Finally committing support for compressed textures on the GPU (DDS+DXT). This patch started out as a patch by me, then cleaned up by Kupoman during his work on Cucumber.

One important thing to keep in mind when using this feature is that you'll need to flip your textures vertically (both the GIMP and Photoshop DDS tools I've seen have support for this on export). This is a quirk in using a texture format originally made for DirectX/DirectDraw, and flipping the compressed data is a real headache. Another quick fix for this issue is to change the Y value for the Size in the Mapping panel in the Texture properties to -1 (default is 1).

1 files changed, 104 insertions, 16 deletions

diff --git a/source/blender/gpu/intern/gpu_draw.c b/source/blender/gpu/intern/gpu_draw.c
index 4197e1a3edb..1e261e10ed6 100644
--- a/source/blender/gpu/intern/gpu_draw.c
+++ b/source/blender/gpu/intern/gpu_draw.c
@@ -427,8 +427,8 @@ int GPU_verify_image(Image *ima, ImageUser *iuser, int tftile, int compare, int
 	ImBuf *ibuf = NULL;
 	unsigned int *bind = NULL;
 	int rectw, recth, tpx=0, tpy=0, y;
-	unsigned int *tilerect= NULL, *scalerect= NULL, *rect= NULL;
-	float *ftilerect= NULL, *fscalerect = NULL, *frect = NULL;
+	unsigned int *tilerect= NULL, *rect= NULL;
+	float *ftilerect= NULL, *frect = NULL;
 	float *srgb_frect = NULL;
 	short texwindx, texwindy, texwinsx, texwinsy;
 	/* flag to determine whether high resolution format is used */
@@ -611,7 +611,32 @@ int GPU_verify_image(Image *ima, ImageUser *iuser, int tftile, int compare, int
 			rect= tilerect;
 		}
 	}
+#ifdef WITH_DDS
+	if (ibuf->ftype & DDS)
+		GPU_create_gl_tex_compressed(bind, rect, rectw, recth, mipmap, ima, ibuf);
+	else
+#endif
+		GPU_create_gl_tex(bind, rect, frect, rectw, recth, mipmap, use_high_bit_depth, ima);
+
+	/* clean up */
+	if (tilerect)
+		MEM_freeN(tilerect);
+	if (ftilerect)
+		MEM_freeN(ftilerect);
+	if (srgb_frect)
+		MEM_freeN(srgb_frect);
 
+	return *bind;
+}
+
+void GPU_create_gl_tex(unsigned int *bind, unsigned int *pix, float * frect, int rectw, int recth, int mipmap, int use_high_bit_depth, Image *ima)
+{
+	unsigned int *scalerect = NULL;
+    float *fscalerect = NULL;
+    
+    int tpx = rectw;
+    int tpy = recth;
+    
 	/* scale if not a power of two. this is not strictly necessary for newer 
 	 * GPUs (OpenGL version >= 2.0) since they support non-power-of-two-textures */
 	if (!is_pow2_limit(rectw) || !is_pow2_limit(recth)) {
@@ -626,9 +651,9 @@ int GPU_verify_image(Image *ima, ImageUser *iuser, int tftile, int compare, int
 		}
 		else {
 			scalerect= MEM_mallocN(rectw*recth*sizeof(*scalerect), "scalerect");
-			gluScaleImage(GL_RGBA, tpx, tpy, GL_UNSIGNED_BYTE, rect, rectw, recth, GL_UNSIGNED_BYTE, scalerect);
+			gluScaleImage(GL_RGBA, tpx, tpy, GL_UNSIGNED_BYTE, pix, rectw, recth, GL_UNSIGNED_BYTE, scalerect);
 
-			rect= scalerect;
+			pix= scalerect;
 		}
 	}
 
@@ -640,7 +665,7 @@ int GPU_verify_image(Image *ima, ImageUser *iuser, int tftile, int compare, int
 		if (use_high_bit_depth)
 			glTexImage2D(GL_TEXTURE_2D, 0,  GL_RGBA16,  rectw, recth, 0, GL_RGBA, GL_FLOAT, frect);
 		else
-			glTexImage2D(GL_TEXTURE_2D, 0,  GL_RGBA,  rectw, recth, 0, GL_RGBA, GL_UNSIGNED_BYTE, rect);
+			glTexImage2D(GL_TEXTURE_2D, 0,  GL_RGBA,  rectw, recth, 0, GL_RGBA, GL_UNSIGNED_BYTE, pix);
 		glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
 		glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, gpu_get_mipmap_filter(1));
 	}
@@ -649,14 +674,14 @@ int GPU_verify_image(Image *ima, ImageUser *iuser, int tftile, int compare, int
 			if (use_high_bit_depth)
 				glTexImage2D(GL_TEXTURE_2D, 0,  GL_RGBA16,  rectw, recth, 0, GL_RGBA, GL_FLOAT, frect);
 			else
-				glTexImage2D(GL_TEXTURE_2D, 0,  GL_RGBA,  rectw, recth, 0, GL_RGBA, GL_UNSIGNED_BYTE, rect);
+				glTexImage2D(GL_TEXTURE_2D, 0,  GL_RGBA,  rectw, recth, 0, GL_RGBA, GL_UNSIGNED_BYTE, pix);
 
 			glGenerateMipmapEXT(GL_TEXTURE_2D);
 		} else {
 			if (use_high_bit_depth)
 				gluBuild2DMipmaps(GL_TEXTURE_2D, GL_RGBA16, rectw, recth, GL_RGBA, GL_FLOAT, frect);
 			else
-				gluBuild2DMipmaps(GL_TEXTURE_2D, GL_RGBA, rectw, recth, GL_RGBA, GL_UNSIGNED_BYTE, rect);
+				gluBuild2DMipmaps(GL_TEXTURE_2D, GL_RGBA, rectw, recth, GL_RGBA, GL_UNSIGNED_BYTE, pix);
 		}
 		glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, gpu_get_mipmap_filter(0));
 		glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, gpu_get_mipmap_filter(1));
@@ -668,21 +693,84 @@ int GPU_verify_image(Image *ima, ImageUser *iuser, int tftile, int compare, int
 		glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_MAX_ANISOTROPY_EXT, GPU_get_anisotropic());
 	/* set to modulate with vertex color */
 	glTexEnvi(GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, GL_MODULATE);
-		
-	/* clean up */
-	if (tilerect)
-		MEM_freeN(tilerect);
-	if (ftilerect)
-		MEM_freeN(ftilerect);
+
 	if (scalerect)
 		MEM_freeN(scalerect);
 	if (fscalerect)
 		MEM_freeN(fscalerect);
-	if (srgb_frect)
-		MEM_freeN(srgb_frect);
-	return *bind;
 }
 
+/**
+ * GPU_upload_dxt_texture() assumes that the texture is already bound and ready to go.
+ * This is so the viewport and the BGE can share some code.
+ * Returns 0 if the provided ImBuf doesn't have a supported DXT compression format
+ */
+int GPU_upload_dxt_texture(ImBuf *ibuf)
+{
+	GLint format, err;
+	int blocksize, height, width, i, size, offset = 0;
+
+	height = ibuf->x;
+	width = ibuf->y;	
+	
+	if (ibuf->dds_data.fourcc == FOURCC_DXT1)
+		format = GL_COMPRESSED_RGBA_S3TC_DXT1_EXT;
+	else if (ibuf->dds_data.fourcc == FOURCC_DXT3)
+		format = GL_COMPRESSED_RGBA_S3TC_DXT3_EXT;
+	else if (ibuf->dds_data.fourcc == FOURCC_DXT5)
+		format = GL_COMPRESSED_RGBA_S3TC_DXT5_EXT;
+	else {
+		printf("Unable to find a suitable DXT compression, falling back to uncompressed\n");
+		return 0;
+	}
+
+	blocksize = (format == GL_COMPRESSED_RGBA_S3TC_DXT1_EXT) ? 8 : 16;
+	for (i=0; i<ibuf->dds_data.nummipmaps && (width||height); ++i) {
+		if (width == 0)
+			width = 1;
+		if (height == 0)
+			height = 1;
+
+		size = ((width+3)/4)*((height+3)/4)*blocksize;
+
+		glCompressedTexImage2D(GL_TEXTURE_2D, i, format, width, height,
+			0, size, ibuf->dds_data.data + offset);
+
+		err = glGetError();
+
+		if (err != GL_NO_ERROR)
+			printf("OpenGL error: %s\nFormat: %x\n", gluErrorString(err), format);
+
+		offset += size;
+		width >>= 1;
+		height >>= 1;
+	}
+
+	return 1;
+}
+
+void GPU_create_gl_tex_compressed(unsigned int *bind, unsigned int *pix, int x, int y, int mipmap, Image *ima, ImBuf *ibuf)
+{
+#ifndef WITH_DDS
+	// Fall back to uncompressed if DDS isn't enabled
+	GPU_create_gl_tex(bind, pix, NULL, x, y, mipmap, 0, ima);
+#else
+
+
+	glGenTextures(1, (GLuint *)bind);
+	glBindTexture(GL_TEXTURE_2D, *bind);
+
+	glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
+	glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, gpu_get_mipmap_filter(1));
+
+	glTexEnvi(GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, GL_MODULATE);
+
+	if (GPU_upload_dxt_texture(ibuf) == 0) {
+		glDeleteTextures(1, (GLuint*)bind);
+		GPU_create_gl_tex(bind, pix, NULL, x, y, mipmap, 0, ima);
+	}
+#endif
+}
 static void gpu_verify_repeat(Image *ima)
 {
 	/* set either clamp or repeat in X/Y */