Raymarched Reflections

[@realcygnus]

A priceless gem ! As always.

[@d74g0n]

funny story; I been building robots n stuff; so someone who wanted to get into coding messaged me about how to start; and they shared me one of your videos and I laughed my guts out. This is no intro to coding channel lol.

[@TheArtofCodeIsCool]

This is shader coding... which is quite different from 'normal' coding. I DO have an absolute beginner tut for shader coding though! kzbin.info/www/bejne/q2arcoyMfdiUqcU

[@ghostriley22]

You make real quality tutorials for shaders. Awesome work!!

[@MartinToernby]

It's really nice, that you're so straight to the point, and go a little bit down the roads, that intuitively would work (such as recursion), and then backstep and implement it the way, that's possible. Superb. Again.

[@trigophers]

:O I didn't know the reflect() function existed until now! Up until watching this video I was just doing it like this: ray = normalize(ray-normal*vec3(2.)*dot(normal, ray)); Yay thanks for teaching me this!

[@daviddawkins]

That was just amazing.

[@dan00b8]

will you do a video about refractions too?

[@TheArtofCodeIsCool]

Yesss :)

[@СергейПавлов-в2е]

Very very cool! Thank fo you!

[@DeanEdis]

Thanks very much! Clear and useful, as ever!

[@DB009]

hey anyway you can cover emissive materials? been trying to make emissive materials for a few weeks now sadly most examples are for pathtracers on shader toy

[@Julian-bh7rh]

Really good video as always!:) Would be really intresting to see a video about this principle with lower optimised reflections as the iteration goes up! or shader rendering optimisation tricks in general;)

[@leftofzen]

@27:20 why do you write clamp(1-dot) and not 1-clamp(dot)?

[@TheArtofCodeIsCool]

You can do either, it comes down to the same thing.

[@imaginingPhysics]

btw. I dont think the clamp is needed: by the way they are constructed vectors n and -rd have an angle between -PI/2 and PI/2, so their dot product is non-negative. Moreover since both are also unit vectors their dot product is always between -1 and 1 (Schwartz inequality).

[@Bleenderhead]

shouldn't you displace the new ray origin along the reflected direction, not the normal?

[@TheArtofCodeIsCool]

Good question! I suppose that could be done, but really the point is to get away from the surface. If we push the ray along the reflected direction then we might have to push it very far from the intersection point at grazing angles.

[@MrlegendOr]

Is raymarching better than other techniques for reflection/Shadow/GI?

[@Igor_Zdrowowicz]

now its time for multiple bounces using raymarching :D

[@bryphi77]

This guy is a little green behind the ears!

[@TheArtofCodeIsCool]

Haha unfortunately, sometimes.. gotta work on that!

[@MartinToernby]

@@TheArtofCodeIsCool Just tint the fresnel ; ) Superb content - you are the only reason I can do anything interesting with shaders at all.

[@trigophers]

Subsurface scattering seems to be anti-green-screen lol