Sign in to follow this  
Followers 0
jdub

Issue with HLSL shader code

4 posts in this topic

I have the following relevant HLSL code:

struct VS_INPUT
{
	float4 Position : POSITION;
	float2 TexCoord : TEXCOORD0;
	float3 Normal : NORMAL; //world space normal
};

struct PS_INPUT
{
	float4 Position : SV_POSITION;
	float2 TexCoord : TEXCOORD0;
	float3 Normal : NORMAL; //view-space normal
};

PS_INPUT VShader(VS_INPUT input, uniform bool useLighting, uniform bool isTransformed, uniform bool isTextured) 
{
	PS_INPUT output = (PS_INPUT)0;

	output.Position = input.Position;
	output.Position.w = 1.0f;

	if(isTextured)
	{
		output.TexCoord = input.TexCoord;
	}

	if(!isTransformed)
	{
		output.Position = mul(output.Position, world);
		output.Position = mul(output.Position, view);
		output.Position = mul(output.Position, projection);
	}

	if(useLighting)
	{
		//transform normal to view space
		output.Normal = mul( input.Normal, (float3x3)world );
		normalize( output.Normal );
	}

	return output;
}

float4 PShader(PS_INPUT input, uniform bool useLighting, uniform bool isTextured) : SV_TARGET
{
	float3 intensity;
	float3 light_dir;

	if(useLighting)
	{
		//apply diffuse lighting
		//----------------------
		//Invert the direction of the light vector so that it can be used to compare with Normals
		//Use dot product to decide intensity of light on each triangle
		light_dir = -light_dir;
		intensity = saturate(dot(input.Normal, light_dir));

		return float4(1.0f, 1.0f, 1.0f, 0.5f);
		//return float4(1.0f, saturate(intensity * light_color), 1.0f, 0.5f); 
	}
	if(isTextured)
	{
		float3 color = Texture.Sample(SampleType, input.TexCoord);

		return float4(color.x, color.y, color.z, 0.5f);
	}

	return float4(0.5f, 0.5f, 0.5f, 0.0f);
}

technique11 RenderTextured
{
	pass P0
	{
		SetVertexShader( CompileShader( vs_4_0, VShader(false, false, true) ) );
		SetGeometryShader( NULL );
		SetPixelShader( CompileShader( ps_4_0, PShader(false, true) ) );
	}
}

 

My first question is:  is it a good idea to organize my code like this (with shader code inside if statements inside one big pixel shader and vertex shader) or should I be organizing each different shader effect into it's own set of pixel/vertex shader functions?

 

I have an issue that occurs when I try and run this code inside the Visual Studio Graphics debugger. When I try to debug the vertex shader, the code starts execution inside of the if(isTextured) block skipping the beginning of the code entirely. the value of all data inside the output variable is all NaN until it executes the statement mul(output.Position, world).  The value of world matrix is the translation (1,0,0) of the identity matrix.  After the mul statement is executed, the output data takes the position of the input value.  However w takes the value 0.5f.  This is completely stumping me.  Any ideas?

0

Share this post


Link to post
Share on other sites

One warning from my experience, don't check isTextured like this:

if(isTextured)
...

 

Because HLSL has a weird way of handling uninitialized variables or false.

 

Do it like this:

if (isTextured == 1)
...

 

and in the other cases do this:

if(isTextured == 0) // the same as !isTextured
...
0

Share this post


Link to post
Share on other sites

if(isTextured)
    {
        output.TexCoord = input.TexCoord;
    }

 

 

In my opinion the branch above isn't needed. You'll need to initialize the texcoord anyway so why not just copy the incoming texture coordinate there instead of making a branch.

 

Cheers!

1

Share this post


Link to post
Share on other sites

+1 what kauna said.

Also instead of "isTextured" you could pass white texture and get same result. IMO you are doing premature optimization.

0

Share this post


Link to post
Share on other sites

My first question is: is it a good idea to organize my code like this (with shader code inside if statements inside one big pixel shader and vertex shader) or should I be organizing each different shader effect into it's own set of pixel/vertex shader functions?

That totally depends if you care about performance more than about your productivity.

 

It is, obviously, much easier for you to put everything into one pixel shader and not handle effects/techniques overhead,  but that also means you are totally abusing your poor gfx card [that most ceretainly did not sign up for that].

 

 

If you do, however, care for the performance a bit, then just bite the bullet and do it right.

0

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now
Sign in to follow this  
Followers 0