << 24)) / 3 - (1 << 24);
> > > > +        lx = (s->current_picture_ptr->motion_val[dir][mv_pos][0] + (3 << 24)) % 3;
> > > > +        ly = (s->current_picture_ptr->motion_val[dir][mv_pos][1] + (3 << 24)) % 3;
> > > > +        dxy = ly*4 + lx;
> > > 
> > > > +        uvmx = chroma_coeffs[((s->current_picture_ptr->motion_val[dir][mv_pos][0] + (3 << 24)) >> 1) % 3];
> > > > +        uvmy = chroma_coeffs[((s->current_picture_ptr->motion_val[dir][mv_pos][0] + (3 << 24)) >> 1) % 3];
> > > 
> > > this is wrong ([0] vs. [1])
> > > also i think (3*(mx&1) + lx)>>1 could be used to avoid the %
> > 
> > Umm, it can't 
> 
> ohh yes it can, heres a test program with a simpler variant without the >>
[code skipped]

My test were flawed then. Sorry.
 
> > > [...]
> > > 
> > > > @@ -422,6 +422,10 @@
> > > >      void (*x8_setup_spatial_compensation)(uint8_t *src, uint8_t *dst, int linesize,
> > > >             int * range, int * sum,  int edges);
> > > >  
> > > > +    /* rv30 functions */
> > > > +    qpel_mc_func put_rv30_tpel_pixels_tab[4][16];
> > > > +    qpel_mc_func avg_rv30_tpel_pixels_tab[4][16];
> > > > +
> > > 
> > > this could be put under CONFIG_RV30_DECODER as well
> > 
> > I won't touch dsputil.* until decoder is fully integrated. 
> 
> well but the hunk above does change dsputil.h

Here it is without dsputil stuff.
 
> [...]
> -- 
> Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
-------------- next part --------------
Index: libavcodec/rv34.c
===================================================================
--- libavcodec/rv34.c	(revision 11361)
+++ libavcodec/rv34.c	(working copy)
@@ -566,6 +559,8 @@
         fill_rectangle(cur_pic->motion_val[!dir][mv_pos], 2, 2, s->b8_stride, 0, 4);
 }
 
+static const int chroma_coeffs[3] = { 8, 5, 3 };
+
 /**
  * generic motion compensation function
  *
@@ -591,21 +586,15 @@
     int is16x16 = 1;
 
     if(thirdpel){
-#if 0 /// todo
         int lx, ly;
 
-        mx = s->current_picture_ptr->motion_val[dir][mv_pos][0] / 3;
-        my = s->current_picture_ptr->motion_val[dir][mv_pos][1] / 3;
-        lx = ((s->current_picture_ptr->motion_val[dir][mv_pos][0] % 3) + 3) % 3;
-        ly = ((s->current_picture_ptr->motion_val[dir][mv_pos][1] % 3) + 3) % 3;
-        dxy = ly*3 + lx;
-        uvmx =
-#endif
-        mx = s->current_picture_ptr->motion_val[dir][mv_pos][0] >> 2;
-        my = s->current_picture_ptr->motion_val[dir][mv_pos][1] >> 2;
-        dxy = ((my & 3) << 2) | (mx & 3);
-        uvmx = mx & 6;
-        uvmy = my & 6;
+        mx = (s->current_picture_ptr->motion_val[dir][mv_pos][0] + (3 << 24)) / 3 - (1 << 24);
+        my = (s->current_picture_ptr->motion_val[dir][mv_pos][1] + (3 << 24)) / 3 - (1 << 24);
+        lx = (s->current_picture_ptr->motion_val[dir][mv_pos][0] + (3 << 24)) % 3;
+        ly = (s->current_picture_ptr->motion_val[dir][mv_pos][1] + (3 << 24)) % 3;
+        dxy = ly*4 + lx;
+        uvmx = chroma_coeffs[(3*(mx&1) + lx) >> 1];
+        uvmy = chroma_coeffs[(3*(my&1) + ly) >> 1];
     }else{
         mx = s->current_picture_ptr->motion_val[dir][mv_pos][0] >> 2;
         my = s->current_picture_ptr->motion_val[dir][mv_pos][1] >> 2;
@@ -662,15 +651,21 @@
                         const int width, const int height, int dir)
 {
     rv34_mc(r, block_type, xoff, yoff, mv_off, width, height, dir, r->rv30,
-            r->s.dsp.put_h264_qpel_pixels_tab, r->s.dsp.put_h264_chroma_pixels_tab);
+            r->rv30 ? r->s.dsp.put_rv30_tpel_pixels_tab
+                    : r->s.dsp.put_h264_qpel_pixels_tab,
+            r->s.dsp.put_h264_chroma_pixels_tab);
 }
 
 static void rv34_mc_2mv(RV34DecContext *r, const int block_type)
 {
     rv34_mc(r, block_type, 0, 0, 0, 2, 2, 0, r->rv30,
-            r->s.dsp.put_h264_qpel_pixels_tab, r->s.dsp.put_h264_chroma_pixels_tab);
+            r->rv30 ? r->s.dsp.put_rv30_tpel_pixels_tab
+                    : r->s.dsp.put_h264_qpel_pixels_tab,
+            r->s.dsp.put_h264_chroma_pixels_tab);
     rv34_mc(r, block_type, 0, 0, 0, 2, 2, 1, r->rv30,
-            r->s.dsp.avg_h264_qpel_pixels_tab, r->s.dsp.avg_h264_chroma_pixels_tab);
+            r->rv30 ? r->s.dsp.avg_rv30_tpel_pixels_tab
+                    : r->s.dsp.avg_h264_qpel_pixels_tab,
+            r->s.dsp.avg_h264_chroma_pixels_tab);
 }
 
 /** number of motion vectors in each macroblock type */
Index: libavcodec/rv30dsp.c
===================================================================
--- libavcodec/rv30dsp.c	(revision 0)
+++ libavcodec/rv30dsp.c	(revision 0)
@@ -0,0 +1,249 @@
+/*
+ * RV30 decoder motion compensation functions
+ * Copyright (c) 2007 Konstantin Shishkov
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file rv30dsp.c
+ * RV30 decoder motion compensation functions
+ */
+
+#include "avcodec.h"
+#include "dsputil.h"
+
+#define RV30_LOWPASS(OPNAME, OP) \
+static av_unused void OPNAME ## rv30_tpel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, const int C1, const int C2){\
+    const int h=8;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
+    int i;\
+    for(i=0; i<h; i++)\
+    {\
+        OP(dst[0], -(src[-1]+src[2]) + src[0]*C1 + src[1]*C2);\
+        OP(dst[1], -(src[ 0]+src[3]) + src[1]*C1 + src[2]*C2);\
+        OP(dst[2], -(src[ 1]+src[4]) + src[2]*C1 + src[3]*C2);\
+        OP(dst[3], -(src[ 2]+src[5]) + src[3]*C1 + src[4]*C2);\
+        OP(dst[4], -(src[ 3]+src[6]) + src[4]*C1 + src[5]*C2);\
+        OP(dst[5], -(src[ 4]+src[7]) + src[5]*C1 + src[6]*C2);\
+        OP(dst[6], -(src[ 5]+src[8]) + src[6]*C1 + src[7]*C2);\
+        OP(dst[7], -(src[ 6]+src[9]) + src[7]*C1 + src[8]*C2);\
+        dst+=dstStride;\
+        src+=srcStride;\
+    }\
+}\
+\
+static void OPNAME ## rv30_tpel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, const int C1, const int C2){\
+    const int w=8;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
+    int i;\
+    for(i=0; i<w; i++)\
+    {\
+        const int srcA= src[-1*srcStride];\
+        const int src0= src[0 *srcStride];\
+        const int src1= src[1 *srcStride];\
+        const int src2= src[2 *srcStride];\
+        const int src3= src[3 *srcStride];\
+        const int src4= src[4 *srcStride];\
+        const int src5= src[5 *srcStride];\
+        const int src6= src[6 *srcStride];\
+        const int src7= src[7 *srcStride];\
+        const int src8= src[8 *srcStride];\
+        const int src9= src[9 *srcStride];\
+        OP(dst[0*dstStride], -(srcA+src2) + src0*C1 + src1*C2);\
+        OP(dst[1*dstStride], -(src0+src3) + src1*C1 + src2*C2);\
+        OP(dst[2*dstStride], -(src1+src4) + src2*C1 + src3*C2);\
+        OP(dst[3*dstStride], -(src2+src5) + src3*C1 + src4*C2);\
+        OP(dst[4*dstStride], -(src3+src6) + src4*C1 + src5*C2);\
+        OP(dst[5*dstStride], -(src4+src7) + src5*C1 + src6*C2);\
+        OP(dst[6*dstStride], -(src5+src8) + src6*C1 + src7*C2);\
+        OP(dst[7*dstStride], -(src6+src9) + src7*C1 + src8*C2);\
+        dst++;\
+        src++;\
+    }\
+}\
+\
+static void OPNAME ## rv30_tpel8_h3_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+    const int h=8+2;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
+    int i;\
+    for(i=0; i<h; i++)\
+    {\
+        OP(dst[0], 6*src[0]+9*src[1]+src[2]);\
+        OP(dst[1], 6*src[1]+9*src[2]+src[3]);\
+        OP(dst[2], 6*src[2]+9*src[3]+src[4]);\
+        OP(dst[3], 6*src[3]+9*src[4]+src[5]);\
+        OP(dst[4], 6*src[4]+9*src[5]+src[6]);\
+        OP(dst[5], 6*src[5]+9*src[6]+src[7]);\
+        OP(dst[6], 6*src[6]+9*src[7]+src[8]);\
+        OP(dst[7], 6*src[7]+9*src[8]+src[9]);\
+        dst+=dstStride;\
+        src+=srcStride;\
+    }\
+}\
+\
+static void OPNAME ## rv30_tpel8_v3_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+    const int w=8;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
+    int i;\
+    for(i=0; i<w; i++)\
+    {\
+        const int src0= src[0 *srcStride];\
+        const int src1= src[1 *srcStride];\
+        const int src2= src[2 *srcStride];\
+        const int src3= src[3 *srcStride];\
+        const int src4= src[4 *srcStride];\
+        const int src5= src[5 *srcStride];\
+        const int src6= src[6 *srcStride];\
+        const int src7= src[7 *srcStride];\
+        const int src8= src[8 *srcStride];\
+        const int src9= src[9 *srcStride];\
+        OP(dst[0*dstStride], 6*src0 + 9*src1 + src2);\
+        OP(dst[1*dstStride], 6*src1 + 9*src2 + src3);\
+        OP(dst[2*dstStride], 6*src2 + 9*src3 + src4);\
+        OP(dst[3*dstStride], 6*src3 + 9*src4 + src5);\
+        OP(dst[4*dstStride], 6*src4 + 9*src5 + src6);\
+        OP(dst[5*dstStride], 6*src5 + 9*src6 + src7);\
+        OP(dst[6*dstStride], 6*src6 + 9*src7 + src8);\
+        OP(dst[7*dstStride], 6*src7 + 9*src8 + src9);\
+        dst ++;\
+        src ++;\
+    }\
+}\
+\
+static void OPNAME ## rv30_tpel8_hv_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+    uint8_t half[8*10];\
+    put_rv30_tpel8_h3_lowpass(half, src, 8, srcStride);\
+    OPNAME ## rv30_tpel8_v3_lowpass(dst, half, dstStride, 8);\
+}\
+\
+static void OPNAME ## rv30_tpel16_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, const int C1, const int C2){\
+    OPNAME ## rv30_tpel8_v_lowpass(dst  , src  , dstStride, srcStride, C1, C2);\
+    OPNAME ## rv30_tpel8_v_lowpass(dst+8, src+8, dstStride, srcStride, C1, C2);\
+    src += 8*srcStride;\
+    dst += 8*dstStride;\
+    OPNAME ## rv30_tpel8_v_lowpass(dst  , src  , dstStride, srcStride, C1, C2);\
+    OPNAME ## rv30_tpel8_v_lowpass(dst+8, src+8, dstStride, srcStride, C1, C2);\
+}\
+\
+static void OPNAME ## rv30_tpel16_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, const int C1, const int C2){\
+    OPNAME ## rv30_tpel8_h_lowpass(dst  , src  , dstStride, srcStride, C1, C2);\
+    OPNAME ## rv30_tpel8_h_lowpass(dst+8, src+8, dstStride, srcStride, C1, C2);\
+    src += 8*srcStride;\
+    dst += 8*dstStride;\
+    OPNAME ## rv30_tpel8_h_lowpass(dst  , src  , dstStride, srcStride, C1, C2);\
+    OPNAME ## rv30_tpel8_h_lowpass(dst+8, src+8, dstStride, srcStride, C1, C2);\
+}\
+\
+static void OPNAME ## rv30_tpel16_hv_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+    OPNAME ## rv30_tpel8_hv_lowpass(dst  , src  , dstStride, srcStride);\
+    OPNAME ## rv30_tpel8_hv_lowpass(dst+8, src+8, dstStride, srcStride);\
+    src += 8*srcStride;\
+    dst += 8*dstStride;\
+    OPNAME ## rv30_tpel8_hv_lowpass(dst  , src  , dstStride, srcStride);\
+    OPNAME ## rv30_tpel8_hv_lowpass(dst+8, src+8, dstStride, srcStride);\
+}\
+\
+
+#define RV30_MC(OPNAME, SIZE) \
+static void OPNAME ## rv30_tpel ## SIZE ## _mc10_c(uint8_t *dst, uint8_t *src, int stride){\
+    OPNAME ## rv30_tpel ## SIZE ## _h_lowpass(dst, src, stride, stride, 12, 6);\
+}\
+\
+static void OPNAME ## rv30_tpel ## SIZE ## _mc20_c(uint8_t *dst, uint8_t *src, int stride){\
+    OPNAME ## rv30_tpel ## SIZE ## _h_lowpass(dst, src, stride, stride, 6, 12);\
+}\
+\
+static void OPNAME ## rv30_tpel ## SIZE ## _mc01_c(uint8_t *dst, uint8_t *src, int stride){\
+    OPNAME ## rv30_tpel ## SIZE ## _v_lowpass(dst, src, stride, stride, 12, 6);\
+}\
+\
+static void OPNAME ## rv30_tpel ## SIZE ## _mc02_c(uint8_t *dst, uint8_t *src, int stride){\
+    OPNAME ## rv30_tpel ## SIZE ## _v_lowpass(dst, src, stride, stride, 6, 12);\
+}\
+\
+static void OPNAME ## rv30_tpel ## SIZE ## _mc11_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t half[SIZE*SIZE];\
+    put_rv30_tpel ## SIZE ## _h_lowpass(half, src, SIZE, stride, 12, 6);\
+    OPNAME ## rv30_tpel ## SIZE ## _v_lowpass(dst, src, stride, stride, 12, 6);\
+}\
+\
+static void OPNAME ## rv30_tpel ## SIZE ## _mc12_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t half[SIZE*SIZE];\
+    put_rv30_tpel ## SIZE ## _h_lowpass(half, src, SIZE, stride, 12, 6);\
+    OPNAME ## rv30_tpel ## SIZE ## _v_lowpass(dst, src, stride, stride, 6, 12);\
+}\
+\
+static void OPNAME ## rv30_tpel ## SIZE ## _mc21_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t half[SIZE*SIZE];\
+    put_rv30_tpel ## SIZE ## _h_lowpass(half, src, SIZE, stride, 6, 12);\
+    OPNAME ## rv30_tpel ## SIZE ## _v_lowpass(dst, src, stride, stride, 12, 6);\
+}\
+\
+static void OPNAME ## rv30_tpel ## SIZE ## _mc22_c(uint8_t *dst, uint8_t *src, int stride){\
+    OPNAME ## rv30_tpel ## SIZE ## _hv_lowpass(dst, src, stride, stride);\
+}\
+\
+
+#define op_avg(a, b)  a = (((a)+cm[((b) + 8)>>4]+1)>>1)
+#define op_put(a, b)  a = cm[((b) + 8)>>4]
+
+RV30_LOWPASS(put_       , op_put)
+RV30_LOWPASS(avg_       , op_avg)
+RV30_MC(put_, 8)
+RV30_MC(put_, 16)
+RV30_MC(avg_, 8)
+RV30_MC(avg_, 16)
+
+void ff_rv30dsp_init(DSPContext* c, AVCodecContext *avctx) {
+    c->put_rv30_tpel_pixels_tab[0][ 0] = c->put_h264_qpel_pixels_tab[0][0];
+    c->put_rv30_tpel_pixels_tab[0][ 1] = put_rv30_tpel16_mc10_c;
+    c->put_rv30_tpel_pixels_tab[0][ 2] = put_rv30_tpel16_mc20_c;
+    c->put_rv30_tpel_pixels_tab[0][ 4] = put_rv30_tpel16_mc01_c;
+    c->put_rv30_tpel_pixels_tab[0][ 5] = put_rv30_tpel16_mc11_c;
+    c->put_rv30_tpel_pixels_tab[0][ 6] = put_rv30_tpel16_mc21_c;
+    c->put_rv30_tpel_pixels_tab[0][ 8] = put_rv30_tpel16_mc02_c;
+    c->put_rv30_tpel_pixels_tab[0][ 9] = put_rv30_tpel16_mc12_c;
+    c->put_rv30_tpel_pixels_tab[0][10] = put_rv30_tpel16_mc22_c;
+    c->avg_rv30_tpel_pixels_tab[0][ 0] = c->avg_h264_qpel_pixels_tab[0][0];
+    c->avg_rv30_tpel_pixels_tab[0][ 1] = avg_rv30_tpel16_mc10_c;
+    c->avg_rv30_tpel_pixels_tab[0][ 2] = avg_rv30_tpel16_mc20_c;
+    c->avg_rv30_tpel_pixels_tab[0][ 4] = avg_rv30_tpel16_mc01_c;
+    c->avg_rv30_tpel_pixels_tab[0][ 5] = avg_rv30_tpel16_mc11_c;
+    c->avg_rv30_tpel_pixels_tab[0][ 6] = avg_rv30_tpel16_mc21_c;
+    c->avg_rv30_tpel_pixels_tab[0][ 8] = avg_rv30_tpel16_mc02_c;
+    c->avg_rv30_tpel_pixels_tab[0][ 9] = avg_rv30_tpel16_mc12_c;
+    c->avg_rv30_tpel_pixels_tab[0][10] = avg_rv30_tpel16_mc22_c;
+    c->put_rv30_tpel_pixels_tab[1][ 0] = c->put_h264_qpel_pixels_tab[1][0];
+    c->put_rv30_tpel_pixels_tab[1][ 1] = put_rv30_tpel8_mc10_c;
+    c->put_rv30_tpel_pixels_tab[1][ 2] = put_rv30_tpel8_mc20_c;
+    c->put_rv30_tpel_pixels_tab[1][ 4] = put_rv30_tpel8_mc01_c;
+    c->put_rv30_tpel_pixels_tab[1][ 5] = put_rv30_tpel8_mc11_c;
+    c->put_rv30_tpel_pixels_tab[1][ 6] = put_rv30_tpel8_mc21_c;
+    c->put_rv30_tpel_pixels_tab[1][ 8] = put_rv30_tpel8_mc02_c;
+    c->put_rv30_tpel_pixels_tab[1][ 9] = put_rv30_tpel8_mc12_c;
+    c->put_rv30_tpel_pixels_tab[1][10] = put_rv30_tpel8_mc22_c;
+    c->avg_rv30_tpel_pixels_tab[1][ 0] = c->avg_h264_qpel_pixels_tab[1][0];
+    c->avg_rv30_tpel_pixels_tab[1][ 1] = avg_rv30_tpel8_mc10_c;
+    c->avg_rv30_tpel_pixels_tab[1][ 2] = avg_rv30_tpel8_mc20_c;
+    c->avg_rv30_tpel_pixels_tab[1][ 4] = avg_rv30_tpel8_mc01_c;
+    c->avg_rv30_tpel_pixels_tab[1][ 5] = avg_rv30_tpel8_mc11_c;
+    c->avg_rv30_tpel_pixels_tab[1][ 6] = avg_rv30_tpel8_mc21_c;
+    c->avg_rv30_tpel_pixels_tab[1][ 8] = avg_rv30_tpel8_mc02_c;
+    c->avg_rv30_tpel_pixels_tab[1][ 9] = avg_rv30_tpel8_mc12_c;
+    c->avg_rv30_tpel_pixels_tab[1][10] = avg_rv30_tpel8_mc22_c;
+}



More information about the ffmpeg-devel mailing list