I am trying to understand the transformations with the ray tracing
renderer in comparison to the transformations that occur in the
scanline renderer.
Please read the following and comment on it.
1. The generated ray is shot from the perspective camera through a
virtual screen and into the scene. The virtual screen I assume has the
same aspect ratio as the specified image to be rendered.
2. The generated ray and the object being intersected are moved into
the same space which could be either world space or object space and
then the intersection is determined.
3. Once the intersection is determined the shaders are invoked,
potentially some more rays are generated and finally the color of the
subpixel/pixel in the virtual screen is determined.
4. After all the subpixels for a pixel have been determined some
filter is applied and the final pixel color is determined.
5. A framebuffer is then updated with this color.
6. The different frame buffers are then merged and then used to update
the final image frame buffer.
As regards transformations:
1. Model transformations are applied to move the object into world
space or the inverse is used to move the ray into object space.
2. The inverse of the view transformations is used to move the ray
into world space.
The following transformations that I assume are NOT used :
1. Transformation to clip space (NDC). I assume clipping takes place
by limiting the primary rays to the frustrum. I assume that the
secondary rays are not limited by the frustrum or clipping planes.
2. Transformation to screen space. I assume this happens when moving
the ray and the eye point to world space.
thanks
