diff --git a/source/gl_skin.c b/source/gl_skin.c
index 53d17c9..ca3ecc0 100644
--- a/source/gl_skin.c
+++ b/source/gl_skin.c
@@ -38,99 +38,91 @@
 
 byte        player_8bit_texels[320 * 200];
 static byte translate[256];
+static unsigned int translate32[256];
 
 void
 Skin_Set_Translate (player_info_t *player)
 {
-	int top;
-	int bottom;
-	int i;
+	int         top;
+	int         bottom;
+	int         i;
 
 	top = bound (0, player->topcolor, 13) * 16;
 	bottom = bound (0, player->bottomcolor, 13) * 16;
 
-	for (i = 0; i < 16; i++) {
-		if (top < 128)	// the artists made some backwards ranges.  sigh.
-			translate[TOP_RANGE + i] = top + i;
+	if (VID_Is8bit ()) {				// 8bit texture upload
+		for (i = 0; i < 16; i++) {
+			if (top < 128)				// the artists made some backwards
+										// ranges.  sigh.
+				translate[TOP_RANGE + i] = top + i;
+			else
+				translate[TOP_RANGE + i] = top + 15 - i;
+
+			if (bottom < 128)
+				translate[BOTTOM_RANGE + i] = bottom + i;
+			else
+				translate[BOTTOM_RANGE + i] = bottom + 15 - i;
+		}
+	} else {
+		if (top < 128)					// the artists made some backwards
+										// ranges.  sigh.
+			memcpy (translate32 + TOP_RANGE, d_8to24table + top,
+
+					16 * sizeof (unsigned int));
 		else
-			translate[TOP_RANGE + i] = top + 15 - i;
+			for (i = 0; i < 16; i++)
+				translate32[TOP_RANGE + i] = top + 15 - i;
 
 		if (bottom < 128)
-			translate[BOTTOM_RANGE + i] = bottom + i;
+			memcpy (translate32 + BOTTOM_RANGE, d_8to24table + bottom,
+					16 * sizeof (unsigned int));
+
 		else
-			translate[BOTTOM_RANGE + i] = bottom + 15 - i;
+			for (i = 0; i < 16; i++)
+				translate32[BOTTOM_RANGE + i] = bottom + 15 - i;
 	}
 }
 
-void
-Skin_Do_Translation (player_info_t *player)
+static void
+build_skin_8 (byte * original, int tinwidth, int tinheight,
+			  unsigned int scaled_width, unsigned int scaled_height,
+			  int inwidth)
 {
-	int         i, j;
-	int         inwidth, inheight;
-	int         tinwidth, tinheight;
-	unsigned int scaled_width, scaled_height;
 	unsigned int frac, fracstep;
-	unsigned int translate32[256];
-	byte       *original;
-	unsigned int pixels[512 * 256], *out;
 	byte       *inrow;
+	byte        pixels[512 * 256], *out;
+	int         i, j;
 
-	// locate the original skin pixels
-	tinwidth = 296;		// real model width
-	tinheight = 194;	// real model height
-
-	if (!player->skin)
-		Skin_Find (player);
-	if ((original = Skin_Cache (player->skin)) != NULL) {
-		// skin data width
-		inwidth = 320;
-		inheight = 200;
-	} else {
-		original = player_8bit_texels;
-		inwidth = 296;
-		inheight = 194;
-	}
-
-	// because this happens during gameplay, do it fast
-	// instead of sending it through GL_Upload8()
-	glBindTexture (GL_TEXTURE_2D, playertextures + (player - cl.players));
-
-	// FIXME deek: This 512x256 limit sucks!
-	scaled_width = min (gl_max_size->int_val, 512);
-	scaled_height = min (gl_max_size->int_val, 256);
-
-	// allow users to crunch sizes down even more if they want
-	scaled_width >>= gl_playermip->int_val;
-	scaled_height >>= gl_playermip->int_val;
-
-	if (VID_Is8bit ()) {			// 8bit texture upload
-		byte	*out2;
-
-		out2 = (byte *) pixels;
-		memset (pixels, 0, sizeof (pixels));
-		fracstep = tinwidth * 0x10000 / scaled_width;
-		for (i = 0; i < scaled_height; i++, out2 += scaled_width) {
-			inrow = original + inwidth * (i * tinheight / scaled_height);
-			frac = fracstep >> 1;
-			for (j = 0; j < scaled_width; j += 4) {
-				out2[j] = translate[inrow[frac >> 16]];
-				frac += fracstep;
-				out2[j + 1] = translate[inrow[frac >> 16]];
-				frac += fracstep;
-				out2[j + 2] = translate[inrow[frac >> 16]];
-				frac += fracstep;
-				out2[j + 3] = translate[inrow[frac >> 16]];
-				frac += fracstep;
-			}
+	out = pixels;
+	memset (pixels, 0, sizeof (pixels));
+	fracstep = tinwidth * 0x10000 / scaled_width;
+	for (i = 0; i < scaled_height; i++, out += scaled_width) {
+		inrow = original + inwidth * (i * tinheight / scaled_height);
+		frac = fracstep >> 1;
+		for (j = 0; j < scaled_width; j += 4) {
+			out[j] = translate[inrow[frac >> 16]];
+			frac += fracstep;
+			out[j + 1] = translate[inrow[frac >> 16]];
+			frac += fracstep;
+			out[j + 2] = translate[inrow[frac >> 16]];
+			frac += fracstep;
+			out[j + 3] = translate[inrow[frac >> 16]];
+			frac += fracstep;
 		}
-
-		GL_Upload8_EXT ((byte *) pixels, scaled_width, scaled_height, false,
-						false);
-		return;
 	}
 
-	for (i = 0; i < 256; i++)
-		translate32[i] = d_8to24table[translate[i]];
+	GL_Upload8_EXT ((byte *) pixels, scaled_width, scaled_height, false, false);
+}
+
+static void
+build_skin_32 (byte * original, int tinwidth, int tinheight,
+			   unsigned int scaled_width, unsigned int scaled_height,
+			   int inwidth)
+{
+	unsigned int frac, fracstep;
+	byte       *inrow;
+	unsigned int pixels[512 * 256], *out;
+	int         i, j;
 
 	out = pixels;
 	memset (pixels, 0, sizeof (pixels));
@@ -158,11 +150,59 @@ Skin_Do_Translation (player_info_t *player)
 	glTexParameterf (GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
 }
 
+void
+Skin_Do_Translation (player_info_t *player)
+{
+	int         inwidth, inheight;
+	int         tinwidth, tinheight;
+	unsigned int scaled_width, scaled_height;
+	byte       *original;
+
+	// locate the original skin pixels
+	tinwidth = 296;						// real model width
+	tinheight = 194;					// real model height
+
+	if (!player->skin)
+		Skin_Find (player);
+	if ((original = Skin_Cache (player->skin)) != NULL) {
+		// skin data width
+		inwidth = 320;
+		inheight = 200;
+	} else {
+		original = player_8bit_texels;
+		inwidth = 296;
+		inheight = 194;
+	}
+
+	// because this happens during gameplay, do it fast
+	// instead of sending it through GL_Upload8()
+	glBindTexture (GL_TEXTURE_2D, playertextures + (player - cl.players));
+
+	// FIXME deek: This 512x256 limit sucks!
+	scaled_width = min (gl_max_size->int_val, 512);
+	scaled_height = min (gl_max_size->int_val, 256);
+
+	// allow users to crunch sizes down even more if they want
+	scaled_width >>= gl_playermip->int_val;
+	scaled_height >>= gl_playermip->int_val;
+
+	if (VID_Is8bit ()) {				// 8bit texture upload
+		build_skin_8 (original, tinwidth, tinheight, scaled_width,
+					  scaled_height, inwidth);
+	} else {
+		build_skin_32 (original, tinwidth, tinheight, scaled_width,
+					   scaled_height, inwidth);
+	}
+
+}
+
 void
 Skin_Init_Translation (void)
 {
-	int i;
+	int         i;
 
-	for (i=0; i < 256; i++)
-		translate [i] = i;
+	for (i = 0; i < 256; i++) {
+		translate[i] = i;
+		translate32[i] = d_8to24table[i];
+	}
 }