[Mplayerxp-cvslog] SF.net SVN: mplayerxp:[465] mplayerxp

[<nic...@us...>]

Revision: 465
          http://mplayerxp.svn.sourceforge.net/mplayerxp/?rev=465&view=rev
Author:   nickols_k
Date:     2012-11-27 15:55:31 +0000 (Tue, 27 Nov 2012)
Log Message:
-----------
clarity: rename osd.* ->osd_render.*

Modified Paths:
--------------
    mplayerxp/libvo/Makefile
    mplayerxp/libvo/sub.cpp
    mplayerxp/libvo/sub.h
    mplayerxp/libvo/video_out.cpp
    mplayerxp/libvo/video_out_internal.h
    mplayerxp/libvo/vidix_system.cpp
    mplayerxp/libvo/vo_fbdev.cpp
    mplayerxp/postproc/libmenu/menu.cpp
    mplayerxp/postproc/vf_expand.cpp
    mplayerxp/postproc/vf_panscan.cpp

Added Paths:
-----------
    mplayerxp/libvo/osd_render.cpp
    mplayerxp/libvo/osd_render.h
    mplayerxp/libvo/osd_render_template.h

Removed Paths:
-------------
    mplayerxp/libvo/osd.cpp
    mplayerxp/libvo/osd.h
    mplayerxp/libvo/osd_template.h

Modified: mplayerxp/libvo/Makefile
===================================================================
--- mplayerxp/libvo/Makefile	2012-11-27 15:26:22 UTC (rev 464)
+++ mplayerxp/libvo/Makefile	2012-11-27 15:55:31 UTC (rev 465)
@@ -2,8 +2,8 @@
 
 LIBNAME = libvo.a
 
-CXXSRCS=video_out.cpp font_load.cpp img_format.cpp screenshot.cpp sub.cpp osd.cpp
-CXXSRCS+=aspect.cpp vo_null.cpp
+CXXSRCS=video_out.cpp font_load.cpp img_format.cpp screenshot.cpp sub.cpp
+CXXSRCS+=osd_render.cpp aspect.cpp vo_null.cpp
 ifeq ($(HAVE_SDL),yes)
 CXXSRCS+=vo_sdl.cpp
 endif

Deleted: mplayerxp/libvo/osd.cpp
===================================================================
--- mplayerxp/libvo/osd.cpp	2012-11-27 15:26:22 UTC (rev 464)
+++ mplayerxp/libvo/osd.cpp	2012-11-27 15:55:31 UTC (rev 465)
@@ -1,261 +0,0 @@
-#include "mp_config.h"
-#include "osdep/mplib.h"
-using namespace mpxp;
-/*
-   Generic alpha renderers for all YUV modes and RGB depths.
-   These are "reference implementations", should be optimized later (MMX, etc)
-   Templating Code from Michael Niedermayer (mic...@gm...) is under GPL
-*/
-
-#include <stdio.h>
-#include <pthread.h>
-
-#include <inttypes.h>
-#include "osdep/cpudetect.h"
-#include "osdep/mangle.h"
-#include "mplayerxp.h"
-#include "osd.h"
-#include "vo_msg.h"
-
-#if defined(__MMX__)
-static const uint64_t bFF __attribute__((used))  __attribute__((aligned(8))) = 0xFFFFFFFFFFFFFFFFULL;
-static const unsigned long long mask24lh  __attribute__((used)) __attribute__((aligned(8))) = 0xFFFF000000000000ULL;
-static const unsigned long long mask24hl  __attribute__((used)) __attribute__((aligned(8))) = 0x0000FFFFFFFFFFFFULL;
-#endif
-
-#define PVECTOR_ACCEL_H "osd_template.h"
-#include "pvector/pvector_inc.h"
-
-#ifdef FAST_OSD_TABLE
-static unsigned short fast_osd_15bpp_table[256];
-static unsigned short fast_osd_16bpp_table[256];
-#endif
-
-static void __FASTCALL__ vo_draw_alpha_rgb15_c(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
-    int y;
-    for(y=0;y<h;y++){
-	register unsigned short *dst = (unsigned short*) dstbase;
-	register int x;
-	for(x=0;x<w;x++){
-	    if(srca[x]){
-#ifdef FAST_OSD
-#ifdef FAST_OSD_TABLE
-		dst[x]=fast_osd_15bpp_table[src[x]];
-#else
-		register unsigned int a=src[x]>>3;
-		dst[x]=(a<<10)|(a<<5)|a;
-#endif
-#else
-		unsigned char r=dst[x]&0x1F;
-		unsigned char g=(dst[x]>>5)&0x1F;
-		unsigned char b=(dst[x]>>10)&0x1F;
-		r=(((r*srca[x])>>5)+src[x])>>3;
-		g=(((g*srca[x])>>5)+src[x])>>3;
-		b=(((b*srca[x])>>5)+src[x])>>3;
-		dst[x]=(b<<10)|(g<<5)|r;
-#endif
-	    }
-	}
-	src+=srcstride;
-	srca+=srcstride;
-	dstbase+=dststride;
-    }
-    return;
-}
-
-static void __FASTCALL__ vo_draw_alpha_rgb16_c(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
-    int y;
-    for(y=0;y<h;y++){
-	register unsigned short *dst = (unsigned short*) dstbase;
-	register int x;
-	for(x=0;x<w;x++){
-	    if(srca[x]){
-#ifdef FAST_OSD
-#ifdef FAST_OSD_TABLE
-		dst[x]=fast_osd_16bpp_table[src[x]];
-#else
-		dst[x]=((src[x]>>3)<<11)|((src[x]>>2)<<5)|(src[x]>>3);
-#endif
-#else
-		unsigned char r=dst[x]&0x1F;
-		unsigned char g=(dst[x]>>5)&0x3F;
-		unsigned char b=(dst[x]>>11)&0x1F;
-		r=(((r*srca[x])>>5)+src[x])>>3;
-		g=(((g*srca[x])>>6)+src[x])>>2;
-		b=(((b*srca[x])>>5)+src[x])>>3;
-		dst[x]=(b<<11)|(g<<5)|r;
-#endif
-	    }
-	}
-	src+=srcstride;
-	srca+=srcstride;
-	dstbase+=dststride;
-    }
-    return;
-}
-
-static void __FASTCALL__ vo_draw_alpha_uyvy_c(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
-    vo_draw_alpha_yuy2_c(w,h,src,srca,srcstride,dstbase+1,dststride,finalize);
-}
-
-static void __FASTCALL__ vo_draw_alpha_null(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
-    UNUSED(w);
-    UNUSED(h);
-    UNUSED(src);
-    UNUSED(srca);
-    UNUSED(srcstride);
-    UNUSED(dstbase);
-    UNUSED(dststride);
-    UNUSED(finalize);
-}
-
-void OSD_Render::render(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize)
-{
-    (*draw_alpha_ptr)(w,h,src,srca,srcstride,dstbase,dststride,finalize);
-}
-
-void OSD_Render::get_draw_alpha(unsigned fmt) {
-    MSG_DBG2("get_draw_alpha(%s)\n",vo_format_name(fmt));
-    switch(fmt) {
-    case IMGFMT_BGR15:
-    case IMGFMT_RGB15:
-	draw_alpha_ptr=draw_alpha_rgb15_ptr;
-	break;
-    case IMGFMT_BGR16:
-    case IMGFMT_RGB16:
-	draw_alpha_ptr=draw_alpha_rgb16_ptr;
-	break;
-    case IMGFMT_BGR24:
-    case IMGFMT_RGB24:
-	draw_alpha_ptr=draw_alpha_rgb24_ptr;
-	break;
-    case IMGFMT_BGR32:
-    case IMGFMT_RGB32:
-	draw_alpha_ptr=draw_alpha_rgb32_ptr;
-	break;
-    case IMGFMT_YV12:
-    case IMGFMT_I420:
-    case IMGFMT_IYUV:
-    case IMGFMT_YVU9:
-    case IMGFMT_IF09:
-    case IMGFMT_Y800:
-    case IMGFMT_Y8:
-	draw_alpha_ptr=draw_alpha_yv12_ptr;
-	break;
-    case IMGFMT_YUY2:
-	draw_alpha_ptr=draw_alpha_yuy2_ptr;
-	break;
-    case IMGFMT_UYVY:
-	draw_alpha_ptr=draw_alpha_uyvy_ptr;
-	break;
-    default:
-	MSG_ERR("draw alpha for %s fourcc not implemented yet!",vo_format_name(fmt));
-	draw_alpha_ptr=vo_draw_alpha_null;
-	break;
-    }
-}
-
-OSD_Render::OSD_Render(unsigned fourcc)
-{
-#ifdef FAST_OSD_TABLE
-    int i;
-    for(i=0;i<256;i++){
-	fast_osd_15bpp_table[i]=((i>>3)<<10)|((i>>3)<<5)|(i>>3);
-	fast_osd_16bpp_table[i]=((i>>3)<<11)|((i>>2)<<5)|(i>>3);
-    }
-#endif
-/*FIXME the optimized stuff is a lie for 15/16bpp as they arent optimized yet*/
-// ordered per speed fasterst first
-#ifdef __AVX__
-if(gCpuCaps.hasSSE41)
-{
-	MSG_V("Using SSE4 Optimized OnScreenDisplay\n");
-	draw_alpha_yv12_ptr=vo_draw_alpha_yv12_AVX;
-	draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_AVX;
-	draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_AVX;
-	draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_AVX;
-}
-else
-#endif
-#ifdef __SSE4_1__
-if(gCpuCaps.hasSSE41)
-{
-	MSG_V("Using SSE4 Optimized OnScreenDisplay\n");
-	draw_alpha_yv12_ptr=vo_draw_alpha_yv12_SSE4;
-	draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_SSE4;
-	draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_SSE4;
-	draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_SSE4;
-}
-else
-#endif
-#ifdef __SSSE3__
-if(gCpuCaps.hasSSSE3)
-{
-	MSG_V("Using SSSE3 Optimized OnScreenDisplay\n");
-	draw_alpha_yv12_ptr=vo_draw_alpha_yv12_SSSE3;
-	draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_SSSE3;
-	draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_SSSE3;
-	draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_SSSE3;
-}
-else
-#endif
-#ifdef __SSE3__
-if(gCpuCaps.hasSSE3)
-{
-	MSG_V("Using SSE3 Optimized OnScreenDisplay\n");
-	draw_alpha_yv12_ptr=vo_draw_alpha_yv12_SSE3;
-	draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_SSE3;
-	draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_SSE3;
-	draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_SSE3;
-}
-else
-#endif
-#ifdef __SSE2__
-if(gCpuCaps.hasSSE2)
-{
-	MSG_V("Using SSE2 Optimized OnScreenDisplay\n");
-	draw_alpha_yv12_ptr=vo_draw_alpha_yv12_SSE2;
-	draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_SSE2;
-	draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_SSE2;
-	draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_SSE2;
-}
-else
-#endif
-#ifndef __x86_64__
-#ifdef __SSE__
-if(gCpuCaps.hasMMX2)
-{
-	MSG_V("Using MMX (with tiny bit MMX2) Optimized OnScreenDisplay\n");
-	draw_alpha_yv12_ptr=vo_draw_alpha_yv12_SSE;
-	draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_SSE;
-	draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_SSE;
-	draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_SSE;
-}
-else
-#endif
-//#ifdef __MMX__
-//if(gCpuCaps.hasMMX)
-//{
-//	MSG_V("Using MMX (with tiny bit MMX2) Optimized OnScreenDisplay\n");
-//	draw_alpha_yv12_ptr=vo_draw_alpha_yv12_MMX;
-//	draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_MMX;
-//	draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_MMX;
-//	draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_MMX;
-//}
-//else
-//#endif
-#endif
-{
-    MSG_V("Using generic OnScreenDisplay\n");
-    draw_alpha_yv12_ptr=vo_draw_alpha_yv12_c;
-    draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_c;
-    draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_c;
-    draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_c;
-}
-    draw_alpha_uyvy_ptr=vo_draw_alpha_uyvy_c;
-    draw_alpha_rgb15_ptr=vo_draw_alpha_rgb15_c;
-    draw_alpha_rgb16_ptr=vo_draw_alpha_rgb16_c;
-    get_draw_alpha(fourcc);
-}
-
-OSD_Render::~OSD_Render(){}

Deleted: mplayerxp/libvo/osd.h
===================================================================
--- mplayerxp/libvo/osd.h	2012-11-27 15:26:22 UTC (rev 464)
+++ mplayerxp/libvo/osd.h	2012-11-27 15:55:31 UTC (rev 465)
@@ -1,25 +0,0 @@
-#ifndef __MPLAYER_OSD_H
-#define __MPLAYER_OSD_H
-
-/* Generic alpha renderers for all YUV modes and RGB depths. */
-/* These are "reference implementations", should be optimized later (MMX, etc) */
-class OSD_Render : public Opaque {
-    public:
-	OSD_Render(unsigned fourcc);
-	virtual ~OSD_Render();
-
-	void render(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize);
-    private:
-	void	get_draw_alpha(unsigned fmt);
-	typedef void (* __FASTCALL__ draw_alpha_f)(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize);
-	draw_alpha_f	draw_alpha_ptr;
-	draw_alpha_f	draw_alpha_yv12_ptr;
-	draw_alpha_f	draw_alpha_yuy2_ptr;
-	draw_alpha_f	draw_alpha_uyvy_ptr;
-	draw_alpha_f	draw_alpha_rgb24_ptr;
-	draw_alpha_f	draw_alpha_rgb32_ptr;
-	draw_alpha_f	draw_alpha_rgb15_ptr;
-	draw_alpha_f	draw_alpha_rgb16_ptr;
-};
-
-#endif

Copied: mplayerxp/libvo/osd_render.cpp (from rev 464, mplayerxp/libvo/osd.cpp)
===================================================================
--- mplayerxp/libvo/osd_render.cpp	                        (rev 0)
+++ mplayerxp/libvo/osd_render.cpp	2012-11-27 15:55:31 UTC (rev 465)
@@ -0,0 +1,261 @@
+#include "mp_config.h"
+#include "osdep/mplib.h"
+using namespace mpxp;
+/*
+   Generic alpha renderers for all YUV modes and RGB depths.
+   These are "reference implementations", should be optimized later (MMX, etc)
+   Templating Code from Michael Niedermayer (mic...@gm...) is under GPL
+*/
+
+#include <stdio.h>
+#include <pthread.h>
+
+#include <inttypes.h>
+#include "osdep/cpudetect.h"
+#include "osdep/mangle.h"
+#include "mplayerxp.h"
+#include "osd_render.h"
+#include "vo_msg.h"
+
+#if defined(__MMX__)
+static const uint64_t bFF __attribute__((used))  __attribute__((aligned(8))) = 0xFFFFFFFFFFFFFFFFULL;
+static const unsigned long long mask24lh  __attribute__((used)) __attribute__((aligned(8))) = 0xFFFF000000000000ULL;
+static const unsigned long long mask24hl  __attribute__((used)) __attribute__((aligned(8))) = 0x0000FFFFFFFFFFFFULL;
+#endif
+
+#define PVECTOR_ACCEL_H "osd_render_template.h"
+#include "pvector/pvector_inc.h"
+
+#ifdef FAST_OSD_TABLE
+static unsigned short fast_osd_15bpp_table[256];
+static unsigned short fast_osd_16bpp_table[256];
+#endif
+
+static void __FASTCALL__ vo_draw_alpha_rgb15_c(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
+    int y;
+    for(y=0;y<h;y++){
+	register unsigned short *dst = (unsigned short*) dstbase;
+	register int x;
+	for(x=0;x<w;x++){
+	    if(srca[x]){
+#ifdef FAST_OSD
+#ifdef FAST_OSD_TABLE
+		dst[x]=fast_osd_15bpp_table[src[x]];
+#else
+		register unsigned int a=src[x]>>3;
+		dst[x]=(a<<10)|(a<<5)|a;
+#endif
+#else
+		unsigned char r=dst[x]&0x1F;
+		unsigned char g=(dst[x]>>5)&0x1F;
+		unsigned char b=(dst[x]>>10)&0x1F;
+		r=(((r*srca[x])>>5)+src[x])>>3;
+		g=(((g*srca[x])>>5)+src[x])>>3;
+		b=(((b*srca[x])>>5)+src[x])>>3;
+		dst[x]=(b<<10)|(g<<5)|r;
+#endif
+	    }
+	}
+	src+=srcstride;
+	srca+=srcstride;
+	dstbase+=dststride;
+    }
+    return;
+}
+
+static void __FASTCALL__ vo_draw_alpha_rgb16_c(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
+    int y;
+    for(y=0;y<h;y++){
+	register unsigned short *dst = (unsigned short*) dstbase;
+	register int x;
+	for(x=0;x<w;x++){
+	    if(srca[x]){
+#ifdef FAST_OSD
+#ifdef FAST_OSD_TABLE
+		dst[x]=fast_osd_16bpp_table[src[x]];
+#else
+		dst[x]=((src[x]>>3)<<11)|((src[x]>>2)<<5)|(src[x]>>3);
+#endif
+#else
+		unsigned char r=dst[x]&0x1F;
+		unsigned char g=(dst[x]>>5)&0x3F;
+		unsigned char b=(dst[x]>>11)&0x1F;
+		r=(((r*srca[x])>>5)+src[x])>>3;
+		g=(((g*srca[x])>>6)+src[x])>>2;
+		b=(((b*srca[x])>>5)+src[x])>>3;
+		dst[x]=(b<<11)|(g<<5)|r;
+#endif
+	    }
+	}
+	src+=srcstride;
+	srca+=srcstride;
+	dstbase+=dststride;
+    }
+    return;
+}
+
+static void __FASTCALL__ vo_draw_alpha_uyvy_c(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
+    vo_draw_alpha_yuy2_c(w,h,src,srca,srcstride,dstbase+1,dststride,finalize);
+}
+
+static void __FASTCALL__ vo_draw_alpha_null(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
+    UNUSED(w);
+    UNUSED(h);
+    UNUSED(src);
+    UNUSED(srca);
+    UNUSED(srcstride);
+    UNUSED(dstbase);
+    UNUSED(dststride);
+    UNUSED(finalize);
+}
+
+void OSD_Render::render(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize)
+{
+    (*draw_alpha_ptr)(w,h,src,srca,srcstride,dstbase,dststride,finalize);
+}
+
+void OSD_Render::get_draw_alpha(unsigned fmt) {
+    MSG_DBG2("get_draw_alpha(%s)\n",vo_format_name(fmt));
+    switch(fmt) {
+    case IMGFMT_BGR15:
+    case IMGFMT_RGB15:
+	draw_alpha_ptr=draw_alpha_rgb15_ptr;
+	break;
+    case IMGFMT_BGR16:
+    case IMGFMT_RGB16:
+	draw_alpha_ptr=draw_alpha_rgb16_ptr;
+	break;
+    case IMGFMT_BGR24:
+    case IMGFMT_RGB24:
+	draw_alpha_ptr=draw_alpha_rgb24_ptr;
+	break;
+    case IMGFMT_BGR32:
+    case IMGFMT_RGB32:
+	draw_alpha_ptr=draw_alpha_rgb32_ptr;
+	break;
+    case IMGFMT_YV12:
+    case IMGFMT_I420:
+    case IMGFMT_IYUV:
+    case IMGFMT_YVU9:
+    case IMGFMT_IF09:
+    case IMGFMT_Y800:
+    case IMGFMT_Y8:
+	draw_alpha_ptr=draw_alpha_yv12_ptr;
+	break;
+    case IMGFMT_YUY2:
+	draw_alpha_ptr=draw_alpha_yuy2_ptr;
+	break;
+    case IMGFMT_UYVY:
+	draw_alpha_ptr=draw_alpha_uyvy_ptr;
+	break;
+    default:
+	MSG_ERR("draw alpha for %s fourcc not implemented yet!",vo_format_name(fmt));
+	draw_alpha_ptr=vo_draw_alpha_null;
+	break;
+    }
+}
+
+OSD_Render::OSD_Render(unsigned fourcc)
+{
+#ifdef FAST_OSD_TABLE
+    int i;
+    for(i=0;i<256;i++){
+	fast_osd_15bpp_table[i]=((i>>3)<<10)|((i>>3)<<5)|(i>>3);
+	fast_osd_16bpp_table[i]=((i>>3)<<11)|((i>>2)<<5)|(i>>3);
+    }
+#endif
+/*FIXME the optimized stuff is a lie for 15/16bpp as they arent optimized yet*/
+// ordered per speed fasterst first
+#ifdef __AVX__
+if(gCpuCaps.hasSSE41)
+{
+	MSG_V("Using SSE4 Optimized OnScreenDisplay\n");
+	draw_alpha_yv12_ptr=vo_draw_alpha_yv12_AVX;
+	draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_AVX;
+	draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_AVX;
+	draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_AVX;
+}
+else
+#endif
+#ifdef __SSE4_1__
+if(gCpuCaps.hasSSE41)
+{
+	MSG_V("Using SSE4 Optimized OnScreenDisplay\n");
+	draw_alpha_yv12_ptr=vo_draw_alpha_yv12_SSE4;
+	draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_SSE4;
+	draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_SSE4;
+	draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_SSE4;
+}
+else
+#endif
+#ifdef __SSSE3__
+if(gCpuCaps.hasSSSE3)
+{
+	MSG_V("Using SSSE3 Optimized OnScreenDisplay\n");
+	draw_alpha_yv12_ptr=vo_draw_alpha_yv12_SSSE3;
+	draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_SSSE3;
+	draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_SSSE3;
+	draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_SSSE3;
+}
+else
+#endif
+#ifdef __SSE3__
+if(gCpuCaps.hasSSE3)
+{
+	MSG_V("Using SSE3 Optimized OnScreenDisplay\n");
+	draw_alpha_yv12_ptr=vo_draw_alpha_yv12_SSE3;
+	draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_SSE3;
+	draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_SSE3;
+	draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_SSE3;
+}
+else
+#endif
+#ifdef __SSE2__
+if(gCpuCaps.hasSSE2)
+{
+	MSG_V("Using SSE2 Optimized OnScreenDisplay\n");
+	draw_alpha_yv12_ptr=vo_draw_alpha_yv12_SSE2;
+	draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_SSE2;
+	draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_SSE2;
+	draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_SSE2;
+}
+else
+#endif
+#ifndef __x86_64__
+#ifdef __SSE__
+if(gCpuCaps.hasMMX2)
+{
+	MSG_V("Using MMX (with tiny bit MMX2) Optimized OnScreenDisplay\n");
+	draw_alpha_yv12_ptr=vo_draw_alpha_yv12_SSE;
+	draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_SSE;
+	draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_SSE;
+	draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_SSE;
+}
+else
+#endif
+//#ifdef __MMX__
+//if(gCpuCaps.hasMMX)
+//{
+//	MSG_V("Using MMX (with tiny bit MMX2) Optimized OnScreenDisplay\n");
+//	draw_alpha_yv12_ptr=vo_draw_alpha_yv12_MMX;
+//	draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_MMX;
+//	draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_MMX;
+//	draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_MMX;
+//}
+//else
+//#endif
+#endif
+{
+    MSG_V("Using generic OnScreenDisplay\n");
+    draw_alpha_yv12_ptr=vo_draw_alpha_yv12_c;
+    draw_alpha_yuy2_ptr=vo_draw_alpha_yuy2_c;
+    draw_alpha_rgb24_ptr=vo_draw_alpha_rgb24_c;
+    draw_alpha_rgb32_ptr=vo_draw_alpha_rgb32_c;
+}
+    draw_alpha_uyvy_ptr=vo_draw_alpha_uyvy_c;
+    draw_alpha_rgb15_ptr=vo_draw_alpha_rgb15_c;
+    draw_alpha_rgb16_ptr=vo_draw_alpha_rgb16_c;
+    get_draw_alpha(fourcc);
+}
+
+OSD_Render::~OSD_Render(){}

Copied: mplayerxp/libvo/osd_render.h (from rev 464, mplayerxp/libvo/osd.h)
===================================================================
--- mplayerxp/libvo/osd_render.h	                        (rev 0)
+++ mplayerxp/libvo/osd_render.h	2012-11-27 15:55:31 UTC (rev 465)
@@ -0,0 +1,25 @@
+#ifndef __MPLAYER_OSD_RENDER_H_INCLUDED
+#define __MPLAYER_OSD_RENDER_H_INCLUDED 1
+
+/* Generic alpha renderers for all YUV modes and RGB depths. */
+/* These are "reference implementations", should be optimized later (MMX, etc) */
+class OSD_Render : public Opaque {
+    public:
+	OSD_Render(unsigned fourcc);
+	virtual ~OSD_Render();
+
+	void render(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize);
+    private:
+	void	get_draw_alpha(unsigned fmt);
+	typedef void (* __FASTCALL__ draw_alpha_f)(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize);
+	draw_alpha_f	draw_alpha_ptr;
+	draw_alpha_f	draw_alpha_yv12_ptr;
+	draw_alpha_f	draw_alpha_yuy2_ptr;
+	draw_alpha_f	draw_alpha_uyvy_ptr;
+	draw_alpha_f	draw_alpha_rgb24_ptr;
+	draw_alpha_f	draw_alpha_rgb32_ptr;
+	draw_alpha_f	draw_alpha_rgb15_ptr;
+	draw_alpha_f	draw_alpha_rgb16_ptr;
+};
+
+#endif

Copied: mplayerxp/libvo/osd_render_template.h (from rev 463, mplayerxp/libvo/osd_template.h)
===================================================================
--- mplayerxp/libvo/osd_render_template.h	                        (rev 0)
+++ mplayerxp/libvo/osd_render_template.h	2012-11-27 15:55:31 UTC (rev 465)
@@ -0,0 +1,318 @@
+// Generic alpha renderers for all YUV modes and RGB depths.
+// Optimized by Nick and Michael
+// Code from Michael Niedermayer (mic...@gm...) is under GPL
+#include "pvector/pvector.h"
+
+#ifdef HAVE_INT_PVECTOR
+static __inline __m64 __attribute__((__gnu_inline__, __always_inline__))
+PVECTOR_RENAME(_m_load)(const any_t*__P)
+{
+    return *(const __m64 *)__P;
+}
+#undef _m_load
+#define _m_load PVECTOR_RENAME(_m_load)
+
+static __inline __m64 __attribute__((__gnu_inline__, __always_inline__))
+PVECTOR_RENAME(_m_load_half)(const any_t*__P)
+{
+    return _mm_cvtsi32_si64 (*(const int *)__P);
+}
+#undef _m_load_half
+#define _m_load_half PVECTOR_RENAME(_m_load_half)
+
+static __inline void __attribute__((__gnu_inline__, __always_inline__))
+PVECTOR_RENAME(_m_store)(any_t*__P, __m64 src)
+{
+    *(__m64 *)__P = src;
+}
+#undef _m_store
+#define _m_store PVECTOR_RENAME(_m_store)
+
+static __inline void __attribute__((__gnu_inline__, __always_inline__))
+PVECTOR_RENAME(_m_store_half)(any_t*__P, __m64 src)
+{
+    *(int *)__P = _mm_cvtsi64_si32(src);
+}
+#undef _m_store_half
+#define _m_store_half PVECTOR_RENAME(_m_store_half)
+
+static __inline void __attribute__((__gnu_inline__, __always_inline__))
+PVECTOR_RENAME(_m_movntq)(any_t*__P, __m64 src)
+{
+#ifdef HAVE_MMX2
+    _mm_stream_pi(__P,src);
+#else
+    _m_store(__P,src);
+#endif
+}
+#undef _m_movntq
+#define _m_movntq PVECTOR_RENAME(_m_movntq)
+
+#endif
+
+static inline void PVECTOR_RENAME(vo_draw_alpha_yv12)(int w,int h,const unsigned char* src,const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
+    unsigned y;
+#ifdef HAVE_INT_PVECTOR
+    __ivec vzero = _ivec_setzero();
+#endif
+PROFILE_START();
+    for(y=0;y<(unsigned)h;y++){
+	int x;
+	x=0;
+#ifdef HAVE_INT_PVECTOR
+	_ivec_prefetchw(&dstbase[x]);
+	_ivec_prefetch(&src[x]);
+	_ivec_prefetch(&srca[x]);
+	/* MOVNTDQ: #GP(0) - If memory operand is not aligned on a 16-byte boundary */
+	if(!_ivec_aligned(dstbase))
+	for(;x<w;x++){
+	    unsigned char *dst=&dstbase[x];
+	    if(srca[x]) *dst=((dstbase[x]*srca[x])>>8)+src[x];
+	    if(_ivec_aligned(dst)) break; /* align on sizeof(MMREG) boundary */
+	}
+	if((w-x)>=_ivec_size())
+	for(;x<w;x+=_ivec_size()){
+	    __ivec vmsk,vdest,vsrc,vsrca,vt[4];
+	    _ivec_prefetchw(&dstbase[x+_ivec_size()*4]);
+	    _ivec_prefetch(&src[x+_ivec_size()*4]);
+	    _ivec_prefetch(&srca[x+_ivec_size()*4]);
+	    vdest = _ivec_loada(&dstbase[x]);
+	    if(_ivec_aligned(&src[x]))
+		vsrc  = _ivec_loada(&src[x]);
+	    else
+		vsrc  = _ivec_loadu(&src[x]);
+	    if(_ivec_aligned(&srca[x]))
+		vsrca = _ivec_loada(&srca[x]);
+	    else
+		vsrca = _ivec_loadu(&srca[x]);
+	    vmsk  = _ivec_not(_ivec_cmpeq_s8(vsrca,vzero));
+	    vt[0] = _ivec_u16_from_u8(vdest,&vt[1]);
+	    vt[2] = _ivec_u16_from_u8(vsrca,&vt[3]);
+	    vt[0] = _ivec_srl_s16_imm(_ivec_mullo_s16(vt[0],vt[2]),8);
+	    vt[1] = _ivec_srl_s16_imm(_ivec_mullo_s16(vt[1],vt[3]),8);
+	    vt[0] = _ivec_add_s8(_ivec_u8_from_u16(vt[0],vt[1]),vsrc);
+	    vt[1] = _ivec_blend_u8(vdest,vt[0],vmsk);
+	    if(finalize)
+		_ivec_stream(&dstbase[x],vt[1]);
+	    else
+		_ivec_storea(&dstbase[x],vt[1]);
+	}
+#endif
+	for(;x<w;x++){
+	    if(srca[x]) dstbase[x]=((dstbase[x]*srca[x])>>8)+src[x];
+	}
+	src+=srcstride;
+	srca+=srcstride;
+	dstbase+=dststride;
+    }
+#ifdef HAVE_INT_PVECTOR
+    _ivec_empty();
+    if(finalize) _ivec_sfence();
+#endif
+PROFILE_END("vo_draw_alpha_yv12");
+    return;
+}
+
+static inline void PVECTOR_RENAME(vo_draw_alpha_yuy2)(int w,int h,const unsigned char* src,const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
+    int y;
+#if defined(FAST_OSD) && !defined(HAVE_MMX)
+    w=w>>1;
+#endif
+PROFILE_START();
+    for(y=0;y<h;y++){
+	register int x=0;
+#ifdef HAVE_INT_PVECTOR
+	/* TODO: port this stuff on PVector */
+	__m64 mm[8];
+	_ivec_prefetchw(&dstbase[x]);
+	_ivec_prefetch(&src[x]);
+	_ivec_prefetch(&srca[x]);
+	mm[7]=_mm_setzero_si64();
+	mm[5]=_mm_set1_pi8(0xFF);
+	mm[4]=mm[5];
+	mm[5]=_m_psllwi(mm[5],8);
+	mm[4]=_m_psrlwi(mm[4],8);
+    for(;x<w;x+=4){
+	    _ivec_prefetchw(&dstbase[x+_ivec_size()*4]);
+	    _ivec_prefetch(&src[x+_ivec_size()*4]);
+	    _ivec_prefetch(&srca[x+_ivec_size()*4]);
+	    mm[0]=_m_load(&(((char *)dstbase)[x*2]));
+	    mm[1]=mm[0];
+	    mm[0]=_m_pand(mm[0],mm[4]);
+	    mm[2]=_m_load_half(&(((char *)srca)[x]));
+	    mm[2]=_m_paddb(mm[2],_m_load(&bFF));
+	    mm[2]=_m_punpcklbw(mm[2],mm[7]);
+	    mm[0]=_m_pmullw(mm[0],mm[2]);
+	    mm[0]=_m_psrlwi(mm[0],8);
+	    mm[1]=_m_pand(mm[1],mm[5]);
+	    mm[2]=_m_load_half(&(((char *)src)[x]));
+	    mm[2]=_m_punpcklbw(mm[2],mm[7]);
+	    mm[0]=_m_por(mm[0],mm[1]);
+	    mm[0]=_m_paddb(mm[0],mm[2]);
+	    if(finalize)
+		_m_movntq(&(((char *)dstbase)[x*2]),mm[0]);
+	    else
+		_m_store(&(((char *)dstbase)[x*2]),mm[0]);
+	}
+#endif
+	for(;x<w;x++){
+#ifdef FAST_OSD
+	    if(srca[2*x+0]) dstbase[4*x+0]=src[2*x+0];
+	    if(srca[2*x+1]) dstbase[4*x+2]=src[2*x+1];
+#else
+	    if(srca[x]) {
+	       dstbase[2*x]=((dstbase[2*x]*srca[x])>>8)+src[x];
+	       dstbase[2*x+1]=((((signed)dstbase[2*x+1]-128)*srca[x])>>8)+128;
+	   }
+#endif
+	}
+	src+=srcstride;
+	srca+=srcstride;
+	dstbase+=dststride;
+    }
+#ifdef HAVE_INT_PVECTOR
+    _ivec_empty();
+    if(finalize) _ivec_sfence();
+#endif
+PROFILE_END("vo_draw_alpha_yuy2");
+    return;
+}
+
+static inline void PVECTOR_RENAME(vo_draw_alpha_rgb24)(int w,int h,const unsigned char* src,const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
+    int y;
+    for(y=0;y<h;y++){
+	register unsigned char *dst = dstbase;
+	register int x=0;
+#ifdef HAVE_INT_PVECTOR
+	__m64 mm[8];
+	_ivec_prefetchw(&dstbase[x]);
+	_ivec_prefetch(&src[x]);
+	_ivec_prefetch(&srca[x]);
+	mm[7]=_mm_setzero_si64();
+	mm[6]=_mm_set1_pi8(0xFF);
+    for(;x<w;x+=2){
+     if(srca[x] || srca[x+1]) {
+	    _ivec_prefetchw(&dstbase[x+_ivec_size()*4]);
+	    _ivec_prefetch(&src[x+_ivec_size()*4]);
+	    _ivec_prefetch(&srca[x+_ivec_size()*4]);
+	    mm[0]=_m_load(&dstbase[0]);
+	    mm[1]=mm[0];
+	    mm[5]=mm[0];
+	    mm[0]=_m_punpcklbw(mm[0],mm[7]);
+	    mm[1]=_m_punpckhbw(mm[1],mm[7]);
+	    mm[2]=_m_load_half(&srca[x]);
+	    mm[2]=_m_paddb(mm[2],mm[6]);
+	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
+	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
+	    mm[3]=mm[2];
+	    mm[2]=_m_punpcklbw(mm[2],mm[7]);
+	    mm[3]=_m_punpckhbw(mm[3],mm[7]);
+	    mm[0]=_m_pmullw(mm[0],mm[2]);
+	    mm[1]=_m_pmullw(mm[1],mm[3]);
+	    mm[0]=_m_psrlwi(mm[0],8);
+	    mm[1]=_m_psrlwi(mm[1],8);
+	    mm[0]=_m_packuswb(mm[0],mm[1]);
+	    mm[2]=_m_load_half(&src[x]);
+	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
+	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
+	    mm[0]=_m_paddb(mm[0],mm[2]);
+	    mm[5]=_m_pand(mm[5],_m_load(&mask24lh));
+	    mm[0]=_m_pand(mm[0],_m_load(&mask24hl));
+	    mm[5]=_m_por(mm[5],mm[0]);
+	    if(finalize)
+		_m_movntq(&dstbase[0],mm[5]);
+	    else
+		_m_store(&dstbase[0],mm[5]);
+	}
+	dst += 6;
+    }
+#endif /* arch_x86 */
+	for(;x<w;x++){
+	    if(srca[x]){
+#ifdef FAST_OSD
+		dst[0]=dst[1]=dst[2]=src[x];
+#else
+		dst[0]=((dst[0]*srca[x])>>8)+src[x];
+		dst[1]=((dst[1]*srca[x])>>8)+src[x];
+		dst[2]=((dst[2]*srca[x])>>8)+src[x];
+#endif
+	    }
+	    dst+=3; // 24bpp
+	}
+	src+=srcstride;
+	srca+=srcstride;
+	dstbase+=dststride;
+    }
+#ifdef HAVE_INT_PVECTOR
+    _ivec_empty();
+    if(finalize) _ivec_sfence();
+#endif
+    return;
+}
+
+static inline void PVECTOR_RENAME(vo_draw_alpha_rgb32)(int w,int h,const unsigned char* src,const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
+    int y;
+PROFILE_START();
+    for(y=0;y<h;y++){
+	register int x=0;
+#ifdef HAVE_INT_PVECTOR
+	__m64 mm[8];
+	_ivec_prefetchw(&dstbase[x]);
+	_ivec_prefetch(&src[x]);
+	_ivec_prefetch(&srca[x]);
+	mm[7]=_mm_setzero_si64();
+	mm[6]=_mm_set1_pi8(0xFF);
+    for(;x<w;x+=2){
+     if(srca[x] || srca[x+1]) {
+	    _ivec_prefetchw(&dstbase[x+_ivec_size()*4]);
+	    _ivec_prefetch(&src[x+_ivec_size()*4]);
+	    _ivec_prefetch(&srca[x+_ivec_size()*4]);
+	    mm[0]=_m_load(&dstbase[4*x]);
+	    mm[1]=mm[0];
+	    mm[0]=_m_punpcklbw(mm[0],mm[7]);
+	    mm[1]=_m_punpckhbw(mm[1],mm[7]);
+	    mm[2]=_m_load_half(&srca[x]);
+	    mm[2]=_m_paddb(mm[2],mm[6]);
+	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
+	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
+	    mm[3]=mm[2];
+	    mm[2]=_m_punpcklbw(mm[2],mm[7]);
+	    mm[3]=_m_punpckhbw(mm[3],mm[7]);
+	    mm[0]=_m_pmullw(mm[0],mm[2]);
+	    mm[1]=_m_pmullw(mm[1],mm[3]);
+	    mm[0]=_m_psrlwi(mm[0],8);
+	    mm[1]=_m_psrlwi(mm[1],8);
+	    mm[0]=_m_packuswb(mm[0],mm[1]);
+	    mm[2]=_m_load_half(&src[x]);
+	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
+	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
+	    mm[0]=_m_paddb(mm[0],mm[2]);
+	    if(finalize)
+		_m_movntq(&dstbase[4*x],mm[0]);
+	    else
+		_m_store(&dstbase[4*x],mm[0]);
+	}
+    }
+#endif /* arch_x86 */
+	for(;x<w;x++){
+	    if(srca[x]){
+#ifdef FAST_OSD
+		dstbase[4*x+0]=dstbase[4*x+1]=dstbase[4*x+2]=src[x];
+#else
+		dstbase[4*x+0]=((dstbase[4*x+0]*srca[x])>>8)+src[x];
+		dstbase[4*x+1]=((dstbase[4*x+1]*srca[x])>>8)+src[x];
+		dstbase[4*x+2]=((dstbase[4*x+2]*srca[x])>>8)+src[x];
+#endif
+	    }
+	}
+	src+=srcstride;
+	srca+=srcstride;
+	dstbase+=dststride;
+    }
+#ifdef HAVE_INT_PVECTOR
+    _ivec_empty();
+    if(finalize) _ivec_sfence();
+#endif
+PROFILE_END("vo_draw_alpha_rgb32");
+    return;
+}

Deleted: mplayerxp/libvo/osd_template.h
===================================================================
--- mplayerxp/libvo/osd_template.h	2012-11-27 15:26:22 UTC (rev 464)
+++ mplayerxp/libvo/osd_template.h	2012-11-27 15:55:31 UTC (rev 465)
@@ -1,318 +0,0 @@
-// Generic alpha renderers for all YUV modes and RGB depths.
-// Optimized by Nick and Michael
-// Code from Michael Niedermayer (mic...@gm...) is under GPL
-#include "pvector/pvector.h"
-
-#ifdef HAVE_INT_PVECTOR
-static __inline __m64 __attribute__((__gnu_inline__, __always_inline__))
-PVECTOR_RENAME(_m_load)(const any_t*__P)
-{
-    return *(const __m64 *)__P;
-}
-#undef _m_load
-#define _m_load PVECTOR_RENAME(_m_load)
-
-static __inline __m64 __attribute__((__gnu_inline__, __always_inline__))
-PVECTOR_RENAME(_m_load_half)(const any_t*__P)
-{
-    return _mm_cvtsi32_si64 (*(const int *)__P);
-}
-#undef _m_load_half
-#define _m_load_half PVECTOR_RENAME(_m_load_half)
-
-static __inline void __attribute__((__gnu_inline__, __always_inline__))
-PVECTOR_RENAME(_m_store)(any_t*__P, __m64 src)
-{
-    *(__m64 *)__P = src;
-}
-#undef _m_store
-#define _m_store PVECTOR_RENAME(_m_store)
-
-static __inline void __attribute__((__gnu_inline__, __always_inline__))
-PVECTOR_RENAME(_m_store_half)(any_t*__P, __m64 src)
-{
-    *(int *)__P = _mm_cvtsi64_si32(src);
-}
-#undef _m_store_half
-#define _m_store_half PVECTOR_RENAME(_m_store_half)
-
-static __inline void __attribute__((__gnu_inline__, __always_inline__))
-PVECTOR_RENAME(_m_movntq)(any_t*__P, __m64 src)
-{
-#ifdef HAVE_MMX2
-    _mm_stream_pi(__P,src);
-#else
-    _m_store(__P,src);
-#endif
-}
-#undef _m_movntq
-#define _m_movntq PVECTOR_RENAME(_m_movntq)
-
-#endif
-
-static inline void PVECTOR_RENAME(vo_draw_alpha_yv12)(int w,int h,const unsigned char* src,const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
-    unsigned y;
-#ifdef HAVE_INT_PVECTOR
-    __ivec vzero = _ivec_setzero();
-#endif
-PROFILE_START();
-    for(y=0;y<(unsigned)h;y++){
-	int x;
-	x=0;
-#ifdef HAVE_INT_PVECTOR
-	_ivec_prefetchw(&dstbase[x]);
-	_ivec_prefetch(&src[x]);
-	_ivec_prefetch(&srca[x]);
-	/* MOVNTDQ: #GP(0) - If memory operand is not aligned on a 16-byte boundary */
-	if(!_ivec_aligned(dstbase))
-	for(;x<w;x++){
-	    unsigned char *dst=&dstbase[x];
-	    if(srca[x]) *dst=((dstbase[x]*srca[x])>>8)+src[x];
-	    if(_ivec_aligned(dst)) break; /* align on sizeof(MMREG) boundary */
-	}
-	if((w-x)>=_ivec_size())
-	for(;x<w;x+=_ivec_size()){
-	    __ivec vmsk,vdest,vsrc,vsrca,vt[4];
-	    _ivec_prefetchw(&dstbase[x+_ivec_size()*4]);
-	    _ivec_prefetch(&src[x+_ivec_size()*4]);
-	    _ivec_prefetch(&srca[x+_ivec_size()*4]);
-	    vdest = _ivec_loada(&dstbase[x]);
-	    if(_ivec_aligned(&src[x]))
-		vsrc  = _ivec_loada(&src[x]);
-	    else
-		vsrc  = _ivec_loadu(&src[x]);
-	    if(_ivec_aligned(&srca[x]))
-		vsrca = _ivec_loada(&srca[x]);
-	    else
-		vsrca = _ivec_loadu(&srca[x]);
-	    vmsk  = _ivec_not(_ivec_cmpeq_s8(vsrca,vzero));
-	    vt[0] = _ivec_u16_from_u8(vdest,&vt[1]);
-	    vt[2] = _ivec_u16_from_u8(vsrca,&vt[3]);
-	    vt[0] = _ivec_srl_s16_imm(_ivec_mullo_s16(vt[0],vt[2]),8);
-	    vt[1] = _ivec_srl_s16_imm(_ivec_mullo_s16(vt[1],vt[3]),8);
-	    vt[0] = _ivec_add_s8(_ivec_u8_from_u16(vt[0],vt[1]),vsrc);
-	    vt[1] = _ivec_blend_u8(vdest,vt[0],vmsk);
-	    if(finalize)
-		_ivec_stream(&dstbase[x],vt[1]);
-	    else
-		_ivec_storea(&dstbase[x],vt[1]);
-	}
-#endif
-	for(;x<w;x++){
-	    if(srca[x]) dstbase[x]=((dstbase[x]*srca[x])>>8)+src[x];
-	}
-	src+=srcstride;
-	srca+=srcstride;
-	dstbase+=dststride;
-    }
-#ifdef HAVE_INT_PVECTOR
-    _ivec_empty();
-    if(finalize) _ivec_sfence();
-#endif
-PROFILE_END("vo_draw_alpha_yv12");
-    return;
-}
-
-static inline void PVECTOR_RENAME(vo_draw_alpha_yuy2)(int w,int h,const unsigned char* src,const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
-    int y;
-#if defined(FAST_OSD) && !defined(HAVE_MMX)
-    w=w>>1;
-#endif
-PROFILE_START();
-    for(y=0;y<h;y++){
-	register int x=0;
-#ifdef HAVE_INT_PVECTOR
-	/* TODO: port this stuff on PVector */
-	__m64 mm[8];
-	_ivec_prefetchw(&dstbase[x]);
-	_ivec_prefetch(&src[x]);
-	_ivec_prefetch(&srca[x]);
-	mm[7]=_mm_setzero_si64();
-	mm[5]=_mm_set1_pi8(0xFF);
-	mm[4]=mm[5];
-	mm[5]=_m_psllwi(mm[5],8);
-	mm[4]=_m_psrlwi(mm[4],8);
-    for(;x<w;x+=4){
-	    _ivec_prefetchw(&dstbase[x+_ivec_size()*4]);
-	    _ivec_prefetch(&src[x+_ivec_size()*4]);
-	    _ivec_prefetch(&srca[x+_ivec_size()*4]);
-	    mm[0]=_m_load(&(((char *)dstbase)[x*2]));
-	    mm[1]=mm[0];
-	    mm[0]=_m_pand(mm[0],mm[4]);
-	    mm[2]=_m_load_half(&(((char *)srca)[x]));
-	    mm[2]=_m_paddb(mm[2],_m_load(&bFF));
-	    mm[2]=_m_punpcklbw(mm[2],mm[7]);
-	    mm[0]=_m_pmullw(mm[0],mm[2]);
-	    mm[0]=_m_psrlwi(mm[0],8);
-	    mm[1]=_m_pand(mm[1],mm[5]);
-	    mm[2]=_m_load_half(&(((char *)src)[x]));
-	    mm[2]=_m_punpcklbw(mm[2],mm[7]);
-	    mm[0]=_m_por(mm[0],mm[1]);
-	    mm[0]=_m_paddb(mm[0],mm[2]);
-	    if(finalize)
-		_m_movntq(&(((char *)dstbase)[x*2]),mm[0]);
-	    else
-		_m_store(&(((char *)dstbase)[x*2]),mm[0]);
-	}
-#endif
-	for(;x<w;x++){
-#ifdef FAST_OSD
-	    if(srca[2*x+0]) dstbase[4*x+0]=src[2*x+0];
-	    if(srca[2*x+1]) dstbase[4*x+2]=src[2*x+1];
-#else
-	    if(srca[x]) {
-	       dstbase[2*x]=((dstbase[2*x]*srca[x])>>8)+src[x];
-	       dstbase[2*x+1]=((((signed)dstbase[2*x+1]-128)*srca[x])>>8)+128;
-	   }
-#endif
-	}
-	src+=srcstride;
-	srca+=srcstride;
-	dstbase+=dststride;
-    }
-#ifdef HAVE_INT_PVECTOR
-    _ivec_empty();
-    if(finalize) _ivec_sfence();
-#endif
-PROFILE_END("vo_draw_alpha_yuy2");
-    return;
-}
-
-static inline void PVECTOR_RENAME(vo_draw_alpha_rgb24)(int w,int h,const unsigned char* src,const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
-    int y;
-    for(y=0;y<h;y++){
-	register unsigned char *dst = dstbase;
-	register int x=0;
-#ifdef HAVE_INT_PVECTOR
-	__m64 mm[8];
-	_ivec_prefetchw(&dstbase[x]);
-	_ivec_prefetch(&src[x]);
-	_ivec_prefetch(&srca[x]);
-	mm[7]=_mm_setzero_si64();
-	mm[6]=_mm_set1_pi8(0xFF);
-    for(;x<w;x+=2){
-     if(srca[x] || srca[x+1]) {
-	    _ivec_prefetchw(&dstbase[x+_ivec_size()*4]);
-	    _ivec_prefetch(&src[x+_ivec_size()*4]);
-	    _ivec_prefetch(&srca[x+_ivec_size()*4]);
-	    mm[0]=_m_load(&dstbase[0]);
-	    mm[1]=mm[0];
-	    mm[5]=mm[0];
-	    mm[0]=_m_punpcklbw(mm[0],mm[7]);
-	    mm[1]=_m_punpckhbw(mm[1],mm[7]);
-	    mm[2]=_m_load_half(&srca[x]);
-	    mm[2]=_m_paddb(mm[2],mm[6]);
-	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
-	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
-	    mm[3]=mm[2];
-	    mm[2]=_m_punpcklbw(mm[2],mm[7]);
-	    mm[3]=_m_punpckhbw(mm[3],mm[7]);
-	    mm[0]=_m_pmullw(mm[0],mm[2]);
-	    mm[1]=_m_pmullw(mm[1],mm[3]);
-	    mm[0]=_m_psrlwi(mm[0],8);
-	    mm[1]=_m_psrlwi(mm[1],8);
-	    mm[0]=_m_packuswb(mm[0],mm[1]);
-	    mm[2]=_m_load_half(&src[x]);
-	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
-	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
-	    mm[0]=_m_paddb(mm[0],mm[2]);
-	    mm[5]=_m_pand(mm[5],_m_load(&mask24lh));
-	    mm[0]=_m_pand(mm[0],_m_load(&mask24hl));
-	    mm[5]=_m_por(mm[5],mm[0]);
-	    if(finalize)
-		_m_movntq(&dstbase[0],mm[5]);
-	    else
-		_m_store(&dstbase[0],mm[5]);
-	}
-	dst += 6;
-    }
-#endif /* arch_x86 */
-	for(;x<w;x++){
-	    if(srca[x]){
-#ifdef FAST_OSD
-		dst[0]=dst[1]=dst[2]=src[x];
-#else
-		dst[0]=((dst[0]*srca[x])>>8)+src[x];
-		dst[1]=((dst[1]*srca[x])>>8)+src[x];
-		dst[2]=((dst[2]*srca[x])>>8)+src[x];
-#endif
-	    }
-	    dst+=3; // 24bpp
-	}
-	src+=srcstride;
-	srca+=srcstride;
-	dstbase+=dststride;
-    }
-#ifdef HAVE_INT_PVECTOR
-    _ivec_empty();
-    if(finalize) _ivec_sfence();
-#endif
-    return;
-}
-
-static inline void PVECTOR_RENAME(vo_draw_alpha_rgb32)(int w,int h,const unsigned char* src,const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
-    int y;
-PROFILE_START();
-    for(y=0;y<h;y++){
-	register int x=0;
-#ifdef HAVE_INT_PVECTOR
-	__m64 mm[8];
-	_ivec_prefetchw(&dstbase[x]);
-	_ivec_prefetch(&src[x]);
-	_ivec_prefetch(&srca[x]);
-	mm[7]=_mm_setzero_si64();
-	mm[6]=_mm_set1_pi8(0xFF);
-    for(;x<w;x+=2){
-     if(srca[x] || srca[x+1]) {
-	    _ivec_prefetchw(&dstbase[x+_ivec_size()*4]);
-	    _ivec_prefetch(&src[x+_ivec_size()*4]);
-	    _ivec_prefetch(&srca[x+_ivec_size()*4]);
-	    mm[0]=_m_load(&dstbase[4*x]);
-	    mm[1]=mm[0];
-	    mm[0]=_m_punpcklbw(mm[0],mm[7]);
-	    mm[1]=_m_punpckhbw(mm[1],mm[7]);
-	    mm[2]=_m_load_half(&srca[x]);
-	    mm[2]=_m_paddb(mm[2],mm[6]);
-	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
-	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
-	    mm[3]=mm[2];
-	    mm[2]=_m_punpcklbw(mm[2],mm[7]);
-	    mm[3]=_m_punpckhbw(mm[3],mm[7]);
-	    mm[0]=_m_pmullw(mm[0],mm[2]);
-	    mm[1]=_m_pmullw(mm[1],mm[3]);
-	    mm[0]=_m_psrlwi(mm[0],8);
-	    mm[1]=_m_psrlwi(mm[1],8);
-	    mm[0]=_m_packuswb(mm[0],mm[1]);
-	    mm[2]=_m_load_half(&src[x]);
-	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
-	    mm[2]=_m_punpcklbw(mm[2],mm[2]);
-	    mm[0]=_m_paddb(mm[0],mm[2]);
-	    if(finalize)
-		_m_movntq(&dstbase[4*x],mm[0]);
-	    else
-		_m_store(&dstbase[4*x],mm[0]);
-	}
-    }
-#endif /* arch_x86 */
-	for(;x<w;x++){
-	    if(srca[x]){
-#ifdef FAST_OSD
-		dstbase[4*x+0]=dstbase[4*x+1]=dstbase[4*x+2]=src[x];
-#else
-		dstbase[4*x+0]=((dstbase[4*x+0]*srca[x])>>8)+src[x];
-		dstbase[4*x+1]=((dstbase[4*x+1]*srca[x])>>8)+src[x];
-		dstbase[4*x+2]=((dstbase[4*x+2]*srca[x])>>8)+src[x];
-#endif
-	    }
-	}
-	src+=srcstride;
-	srca+=srcstride;
-	dstbase+=dststride;
-    }
-#ifdef HAVE_INT_PVECTOR
-    _ivec_empty();
-    if(finalize) _ivec_sfence();
-#endif
-PROFILE_END("vo_draw_alpha_rgb32");
-    return;
-}

Modified: mplayerxp/libvo/sub.cpp
===================================================================
--- mplayerxp/libvo/sub.cpp	2012-11-27 15:26:22 UTC (rev 464)
+++ mplayerxp/libvo/sub.cpp	2012-11-27 15:55:31 UTC (rev 465)
@@ -11,7 +11,6 @@
 #include "video_out.h"
 #include "font_load.h"
 #include "sub.h"
-#include "osd.h"
 #include "libmpsub/spudec.h"
 #include "libmpsub/vobsub.h"
 #define MSGT_CLASS MSGT_OSD

Modified: mplayerxp/libvo/sub.h
===================================================================
--- mplayerxp/libvo/sub.h	2012-11-27 15:26:22 UTC (rev 464)
+++ mplayerxp/libvo/sub.h	2012-11-27 15:55:31 UTC (rev 465)
@@ -1,7 +1,7 @@
 #ifndef __MPLAYER_SUB_H
 #define __MPLAYER_SUB_H
 #include <inttypes.h>
-#include "osd.h"
+#include "osd_render.h"
 
 typedef struct mp_osd_bbox_s {
     int x1,y1,x2,y2;

Modified: mplayerxp/libvo/video_out.cpp
===================================================================
--- mplayerxp/libvo/video_out.cpp	2012-11-27 15:26:22 UTC (rev 464)
+++ mplayerxp/libvo/video_out.cpp	2012-11-27 15:55:31 UTC (rev 465)
@@ -44,7 +44,7 @@
 #include "screenshot.h"
 #include "osdep/bswap.h"
 #include "dri_vo.h"
-#include "osd.h"
+#include "osd_render.h"
 #include "sub.h"
 #include "vo_msg.h"
 

Modified: mplayerxp/libvo/video_out_internal.h
===================================================================
--- mplayerxp/libvo/video_out_internal.h	2012-11-27 15:26:22 UTC (rev 464)
+++ mplayerxp/libvo/video_out_internal.h	2012-11-27 15:55:31 UTC (rev 465)
@@ -52,6 +52,4 @@
 	virtual MPXP_Rc ctrl(uint32_t request, any_t*data) = 0;
 };
 
-#include "osd.h"
-
 #endif

Modified: mplayerxp/libvo/vidix_system.cpp
===================================================================
--- mplayerxp/libvo/vidix_system.cpp	2012-11-27 15:26:22 UTC (rev 464)
+++ mplayerxp/libvo/vidix_system.cpp	2012-11-27 15:55:31 UTC (rev 465)
@@ -27,8 +27,6 @@
 
 #include "vidix_system.h"
 #include "osdep/fastmemcpy.h"
-#include "osd.h"
-#include "sub.h"
 #include "xmpcore/mp_image.h"
 #include "vo_msg.h"
 

Modified: mplayerxp/libvo/vo_fbdev.cpp
===================================================================
--- mplayerxp/libvo/vo_fbdev.cpp	2012-11-27 15:26:22 UTC (rev 464)
+++ mplayerxp/libvo/vo_fbdev.cpp	2012-11-27 15:55:31 UTC (rev 465)
@@ -34,7 +34,6 @@
 #include "vidix_system.h"
 #endif
 #include "aspect.h"
-#include "osd.h"
 #include "dri_vo.h"
 #include "libmpstream/mrl.h"
 #include "vo_msg.h"

Modified: mplayerxp/postproc/libmenu/menu.cpp
===================================================================
--- mplayerxp/postproc/libmenu/menu.cpp	2012-11-27 15:26:22 UTC (rev 464)
+++ mplayerxp/postproc/libmenu/menu.cpp	2012-11-27 15:55:31 UTC (rev 465)
@@ -11,7 +11,7 @@
 #include <fcntl.h>
 #include <unistd.h>
 
-#include "libvo/osd.h"
+#include "libvo/osd_render.h"
 #include "libvo/font_load.h"
 #include "libvo/sub.h"
 #include "osdep/keycodes.h"

Modified: mplayerxp/postproc/vf_expand.cpp
===================================================================
--- mplayerxp/postproc/vf_expand.cpp	2012-11-27 15:26:22 UTC (rev 464)
+++ mplayerxp/postproc/vf_expand.cpp	2012-11-27 15:55:31 UTC (rev 465)
@@ -16,7 +16,6 @@
 #ifdef OSD_SUPPORT
 #include "libvo/video_out.h"
 #include "libvo/sub.h"
-#include "libvo/osd.h"
 #endif
 #include "pp_msg.h"
 #include "mplayerxp.h" // vo_data

Modified: mplayerxp/postproc/vf_panscan.cpp
===================================================================
--- mplayerxp/postproc/vf_panscan.cpp	2012-11-27 15:26:22 UTC (rev 464)
+++ mplayerxp/postproc/vf_panscan.cpp	2012-11-27 15:55:31 UTC (rev 465)
@@ -17,7 +17,6 @@
 
 #ifdef OSD_SUPPORT
 #include "libvo/sub.h"
-#include "libvo/osd.h"
 #endif
 #include "pp_msg.h"
 

This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.