--- image.c	2002/11/27 14:29:34	1.20.2.8
+++ image.c	2003/03/16 12:04:14	1.26.2.2
@@ -59,7 +59,7 @@
 #include "image.h"
 #include "colorspace.h"
 #include "interpolate8x8.h"
-#include "../divx4.h"
+#include "reduced.h"
 #include "../utils/mem_align.h"
 
 #include "font.h"		// XXX: remove later
@@ -131,19 +131,9 @@
 image_swap(IMAGE * image1,
 		   IMAGE * image2)
 {
-	uint8_t *tmp;
-
-	tmp = image1->y;
-	image1->y = image2->y;
-	image2->y = tmp;
-
-	tmp = image1->u;
-	image1->u = image2->u;
-	image2->u = tmp;
-
-	tmp = image1->v;
-	image1->v = image2->v;
-	image2->v = tmp;
+    SWAP(uint8_t*, image1->y, image2->y);
+    SWAP(uint8_t*, image1->u, image2->u);
+    SWAP(uint8_t*, image1->v, image2->v);
 }
 
 
@@ -475,6 +465,48 @@
 }
 
 
+/*
+chroma optimize filter, invented by mf
+a chroma pixel is average from the surrounding pixels, when the
+correpsonding luma pixels are pure black or white.
+*/
+
+void
+image_chroma_optimize(IMAGE * img, int width, int height, int edged_width)
+{
+	int x,y;
+	int pixels = 0;
+
+	for (y = 1; y < height/2 - 1; y++)
+	for (x = 1; x < width/2 - 1; x++)
+	{
+#define IS_PURE(a)  ((a)<=16||(a)>=235)
+#define IMG_Y(Y,X)	img->y[(Y)*edged_width + (X)]
+#define IMG_U(Y,X)	img->u[(Y)*edged_width/2 + (X)]
+#define IMG_V(Y,X)	img->v[(Y)*edged_width/2 + (X)]
+
+		if (IS_PURE(IMG_Y(y*2  ,x*2  )) && 
+			IS_PURE(IMG_Y(y*2  ,x*2+1)) &&
+			IS_PURE(IMG_Y(y*2+1,x*2  )) && 
+			IS_PURE(IMG_Y(y*2+1,x*2+1)))
+		{
+			IMG_U(y,x) = (IMG_U(y,x-1) + IMG_U(y-1, x) + IMG_U(y, x+1) + IMG_U(y+1, x)) / 4;
+			IMG_V(y,x) = (IMG_V(y,x-1) + IMG_V(y-1, x) + IMG_V(y, x+1) + IMG_V(y+1, x)) / 4;
+			pixels++;
+		}
+
+#undef IS_PURE
+#undef IMG_Y
+#undef IMG_U
+#undef IMG_V
+	}
+	
+	DPRINTF(DPRINTF_DEBUG,"chroma_optimized_pixels = %i/%i", pixels, width*height/4);
+}
+
+
+
+
 
 /*
   perform safe packed colorspace conversion, by splitting
@@ -521,8 +553,8 @@
 			uint32_t width,
 			int height,
 			uint32_t edged_width,
-			uint8_t * src,
-			int src_stride,
+			uint8_t * src[4],
+			int src_stride[4],
 			int csp,
 			int interlacing)
 {
@@ -531,37 +563,10 @@
 	const int height2 = height/2;
 	//const int height_signed = (csp & XVID_CSP_VFLIP) ? -height : height;
 
-	
-	//	int src_stride = width;
-
-	// --- xvid 2.1 compatiblity patch ---
-	// --- remove when xvid_dec_frame->stride equals real stride
-	/*
-	if ((csp & ~XVID_CSP_VFLIP) == XVID_CSP_RGB555 ||
-		(csp & ~XVID_CSP_VFLIP) == XVID_CSP_RGB565 ||
-		(csp & ~XVID_CSP_VFLIP) == XVID_CSP_YUY2 ||
-		(csp & ~XVID_CSP_VFLIP) == XVID_CSP_YVYU ||
-		(csp & ~XVID_CSP_VFLIP) == XVID_CSP_UYVY)
-	{
-		src_stride *= 2;
-	} 
-	else if ((csp & ~XVID_CSP_VFLIP) == XVID_CSP_RGB24)
-	{
-		src_stride *= 3;
-	}
-	else if ((csp & ~XVID_CSP_VFLIP) == XVID_CSP_RGB32 ||
-		(csp & ~XVID_CSP_VFLIP) == XVID_CSP_ABGR ||
-		(csp & ~XVID_CSP_VFLIP) == XVID_CSP_RGBA)
-	{
-		src_stride *= 4;
-	}
-	*/
-	// ^--- xvid 2.1 compatiblity fix ---^
-
 	switch (csp & ~XVID_CSP_VFLIP) {
 	case XVID_CSP_RGB555:
 		safe_packed_conv(
-			src, src_stride, image->y, image->u, image->v, 
+			src[0], src_stride[0], image->y, image->u, image->v, 
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?rgb555i_to_yv12  :rgb555_to_yv12,
 			interlacing?rgb555i_to_yv12_c:rgb555_to_yv12_c, 2);
@@ -569,24 +574,24 @@
 
 	case XVID_CSP_RGB565:
 		safe_packed_conv(
-			src, src_stride, image->y, image->u, image->v, 
+			src[0], src_stride[0], image->y, image->u, image->v, 
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?rgb565i_to_yv12  :rgb565_to_yv12,
 			interlacing?rgb565i_to_yv12_c:rgb565_to_yv12_c, 2);
 		break;
 
 
-	case XVID_CSP_RGB24:
+	case XVID_CSP_BGR:
 		safe_packed_conv(
-			src, src_stride, image->y, image->u, image->v, 
+			src[0], src_stride[0], image->y, image->u, image->v, 
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?bgri_to_yv12  :bgr_to_yv12,
 			interlacing?bgri_to_yv12_c:bgr_to_yv12_c, 3);
 		break;
 
-	case XVID_CSP_RGB32:
+	case XVID_CSP_BGRA:
 		safe_packed_conv(
-			src, src_stride, image->y, image->u, image->v, 
+			src[0], src_stride[0], image->y, image->u, image->v, 
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?bgrai_to_yv12  :bgra_to_yv12,
 			interlacing?bgrai_to_yv12_c:bgra_to_yv12_c, 4);
@@ -594,7 +599,7 @@
 
 	case XVID_CSP_ABGR :
 		safe_packed_conv(
-			src, src_stride, image->y, image->u, image->v, 
+			src[0], src_stride[0], image->y, image->u, image->v, 
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?abgri_to_yv12  :abgr_to_yv12,
 			interlacing?abgri_to_yv12_c:abgr_to_yv12_c, 4);
@@ -602,7 +607,7 @@
 
 	case XVID_CSP_RGBA :
 		safe_packed_conv(
-			src, src_stride, image->y, image->u, image->v, 
+			src[0], src_stride[0], image->y, image->u, image->v, 
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?rgbai_to_yv12  :rgba_to_yv12,
 			interlacing?rgbai_to_yv12_c:rgba_to_yv12_c, 4);
@@ -610,7 +615,7 @@
 
 	case XVID_CSP_YUY2:
 		safe_packed_conv(
-			src, src_stride, image->y, image->u, image->v, 
+			src[0], src_stride[0], image->y, image->u, image->v, 
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?yuyvi_to_yv12  :yuyv_to_yv12,
 			interlacing?yuyvi_to_yv12_c:yuyv_to_yv12_c, 2);
@@ -618,7 +623,7 @@
 
 	case XVID_CSP_YVYU:		/* u/v swapped */
 		safe_packed_conv(
-			src, src_stride, image->y, image->v, image->y, 
+			src[0], src_stride[0], image->y, image->v, image->y, 
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?yuyvi_to_yv12  :yuyv_to_yv12,
 			interlacing?yuyvi_to_yv12_c:yuyv_to_yv12_c, 2);
@@ -626,7 +631,7 @@
 
 	case XVID_CSP_UYVY:
 		safe_packed_conv(
-			src, src_stride, image->y, image->u, image->v, 
+			src[0], src_stride[0], image->y, image->u, image->v, 
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?uyvyi_to_yv12  :uyvy_to_yv12,
 			interlacing?uyvyi_to_yv12_c:uyvy_to_yv12_c, 2);
@@ -634,23 +639,21 @@
 
 	case XVID_CSP_I420:
 		yv12_to_yv12(image->y, image->u, image->v, edged_width, edged_width2,
-			src, src + width*height, src + width*height + width2*height2,
-			width, width2, width, height, (csp & XVID_CSP_VFLIP));
+			src[0], src[0] + src_stride[0]*height, src[0] + src_stride[0]*height + (src_stride[0]/2)*height2,
+			src_stride[0], src_stride[0]/2, width, height, (csp & XVID_CSP_VFLIP));
 		break
 			;
 	case XVID_CSP_YV12:		/* u/v swapped */
 		yv12_to_yv12(image->y, image->v, image->u, edged_width, edged_width2,
-			src, src + width*height, src + width*height + width2*height2,
-			width, width2, width, height, (csp & XVID_CSP_VFLIP));
+			src[0], src[0] + src_stride[0]*height, src[0] + src_stride[0]*height + (src_stride[0]/2)*height2,
+			src_stride[0], src_stride[0]/2, width, height, (csp & XVID_CSP_VFLIP));
 		break;
 
 	case XVID_CSP_USER:
-		{
-			DEC_PICTURE * pic = (DEC_PICTURE*)src;
-			yv12_to_yv12(image->y, image->u, image->v, edged_width, edged_width2,
-				pic->y, pic->u, pic->v, pic->stride_y, pic->stride_y,
-				width, height, (csp & XVID_CSP_VFLIP));
-		}
+        /*XXX: support for different u & v strides */
+		yv12_to_yv12(image->y, image->u, image->v, edged_width, edged_width2,
+			src[0], src[1], src[2], src_stride[0], src_stride[1],
+			width, height, (csp & XVID_CSP_VFLIP));
 		break;
 
 	case XVID_CSP_NULL:
@@ -716,13 +719,12 @@
 			 uint32_t width,
 			 int height,
 			 uint32_t edged_width,
-			 uint8_t * dst,
-			 uint32_t dst_stride,
+			 uint8_t * dst[4],
+			 uint32_t dst_stride[4],
 			 int csp,
 			 int interlacing)
 {
 	const int edged_width2 = edged_width/2;
-	int width2 = width/2;
 	int height2 = height/2;
 
 /*
@@ -731,36 +733,10 @@
 	image_dump_yuvpgm(image, edged_width, width, height, "\\decode.pgm");
 */
 
-
-	// --- xvid 2.1 compatiblity patch ---
-	// --- remove when xvid_dec_frame->stride equals real stride
-	/*
-	if ((csp & ~XVID_CSP_VFLIP) == XVID_CSP_RGB555 ||
-		(csp & ~XVID_CSP_VFLIP) == XVID_CSP_RGB565 ||
-		(csp & ~XVID_CSP_VFLIP) == XVID_CSP_YUY2 ||
-		(csp & ~XVID_CSP_VFLIP) == XVID_CSP_YVYU ||
-		(csp & ~XVID_CSP_VFLIP) == XVID_CSP_UYVY)
-	{
-		dst_stride *= 2;
-	} 
-	else if ((csp & ~XVID_CSP_VFLIP) == XVID_CSP_RGB24)
-	{
-		dst_stride *= 3;
-	}
-	else if ((csp & ~XVID_CSP_VFLIP) == XVID_CSP_RGB32 ||
-		(csp & ~XVID_CSP_VFLIP) == XVID_CSP_ABGR ||
-		(csp & ~XVID_CSP_VFLIP) == XVID_CSP_RGBA)
-	{
-		dst_stride *= 4;
-	}
-	*/
-	// ^--- xvid 2.1 compatiblity fix ---^
-
-	
 	switch (csp & ~XVID_CSP_VFLIP) {
 	case XVID_CSP_RGB555:
 		safe_packed_conv(
-			dst, dst_stride, image->y, image->u, image->v,
+			dst[0], dst_stride[0], image->y, image->u, image->v,
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?yv12_to_rgb555i  :yv12_to_rgb555,
 			interlacing?yv12_to_rgb555i_c:yv12_to_rgb555_c, 2);
@@ -768,23 +744,23 @@
 
 	case XVID_CSP_RGB565:
 		safe_packed_conv(
-			dst, dst_stride, image->y, image->u, image->v,
+			dst[0], dst_stride[0], image->y, image->u, image->v,
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?yv12_to_rgb565i  :yv12_to_rgb565,
 			interlacing?yv12_to_rgb565i_c:yv12_to_rgb565_c, 2);
 		return 0;
 
-	case XVID_CSP_RGB24:
+    case XVID_CSP_BGR:
 		safe_packed_conv(
-			dst, dst_stride, image->y, image->u, image->v,
+			dst[0], dst_stride[0], image->y, image->u, image->v,
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?yv12_to_bgri  :yv12_to_bgr,
 			interlacing?yv12_to_bgri_c:yv12_to_bgr_c, 3);
 		return 0;
 
-	case XVID_CSP_RGB32:
+	case XVID_CSP_BGRA:
 		safe_packed_conv(
-			dst, dst_stride, image->y, image->u, image->v,
+			dst[0], dst_stride[0], image->y, image->u, image->v,
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?yv12_to_bgrai  :yv12_to_bgra,
 			interlacing?yv12_to_bgrai_c:yv12_to_bgra_c, 4);
@@ -792,7 +768,7 @@
 
 	case XVID_CSP_ABGR:
 		safe_packed_conv(
-			dst, dst_stride, image->y, image->u, image->v,
+			dst[0], dst_stride[0], image->y, image->u, image->v,
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?yv12_to_abgri  :yv12_to_abgr,
 			interlacing?yv12_to_abgri_c:yv12_to_abgr_c, 4);
@@ -800,7 +776,7 @@
 
 	case XVID_CSP_RGBA:
 		safe_packed_conv(
-			dst, dst_stride, image->y, image->u, image->v,
+			dst[0], dst_stride[0], image->y, image->u, image->v,
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?yv12_to_rgbai  :yv12_to_rgba,
 			interlacing?yv12_to_rgbai_c:yv12_to_rgba_c, 4);
@@ -808,7 +784,7 @@
 
 	case XVID_CSP_YUY2:
 		safe_packed_conv(
-			dst, dst_stride, image->y, image->u, image->v,
+			dst[0], dst_stride[0], image->y, image->u, image->v,
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?yv12_to_yuyvi  :yv12_to_yuyv,
 			interlacing?yv12_to_yuyvi_c:yv12_to_yuyv_c, 2);
@@ -816,7 +792,7 @@
 
 	case XVID_CSP_YVYU:		// u,v swapped
 		safe_packed_conv(
-			dst, dst_stride, image->y, image->v, image->u,
+			dst[0], dst_stride[0], image->y, image->v, image->u,
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?yv12_to_yuyvi  :yv12_to_yuyv,
 			interlacing?yv12_to_yuyvi_c:yv12_to_yuyv_c, 2);
@@ -824,39 +800,44 @@
 
 	case XVID_CSP_UYVY:
 		safe_packed_conv(
-			dst, dst_stride, image->y, image->u, image->v,
+			dst[0], dst_stride[0], image->y, image->u, image->v,
 			edged_width, edged_width2, width, height, (csp & XVID_CSP_VFLIP),
 			interlacing?yv12_to_uyvyi  :yv12_to_uyvy,
 			interlacing?yv12_to_uyvyi_c:yv12_to_uyvy_c, 2);
 		return 0;
 
 	case XVID_CSP_I420:
-		yv12_to_yv12(dst, dst + width*height, dst + width*height + width2*height2,
-			width, width2,
+		yv12_to_yv12(dst[0], dst[0] + dst_stride[0]*height, dst[0] + dst_stride[0]*height + (dst_stride[0]/2)*height2,
+			dst_stride[0], dst_stride[0]/2,
 			image->y, image->u, image->v, edged_width, edged_width2,
 			width, height, (csp & XVID_CSP_VFLIP));
 		return 0;
 
 	case XVID_CSP_YV12:		// u,v swapped
-		yv12_to_yv12(dst, dst + width*height, dst + width*height + width2*height2,
-			width, width2,
+		yv12_to_yv12(dst[0], dst[0] + dst_stride[0]*height, dst[0] + dst_stride[0]*height + (dst_stride[0]/2)*height2,
+			dst_stride[0], dst_stride[0]/2,
 			image->y, image->v, image->u, edged_width, edged_width2,
 			width, height, (csp & XVID_CSP_VFLIP));
 		return 0;
 
-	case XVID_CSP_USER:
-		{
-			DEC_PICTURE * pic = (DEC_PICTURE*)dst;
-			pic->y = image->y;
-			pic->u = image->u;
-			pic->v = image->v;
-			pic->stride_y = edged_width;
-			pic->stride_uv = edged_width / 2;
-		}
+	case XVID_CSP_USER :		// u,v swapped
+		yv12_to_yv12(dst[0], dst[1], dst[2],
+			dst_stride[0], dst_stride[1],	/* v: dst_stride[2] */
+			image->y, image->v, image->u, edged_width, edged_width2,
+			width, height, (csp & XVID_CSP_VFLIP));
+		return 0;
+
+	case XVID_CSP_INTERNAL :
+		dst[0] = image->y;
+		dst[1] = image->u;
+		dst[2] = image->v;
+		dst_stride[0] = edged_width;
+		dst_stride[1] = edged_width/2;
+		dst_stride[2] = edged_width/2;
 		return 0;
 
 	case XVID_CSP_NULL:
-	case XVID_CSP_EXTERN:
+	case XVID_CSP_SLICE:
 		return 0;
 
 	}
@@ -896,6 +877,36 @@
 	return psnr_y;
 }
 
+
+float sse_to_PSNR(long sse, int pixels)
+{
+        if (sse==0)
+                return 99.99F;
+
+        return 48.131F - 10*(float)log10((float)sse/(float)(pixels));   // log10(255*255)=4.8131
+
+}
+
+long plane_sse(uint8_t * orig,
+		   uint8_t * recon,
+		   uint16_t stride,
+		   uint16_t width,
+		   uint16_t height)
+{
+	int diff, x, y;
+	long sse=0;
+
+	for (y = 0; y < height; y++) {
+		for (x = 0; x < width; x++) {
+			diff = *(orig + x) - *(recon + x);
+			sse += diff * diff;
+		}
+		orig += stride;
+		recon += stride;
+	}
+	return sse;
+}
+
 /*
 
 #include <stdio.h>
@@ -992,7 +1003,6 @@
 }
 
 
-#define ABS(X)    (((X)>0)?(X):-(X))
 float
 image_mad(const IMAGE * img1,
 		  const IMAGE * img2,
@@ -1023,7 +1033,7 @@
 }
 
 void
-output_slice(IMAGE * cur, int std, int width, XVID_DEC_PICTURE* out_frm, int mbx, int mby,int mbl) {
+output_slice(IMAGE * cur, int std, int width, xvid_image_t* out_frm, int mbx, int mby,int mbl) {
   uint8_t *dY,*dU,*dV,*sY,*sU,*sV;
   int std2 = std >> 1;
   int w = mbl << 4, w2,i;
@@ -1032,26 +1042,128 @@
     w = width;
   w2 = w >> 1;
 
-  dY = (uint8_t*)out_frm->y + (mby << 4) * out_frm->stride_y + (mbx << 4);
-  dU = (uint8_t*)out_frm->u + (mby << 3) * out_frm->stride_u + (mbx << 3);
-  dV = (uint8_t*)out_frm->v + (mby << 3) * out_frm->stride_v + (mbx << 3);
+  dY = (uint8_t*)out_frm->plane[0] + (mby << 4) * out_frm->stride[0] + (mbx << 4);
+  dU = (uint8_t*)out_frm->plane[1] + (mby << 3) * out_frm->stride[1] + (mbx << 3);
+  dV = (uint8_t*)out_frm->plane[2] + (mby << 3) * out_frm->stride[2] + (mbx << 3);
   sY = cur->y + (mby << 4) * std + (mbx << 4);
   sU = cur->u + (mby << 3) * std2 + (mbx << 3);
   sV = cur->v + (mby << 3) * std2 + (mbx << 3);
 
   for(i = 0 ; i < 16 ; i++) {
     memcpy(dY,sY,w);
-    dY += out_frm->stride_y;
+    dY += out_frm->stride[0];
     sY += std;
   }
   for(i = 0 ; i < 8 ; i++) {
     memcpy(dU,sU,w2);
-    dU += out_frm->stride_u;
+    dU += out_frm->stride[1];
     sU += std2;
   }
   for(i = 0 ; i < 8 ; i++) {
     memcpy(dV,sV,w2);
-    dV += out_frm->stride_v;
+    dV += out_frm->stride[2];
     sV += std2;
   }
 }
+
+
+void
+image_clear(IMAGE * img, int width, int height, int edged_width,
+					int y, int u, int v)
+{
+	uint8_t * p;
+	int i;
+
+	p = img->y;
+	for (i = 0; i < height; i++) {
+		memset(p, y, width);
+		p += edged_width;
+	}
+
+	p = img->u;
+	for (i = 0; i < height/2; i++) {
+		memset(p, u, width/2);
+		p += edged_width/2;
+	}
+
+	p = img->v;
+	for (i = 0; i < height/2; i++) {
+		memset(p, v, width/2);
+		p += edged_width/2;
+	}
+}
+
+
+/* reduced resolution deblocking filter 
+	block = block size (16=rrv, 8=full resolution)
+	flags = XVID_DEC_YDEBLOCK|XVID_DEC_UVDEBLOCK
+*/
+void
+image_deblock_rrv(IMAGE * img, int edged_width,
+				const MACROBLOCK * mbs, int mb_width, int mb_height, int mb_stride,
+				int block, int flags)
+{
+	const int edged_width2 = edged_width /2;
+	const int nblocks = block / 8;	/* skals code uses 8pixel block uints */
+	int i,j;
+
+	/* luma: j,i in block units */
+
+		for (j = 1; j < mb_height*2; j++)		/* horizontal deblocking */
+		for (i = 0; i < mb_width*2; i++)
+		{
+			if (mbs[(j-1)/2*mb_stride + (i/2)].mode != MODE_NOT_CODED ||
+				mbs[(j+0)/2*mb_stride + (i/2)].mode != MODE_NOT_CODED)
+			{
+				hfilter_31(img->y + (j*block - 1)*edged_width + i*block,
+								  img->y + (j*block + 0)*edged_width + i*block, nblocks);
+			}
+		}
+
+		for (j = 0; j < mb_height*2; j++)		/* vertical deblocking */
+		for (i = 1; i < mb_width*2; i++)
+		{
+			if (mbs[(j/2)*mb_stride + (i-1)/2].mode != MODE_NOT_CODED ||
+				mbs[(j/2)*mb_stride + (i+0)/2].mode != MODE_NOT_CODED)
+			{
+				vfilter_31(img->y + (j*block)*edged_width + i*block - 1,
+						   img->y + (j*block)*edged_width + i*block + 0,
+						   edged_width, nblocks);
+			}
+		}
+
+
+
+	/* chroma */
+
+		for (j = 1; j < mb_height; j++)		/* horizontal deblocking */
+		for (i = 0; i < mb_width; i++)
+		{
+			if (mbs[(j-1)*mb_stride + i].mode != MODE_NOT_CODED || 
+				mbs[(j+0)*mb_stride + i].mode != MODE_NOT_CODED)
+			{
+				hfilter_31(img->u + (j*block - 1)*edged_width2 + i*block,
+						   img->u + (j*block + 0)*edged_width2 + i*block, nblocks);
+				hfilter_31(img->v + (j*block - 1)*edged_width2 + i*block,
+						   img->v + (j*block + 0)*edged_width2 + i*block, nblocks);
+			}
+		}
+
+		for (j = 0; j < mb_height; j++)		/* vertical deblocking */	
+		for (i = 1; i < mb_width; i++)
+		{
+			if (mbs[j*mb_stride + i - 1].mode != MODE_NOT_CODED ||
+				mbs[j*mb_stride + i + 0].mode != MODE_NOT_CODED) 
+			{
+				vfilter_31(img->u + (j*block)*edged_width2 + i*block - 1,
+						   img->u + (j*block)*edged_width2 + i*block + 0,
+						   edged_width2, nblocks);
+				vfilter_31(img->v + (j*block)*edged_width2 + i*block - 1,
+						   img->v + (j*block)*edged_width2 + i*block + 0,
+						   edged_width2, nblocks);
+			}
+		}
+
+
+}
+