Cosine-based specular env map LOD selection #7414
Conversation
bhouston
changed the title
|
@WestLangley thoughts? |
|
After things settle down, please. |
|
I've updated this to work with the latest DEV branch and the other changes. (Although I am very confused as to why a number of changes by @mrdoob in the dev branch appear in this PR -- is this github screwing up? I've never seen this before? Or is there a revert in the history?) |
|
The weirdest thing is that if I do a straight compare between these two branches, it doesn't show these @mrdoob commits on dev branch -- it is proper, so I am fairly confused here: |
|
I think it's github screwing up. It's been doing it for a while. I only look at the "Files Changed" section. |
|
Which, actually, shows that you're including the builds somehow. |
|
I submitted some extra stuff by accident as well. I'll remove it now. |
|
Yeah. Try rebasing. |
|
I removed two changes that probably are not necessary in this PR (although I didn't remove any builds, because there doesn't seem to be any) and it seems to be focused now: dev...bhouston:cosine_envMap_lod |
|
There is still a lot of noise in "Files changed". Hard to review until that it's clean... |
bhouston
force-pushed
the
cosine_envMap_lod
branch
from
a7da2c7 to
6e3c2e8
Compare
|
Rebase done. |
bhouston
changed the title
|
/ping @WestLangley |
|
This PR looks OK to me. |
Cosine-based specular env map LOD selection
|
Thanks guys! |
| @@ -1,4 +1,4 @@ | |||
| float calcLightAttenuation( float lightDistance, float cutoffDistance, float decayExponent ) { | |||
| float calcLightAttenuation( const in float lightDistance, const in float cutoffDistance, const in float decayExponent ) { | |||
There was a problem hiding this comment.
no need to have "const in" here really, it just leads to more verbose code and I've never seen it make a performance difference.
Built upon the simplified lighting PR: #7324
I've implemented the cosine env map lod queries in this PR based on the ideas here: #7411