Pixel Shaders To Create Predator Effect

Question

I am learning shaders, and I figured I would try to use them to create sort of a predator effect. I got to the point where I rendered my level to a rendertarget and passed that to the shader file, and attempted to get the color from it. This, by my understanding, should render the player completely invisible from the code I have. I have included relevant portions of my code.

Shader Code:

float4x4 World;
float4x4 View;
float4x4 Projection;

sampler s0;
texture land;
sampler land_sample = sampler_state{ Texture = land; };

float4 PixelShaderFunction(float2 coords: TEXCOORD0) : COLOR0
{
    float4 color = tex2D(s0, coords);
    float4 land_color = tex2D(land_sample, coords);

    if (color.a)
        return land_color;

    return color;
}

Main Draw Method:

protected override void Draw(GameTime gameTime)
{
    //First draw the level to the land rendertarget
    GraphicsDevice.SetRenderTarget(land);
    GraphicsDevice.Clear(Color.Black);

    spriteBatch.Begin(SpriteSortMode.Immediate, null, SamplerState.PointClamp, null, null, null, camera.Transform);
    levelOne.Draw(spriteBatch, gameTime);
    spriteBatch.End();

    GraphicsDevice.SetRenderTarget(null);

    spriteBatch.Begin();

    //Draw land
    spriteBatch.Draw(land, new Vector2(0, 0), Color.White);

    spriteBatch.End();

    //Draw the player and npcs normally
    spriteBatch.Begin(SpriteSortMode.Immediate, null, SamplerState.PointClamp, null, null, null, camera.Transform);
    npcs.ForEach(n => n.Draw(spriteBatch, gameTime));

    effect.Parameters["land"].SetValue(land);
    effect.CurrentTechnique.Passes[0].Apply();
    player.Draw(spriteBatch, gameTime);

    spriteBatch.End();

    base.Draw(gameTime);
}

Level Draw Code:

public void Draw(SpriteBatch spriteBatch, GameTime gameTime)
{
    //Will cull tiles from level based on camera position later
    for (int i = 0; i < tiles.Length; i++)
    {
        for (int j = 0; j < tiles[i].Length; j++)
        {
            tiles[i][j].Draw(spriteBatch, gameTime);
        }
    }
}

Based on the tutorial I was following, this should replace the color in the player's sprite with that of the render target, but it does not. Furthermore, the 'level' drawn in the player's sprite is scaled down a bit.

My question is why this isn't working? Is there something I'm missing in the shader code? Also, as a beginner shader learner please try to keep your responses rather simple, or point me to a reference. Thanks! :)

Answer 1

float4 color = tex2D(s0, coords);
float4 land_color = tex2D(land_sample, coords);

You are sampling the "world" using the SpriteBatch quad's UV's. The blue UV's are used to sample the color-texture; the yellows are used to sample the world-texture. A Draw() parameter and/or the camera matrix is inverting the Y-axis, so it is using the bottom corner as (0,0) instead of the top, with Y increasing as you go "up".

UV.X = (pixelPosition.X / (TextureWidth - 1));
UV.Y = (pixelPosition.Y / (TextureHeight - 1));
UV.Y = 1.0f - UV.y; //invert Y

//I'm starting to forget XNA... (I think SpriteBatch does this for you)
//If it looks very close, but not quite...
//UV.X = ((pixelPosition.X + 0.5f) / (TextureWidth - 1)); //May also need to offset all
//UV.Y = ((pixelPosition.Y + 0.5f) / (TextureHeight - 1));//samples by one half-texel

Texel-alignment:
In case you need this, an example 2x2 texture. It also doesn't show the inverted Y, but the idea is the same; I'm willing to generate the inverted diagrams if you need additional assistance.

//screen-coordinates are -1:1 range
pixelPosition += 1; //move to 0:2 range
pixelPosition /= 2; //scale to 0:1 range
pixelPosition.Y = 1.0f - pixelPosition.Y; //invert Y
//for screenPosition (-1,1), pixelPosition calculates to (0,0)
UV.X = ((pixelPosition.X + 0.5f) / (TextureWidth - 1));
//UV.X = ((0 + 0.5f) / (2 - 1));
//UV.X = (0.5f / 1);
UV.Y = ((pixelPosition.Y + 0.5f) / (TextureHeight - 1));
//UV.Y = ((0 + 0.5f) / (2 - 1));
//UV.Y = (0.5f / 1);