mplayer idependant (not really yet) swscale example

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9922 b3059339-0415-0410-9bf9-f77b7e298cf2

4 files changed, 207 insertions, 133 deletions

diff --git a/postproc/Makefile b/postproc/Makefile
index 6537ef7aa4..deb1e9a263 100644
--- a/postproc/Makefile
+++ b/postproc/Makefile
@@ -36,6 +36,8 @@ depend:
 cs_test: $(CS_TEST_OBJS)
 	$(CC) $(CS_TEST_OBJS) -o cs_test
 
+swscale-example: swscale-example.o $(SWSLIB)
+	$(CC) swscale-example.o $(SWSLIB) ../cpudetect.o  ../libmpcodecs/img_format.o -lm -o swscale-example -W -Wall
 #
 # include dependency files if they exist
 #
diff --git a/postproc/swscale-example.c b/postproc/swscale-example.c
new file mode 100644
index 0000000000..f83fe0ce07
--- /dev/null
+++ b/postproc/swscale-example.c
@@ -0,0 +1,204 @@
+/*
+    Copyright (C) 2003 Michael Niedermayer <michaelni@gmx.at>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program 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 General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+*/
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <inttypes.h>
+#include <stdarg.h>
+
+#include "swscale.h"
+#include "../libvo/img_format.h"
+#include "../cpudetect.h"
+
+static int testFormat[]={
+IMGFMT_YVU9,
+IMGFMT_YV12,
+//IMGFMT_IYUV,
+IMGFMT_I420,
+IMGFMT_BGR15,
+IMGFMT_BGR16,
+IMGFMT_BGR24,
+IMGFMT_BGR32,
+IMGFMT_RGB24,
+IMGFMT_RGB32,
+//IMGFMT_Y8,
+IMGFMT_Y800,
+//IMGFMT_YUY2,
+0
+};
+
+static uint64_t getSSD(uint8_t *src1, uint8_t *src2, int stride1, int stride2, int w, int h){
+	int x,y;
+	uint64_t ssd=0;
+
+//printf("%d %d\n", w, h);
+        
+	for(y=0; y<h; y++){
+		for(x=0; x<w; x++){
+			int d= src1[x + y*stride1] - src2[x + y*stride2];
+			ssd+= d*d;
+//printf("%d", abs(src1[x + y*stride1] - src2[x + y*stride2])/26 );
+		}
+//printf("\n");
+	}
+	return ssd;
+}
+
+// test by ref -> src -> dst -> out & compare out against ref
+// ref & out are YV12
+static void doTest(uint8_t *ref[3], int refStride[3], int w, int h, int srcFormat, int dstFormat, 
+                   int srcW, int srcH, int dstW, int dstH, int flags){
+	uint8_t *src[3];
+	uint8_t *dst[3];
+	uint8_t *out[3];
+	int srcStride[3], dstStride[3];
+	int i;
+	uint64_t ssdY, ssdU, ssdV;
+	struct SwsContext *srcContext, *dstContext, *outContext;
+	
+	for(i=0; i<3; i++){
+		// avoid stride % bpp != 0
+		if(srcFormat==IMGFMT_RGB24 || srcFormat==IMGFMT_BGR24)
+			srcStride[i]= srcW*3;
+		else
+			srcStride[i]= srcW*4;
+		
+		if(dstFormat==IMGFMT_RGB24 || dstFormat==IMGFMT_BGR24)
+			dstStride[i]= dstW*3;
+		else
+			dstStride[i]= dstW*4;
+	
+		src[i]= (uint8_t*) malloc(srcStride[i]*srcH);
+		dst[i]= (uint8_t*) malloc(dstStride[i]*dstH);
+		out[i]= (uint8_t*) malloc(refStride[i]*h);
+	}
+
+	srcContext= sws_getContext(w, h, IMGFMT_YV12, srcW, srcH, srcFormat, flags, NULL, NULL);
+	dstContext= sws_getContext(srcW, srcH, srcFormat, dstW, dstH, dstFormat, flags, NULL, NULL);
+	outContext= sws_getContext(dstW, dstH, dstFormat, w, h, IMGFMT_YV12, flags, NULL, NULL);
+	if(srcContext==NULL ||dstContext==NULL ||outContext==NULL){
+		printf("Failed allocating swsContext\n");
+		goto end;
+	}
+//	printf("test %X %X %X -> %X %X %X\n", (int)ref[0], (int)ref[1], (int)ref[2],
+//		(int)src[0], (int)src[1], (int)src[2]);
+
+	sws_scale(srcContext, ref, refStride, 0, h   , src, srcStride);
+	sws_scale(dstContext, src, srcStride, 0, srcH, dst, dstStride);
+	sws_scale(outContext, dst, dstStride, 0, dstH, out, refStride);
+asm volatile ("emms\n\t");
+	     
+	ssdY= getSSD(ref[0], out[0], refStride[0], refStride[0], w, h);
+	ssdU= getSSD(ref[1], out[1], refStride[1], refStride[1], (w+1)>>1, (h+1)>>1);
+	ssdV= getSSD(ref[2], out[2], refStride[2], refStride[2], (w+1)>>1, (h+1)>>1);
+	
+	if(srcFormat == IMGFMT_Y800 || dstFormat==IMGFMT_Y800) ssdU=ssdV=0; //FIXME check that output is really gray
+	
+	ssdY/= w*h;
+	ssdU/= w*h/4;
+	ssdV/= w*h/4;
+	
+	if(ssdY>100 || ssdU>100 || ssdV>100){
+		printf(" %s %dx%d -> %s %4dx%4d flags=%2d SSD=%5lld,%5lld,%5lld\n", 
+			vo_format_name(srcFormat), srcW, srcH, 
+			vo_format_name(dstFormat), dstW, dstH,
+			flags,
+			ssdY, ssdU, ssdV);
+	}
+
+	end:
+	
+	sws_freeContext(srcContext);
+	sws_freeContext(dstContext);
+	sws_freeContext(outContext);
+
+	for(i=0; i<3; i++){
+		free(src[i]);
+		free(dst[i]);
+		free(out[i]);
+	}
+}
+
+void mp_msg_c( int x, const char *format, ... ){
+    va_list va;
+    va_start(va, format);
+    vfprintf(stderr, format, va);
+    va_end(va);
+}
+
+int verbose=0; //FIXME
+void fast_memcpy(void *a, void *b, int s){ //FIXME
+    memcpy(a, b, s);
+}
+
+static void selfTest(uint8_t *src[3], int stride[3], int w, int h){
+	int srcFormat, dstFormat, srcFormatIndex, dstFormatIndex;
+	int srcW, srcH, dstW, dstH;
+	int flags;
+
+	for(srcFormatIndex=0; ;srcFormatIndex++){
+		srcFormat= testFormat[srcFormatIndex];
+		if(!srcFormat) break;
+		for(dstFormatIndex=0; ;dstFormatIndex++){
+			dstFormat= testFormat[dstFormatIndex];
+			if(!dstFormat) break;
+//			if(!isSupportedOut(dstFormat)) continue;
+printf("%s -> %s\n", 
+	vo_format_name(srcFormat),
+	vo_format_name(dstFormat));
+ 
+			srcW= w;
+			srcH= h;
+			for(dstW=w - w/3; dstW<= 4*w/3; dstW+= w/3){
+				for(dstH=h - h/3; dstH<= 4*h/3; dstH+= h/3){
+					for(flags=1; flags<33; flags*=2)
+						doTest(src, stride, w, h, srcFormat, dstFormat,
+							srcW, srcH, dstW, dstH, flags);
+				}
+			}
+		}
+	}
+}
+
+#define W 96
+#define H 96
+
+int main(int argc, char **argv){
+	uint8_t rgb_data[W*H*4];
+	uint8_t *rgb_src[3]= {rgb_data, NULL, NULL};
+	int rgb_stride[3]={4*W, 0, 0};
+	uint8_t data[3][W*H];
+	uint8_t *src[3]= {data[0], data[1], data[2]};
+	int stride[3]={W, W, W};
+	int x, y;
+	struct SwsContext *sws;
+	GetCpuCaps(&gCpuCaps);
+
+	sws= sws_getContext(W/12, H/12, IMGFMT_BGR32, W, H, IMGFMT_YV12, 2, NULL, NULL);
+        
+	for(y=0; y<H; y++){
+		for(x=0; x<W*4; x++){
+			rgb_data[ x + y*4*W]= random();
+		}
+	}
+
+	sws_scale(sws, rgb_src, rgb_stride, 0, H   , src, stride);
+asm volatile ("emms\n\t");
+	selfTest(src,  stride, W, H);
+}
diff --git a/postproc/swscale.c b/postproc/swscale.c
index a3b20cfe67..910fa3f3a7 100644
--- a/postproc/swscale.c
+++ b/postproc/swscale.c
@@ -247,139 +247,6 @@ void in_asm_used_var_warning_killer()
 }
 #endif
 
-static int testFormat[]={
-IMGFMT_YVU9,
-IMGFMT_YV12,
-//IMGFMT_IYUV,
-IMGFMT_I420,
-IMGFMT_BGR15,
-IMGFMT_BGR16,
-IMGFMT_BGR24,
-IMGFMT_BGR32,
-IMGFMT_RGB24,
-IMGFMT_RGB32,
-//IMGFMT_Y8,
-IMGFMT_Y800,
-//IMGFMT_YUY2,
-0
-};
-
-static uint64_t getSSD(uint8_t *src1, uint8_t *src2, int stride1, int stride2, int w, int h){
-	int x,y;
-	uint64_t ssd=0;
-
-	for(y=0; y<h; y++){
-		for(x=0; x<w; x++){
-			int d= src1[x + y*stride1] - src2[x + y*stride2];
-			ssd+= d*d;
-		}
-	}
-	return ssd;
-}
-
-// test by ref -> src -> dst -> out & compare out against ref
-// ref & out are YV12
-static void doTest(uint8_t *ref[3], int refStride[3], int w, int h, int srcFormat, int dstFormat, 
-                   int srcW, int srcH, int dstW, int dstH, int flags){
-	uint8_t *src[3];
-	uint8_t *dst[3];
-	uint8_t *out[3];
-	int srcStride[3], dstStride[3];
-	int i;
-	uint64_t ssdY, ssdU, ssdV;
-	SwsContext *srcContext, *dstContext, *outContext;
-	
-	for(i=0; i<3; i++){
-		// avoid stride % bpp != 0
-		if(srcFormat==IMGFMT_RGB24 || srcFormat==IMGFMT_BGR24)
-			srcStride[i]= srcW*3;
-		else
-			srcStride[i]= srcW*4;
-		
-		if(dstFormat==IMGFMT_RGB24 || dstFormat==IMGFMT_BGR24)
-			dstStride[i]= dstW*3;
-		else
-			dstStride[i]= dstW*4;
-	
-		src[i]= malloc(srcStride[i]*srcH);
-		dst[i]= malloc(dstStride[i]*dstH);
-		out[i]= malloc(refStride[i]*h);
-	}
-
-	srcContext= sws_getContext(w, h, IMGFMT_YV12, srcW, srcH, srcFormat, flags, NULL, NULL);
-	dstContext= sws_getContext(srcW, srcH, srcFormat, dstW, dstH, dstFormat, flags, NULL, NULL);
-	outContext= sws_getContext(dstW, dstH, dstFormat, w, h, IMGFMT_YV12, flags, NULL, NULL);
-	if(srcContext==NULL ||dstContext==NULL ||outContext==NULL){
-		printf("Failed allocating swsContext\n");
-		goto end;
-	}
-//	printf("test %X %X %X -> %X %X %X\n", (int)ref[0], (int)ref[1], (int)ref[2],
-//		(int)src[0], (int)src[1], (int)src[2]);
-
-	srcContext->swScale(srcContext, ref, refStride, 0, h   , src, srcStride);
-	dstContext->swScale(dstContext, src, srcStride, 0, srcH, dst, dstStride);
-	outContext->swScale(outContext, dst, dstStride, 0, dstH, out, refStride);
-	     
-	ssdY= getSSD(ref[0], out[0], refStride[0], refStride[0], w, h);
-	ssdU= getSSD(ref[1], out[1], refStride[1], refStride[1], (w+1)>>1, (h+1)>>1);
-	ssdV= getSSD(ref[2], out[2], refStride[2], refStride[2], (w+1)>>1, (h+1)>>1);
-	
-	if(isGray(srcFormat) || isGray(dstFormat)) ssdU=ssdV=0; //FIXME check that output is really gray
-	
-	ssdY/= w*h;
-	ssdU/= w*h/4;
-	ssdV/= w*h/4;
-	
-	if(ssdY>100 || ssdU>50 || ssdV>50){
-		printf(" %s %dx%d -> %s %4dx%4d flags=%2d SSD=%5lld,%5lld,%5lld\n", 
-			vo_format_name(srcFormat), srcW, srcH, 
-			vo_format_name(dstFormat), dstW, dstH,
-			flags,
-			ssdY, ssdU, ssdV);
-	}
-
-	end:
-	
-	sws_freeContext(srcContext);
-	sws_freeContext(dstContext);
-	sws_freeContext(outContext);
-
-	for(i=0; i<3; i++){
-		free(src[i]);
-		free(dst[i]);
-		free(out[i]);
-	}
-}
-
-static void selfTest(uint8_t *src[3], int stride[3], int w, int h){
-	int srcFormat, dstFormat, srcFormatIndex, dstFormatIndex;
-	int srcW, srcH, dstW, dstH;
-	int flags;
-
-	for(srcFormatIndex=0; ;srcFormatIndex++){
-		srcFormat= testFormat[srcFormatIndex];
-		if(!srcFormat) break;
-		for(dstFormatIndex=0; ;dstFormatIndex++){
-			dstFormat= testFormat[dstFormatIndex];
-			if(!dstFormat) break;
-			if(!isSupportedOut(dstFormat)) continue;
-printf("%s -> %s\n", 
-	vo_format_name(srcFormat),
-	vo_format_name(dstFormat));
-
-			srcW= w+w/3;
-			srcH= h+h/3;
-			for(dstW=w; dstW<w*2; dstW+= dstW/3){
-				for(dstH=h; dstH<h*2; dstH+= dstH/3){
-					for(flags=1; flags<33; flags*=2)
-						doTest(src, stride, w, h, srcFormat, dstFormat,
-							srcW, srcH, dstW, dstH, flags);
-				}
-			}
-		}
-	}
-}
-
 static inline void yuv2yuvXinC(int16_t *lumFilter, int16_t **lumSrc, int lumFilterSize,
 				    int16_t *chrFilter, int16_t **chrSrc, int chrFilterSize,
 				    uint8_t *dest, uint8_t *uDest, uint8_t *vDest, int dstW, int chrDstW)
diff --git a/postproc/swscale_template.c b/postproc/swscale_template.c
index d3c7e1d610..760a4dc3d5 100644
--- a/postproc/swscale_template.c
+++ b/postproc/swscale_template.c
@@ -2005,6 +2005,7 @@ static inline void RENAME(hScale)(int16_t *dst, int dstW, uint8_t *src, int srcW
 				  int16_t *filter, int16_t *filterPos, int filterSize)
 {
 #ifdef HAVE_MMX
+	assert(filterSize % 4 == 0 && filterSize>0);
 	if(filterSize==4) // allways true for upscaling, sometimes for down too
 	{
 		int counter= -2*dstW;