diff --git a/src/Renderer/Blitter.cpp b/src/Renderer/Blitter.cpp
index a6773fc..7e7ca5e 100644
--- a/src/Renderer/Blitter.cpp
+++ b/src/Renderer/Blitter.cpp
@@ -30,7 +30,7 @@
 		delete blitCache;
 	}
 
-	void Blitter::clear(void* pixel, sw::Format format, Surface *dest, const SliceRect &dRect, unsigned int rgbaMask)
+	void Blitter::clear(void *pixel, sw::Format format, Surface *dest, const SliceRect &dRect, unsigned int rgbaMask)
 	{
 		if(fastClear(pixel, format, dest, dRect, rgbaMask))
 		{
@@ -44,7 +44,7 @@
 		delete color;
 	}
 
-	bool Blitter::fastClear(void* pixel, sw::Format format, Surface *dest, const SliceRect &dRect, unsigned int rgbaMask)
+	bool Blitter::fastClear(void *pixel, sw::Format format, Surface *dest, const SliceRect &dRect, unsigned int rgbaMask)
 	{
 		if(format != FORMAT_A32B32G32R32F)
 		{
@@ -99,26 +99,33 @@
 			return false;
 		}
 
-		uint8_t *d = (uint8_t*)dest->lockInternal(dRect.x0, dRect.y0, dRect.slice, sw::LOCK_WRITEONLY, sw::PUBLIC);
+		uint8_t *slice = (uint8_t*)dest->lockInternal(dRect.x0, dRect.y0, dRect.slice, sw::LOCK_WRITEONLY, sw::PUBLIC);
 
-		switch(Surface::bytes(dest->getFormat()))
+		for(int j = 0; j < dest->getSamples(); j++)
 		{
-		case 2:
-			for(int i = dRect.y0; i < dRect.y1; i++)
+			uint8_t *d = slice;
+
+			switch(Surface::bytes(dest->getFormat()))
 			{
-				sw::clear((uint16_t*)d, packed, dRect.x1 - dRect.x0);
-				d += dest->getInternalPitchB();
+			case 2:
+				for(int i = dRect.y0; i < dRect.y1; i++)
+				{
+					sw::clear((uint16_t*)d, packed, dRect.x1 - dRect.x0);
+					d += dest->getInternalPitchB();
+				}
+				break;
+			case 4:
+				for(int i = dRect.y0; i < dRect.y1; i++)
+				{
+					sw::clear((uint32_t*)d, packed, dRect.x1 - dRect.x0);
+					d += dest->getInternalPitchB();
+				}
+				break;
+			default:
+				assert(false);
 			}
-			break;
-		case 4:
-			for(int i = dRect.y0; i < dRect.y1; i++)
-			{
-				sw::clear((uint32_t*)d, packed, dRect.x1 - dRect.x0);
-				d += dest->getInternalPitchB();
-			}
-			break;
-		default:
-			assert(false);
+
+			slice += dest->getInternalSliceB();
 		}
 
 		dest->unlockInternal();
@@ -1201,6 +1208,7 @@
 				For(Int i = x0d, i < x1d, i++)
 				{
 					Pointer<Byte> d = destLine + (dstQuadLayout ? (((j & Int(1)) << 1) + (i * 2) - (i & Int(1))) : RValue<Int>(i)) * dstBytes;
+
 					if(hasConstantColorI)
 					{
 						if(!write(constantColorI, d, state.destFormat, state.options))
@@ -1281,10 +1289,20 @@
 							        (c10 * ix + c11 * fx) * fy;
 						}
 
-						if(!ApplyScaleAndClamp(color, state) || !write(color, d, state.destFormat, state.options))
+						if(!ApplyScaleAndClamp(color, state))
 						{
 							return nullptr;
 						}
+
+						for(int s = 0; s < state.destSamples; s++)
+						{
+							if(!write(color, d, state.destFormat, state.options))
+							{
+								return nullptr;
+							}
+
+							d += *Pointer<Int>(blit + OFFSET(BlitData,dSliceB));
+						}
 					}
 
 					if(!hasConstantColorI && !hasConstantColorF) { x += w; }
@@ -1322,6 +1340,7 @@
 
 		state.sourceFormat = isStencil ? source->getStencilFormat() : source->getFormat(useSourceInternal);
 		state.destFormat = isStencil ? dest->getStencilFormat() : dest->getFormat(useDestInternal);
+		state.destSamples = dest->getSamples();
 		state.options = options;
 
 		criticalSection.lock();
@@ -1355,6 +1374,7 @@
 		                        dest->lock(0, 0, destRect.slice, isRGBA ? (isEntireDest ? sw::LOCK_DISCARD : sw::LOCK_WRITEONLY) : sw::LOCK_READWRITE, sw::PUBLIC, useDestInternal);
 		data.sPitchB = isStencil ? source->getStencilPitchB() : source->getPitchB(useSourceInternal);
 		data.dPitchB = isStencil ? dest->getStencilPitchB() : dest->getPitchB(useDestInternal);
+		data.dSliceB = isStencil ? dest->getStencilSliceB() : dest->getSliceB(useDestInternal);
 
 		data.w = sRect.width() / dRect.width();
 		data.h = sRect.height() / dRect.height();
diff --git a/src/Renderer/Blitter.hpp b/src/Renderer/Blitter.hpp
index ce7c789..6e87db8 100644
--- a/src/Renderer/Blitter.hpp
+++ b/src/Renderer/Blitter.hpp
@@ -47,6 +47,7 @@
 
 			Format sourceFormat;
 			Format destFormat;
+			int destSamples;
 			Blitter::Options options;
 		};
 
@@ -56,6 +57,7 @@
 			void *dest;
 			int sPitchB;
 			int dPitchB;
+			int dSliceB;
 
 			float x0;
 			float y0;
@@ -75,12 +77,12 @@
 		Blitter();
 		virtual ~Blitter();
 
-		void clear(void* pixel, sw::Format format, Surface *dest, const SliceRect &dRect, unsigned int rgbaMask);
+		void clear(void *pixel, sw::Format format, Surface *dest, const SliceRect &dRect, unsigned int rgbaMask);
 		void blit(Surface *source, const SliceRectF &sRect, Surface *dest, const SliceRect &dRect, bool filter, bool isStencil = false);
 		void blit3D(Surface *source, Surface *dest);
 
 	private:
-		bool fastClear(void* pixel, sw::Format format, Surface *dest, const SliceRect &dRect, unsigned int rgbaMask);
+		bool fastClear(void *pixel, sw::Format format, Surface *dest, const SliceRect &dRect, unsigned int rgbaMask);
 
 		bool read(Float4 &color, Pointer<Byte> element, Format format);
 		bool write(Float4 &color, Pointer<Byte> element, Format format, const Blitter::Options& options);
diff --git a/src/Renderer/Renderer.cpp b/src/Renderer/Renderer.cpp
index fd0c340..0b95c72 100644
--- a/src/Renderer/Renderer.cpp
+++ b/src/Renderer/Renderer.cpp
@@ -606,7 +606,8 @@
 
 					if(draw->renderTarget[index])
 					{
-						data->colorBuffer[index] = (unsigned int*)context->renderTarget[index]->lockInternal(0, 0, q * ms, LOCK_READWRITE, MANAGED);
+						data->colorBuffer[index] = (unsigned int*)context->renderTarget[index]->lockInternal(0, 0, 0, LOCK_READWRITE, MANAGED);
+						data->colorBuffer[index] += q * ms * context->renderTarget[index]->getSliceB(true);
 						data->colorPitchB[index] = context->renderTarget[index]->getInternalPitchB();
 						data->colorSliceB[index] = context->renderTarget[index]->getInternalSliceB();
 					}
@@ -617,14 +618,16 @@
 
 				if(draw->depthBuffer)
 				{
-					data->depthBuffer = (float*)context->depthBuffer->lockInternal(0, 0, q * ms, LOCK_READWRITE, MANAGED);
+					data->depthBuffer = (float*)context->depthBuffer->lockInternal(0, 0, 0, LOCK_READWRITE, MANAGED);
+					data->depthBuffer += q * ms * context->depthBuffer->getSliceB(true);
 					data->depthPitchB = context->depthBuffer->getInternalPitchB();
 					data->depthSliceB = context->depthBuffer->getInternalSliceB();
 				}
 
 				if(draw->stencilBuffer)
 				{
-					data->stencilBuffer = (unsigned char*)context->stencilBuffer->lockStencil(0, 0, q * ms, MANAGED);
+					data->stencilBuffer = (unsigned char*)context->stencilBuffer->lockStencil(0, 0, 0, MANAGED);
+					data->stencilBuffer += q * ms * context->stencilBuffer->getSliceB(true);
 					data->stencilPitchB = context->stencilBuffer->getStencilPitchB();
 					data->stencilSliceB = context->stencilBuffer->getStencilSliceB();
 				}
@@ -673,13 +676,7 @@
 
 	void Renderer::clear(void *value, Format format, Surface *dest, const Rect &clearRect, unsigned int rgbaMask)
 	{
-		SliceRect rect = clearRect;
-		int samples = dest->getDepth();
-
-		for(rect.slice = 0; rect.slice < samples; rect.slice++)
-		{
-			blitter->clear(value, format, dest, rect, rgbaMask);
-		}
+		blitter->clear(value, format, dest, clearRect, rgbaMask);
 	}
 
 	void Renderer::blit(Surface *source, const SliceRectF &sRect, Surface *dest, const SliceRect &dRect, bool filter, bool isStencil)
diff --git a/src/Renderer/Surface.cpp b/src/Renderer/Surface.cpp
index b15102d..b758bad 100644
--- a/src/Renderer/Surface.cpp
+++ b/src/Renderer/Surface.cpp
@@ -44,16 +44,24 @@
 
 	void Surface::Buffer::write(int x, int y, int z, const Color<float> &color)
 	{
-		void *element = (unsigned char*)buffer + (x + border) * bytes + (y + border) * pitchB + z * sliceB;
+		byte *element = (byte*)buffer + (x + border) * bytes + (y + border) * pitchB + z * samples * sliceB;
 
-		write(element, color);
+		for(int i = 0; i < samples; i++)
+		{
+			write(element, color);
+			element += sliceB;
+		}
 	}
 
 	void Surface::Buffer::write(int x, int y, const Color<float> &color)
 	{
-		void *element = (unsigned char*)buffer + (x + border) * bytes + (y + border) * pitchB;
+		byte *element = (byte*)buffer + (x + border) * bytes + (y + border) * pitchB;
 
-		write(element, color);
+		for(int i = 0; i < samples; i++)
+		{
+			write(element, color);
+			element += sliceB;
+		}
 	}
 
 	inline void Surface::Buffer::write(void *element, const Color<float> &color)
@@ -365,7 +373,7 @@
 
 	Color<float> Surface::Buffer::read(int x, int y, int z) const
 	{
-		void *element = (unsigned char*)buffer + (x + border) * bytes + (y + border) * pitchB + z * sliceB;
+		void *element = (unsigned char*)buffer + (x + border) * bytes + (y + border) * pitchB + z * samples * sliceB;
 
 		return read(element);
 	}
@@ -1039,7 +1047,7 @@
 		return c000 + c100 + c010 + c110 + c001 + c101 + c011 + c111;
 	}
 
-	Color<float> Surface::Buffer::sample(float x, float y) const
+	Color<float> Surface::Buffer::sample(float x, float y, int layer) const
 	{
 		x -= 0.5f;
 		y -= 0.5f;
@@ -1050,10 +1058,10 @@
 		int y0 = clamp((int)y, 0, height - 1);
 		int y1 = (y0 + 1 >= height) ? y0 : y0 + 1;
 
-		Color<float> c00 = read(x0, y0);
-		Color<float> c10 = read(x1, y0);
-		Color<float> c01 = read(x0, y1);
-		Color<float> c11 = read(x1, y1);
+		Color<float> c00 = read(x0, y0, layer);
+		Color<float> c10 = read(x1, y0, layer);
+		Color<float> c01 = read(x0, y1, layer);
+		Color<float> c11 = read(x1, y1, layer);
 
 		float fx = x - x0;
 		float fy = y - y0;
@@ -1156,11 +1164,11 @@
 			case FORMAT_ATI2:
 				return (unsigned char*)buffer + 16 * (x / 4) + (y / 4) * pitchB + z * sliceB;
 			default:
-				return (unsigned char*)buffer + x * bytes + y * pitchB + z * sliceB;
+				return (unsigned char*)buffer + x * bytes + y * pitchB + z * samples * sliceB;
 			}
 		}
 
-		return 0;
+		return nullptr;
 	}
 
 	void Surface::Buffer::unlockRect()
@@ -1173,8 +1181,8 @@
 	public:
 		SurfaceImplementation(int width, int height, int depth, Format format, void *pixels, int pitch, int slice)
 			: Surface(width, height, depth, format, pixels, pitch, slice) {}
-		SurfaceImplementation(Resource *texture, int width, int height, int depth, int border, Format format, bool lockable, bool renderTarget, int pitchP = 0)
-			: Surface(texture, width, height, depth, border, format, lockable, renderTarget, pitchP) {}
+		SurfaceImplementation(Resource *texture, int width, int height, int depth, int border, int samples, Format format, bool lockable, bool renderTarget, int pitchP = 0)
+			: Surface(texture, width, height, depth, border, samples, format, lockable, renderTarget, pitchP) {}
 		~SurfaceImplementation() override {};
 
 		void *lockInternal(int x, int y, int z, Lock lock, Accessor client) override
@@ -1193,9 +1201,9 @@
 		return new SurfaceImplementation(width, height, depth, format, pixels, pitch, slice);
 	}
 
-	Surface *Surface::create(Resource *texture, int width, int height, int depth, int border, Format format, bool lockable, bool renderTarget, int pitchPprovided)
+	Surface *Surface::create(Resource *texture, int width, int height, int depth, int border, int samples, Format format, bool lockable, bool renderTarget, int pitchPprovided)
 	{
-		return new SurfaceImplementation(texture, width, height, depth, border, format, lockable, renderTarget, pitchPprovided);
+		return new SurfaceImplementation(texture, width, height, depth, border, samples, format, lockable, renderTarget, pitchPprovided);
 	}
 
 	Surface::Surface(int width, int height, int depth, Format format, void *pixels, int pitch, int slice) : lockable(true), renderTarget(false)
@@ -1209,6 +1217,7 @@
 		external.width = width;
 		external.height = height;
 		external.depth = depth;
+		external.samples = 1;
 		external.format = format;
 		external.bytes = bytes(external.format);
 		external.pitchB = pitch;
@@ -1223,6 +1232,7 @@
 		internal.width = width;
 		internal.height = height;
 		internal.depth = depth;
+		internal.samples = 1;
 		internal.format = selectInternalFormat(format);
 		internal.bytes = bytes(internal.format);
 		internal.pitchB = pitchB(internal.width, 0, internal.format, false);
@@ -1237,6 +1247,7 @@
 		stencil.width = width;
 		stencil.height = height;
 		stencil.depth = depth;
+		stencil.samples = 1;
 		stencil.format = isStencil(format) ? FORMAT_S8 : FORMAT_NULL;
 		stencil.bytes = bytes(stencil.format);
 		stencil.pitchB = pitchB(stencil.width, 0, stencil.format, false);
@@ -1251,17 +1262,19 @@
 		paletteUsed = 0;
 	}
 
-	Surface::Surface(Resource *texture, int width, int height, int depth, int border, Format format, bool lockable, bool renderTarget, int pitchPprovided) : lockable(lockable), renderTarget(renderTarget)
+	Surface::Surface(Resource *texture, int width, int height, int depth, int border, int samples, Format format, bool lockable, bool renderTarget, int pitchPprovided) : lockable(lockable), renderTarget(renderTarget)
 	{
 		resource = texture ? texture : new Resource(0);
 		hasParent = texture != nullptr;
 		ownExternal = true;
 		depth = max(1, depth);
+		samples = max(1, samples);
 
 		external.buffer = nullptr;
 		external.width = width;
 		external.height = height;
 		external.depth = depth;
+		external.samples = (short)samples;
 		external.format = format;
 		external.bytes = bytes(external.format);
 		external.pitchB = pitchB(external.width, 0, external.format, renderTarget && !texture);
@@ -1276,13 +1289,14 @@
 		internal.width = width;
 		internal.height = height;
 		internal.depth = depth;
+		internal.samples = (short)samples;
 		internal.format = selectInternalFormat(format);
 		internal.bytes = bytes(internal.format);
 		internal.pitchB = !pitchPprovided ? pitchB(internal.width, border, internal.format, renderTarget) : pitchPprovided * internal.bytes;
 		internal.pitchP = !pitchPprovided ? pitchP(internal.width, border, internal.format, renderTarget) : pitchPprovided;
 		internal.sliceB = sliceB(internal.width, internal.height, border, internal.format, renderTarget);
 		internal.sliceP = sliceP(internal.width, internal.height, border, internal.format, renderTarget);
-		internal.border = border;
+		internal.border = (short)border;
 		internal.lock = LOCK_UNLOCKED;
 		internal.dirty = false;
 
@@ -1290,6 +1304,7 @@
 		stencil.width = width;
 		stencil.height = height;
 		stencil.depth = depth;
+		stencil.samples = (short)samples;
 		stencil.format = isStencil(format) ? FORMAT_S8 : FORMAT_NULL;
 		stencil.bytes = bytes(stencil.format);
 		stencil.pitchB = pitchB(stencil.width, 0, stencil.format, renderTarget);
@@ -1344,7 +1359,7 @@
 			}
 			else
 			{
-				external.buffer = allocateBuffer(external.width, external.height, external.depth, external.border, external.format);
+				external.buffer = allocateBuffer(external.width, external.height, external.depth, external.border, external.samples, external.format);
 			}
 		}
 
@@ -1396,7 +1411,7 @@
 			}
 			else
 			{
-				internal.buffer = allocateBuffer(internal.width, internal.height, internal.depth, internal.border, internal.format);
+				internal.buffer = allocateBuffer(internal.width, internal.height, internal.depth, internal.border, internal.samples, internal.format);
 			}
 		}
 
@@ -1471,7 +1486,7 @@
 
 		if(!stencil.buffer)
 		{
-			stencil.buffer = allocateBuffer(stencil.width, stencil.height, stencil.depth, stencil.border, stencil.format);
+			stencil.buffer = allocateBuffer(stencil.width, stencil.height, stencil.depth, stencil.border, stencil.samples, stencil.format);
 		}
 
 		return stencil.lockRect(x, y, front, LOCK_READWRITE);   // FIXME
@@ -1931,17 +1946,21 @@
 		unsigned char *sourceSlice = (unsigned char*)source.lockRect(0, 0, 0, sw::LOCK_READONLY);
 		unsigned char *destinationSlice = (unsigned char*)destination.lockRect(0, 0, 0, sw::LOCK_WRITEONLY);
 
-		for(int z = 0; z < destination.depth && z < source.depth; z++)
+		int depth = min(destination.depth, source.depth);
+		int height = min(destination.height, source.height);
+		int width = min(destination.width, source.width);
+
+		for(int z = 0; z < depth; z++)
 		{
 			unsigned char *sourceRow = sourceSlice;
 			unsigned char *destinationRow = destinationSlice;
 
-			for(int y = 0; y < destination.height && y < source.height; y++)
+			for(int y = 0; y < height; y++)
 			{
 				unsigned char *sourceElement = sourceRow;
 				unsigned char *destinationElement = destinationRow;
 
-				for(int x = 0; x < destination.width && x < source.width; x++)
+				for(int x = 0; x < width; x++)
 				{
 					unsigned int b = sourceElement[0];
 					unsigned int g = sourceElement[1];
@@ -1970,17 +1989,21 @@
 		unsigned char *sourceSlice = (unsigned char*)source.lockRect(0, 0, 0, sw::LOCK_READONLY);
 		unsigned char *destinationSlice = (unsigned char*)destination.lockRect(0, 0, 0, sw::LOCK_WRITEONLY);
 
-		for(int z = 0; z < destination.depth && z < source.depth; z++)
+		int depth = min(destination.depth, source.depth);
+		int height = min(destination.height, source.height);
+		int width = min(destination.width, source.width);
+
+		for(int z = 0; z < depth; z++)
 		{
 			unsigned char *sourceRow = sourceSlice;
 			unsigned char *destinationRow = destinationSlice;
 
-			for(int y = 0; y < destination.height && y < source.height; y++)
+			for(int y = 0; y < height; y++)
 			{
 				unsigned char *sourceElement = sourceRow;
 				unsigned char *destinationElement = destinationRow;
 
-				for(int x = 0; x < destination.width && x < source.width; x++)
+				for(int x = 0; x < width; x++)
 				{
 					unsigned int xrgb = *(unsigned short*)sourceElement;
 
@@ -2011,17 +2034,21 @@
 		unsigned char *sourceSlice = (unsigned char*)source.lockRect(0, 0, 0, sw::LOCK_READONLY);
 		unsigned char *destinationSlice = (unsigned char*)destination.lockRect(0, 0, 0, sw::LOCK_WRITEONLY);
 
-		for(int z = 0; z < destination.depth && z < source.depth; z++)
+		int depth = min(destination.depth, source.depth);
+		int height = min(destination.height, source.height);
+		int width = min(destination.width, source.width);
+
+		for(int z = 0; z < depth; z++)
 		{
 			unsigned char *sourceRow = sourceSlice;
 			unsigned char *destinationRow = destinationSlice;
 
-			for(int y = 0; y < destination.height && y < source.height; y++)
+			for(int y = 0; y < height; y++)
 			{
 				unsigned char *sourceElement = sourceRow;
 				unsigned char *destinationElement = destinationRow;
 
-				for(int x = 0; x < destination.width && x < source.width; x++)
+				for(int x = 0; x < width; x++)
 				{
 					unsigned int argb = *(unsigned short*)sourceElement;
 
@@ -2053,17 +2080,21 @@
 		unsigned char *sourceSlice = (unsigned char*)source.lockRect(0, 0, 0, sw::LOCK_READONLY);
 		unsigned char *destinationSlice = (unsigned char*)destination.lockRect(0, 0, 0, sw::LOCK_WRITEONLY);
 
-		for(int z = 0; z < destination.depth && z < source.depth; z++)
+		int depth = min(destination.depth, source.depth);
+		int height = min(destination.height, source.height);
+		int width = min(destination.width, source.width);
+
+		for(int z = 0; z < depth; z++)
 		{
 			unsigned char *sourceRow = sourceSlice;
 			unsigned char *destinationRow = destinationSlice;
 
-			for(int y = 0; y < destination.height && y < source.height; y++)
+			for(int y = 0; y < height; y++)
 			{
 				unsigned char *sourceElement = sourceRow;
 				unsigned char *destinationElement = destinationRow;
 
-				for(int x = 0; x < destination.width && x < source.width; x++)
+				for(int x = 0; x < width; x++)
 				{
 					unsigned int xrgb = *(unsigned short*)sourceElement;
 
@@ -2094,17 +2125,21 @@
 		unsigned char *sourceSlice = (unsigned char*)source.lockRect(0, 0, 0, sw::LOCK_READONLY);
 		unsigned char *destinationSlice = (unsigned char*)destination.lockRect(0, 0, 0, sw::LOCK_WRITEONLY);
 
-		for(int z = 0; z < destination.depth && z < source.depth; z++)
+		int depth = min(destination.depth, source.depth);
+		int height = min(destination.height, source.height);
+		int width = min(destination.width, source.width);
+
+		for(int z = 0; z < depth; z++)
 		{
 			unsigned char *sourceRow = sourceSlice;
 			unsigned char *destinationRow = destinationSlice;
 
-			for(int y = 0; y < destination.height && y < source.height; y++)
+			for(int y = 0; y < height; y++)
 			{
 				unsigned char *sourceElement = sourceRow;
 				unsigned char *destinationElement = destinationRow;
 
-				for(int x = 0; x < destination.width && x < source.width; x++)
+				for(int x = 0; x < width; x++)
 				{
 					unsigned int argb = *(unsigned short*)sourceElement;
 
@@ -2136,17 +2171,21 @@
 		unsigned char *sourceSlice = (unsigned char*)source.lockRect(0, 0, 0, sw::LOCK_READONLY);
 		unsigned char *destinationSlice = (unsigned char*)destination.lockRect(0, 0, 0, sw::LOCK_WRITEONLY);
 
-		for(int z = 0; z < destination.depth && z < source.depth; z++)
+		int depth = min(destination.depth, source.depth);
+		int height = min(destination.height, source.height);
+		int width = min(destination.width, source.width);
+
+		for(int z = 0; z < depth; z++)
 		{
 			unsigned char *sourceRow = sourceSlice;
 			unsigned char *destinationRow = destinationSlice;
 
-			for(int y = 0; y < destination.height && y < source.height; y++)
+			for(int y = 0; y < height; y++)
 			{
 				unsigned char *sourceElement = sourceRow;
 				unsigned char *destinationElement = destinationRow;
 
-				for(int x = 0; x < destination.width && x < source.width; x++)
+				for(int x = 0; x < width; x++)
 				{
 					unsigned int abgr = palette[*(unsigned char*)sourceElement];
 
@@ -2581,7 +2620,7 @@
 	{
 	}
 
-	unsigned int Surface::size(int width, int height, int depth, int border, Format format)
+	unsigned int Surface::size(int width, int height, int depth, int border, int samples, Format format)
 	{
 		width += 2 * border;
 		height += 2 * border;
@@ -2667,7 +2706,7 @@
 				return YSize + 2 * CSize;
 			}
 		default:
-			return bytes(format) * width * height * depth;
+			return bytes(format) * width * height * depth * samples;
 		}
 	}
 
@@ -3184,7 +3223,7 @@
 		return 1;
 	}
 
-	void *Surface::allocateBuffer(int width, int height, int depth, int border, Format format)
+	void *Surface::allocateBuffer(int width, int height, int depth, int border, int samples, Format format)
 	{
 		// Render targets require 2x2 quads
 		int width2 = (width + 1) & ~1;
@@ -3193,7 +3232,7 @@
 		// FIXME: Unpacking byte4 to short4 in the sampler currently involves reading 8 bytes,
 		// and stencil operations also read 8 bytes per four 8-bit stencil values,
 		// so we have to allocate 4 extra bytes to avoid buffer overruns.
-		return allocate(size(width2, height2, depth, border, format) + 4);
+		return allocate(size(width2, height2, depth, border, samples, format) + 4);
 	}
 
 	void Surface::memfill4(void *buffer, int pattern, int bytes)
@@ -3308,7 +3347,7 @@
 		{
 			float *target = (float*)lockInternal(0, 0, 0, lock, PUBLIC) + x0 + width2 * y0;
 
-			for(int z = 0; z < internal.depth; z++)
+			for(int z = 0; z < internal.samples; z++)
 			{
 				for(int y = y0; y < y1; y++)
 				{
@@ -3333,7 +3372,7 @@
 			int evenX0 = ((x0 + 1) & ~1) * 2;
 			int evenBytes = (oddX1 - evenX0) * sizeof(float);
 
-			for(int z = 0; z < internal.depth; z++)
+			for(int z = 0; z < internal.samples; z++)
 			{
 				for(int y = y0; y < y1; y++)
 				{
@@ -3437,7 +3476,7 @@
 		char *buffer = (char*)lockStencil(0, 0, 0, PUBLIC);
 
 		// Stencil buffers are assumed to use quad layout
-		for(int z = 0; z < stencil.depth; z++)
+		for(int z = 0; z < stencil.samples; z++)
 		{
 			for(int y = y0; y < y1; y++)
 			{
@@ -3534,7 +3573,7 @@
 		}
 	}
 
-	void Surface::copyInternal(const Surface* source, int x, int y, float srcX, float srcY, bool filter)
+	void Surface::copyInternal(const Surface *source, int x, int y, float srcX, float srcY, bool filter)
 	{
 		ASSERT(internal.lock != LOCK_UNLOCKED && source && source->internal.lock != LOCK_UNLOCKED);
 
@@ -3542,17 +3581,17 @@
 
 		if(!filter)
 		{
-			color = source->internal.read((int)srcX, (int)srcY);
+			color = source->internal.read((int)srcX, (int)srcY, 0);
 		}
 		else   // Bilinear filtering
 		{
-			color = source->internal.sample(srcX, srcY);
+			color = source->internal.sample(srcX, srcY, 0);
 		}
 
 		internal.write(x, y, color);
 	}
 
-	void Surface::copyInternal(const Surface* source, int x, int y, int z, float srcX, float srcY, float srcZ, bool filter)
+	void Surface::copyInternal(const Surface *source, int x, int y, int z, float srcX, float srcY, float srcZ, bool filter)
 	{
 		ASSERT(internal.lock != LOCK_UNLOCKED && source && source->internal.lock != LOCK_UNLOCKED);
 
@@ -3688,7 +3727,8 @@
 		       external.depth  == internal.depth &&
 		       external.pitchB == internal.pitchB &&
 		       external.sliceB == internal.sliceB &&
-		       external.border == internal.border;
+		       external.border == internal.border &&
+		       external.samples == internal.samples;
 	}
 
 	Format Surface::selectInternalFormat(Format format) const
@@ -3947,11 +3987,13 @@
 
 	void Surface::resolve()
 	{
-		if(internal.depth <= 1 || !internal.dirty || !renderTarget || internal.format == FORMAT_NULL)
+		if(internal.samples <= 1 || !internal.dirty || !renderTarget || internal.format == FORMAT_NULL)
 		{
 			return;
 		}
 
+		ASSERT(internal.depth == 1);  // Unimplemented
+
 		void *source = internal.lockRect(0, 0, 0, LOCK_READWRITE);
 
 		int width = internal.width;
@@ -3983,7 +4025,7 @@
 			#if defined(__i386__) || defined(__x86_64__)
 				if(CPUID::supportsSSE2() && (width % 4) == 0)
 				{
-					if(internal.depth == 2)
+					if(internal.samples == 2)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4001,7 +4043,7 @@
 							source1 += pitch;
 						}
 					}
-					else if(internal.depth == 4)
+					else if(internal.samples == 4)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4025,7 +4067,7 @@
 							source3 += pitch;
 						}
 					}
-					else if(internal.depth == 8)
+					else if(internal.samples == 8)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4061,7 +4103,7 @@
 							source7 += pitch;
 						}
 					}
-					else if(internal.depth == 16)
+					else if(internal.samples == 16)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4128,7 +4170,7 @@
 			{
 				#define AVERAGE(x, y) (((x) & (y)) + ((((x) ^ (y)) >> 1) & 0x7F7F7F7F) + (((x) ^ (y)) & 0x01010101))
 
-				if(internal.depth == 2)
+				if(internal.samples == 2)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -4146,7 +4188,7 @@
 						source1 += pitch;
 					}
 				}
-				else if(internal.depth == 4)
+				else if(internal.samples == 4)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -4170,7 +4212,7 @@
 						source3 += pitch;
 					}
 				}
-				else if(internal.depth == 8)
+				else if(internal.samples == 8)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -4206,7 +4248,7 @@
 						source7 += pitch;
 					}
 				}
-				else if(internal.depth == 16)
+				else if(internal.samples == 16)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -4277,7 +4319,7 @@
 			#if defined(__i386__) || defined(__x86_64__)
 				if(CPUID::supportsSSE2() && (width % 4) == 0)
 				{
-					if(internal.depth == 2)
+					if(internal.samples == 2)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4295,7 +4337,7 @@
 							source1 += pitch;
 						}
 					}
-					else if(internal.depth == 4)
+					else if(internal.samples == 4)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4319,7 +4361,7 @@
 							source3 += pitch;
 						}
 					}
-					else if(internal.depth == 8)
+					else if(internal.samples == 8)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4355,7 +4397,7 @@
 							source7 += pitch;
 						}
 					}
-					else if(internal.depth == 16)
+					else if(internal.samples == 16)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4422,7 +4464,7 @@
 			{
 				#define AVERAGE(x, y) (((x) & (y)) + ((((x) ^ (y)) >> 1) & 0x7FFF7FFF) + (((x) ^ (y)) & 0x00010001))
 
-				if(internal.depth == 2)
+				if(internal.samples == 2)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -4440,7 +4482,7 @@
 						source1 += pitch;
 					}
 				}
-				else if(internal.depth == 4)
+				else if(internal.samples == 4)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -4464,7 +4506,7 @@
 						source3 += pitch;
 					}
 				}
-				else if(internal.depth == 8)
+				else if(internal.samples == 8)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -4500,7 +4542,7 @@
 						source7 += pitch;
 					}
 				}
-				else if(internal.depth == 16)
+				else if(internal.samples == 16)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -4570,7 +4612,7 @@
 			#if defined(__i386__) || defined(__x86_64__)
 				if(CPUID::supportsSSE2() && (width % 2) == 0)
 				{
-					if(internal.depth == 2)
+					if(internal.samples == 2)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4588,7 +4630,7 @@
 							source1 += pitch;
 						}
 					}
-					else if(internal.depth == 4)
+					else if(internal.samples == 4)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4612,7 +4654,7 @@
 							source3 += pitch;
 						}
 					}
-					else if(internal.depth == 8)
+					else if(internal.samples == 8)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4648,7 +4690,7 @@
 							source7 += pitch;
 						}
 					}
-					else if(internal.depth == 16)
+					else if(internal.samples == 16)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4715,7 +4757,7 @@
 			{
 				#define AVERAGE(x, y) (((x) & (y)) + ((((x) ^ (y)) >> 1) & 0x7FFF7FFF) + (((x) ^ (y)) & 0x00010001))
 
-				if(internal.depth == 2)
+				if(internal.samples == 2)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -4733,7 +4775,7 @@
 						source1 += pitch;
 					}
 				}
-				else if(internal.depth == 4)
+				else if(internal.samples == 4)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -4757,7 +4799,7 @@
 						source3 += pitch;
 					}
 				}
-				else if(internal.depth == 8)
+				else if(internal.samples == 8)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -4793,7 +4835,7 @@
 						source7 += pitch;
 					}
 				}
-				else if(internal.depth == 16)
+				else if(internal.samples == 16)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -4863,7 +4905,7 @@
 			#if defined(__i386__) || defined(__x86_64__)
 				if(CPUID::supportsSSE() && (width % 4) == 0)
 				{
-					if(internal.depth == 2)
+					if(internal.samples == 2)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4882,7 +4924,7 @@
 							source1 += pitch;
 						}
 					}
-					else if(internal.depth == 4)
+					else if(internal.samples == 4)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4907,7 +4949,7 @@
 							source3 += pitch;
 						}
 					}
-					else if(internal.depth == 8)
+					else if(internal.samples == 8)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -4944,7 +4986,7 @@
 							source7 += pitch;
 						}
 					}
-					else if(internal.depth == 16)
+					else if(internal.samples == 16)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -5010,7 +5052,7 @@
 				else
 			#endif
 			{
-				if(internal.depth == 2)
+				if(internal.samples == 2)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -5029,7 +5071,7 @@
 						source1 += pitch;
 					}
 				}
-				else if(internal.depth == 4)
+				else if(internal.samples == 4)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -5054,7 +5096,7 @@
 						source3 += pitch;
 					}
 				}
-				else if(internal.depth == 8)
+				else if(internal.samples == 8)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -5091,7 +5133,7 @@
 						source7 += pitch;
 					}
 				}
-				else if(internal.depth == 16)
+				else if(internal.samples == 16)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -5160,7 +5202,7 @@
 			#if defined(__i386__) || defined(__x86_64__)
 				if(CPUID::supportsSSE() && (width % 2) == 0)
 				{
-					if(internal.depth == 2)
+					if(internal.samples == 2)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -5179,7 +5221,7 @@
 							source1 += pitch;
 						}
 					}
-					else if(internal.depth == 4)
+					else if(internal.samples == 4)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -5204,7 +5246,7 @@
 							source3 += pitch;
 						}
 					}
-					else if(internal.depth == 8)
+					else if(internal.samples == 8)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -5241,7 +5283,7 @@
 							source7 += pitch;
 						}
 					}
-					else if(internal.depth == 16)
+					else if(internal.samples == 16)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -5307,7 +5349,7 @@
 				else
 			#endif
 			{
-				if(internal.depth == 2)
+				if(internal.samples == 2)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -5326,7 +5368,7 @@
 						source1 += pitch;
 					}
 				}
-				else if(internal.depth == 4)
+				else if(internal.samples == 4)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -5351,7 +5393,7 @@
 						source3 += pitch;
 					}
 				}
-				else if(internal.depth == 8)
+				else if(internal.samples == 8)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -5388,7 +5430,7 @@
 						source7 += pitch;
 					}
 				}
-				else if(internal.depth == 16)
+				else if(internal.samples == 16)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -5457,7 +5499,7 @@
 			#if defined(__i386__) || defined(__x86_64__)
 				if(CPUID::supportsSSE())
 				{
-					if(internal.depth == 2)
+					if(internal.samples == 2)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -5476,7 +5518,7 @@
 							source1 += pitch;
 						}
 					}
-					else if(internal.depth == 4)
+					else if(internal.samples == 4)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -5501,7 +5543,7 @@
 							source3 += pitch;
 						}
 					}
-					else if(internal.depth == 8)
+					else if(internal.samples == 8)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -5538,7 +5580,7 @@
 							source7 += pitch;
 						}
 					}
-					else if(internal.depth == 16)
+					else if(internal.samples == 16)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -5604,7 +5646,7 @@
 				else
 			#endif
 			{
-				if(internal.depth == 2)
+				if(internal.samples == 2)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -5623,7 +5665,7 @@
 						source1 += pitch;
 					}
 				}
-				else if(internal.depth == 4)
+				else if(internal.samples == 4)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -5648,7 +5690,7 @@
 						source3 += pitch;
 					}
 				}
-				else if(internal.depth == 8)
+				else if(internal.samples == 8)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -5685,7 +5727,7 @@
 						source7 += pitch;
 					}
 				}
-				else if(internal.depth == 16)
+				else if(internal.samples == 16)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -5754,7 +5796,7 @@
 			#if defined(__i386__) || defined(__x86_64__)
 				if(CPUID::supportsSSE2() && (width % 8) == 0)
 				{
-					if(internal.depth == 2)
+					if(internal.samples == 2)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -5783,7 +5825,7 @@
 							source1 += pitch;
 						}
 					}
-					else if(internal.depth == 4)
+					else if(internal.samples == 4)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -5824,7 +5866,7 @@
 							source3 += pitch;
 						}
 					}
-					else if(internal.depth == 8)
+					else if(internal.samples == 8)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -5889,7 +5931,7 @@
 							source7 += pitch;
 						}
 					}
-					else if(internal.depth == 16)
+					else if(internal.samples == 16)
 					{
 						for(int y = 0; y < height; y++)
 						{
@@ -6009,7 +6051,7 @@
 			{
 				#define AVERAGE(x, y) (((x) & (y)) + ((((x) ^ (y)) >> 1) & 0x7BEF) + (((x) ^ (y)) & 0x0821))
 
-				if(internal.depth == 2)
+				if(internal.samples == 2)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -6027,7 +6069,7 @@
 						source1 += pitch;
 					}
 				}
-				else if(internal.depth == 4)
+				else if(internal.samples == 4)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -6051,7 +6093,7 @@
 						source3 += pitch;
 					}
 				}
-				else if(internal.depth == 8)
+				else if(internal.samples == 8)
 				{
 					for(int y = 0; y < height; y++)
 					{
@@ -6087,7 +6129,7 @@
 						source7 += pitch;
 					}
 				}
-				else if(internal.depth == 16)
+				else if(internal.samples == 16)
 				{
 					for(int y = 0; y < height; y++)
 					{
diff --git a/src/Renderer/Surface.hpp b/src/Renderer/Surface.hpp
index 16aa24e..7ea7bee 100644
--- a/src/Renderer/Surface.hpp
+++ b/src/Renderer/Surface.hpp
@@ -48,7 +48,7 @@
 	typedef RectT<int> Rect;
 	typedef RectT<float> RectF;
 
-	template <typename T> struct SliceRectT : public RectT<T>
+	template<typename T> struct SliceRectT : public RectT<T>
 	{
 		SliceRectT() : slice(0) {}
 		SliceRectT(const RectT<T>& rect) : RectT<T>(rect), slice(0) {}
@@ -237,7 +237,9 @@
 	private:
 		struct Buffer
 		{
-		public:
+			friend Surface;
+
+		private:
 			void write(int x, int y, int z, const Color<float> &color);
 			void write(int x, int y, const Color<float> &color);
 			void write(void *element, const Color<float> &color);
@@ -245,7 +247,7 @@
 			Color<float> read(int x, int y) const;
 			Color<float> read(void *element) const;
 			Color<float> sample(float x, float y, float z) const;
-			Color<float> sample(float x, float y) const;
+			Color<float> sample(float x, float y, int layer) const;
 
 			void *lockRect(int x, int y, int z, Lock lock);
 			void unlockRect();
@@ -254,12 +256,15 @@
 			int width;
 			int height;
 			int depth;
+			short border;
+			short samples;
+
 			int bytes;
 			int pitchB;
 			int pitchP;
 			int sliceB;
 			int sliceP;
-			int border;
+
 			Format format;
 			AtomicInt lock;
 
@@ -268,11 +273,11 @@
 
 	protected:
 		Surface(int width, int height, int depth, Format format, void *pixels, int pitch, int slice);
-		Surface(Resource *texture, int width, int height, int depth, int border, Format format, bool lockable, bool renderTarget, int pitchP = 0);
+		Surface(Resource *texture, int width, int height, int depth, int border, int samples, Format format, bool lockable, bool renderTarget, int pitchP = 0);
 
 	public:
 		static Surface *create(int width, int height, int depth, Format format, void *pixels, int pitch, int slice);
-		static Surface *create(Resource *texture, int width, int height, int depth, int border, Format format, bool lockable, bool renderTarget, int pitchP = 0);
+		static Surface *create(Resource *texture, int width, int height, int depth, int border, int samples, Format format, bool lockable, bool renderTarget, int pitchP = 0);
 
 		virtual ~Surface() = 0;
 
@@ -313,6 +318,7 @@
 		void sync();                      // Wait for lock(s) to be released.
 		inline bool isUnlocked() const;   // Only reliable after sync().
 
+		inline int getSamples() const;
 		inline int getMultiSampleCount() const;
 		inline int getSuperSampleCount() const;
 
@@ -351,7 +357,7 @@
 		static int pitchP(int width, int border, Format format, bool target);
 		static int sliceB(int width, int height, int border, Format format, bool target);
 		static int sliceP(int width, int height, int border, Format format, bool target);
-		static unsigned int size(int width, int height, int depth, int border, Format format);   // FIXME: slice * depth
+		static unsigned int size(int width, int height, int depth, int border, int samples, Format format);   // FIXME: slice * depth
 
 		static bool isStencil(Format format);
 		static bool isDepth(Format format);
@@ -474,7 +480,7 @@
 
 		static void update(Buffer &destination, Buffer &source);
 		static void genericUpdate(Buffer &destination, Buffer &source);
-		static void *allocateBuffer(int width, int height, int depth, int border, Format format);
+		static void *allocateBuffer(int width, int height, int depth, int border, int samples, Format format);
 		static void memfill4(void *buffer, int pattern, int bytes);
 
 		bool identicalFormats() const;
@@ -625,14 +631,19 @@
 		return stencil.sliceB;
 	}
 
+	int Surface::getSamples() const
+	{
+		return internal.samples;
+	}
+
 	int Surface::getMultiSampleCount() const
 	{
-		return sw::min(internal.depth, 4);
+		return sw::min((int)internal.samples, 4);
 	}
 
 	int Surface::getSuperSampleCount() const
 	{
-		return internal.depth > 4 ? internal.depth / 4 : 1;
+		return internal.samples > 4 ? internal.samples / 4 : 1;
 	}
 
 	bool Surface::isUnlocked() const