8 replies to this topic

[Rexxaw(Forgrim's mate)]

[source lang="cpp"]cbuffer cbPerObject{ float4x4 gWorldViewProj;};Texture2D CubesTexture;SamplerState CubesTexSamplerState;/////////////////////////////struct VS_inputVPCT{ float3 Pos : POSITION; float4 Col : COLOUR; float2 Tex : TEXTURE;};struct VS_outputVPCT{ float4 Pos : SV_POSITION; float4 Col : COLOUR; float2 Tex : TEXTURE;};struct PS_outputVPCT{ float4 Col : COLOUR; float2 Tex : TEXTURE;};///////////////////////////////////////////////////////////////////////////////////VS_outputVPCT VSfuncVPCT(VS_inputVPCT input){ VS_outputVPCT output; output.Pos = mul(float4(input.Pos, 1.0f), gWorldViewProj); output.Col = input.Col; output.Tex = input.Tex; return output;}PS_outputVPCT PSfuncVPCT(VS_outputVPCT input) : SV_Target{ PS_outputVPCT output; output.Col = input.Col; output.Tex = CubesTexture.Sample(CubesTexSamplerState, input.Tex); return output;}///////////////////////////////////////////////////////////////[/source]




Thankyou for any assistance!

[Ashaman73]

Does my shader code look alright?

Well, it looks like shader code, when it compiles, it is most likely ok.

But if you want to check, if the code is doing what you want it to do, then you should at least tell us, what your goal is.

[Rexxaw(Forgrim's mate)]

My vertices contain position, colour and tex information. I have 1 constant buffer keeping track of world/view/proj, and that's about it. I am drawing a grid of vertices and expecting to see them, which i don't. I didn't think the problem was with my structs and functions here, so i just wanted a second opinion.
From this code would you expect a minimal amount of vertices of type Pos/Col/Tex to be transformed correctly, and passed successfully through the vertex and pixel shader?

[GuyWithBeard]

Your semantics look weird to me. To be standard code COLOUR should be COLOR, and TEXTURE should be TEXCOORD. However, maybe it works like that as well if you have declared your input layouts accordingly. I would try changing them though.

[Rexxaw(Forgrim's mate)]

I've used those semantics before no prob, afaik as long as the input layout matches it's good to go.

[SamiHuutoniemi]

Why would you want to return both texture color and the "other" color from the pixel shader? Don't you want to either blend them in the shader, or somewhere before this point decide what color you want to output?

[Rexxaw(Forgrim's mate)]

Yeah it's redundant isn't it? I should remove the colour since it's not needed, but will it cause problems? It looks like the code is ok? Thank's for the feedback.

[GuyWithBeard]

Ahh, that might be the problem. SV_Target maps to a float4. You should generally return a float4 (or nothing at all) from the pixel shader. Try removing the texture and see if that helps.

[kauna]

Hello,

output.Pos = mul(float4(input.Pos, 1.0f), gWorldViewProj);

Casting input.Pos to float4 with the 4th component as 1.0 is unnecessary if you define your input vertex structure as :

struct VS_inputVPCT
{
float4 Pos : POSITION;
...
}

Remember that the program side structure needs to be still 3 component vector and the vertex declaration needs to be DXGI_FORMAT_R32G32B32_FLOAT for the position.


Otherwise, you may use what ever name for your semantics as long as you use the same name in the vertex declaration.


Cheers!