fabricecastel / scala-raytracer Goto Github PK

https://www.cs.utexas.edu/~whunt/papers/hunt-pers.pdf

The javax code does some pretty nasty stuff like:

v1.cross(v2, v3)

will set the v1 vector's value to be the result of v2 cross v3. Would be better to have a method defined in the Vector3D wrapper method that takes in a Vector3D and returns a new Vector3D rather than this mutation bullshit.

the wrapper is vecmath.scala

https://en.wikipedia.org/wiki/M%C3%B6ller%E2%80%93Trumbore_intersection_algorithm

Requires the polynomial solver to be implemented beforehand

Used to be in Lua, could be just about anything - I'm not picky

Do a second pass after the initial aliased render and average N squared new rays per pixel whose colour value differs from its neighbours beyond an adjustable threshold

To run tests, use the "sbt test" command. There's a sample test suite under src>main>test>scala>TriangleSpec.scala for reference

A quick google search turned this up: http://stackoverflow.com/questions/14094042/calculating-camera-ray-direction-to-3d-world-pixel

May or may not be enough to implement this.

Seems to do the job. Haven't tested it much. Would be nice for it not to look like such a mess though.

For now, we only need to be able to solve roots for polynomials of degree 2, though if we can also have higher degrees as a freebie I can't see why we shouldn't do it.

Implement a new Polysolver class for this purpose with a simple solve(a: Int, b: Int, c: Int) API that returns a List[Int] and disregards complex solutions

maybe a bit overkill, but should have all the relevant info

http://image.diku.dk/projects/media/morten.mikkelsen.08.pdf