Vex: How to use Snell's Law for Refraction Vector?

Forums Technical Discussion Vex: How to use Snell's Law for Refraction Vector?

1031 11 0


olivierth: Member; 1007 posts; Joined: April 2017; Offline

Feb. 5, 2024 1:11 p.m.

Hi!

I came across a very informative video explaining Snell's Law. The trouble I'm having is how to solve for "r" (angle of outgoing ray). Can anybody tell me how to switch the formula around to find "r" if I have all the other values? I'm bad at math...

Edited by olivierth - Feb. 5, 2024 13:13:00

Attachments:
Houdini_SnellsLaw_01.JPG (66.3 KB)


rafaels: Member; 696 posts; Joined: March 2009; Offline

Feb. 5, 2024 2:02 p.m.

r = asin(n2 * sin(i)/n1))... me thinks

Toronto - ON
My Houdini playground [renderfarm.tumblr.com]
“As technology advances, the rendering time remains constant.”


Konstantin Magnus: Member; 670 posts; Joined: Sept. 2013; Offline

Feb. 5, 2024 4:38 p.m.

https://www.sidefx.com/docs/houdini/vex/functions/refract.html [www.sidefx.com]

https://procegen.konstantinmagnus.de/ [procegen.konstantinmagnus.de]


olivierth: Member; 1007 posts; Joined: April 2017; Offline

Feb. 5, 2024 5:05 p.m.

What! There's already a function for that? Super!

-Olivier


olivierth: Member; 1007 posts; Joined: April 2017; Offline

Feb. 5, 2024 5:58 p.m.

Looks like the refract function works BUT, I think the surface normals have to be inverted for it to bend the rays in the logical direction...

-Olviier

Edited by olivierth - Feb. 5, 2024 17:59:58


animatrix_: Member; 4516 posts; Joined: Feb. 2012; Online

Feb. 6, 2024 1:55 a.m.

You shouldn't have to invert the normals. I use the regular surface normal of a grid here to refract rays on to the floor:

Senior FX TD @ Industrial Light & Magic
Get to the NEXT level in Houdini & VEX with Pragmatic VEX! [www.pragmatic-vfx.com]

youtube.com/@pragmaticvfx | patreon.com/animatrix | animatrix2k7.gumroad.com


olivierth: Member; 1007 posts; Joined: April 2017; Offline

Feb. 6, 2024 10:14 a.m.

hmm... that's odd. we do things differently. I start with intersect to get the surface normal before I use refract.

EDIT: I'm confused. In your example, the IOR is below 1...?

Edited by olivierth - Feb. 6, 2024 12:02:00

Attachments:
Houdini_SnellsLaw_02.JPG (164.4 KB)


jsmack: Member; 7771 posts; Joined: Sept. 2011; Offline

Feb. 6, 2024 12:18 p.m.

olivierth
EDIT: I'm confused. In your example, the IOR is below 1...?

I think it's N1/N2, eg 1.0/1.5


olivierth: Member; 1007 posts; Joined: April 2017; Offline

Feb. 6, 2024 1:50 p.m.

I'm trying a simpler example and I still get weird results. My InAngle (theta 1) is 45 degrees (makes sens) but the resulting OutAngle (theta 2) is 135 when it should be less than 45. What am I not doing right? What am I not understanding?

-Olivier

Edited by olivierth - Feb. 6, 2024 13:53:16

Attachments:
Houdini_SnellsLaw_03.JPG (129.8 KB)


jsmack: Member; 7771 posts; Joined: Sept. 2011; Offline

Feb. 6, 2024 1:56 p.m.

olivierth
I'm trying a simpler example and I still get weird results. My InAngle (theta 1) is 45 degrees (makes sens) but the resulting OutAngle (theta 2) is 135 when it should be less than 45. What am I not doing right? What am I not understanding?

-Olivier

the refract function doesn't take the IOR of the medium as an argument, which you've put in as 1.6

It takes the ratio of the outside medium to the inside medium, for air and glass it would be 1.0 divided by 1.6

https://www.sidefx.com/docs/houdini/vex/functions/refract.html [www.sidefx.com]


olivierth: Member; 1007 posts; Joined: April 2017; Offline

Feb. 6, 2024 2:17 p.m.

Oh, that "N1/N2" might be where I'm goin' wrong!


olivierth: Member; 1007 posts; Joined: April 2017; Offline

Feb. 7, 2024 3:40 p.m.

Works like a charm! I baked the result onto a grid with some dispersion effect (not physically accurate).

Attachments:
Houdini_SnellsLaw_04.JPG (65.4 KB)

Quick Links

                    
                        Search links
                        Show recent posts
                        Show unanswered posts