Hi all,

I'm working on a 3D space shooter with XNA 4.0 and have downloaded a temporary skybox(actually a sphere) with texture which I would like to replace with another scene.

So I finished creating the cube map with the photoshop CS3 Nvidia dds plugin and exported as a cubemap (other settings default) and, while I can import the texture and it displays well, no seams etc, I was forced to scale down to a size below 4096 for use with XNA.

However this results in the skybox feeling very "close" to the scene due to the smaller texture size compared to the original.

After checking the dimensions of the original texture in photoshop (even though it only displays a black background?) it seems to be a 9000 x 1500 texture size and the file size is just over 100 Mb!

My own texture is a paltry 3072 x 512 with a file size of about 1 Mb by comparison.

So my question is are there some export settings I am ignoring? Or, why exactly does XNA turn it's nose up at my own texture when it is over 4096 and not another?

I have checked the content importer and processor settings for each texture and they appear the same so is there something I'm missing?

If anyone can shed some light I would be much obliged.

Thanks.

Yeah that was my understanding as well of the way video memory was handled on the GPU, but there is still the fact that the texture being referenced is of that size and those dimensions in the content folder.. So the only thing I can think of is that its being either scaled down or sampled in some way along the content pipeline that I am unaware of, or perhaps that there is some sort of compression technique that was applied to the original texture during its creation that enables the image data to be referenced at runtime... Im just fishing here because, as you say, it doesn't conform to my understanding of the the addresing limitations of the environment..

Cheers though I appreciate the response!

Yeah I suppose I could have a better grasp though I am aware of the
byte code compilation process during the build process.. The thing is,
that is exactly why I don't understand the differences between the on
screen results I am achieving. I mean, if the content pipeline is
supposed to pretty much automatically apply scale, sampling and other
pertinent texture info at compile time and imbedded in the model, all of which is sent to the GPU to be rendered, then what is causing the scaling issues I am having? If you know what mean.

I
dont mean to sound like an idiot either, but I am not really a guy who
has often tried to stitch his own textures together in photoshop. My
reference to the file sizes are simply what I see on the screen and in
the properties of the files, I was just including them in case it had
context for the problem. I not really an artist. I dont really know how to measure what size an image is except by what I see on the screen. When I create a new photoshop document I set it to 6144 x 1024 pixels at 300 pixels per inch in 32 bit RGB mode. Then exported it to .dds with default options bar changing it from 2D texture to cubemap. The resulting mipmapped file appears to be 1Mb in the file explorer then.

Also the shader operating on the texture is as follows skybox.fx

uniform extern float4x4 ViewMatrix;
uniform extern float4x4 ProjectionMatrix;

void SkyboxVertexShader( float3 pos : POSITION0,
                         out float4 SkyPos : POSITION0,
                         out float3 SkyCoord : TEXCOORD0 )
{
    // Calculate rotation. Using a float3 result, so translation is ignored
    float3 rotatedPosition = mul(pos, ViewMatrix);           
    // Calculate projection, moving all vertices to the far clip plane 
    // (w and z both 1.0)
    SkyPos = mul(float4(rotatedPosition, 1), ProjectionMatrix).xyww;    

    SkyCoord = pos;
};
uniform extern texture SkyboxTexture;
sampler SkyboxS = sampler_state
{
    Texture = <SkyboxTexture>;
    MinFilter = LINEAR;
    MagFilter = LINEAR;
    MipFilter = LINEAR;
    AddressU = CLAMP;
    AddressV = CLAMP;
};
float4 SkyboxPixelShader( float3 SkyCoord : TEXCOORD0 ) : COLOR
{
    // grab the pixel color value from the skybox cube map
    return texCUBE(SkyboxS, SkyCoord);
};
technique SkyboxTechnique
{
    pass P0
    {
        vertexShader = compile vs_2_0 SkyboxVertexShader();
        pixelShader = compile ps_2_0 SkyboxPixelShader();

        // We're drawing the inside of a model
        CullMode = None;  
        // We don't want it to obscure objects with a Z < 1
        ZWriteEnable = false; 
    }
}

And the draw and update methods in my skysphere.cs file.

 public override void LoadContent()
        {

            SkySphereModel = XNAGame.Instance().Content.Load<Model>("SkySphere/SphereHighPoly");
            TextureCube SkySphereTexture = XNAGame.Instance().Content.Load<TextureCube>("SkySphere/SkySphereTexture");
            SkySphereEffect = XNAGame.Instance().Content.Load<Effect>("SkySphere");
            SkySphereEffect.Parameters["ViewMatrix"].SetValue(XNAGame.Instance().Camera.view);
            SkySphereEffect.Parameters["ProjectionMatrix"].SetValue(XNAGame.Instance().Camera.projection);
            SkySphereEffect.Parameters["SkyboxTexture"].SetValue(SkySphereTexture);

            SkySphereEffect = XNAGame.Instance().Content.Load<Effect>("SkySphere");

            foreach (ModelMesh mesh in SkySphereModel.Meshes)
            {
                foreach (ModelMeshPart part in mesh.MeshParts)
                {
                    part.Effect = SkySphereEffect;
                }
            }
        }

        public override void Draw(GameTime gameTime)
        {
            SkySphereEffect.Parameters["ViewMatrix"].SetValue(XNAGame.Instance().Camera.view);
            SkySphereEffect.Parameters["ProjectionMatrix"].SetValue(XNAGame.Instance().Camera.projection);

            DepthStencilState dss = new DepthStencilState();
            dss.DepthBufferEnable = false;
            XNAGame.Instance().GraphicsDevice.DepthStencilState = dss;
            foreach (ModelMesh mesh in SkySphereModel.Meshes)
            {

                mesh.Draw();
            }

            dss = new DepthStencilState();
            dss.DepthBufferEnable = true;
            XNAGame.Instance().GraphicsDevice.DepthStencilState = dss;
        }
    }