diff --git a/plugins/GSdx/GSDrawScanline.cpp b/plugins/GSdx/GSDrawScanline.cpp
index ff9ab080aa..b546facfaf 100644
--- a/plugins/GSdx/GSDrawScanline.cpp
+++ b/plugins/GSdx/GSDrawScanline.cpp
@@ -166,6 +166,7 @@ void GSDrawScanline::BeginDraw(const GSRasterizerData* data, Functions* f)
 		m_env.t.min = m_env.t.min.xxxxlh();
 		m_env.t.max = m_env.t.max.xxxxlh();
 		m_env.t.mask = m_env.t.mask.xxzz();
+		m_env.t.invmask = ~m_env.t.mask;
 	}
 
 	//
diff --git a/plugins/GSdx/GSDrawScanlineCodeGenerator.cpp b/plugins/GSdx/GSDrawScanlineCodeGenerator.cpp
index cbe6fda70d..1b35ab577c 100644
--- a/plugins/GSdx/GSDrawScanlineCodeGenerator.cpp
+++ b/plugins/GSdx/GSDrawScanlineCodeGenerator.cpp
@@ -164,7 +164,7 @@ L("@@");
 	// esi = fzbr
 	// edi = fzbc
 	// ebp = za
-	//xmm2 = fd
+	// xmm2 = fd
 	// xmm3 = fm
 	// xmm4 = zm
 	// xmm5 = rb
@@ -768,17 +768,17 @@ void GSDrawScanlineCodeGenerator::SampleTexture()
 		pcmpeqd(xmm1, xmm1);
 		psrlw(xmm1, 15);
 		paddw(xmm3, xmm1);
-	}
 
-	// uv0 = Wrap(uv0);
-
-	Wrap(xmm2, xmm1);
-
-	if(m_env.sel.ltf)
-	{
+		// uv0 = Wrap(uv0);
 		// uv1 = Wrap(uv1);
 
-		Wrap(xmm3, xmm1);
+		Wrap(xmm2, xmm3);
+	}
+	else
+	{
+		// uv0 = Wrap(uv0);
+
+		Wrap(xmm2);
 	}
 
 	// xmm2 = uv0
@@ -983,6 +983,118 @@ void GSDrawScanlineCodeGenerator::SampleTexture()
 	}
 }
 
+void GSDrawScanlineCodeGenerator::Wrap(const Xmm& uv)
+{
+	// xmm0, xmm1, xmm4, xmm5, xmm6 = free
+
+	if(m_env.sel.wms == m_env.sel.wmt)
+	{
+		if(m_env.sel.wms)
+		{
+			pmaxsw(uv, xmmword[&m_env.t.min]);
+			pminsw(uv, xmmword[&m_env.t.max]);
+		}
+		else
+		{
+			pand(uv, xmmword[&m_env.t.min]);
+			por(uv, xmmword[&m_env.t.max]);
+		}
+	}
+	else
+	{
+		movdqa(xmm1, uv);
+
+		movdqa(xmm4, xmmword[&m_env.t.min]);
+		movdqa(xmm5, xmmword[&m_env.t.max]);
+
+		// GSVector4i clamp = t.sat_i16(m_env.t.min, m_env.t.max);
+		
+		pmaxsw(uv, xmm4);
+		pminsw(uv, xmm5);
+
+		// GSVector4i repeat = (t & m_env.t.min) | m_env.t.max;
+
+		pand(xmm1, xmm4);
+		por(xmm1, xmm5);
+
+		// clamp.blend8(repeat, m_env.t.mask);
+
+		movdqa(xmm0, xmmword[&m_env.t.mask]);
+		blend8(uv, xmm1);
+	}
+}
+
+void GSDrawScanlineCodeGenerator::Wrap(const Xmm& uv0, const Xmm& uv1)
+{
+	// xmm0, xmm1, xmm4, xmm5, xmm6 = free
+
+	if(m_env.sel.wms == m_env.sel.wmt)
+	{
+		movdqa(xmm4, xmmword[&m_env.t.min]);
+		movdqa(xmm5, xmmword[&m_env.t.max]);
+
+		if(m_env.sel.wms)
+		{
+			pmaxsw(uv0, xmm4);
+			pminsw(uv0, xmm5);
+			pmaxsw(uv1, xmm4);
+			pminsw(uv1, xmm5);
+		}
+		else
+		{
+			pand(uv0, xmm4);
+			por(uv0, xmm5);
+			pand(uv1, xmm4);
+			por(uv1, xmm5);
+		}
+	}
+	else
+	{
+		movdqa(xmm1, uv0);
+		movdqa(xmm6, uv1);
+
+		movdqa(xmm4, xmmword[&m_env.t.min]);
+		movdqa(xmm5, xmmword[&m_env.t.max]);
+
+		// GSVector4i clamp = t.sat_i16(m_env.t.min, m_env.t.max);
+		
+		pmaxsw(uv0, xmm4);
+		pminsw(uv0, xmm5);
+		pmaxsw(uv1, xmm4);
+		pminsw(uv1, xmm5);
+
+		// GSVector4i repeat = (t & m_env.t.min) | m_env.t.max;
+
+		pand(xmm1, xmm4);
+		por(xmm1, xmm5);
+		pand(xmm6, xmm4);
+		por(xmm6, xmm5);
+
+		// clamp.blend8(repeat, m_env.t.mask);
+
+		if(m_cpu.has(util::Cpu::tSSE41))
+		{
+			movdqa(xmm0, xmmword[&m_env.t.mask]);
+
+			pblendvb(uv0, xmm1);
+			pblendvb(uv1, xmm6);
+		}
+		else
+		{
+			movdqa(xmm0, xmmword[&m_env.t.invmask]);
+			movdqa(xmm4, xmm0);
+
+			pand(uv0, xmm0);
+			pandn(xmm0, xmm1);
+			por(uv0, xmm0);
+
+			pand(uv1, xmm4);
+			pandn(xmm4, xmm6);
+			por(uv1, xmm4);
+		}
+	}
+}
+
 void GSDrawScanlineCodeGenerator::AlphaTFX()
 {
 	if(!m_env.sel.fb)
@@ -1288,9 +1400,32 @@ void GSDrawScanlineCodeGenerator::TestDestAlpha()
 	// test |= ((fd [<< 16]) ^ m_env.datm).sra32(31);
 
 	movdqa(xmm1, xmm2);
-	if(m_env.sel.fpsm == 2) pslld(xmm1, 16);
-	pxor(xmm1, xmmword[&m_env.datm]);
-	psrad(xmm1, 31);
+
+	if(m_env.sel.datm)
+	{
+		if(m_env.sel.fpsm == 2)
+		{
+			pxor(xmm0, xmm0);
+			psrld(xmm1, 15);
+			pcmpeqd(xmm1, xmm0);
+		}
+		else
+		{
+			pcmpeqd(xmm0, xmm0);
+			pxor(xmm1, xmm0);
+			psrad(xmm1, 31);
+		}
+	}
+	else
+	{
+		if(m_env.sel.fpsm == 2)
+		{
+			pslld(xmm1, 16);
+		}
+
+		psrad(xmm1, 31);
+	}
+
 	por(xmm7, xmm1);
 
 	alltrue(xmm7, eax, "step");
@@ -1826,42 +1961,6 @@ void GSDrawScanlineCodeGenerator::ReadTexel(const Xmm& dst, const Xmm& addr, con
 	else pinsrd(dst, src, i);
 }
 
-void GSDrawScanlineCodeGenerator::Wrap(const Xmm& uv, const Xmm& temp)
-{
-	if(m_env.sel.wms == m_env.sel.wmt)
-	{
-		if(m_env.sel.wms)
-		{
-			pmaxsw(uv, xmmword[&m_env.t.min]);
-			pminsw(uv, xmmword[&m_env.t.max]);
-		}
-		else
-		{
-			pand(uv, xmmword[&m_env.t.min]);
-			por(uv, xmmword[&m_env.t.max]);
-		}
-	}
-	else
-	{
-		movdqa(temp, uv);
-
-		// GSVector4i clamp = t.sat_i16(m_env.t.min, m_env.t.max);
-		
-		pmaxsw(uv, xmmword[&m_env.t.min]);
-		pminsw(uv, xmmword[&m_env.t.max]);
-
-		// GSVector4i repeat = (t & m_env.t.min) | m_env.t.max;
-
-		pand(temp, xmmword[&m_env.t.min]);
-		por(temp, xmmword[&m_env.t.max]);
-
-		// clamp.blend8(repeat, m_env.t.mask);
-
-		movdqa(xmm0, xmmword[&m_env.t.mask]);
-		blend8(uv, temp);
-	}
-}
-
 template<int shift> 
 void GSDrawScanlineCodeGenerator::modulate16(const Xmm& a, const Operand& f)
 {
@@ -1871,7 +1970,7 @@ void GSDrawScanlineCodeGenerator::modulate16(const Xmm& a, const Operand& f)
 	}
 	else
 	{
-		pslld(a, shift + 1);
+		psllw(a, shift + 1);
 		pmulhw(a, f);
 	}
 }
diff --git a/plugins/GSdx/GSDrawScanlineCodeGenerator.h b/plugins/GSdx/GSDrawScanlineCodeGenerator.h
index 0bf01fd506..96b47250a8 100644
--- a/plugins/GSdx/GSDrawScanlineCodeGenerator.h
+++ b/plugins/GSdx/GSDrawScanlineCodeGenerator.h
@@ -43,6 +43,8 @@ class GSDrawScanlineCodeGenerator : public CodeGenerator
 	void Step();
 	void TestZ(const Xmm& temp1, const Xmm& temp2);
 	void SampleTexture();
+	void Wrap(const Xmm& uv0);
+	void Wrap(const Xmm& uv0, const Xmm& uv1);
 	void AlphaTFX();
 	void TestAlpha();
 	void ColorTFX();
@@ -57,8 +59,7 @@ class GSDrawScanlineCodeGenerator : public CodeGenerator
 	void WritePixel(const Xmm& src, const Xmm& temp, const Reg32& addr, uint8 i, int psm);
 	void ReadTexel(const Xmm& dst, const Xmm& addr, const Reg32& base, const Xmm& temp1, const Xmm& temp2);
 	void ReadTexel(const Xmm& dst, const Xmm& addr, const Reg32& base, uint8 i);
-	void Wrap(const Xmm& uv, const Xmm& temp);
-	
+
 	template<int shift> void modulate16(const Xmm& a, const Operand& f);
 	template<int shift> void lerp16(const Xmm& a, const Xmm& b, const Xmm& f);
 	void mix16(const Xmm& a, const Xmm& b, const Xmm& temp);
diff --git a/plugins/GSdx/GSRendererSW.h b/plugins/GSdx/GSRendererSW.h
index 521e0fce7b..da9cbf6cc3 100644
--- a/plugins/GSdx/GSRendererSW.h
+++ b/plugins/GSdx/GSRendererSW.h
@@ -446,6 +446,7 @@ protected:
 			if(context->FRAME.PSM != PSM_PSMCT24)
 			{
 				p.sel.date = context->TEST.DATE;
+				p.sel.datm = context->TEST.DATM;
 			}
 
 			if(PRIM->ABE)
diff --git a/plugins/GSdx/GSScanlineEnvironment.h b/plugins/GSdx/GSScanlineEnvironment.h
index b88e7d1eb2..7bf7c1201c 100644
--- a/plugins/GSdx/GSScanlineEnvironment.h
+++ b/plugins/GSdx/GSScanlineEnvironment.h
@@ -55,9 +55,10 @@ union GSScanlineSelector
 		DWORD ztest:1; // 35
 		DWORD wms:1; // 36 (0: repeat, 1: clamp)
 		DWORD wmt:1; // 37
-		DWORD colclamp:1; // 38
-		DWORD fba:1; // 39
-		DWORD dthe:1; // 40
+		DWORD datm:1; // 38
+		DWORD colclamp:1; // 39
+		DWORD fba:1; // 40
+		DWORD dthe:1; // 41
 	};
 
 	struct
@@ -128,7 +129,7 @@ __declspec(align(16)) struct GSScanlineEnvironment
 	GSVector4i* dimx;
 
 	GSVector4i fm, zm;
-	struct {GSVector4i min, max, mask;} t; // [u] x 4 [v] x 4
+	struct {GSVector4i min, max, mask, invmask;} t; // [u] x 4 [v] x 4
 	GSVector4i datm;
 	GSVector4i colclamp;
 	GSVector4i fba;
diff --git a/plugins/GSdx/GSVector.h b/plugins/GSdx/GSVector.h
index a160a36d0a..eebf06e3ac 100644
--- a/plugins/GSdx/GSVector.h
+++ b/plugins/GSdx/GSVector.h
@@ -327,7 +327,15 @@ public:
 
 	GSVector4i blend8(const GSVector4i& a, const GSVector4i& mask) const
 	{
+		#if _M_SSE >= 0x401
+
 		return GSVector4i(_mm_blendv_epi8(m, a, mask));
+
+		#else
+
+		return GSVector4i(_mm_or_si128(_mm_andnot_si128(mask, m), _mm_and_si128(mask, a)));
+
+		#endif
 	}
 
 	#if _M_SSE >= 0x401
@@ -2310,7 +2318,15 @@ public:
 
 	GSVector4 blend8(const GSVector4& a, const GSVector4& mask)  const
 	{
+		#if _M_SSE >= 0x401
+
 		return GSVector4(_mm_blendv_ps(m, a, mask));
+
+		#else
+
+		return GSVector4(_mm_or_ps(_mm_andnot_ps(mask, m), _mm_and_ps(mask, a)));
+
+		#endif
 	}
 
 	GSVector4 upl(const GSVector4& a) const
diff --git a/plugins/GSdx/sse.h b/plugins/GSdx/sse.h
index b8af513a10..d261bfbff6 100644
--- a/plugins/GSdx/sse.h
+++ b/plugins/GSdx/sse.h
@@ -103,11 +103,6 @@
 
 #else
 
-	// not an equal replacement for sse4's blend but for our needs it is ok
-
-	#define _mm_blendv_ps(a, b, mask) _mm_or_ps(_mm_andnot_ps(mask, a), _mm_and_ps(mask, b))
-	#define _mm_blendv_epi8(a, b, mask) _mm_or_si128(_mm_andnot_si128(mask, a), _mm_and_si128(mask, b))
-
 	__forceinline __m128 _mm_round_ps(__m128 x)
 	{
 		__m128 t = _mm_or_ps(_mm_and_ps(ps_80000000, x), ps_4b000000);