What am I doing wrong here? I have the following errors: error X4502: invalid output semantic 'POSITION': Legal indices are in [0,0] error X4502: invalid vs_1_1 output semantic 'POSITION' error X4502: invalid output semantic 'COLOR': Legal indices are in [0,1] error X4502: invalid vs_1_1 output semantic 'COLOR' I'm using the following FVF: struct NormalTexVertex { float x, y, z; // Position. float nx, ny, nz; // Normal vector. float u, v; // Texture coordinates. }; #define FVF_NORMAL_TEX (D3DFVF_XYZ | D3DFVF_NORMAL | D3DFVF_TEX1) I'm using the following output struct: struct VS_OUTPUT { float position[3] : POSITION; float diffuse[3] : COLOR; float uv[2] : TEXCOORD; }; [Edited by - Timptation on May 16, 2009 9:07:05 PM]

They're not arrays. They're vectors and can be accessed using .xyzw or .rgba (or any combination, single or otherwise)

struct VS_OUTPUT
{
float3 position : POSITION;
float3 diffuse : COLOR;
float2 uv : TEXCOORD0;
};

That is how you would correctly access vertex position/color/uv. Note the float3/float2 the number is the size of the vector. Hope that helps.

That did help, but now I've got:

Direct3D9: (ERROR) :Vertex shader function usage (D3DDECLUSAGE_TEXCOORD,0) does not have corresponding usage in the current vertex declaration

But I DO have that don't I?

SHADER:

struct VS_INPUT{
float3 position : POSITION;
float3 normal : NORMAL;
float2 uv : TEXCOORD;
};

struct VS_OUTPUT{
float4 position : POSITION;
float3 diffuse : COLOR;
float2 uv : TEXCOORD;
};

CODE:

struct NormalTexVertex{
float x, y, z; // Position.
float nx, ny, nz; // Normal vector.
float u, v; // Texture coordinates.
};
#define FVF_NORMAL_TEX (D3DFVF_XYZ | D3DFVF_NORMAL | D3DFVF_TEX1)

I tried changing D3DFVF_TEX1 to D3DFVF_TEX0, but got the same error. I'm using D3DFVF_TEX1 because D3DFVF_TEX0 is 0 and so my define conflicts with another FVF define i have that's just (D3DFVF_XYZ | D3DFVF_NORMAL).

sirpalee: I tried defining them that way, but still got the same error. The DirectX book I have in front of me says that leaving off the number implies 0. It's a 9.0c book though, and maybe its different in 10?

I've not heard of fxc.exe, but I see it and am looking into it now. I'm aware of vertex declarations, but I also understand that FVFs are internally converted to vertex declarations, so I was hoping to get away with not having to write the vertex declaration. I probably should just to get the experience though. I'm always wanting something for nothing...

This may not work and i know its not optimized the way I'm converteing from float3 to vector and all, but this is the shader main body:

VS_OUTPUT output = (VS_OUTPUT)0;

output.uv[0] = input.uv[0];
output.uv[1] = input.uv[1];

vector outPos;
outPos.x = input.position[0];
outPos.y = input.position[1];
outPos.z = input.position[2];
outPos.w = 1.0f;

outPos = mul(outPos, ViewProjMatrix);

output.position[0] = outPos.x;
output.position[1] = outPos.y;
output.position[2] = outPos.z;
output.position[3] = 0.0f;

LightDirection.w = 0.0f;
vector inNorm;
inNorm.x = input.normal[0];
inNorm.y = input.normal[1];
inNorm.z = input.normal[2];
inNorm.w = 0.0f;
LightDirection = mul(LightDirection, ViewMatrix);
inNorm = mul(inNorm, ViewMatrix);

float s = dot(LightDirection, inNorm);

if( s < 0.0f )
s = 0.0f;

vector outDiff;
outDiff = (AmbientMtrl * AmbientLightIntensity) +
(s * (DiffuseLightIntensity * DiffuseMtrl));

output.diffuse[0] = outDiff.x;
output.diffuse[1] = outDiff.y;
output.diffuse[2] = outDiff.z;

return output;

I'm running, but I don't see anything but my cleared screen. Rather than use FVFs I have this now:

// Trying out vertex declaration.
D3DVERTEXELEMENT9 vdecl[] = {
{0, 0, D3DDECLTYPE_FLOAT3, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_POSITION, 0},
{0, 12, D3DDECLTYPE_FLOAT3, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_NORMAL, 0},
{0, 24, D3DDECLTYPE_FLOAT2, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_TEXCOORD, 0},
D3DDECL_END() };

device->CreateVertexDeclaration(vdecl, &g_vertDecl);
device->SetVertexDeclaration(g_vertDecl);

My input struct is:
struct VS_INPUT{
float3 position : POSITION;
float3 normal : NORMAL;
float2 uv : TEXCOORD0;
};

"float3" being a predefined type for "vector<float, 3>" for example. Is this wrong? Changing it to the "vector<float, x>" format didnt change anything...

Feel so awesome for figuring this out lol. The problem was that I was multiplying a 1x3 vector with a 4x4 matrix. I can't believe that wasn't an error. Perhaps it automatically filled in the extra space with zeros when i needed ones. Anyway, thanks a bunch for the help guys!