diff --git a/postproc/rgb2rgb.c b/postproc/rgb2rgb.c
index f83d565dbb..923837a134 100644
--- a/postproc/rgb2rgb.c
+++ b/postproc/rgb2rgb.c
@@ -181,6 +181,27 @@ void palette8torgb32(uint8_t *src, uint8_t *dst, int src_size, uint8_t *palette)
 		((uint32_t *)dst)[i] = ((uint32_t *)palette)[ src[i] ];
 }
 
+/**
+ * Pallete is assumed to contain bgr32
+ */
+void palette8torgb24(uint8_t *src, uint8_t *dst, int src_size, uint8_t *palette)
+{
+	int i;
+/*
+	writes 1 byte o much and might cause alignment issues on some architectures?
+	for(i=0; i<src_size; i++)
+		((uint32_t *)(&dst[i*3])) = ((uint32_t *)palette)[ src[i] ];
+*/
+	for(i=0; i<src_size; i++)
+	{
+		//FIXME slow?
+		dst[0]= palette[ src[i]*4+0 ];
+		dst[1]= palette[ src[i]*4+1 ];
+		dst[2]= palette[ src[i]*4+2 ];
+		dst+= 3;
+	}
+}
+
 void rgb32to16(uint8_t *src, uint8_t *dst, int src_size)
 {
 	int i;
@@ -226,4 +247,4 @@ void palette8torgb15(uint8_t *src, uint8_t *dst, int src_size, uint8_t *palette)
 	int i;
 	for(i=0; i<src_size; i++)
 		((uint16_t *)dst)[i] = ((uint16_t *)palette)[ src[i] ];
-}
\ No newline at end of file
+}
diff --git a/postproc/rgb2rgb.h b/postproc/rgb2rgb.h
index db6aff5f80..0f7c8141a0 100644
--- a/postproc/rgb2rgb.h
+++ b/postproc/rgb2rgb.h
@@ -16,6 +16,6 @@ void rgb32to15(uint8_t *src, uint8_t *dst, int src_size);
 void palette8torgb32(uint8_t *src, uint8_t *dst, int src_size, uint8_t *palette);
 void palette8torgb16(uint8_t *src, uint8_t *dst, int src_size, uint8_t *palette);
 void palette8torgb15(uint8_t *src, uint8_t *dst, int src_size, uint8_t *palette);
-
+void palette8torgb24(uint8_t *src, uint8_t *dst, int src_size, uint8_t *palette);
 
 #endif
diff --git a/postproc/rgb2rgb_template.c b/postproc/rgb2rgb_template.c
index f83d565dbb..923837a134 100644
--- a/postproc/rgb2rgb_template.c
+++ b/postproc/rgb2rgb_template.c
@@ -181,6 +181,27 @@ void palette8torgb32(uint8_t *src, uint8_t *dst, int src_size, uint8_t *palette)
 		((uint32_t *)dst)[i] = ((uint32_t *)palette)[ src[i] ];
 }
 
+/**
+ * Pallete is assumed to contain bgr32
+ */
+void palette8torgb24(uint8_t *src, uint8_t *dst, int src_size, uint8_t *palette)
+{
+	int i;
+/*
+	writes 1 byte o much and might cause alignment issues on some architectures?
+	for(i=0; i<src_size; i++)
+		((uint32_t *)(&dst[i*3])) = ((uint32_t *)palette)[ src[i] ];
+*/
+	for(i=0; i<src_size; i++)
+	{
+		//FIXME slow?
+		dst[0]= palette[ src[i]*4+0 ];
+		dst[1]= palette[ src[i]*4+1 ];
+		dst[2]= palette[ src[i]*4+2 ];
+		dst+= 3;
+	}
+}
+
 void rgb32to16(uint8_t *src, uint8_t *dst, int src_size)
 {
 	int i;
@@ -226,4 +247,4 @@ void palette8torgb15(uint8_t *src, uint8_t *dst, int src_size, uint8_t *palette)
 	int i;
 	for(i=0; i<src_size; i++)
 		((uint16_t *)dst)[i] = ((uint16_t *)palette)[ src[i] ];
-}
\ No newline at end of file
+}