Well, first the decrease in resolution isn't noticeable.  

The fisheye file is a post-process shader.  For every pixel in a view, the shader is called, if it's set to in the view configuration.  The function in the file runs once every time for each pixel.  And the output is a new color for the pixel.  You can access the current image for the view by uTDim.x and .y.  I am taking a color from another location in the image and putting it somewhere else, pixel by pixel.

Technically they aren't pixels, but texels.  texels are like pixels, but on a texture that has a height and a width.  Shaders handle the view image as texture, which can be stretched and compressed, etc.  

 

So it is called by directx if you set it to call it that psh file, using the xml file and selecting it in your custom camera.  For each texel in the texture painted on to the final camera view, you can output a color, or a depth value (depth being how deep in the view that pixel originates from, for fog and other effects, as far as I understand it now, which is probably pretty shaky on that depth thing).

If your shaky, I'm positively quivering lol

 

 

 

 

Just be quiet and take this stuff down.