Image not available

1920x1080

CGI.png

๐Ÿงต Why aren't AAA studios doing this?

Anonymous No. 965230

https://www.youtube.com/watch?v=CKp2YnbEnxc
Do any of you use the roughness truncation technique, talked about in the video, in your renders?

Anonymous No. 965236

He claims that fresnel doesn't incorporate this effect. That doesn't make sense because all it does is give you a value that's based on the surface normal and the camera direction.

Anonymous No. 965239

>>965236
Well, not on its own, no. As you said, it's literally just a value based on camera angle.
Implementing this would be trivial in any shader. Plug it into a multiplier/ramp for roughness and there you have it.

>>965230
>why aren't AAA studios doing this?
Is that what the faggot grifter said in his video to prop himself up?
This is only mind bogglingly revolutionary for people who get their knowledge from YouTube.
>hurr durr but why isn't it included in the didney principled shader, I read all the papers and didn't understand jack shit, I must be so much better than the didney engineers
It's written right there.
>"Intuitive rather than physical parameters"
>"As few parameters as possible"
>"Parameters should be zero to one"

The principled shader is meant to be artist-friendly while still being physically correct.
A slider that goes from "very reflective" to "very rough" is as friendly as it gets, and works for 99.9% of cases.
You only appreciate fresnel-dependent variable roughness when dealing with large, flat surfaces of uniform roughness.
If you were to add that to the main shader, it would violate the three rules above.
How do you decide what angle range to remap to 0-1? Falloff amount? What's their default values?
Not worth it for something that isn't truly visible in most cases. Better to leave it as something to wire yourself when you need it.

Yet another sensationalist trite for clicks dumb people fall head over heels over.

Image not available

1837x1895

fresnel_roughness.png

Anonymous No. 965241

>>965230
>fresnel doesn't work
it does
>it's a linear effect
smooth step makes it better
>conclusion
picrel

Anonymous No. 965242

>>965241
>smooth step makes it better
Not even, literally just use a ramp. You can make it zig zag if you so wished.
OP is a double faggot. Snsvkj

Anonymous No. 965243

>>965242
You are right, it exaggerates the effect and in reality the apparent smoothness only happens at extreme angles.