Great work, source looks interesting. Die not know that indirect diffuse can be gathered via horizon search. Why did you still need AO if you have recursive multi bounce light transport? For cubemap influence?
-
-
-
Cheers! I didn't know either, so had to try it. Looks like
@Patapom2 beat me to, but I think I have a small improvement in there :) It seems to match "reference mode" done via ray-tracing that samples screen color. AO is for (pre-convolved) envmap indeed; should be bent cones. - Još 6 drugih odgovora
Novi razgovor -
-
-
Have you read the HBIL paper? It should be of interest for ya'

-
I have not, though I probably should have! Thanks :) Section 2.3.2 sounds similar to what I'm doing, though I have an extra term to clip the gathered samples' influences to their positive hemispheres. I believe it would fix the light bleeding in Figure 18.
Kraj razgovora
Novi razgovor -
-
-
haha you just had to do it after that silly other screenshot? :)
-
Silly is essential sometimes!
Kraj razgovora
Novi razgovor -
-
-
I would love to see more details on the implementation for this

-
- Još 1 odgovor
Novi razgovor -
-
-
vulcan't be bothered?
- Kraj razgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.