I have written a code of diffuse shade, which using lowpoly and tangentspace normalmap.I have used the same code in rendermonkey and A6.2,but I got better effect in rendermonkey than in the A6.2.Who can tell me what the reason is? Is it the problem of this code or A6.2? I would appreciate if anyone can tell me the reason.

Code:

---

 material mat_normalmap

{  

   	flags = tangent;

	effect =

	"  

		texture entSkin1;

		texture entSkin2;

		matrix matWorldViewProj;

		matrix matWorld;

		vector vecSkill41;

		vector vecViewPos;

		

		technique specular

		{

			pass p0

			{

				VertexShaderConstant[0] = <matWorld>; 	//World Matrix

				VertexShaderConstant[4] = <matWorldViewProj>;	//World*View*Proj Matrix

			        VertexShaderConstant[16] = <vecSkill41>;	//Light Direction		

				VertexShaderConstant[33] = (0.5f,1.2f,0.4f,0.0f);

				

				Texture[0]   = <entSkin1>;  //basecolor

			        Texture[1]   = <entSkin2>;  //normal

          

				VertexShader=

				decl

				{

					stream 0;

					float v0[3]; //position		

					float v1[3]; //normal

					float v2[3]; //uv

					float v3[3]; //tangent

				}

				asm

				{

			   	vs.1.1

				   m4x4 oPos,v0,c4

				   

				   mov oT0,v2				// output uvs

				   mov oT1,v2				// output uvs

							    

                                    mov r3,v3

                                    mov r5,v1

              

				   //gen binormal via Cross product

				   mul r0,-r3.zxyw,r5.yzxw;

				   mad r4,-r3.yzxw,r5.zxyw,-r0;

   			         

				   //lightvector				   

                                   mov r10,c16

				   

				   dp3 r11.x, r10.xyz,r10.xyz

				   rsq r11.xyz,r11.x

				   mul r10.xyz,r10.xyz,r11.xyz

				 

				   dp3 r8.x,r3,-r10

				   dp3 r8.y,r4,-r10

				   dp3 r8.z,r5,-r10

				   

				   mad oD1.xyz,r8.xyz,c33.x,c33.x

			  

				};

			

				PixelShader= 

				asm

				{

				      ps.1.1

				  	

				      tex t0					//sample base map

				      tex t1					//sample normal

				   

				      dp3 r0,v1_bx2,v1_bx2

		                      mad r0,v1_bias,1-r0,v1_bx2 //lightvector normalizer

		                      dp3 r1,r0,t1_bx2 //L.N

				      mul r0,t0,r1

				   			  					

				};

			}

		}

	";

}  





ACTION model1

{

   my.skill41=float(-500); 		

	my.skill42=float(153); 		

	my.skill43=float(-57);  		

	my.skill44=float(1);

	MY.material= mat_normalmap;

	while(1) 

  {

   my.PAN+= 0.4;

   wait(1); 

  }

				

}   

---

The first thing i notice is that the uv's are in v7 and v8 (tangent). I dont think you can override them. Try changing your v2 -> v7, and v3 -> v8 and see if it works.

NORMAL MAPPING...DROOL.... please post final code if possible!!

What's so funny?

querty, at the moment this doesnot matter. the stream gives us the data in a certain order (position, normal/uv/tangent) no matter which numbers (i have tested this out).

but in books you find different information , there are up to 16 different datas in the stream, which btw. would be very useful to have.

Thank your reply. I have tested V2->V7,V3->V8,but I got the same result. I will post the pictures, and you will understand my meaning.
Additionaly, IN A6.2,I found an interesting thing, this code works well for the simple sphere and cube,but it doesn't work for the complex object,such as the character. But in RENDERMONKEY, the same code,texture and model can work well, why?

to DREW:because of this problem,then I can't finish the final code.

[image]http://pic8.album.tom.com/album_pic/2004/5/6/1083842260_34274.jpg?%75%4f%65%41%66%77%45%73%6b%7c[/image]

hey, i use this model too for my experiments. its very good for such purposes

maybe you should mention this in the bug forum since it could be a bug

Try using a 32bit TGA for your normalmap.