Announcement

Collapse
No announcement yet.

Treat glossy rays as GI rays = ALWAYS

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Treat glossy rays as GI rays = ALWAYS

    Vlado posted something with regard to "Treat glossy rays as GI rays" in the following post that got me experimenting with this today.

    http://www.chaosgroup.com/forum/phpB...ic.php?t=21340

    There is very little info with regard to this in the forums and docs but my tests led to some very interesting and positive results. GI is LC + Brute or IM. When this option is set to "always", we witnessed the following which seems almost too good to be true.

    First, it did seem to clear up overbright dots reulting from extremely high dynamic range images which we use almost exclusively for lighting without having to resort to subpixel mapping which we feel seems to dull and wash everything out.

    Second, the materials just seemed to look better with a lot more "pop" in the glossy reflections.

    Third, It seemed to render faster.

    So what are the drawbacks of this, if any, over the default settings? Also, there does not seem to be an override for this but if this seems as good as it appears at first look, a global override activating this for all shaders would be a very welcome addition to the SP if it isn't locked already. Am I missing something because as of now, this seems too good to be true. Why is this not the "default"?

  • #2
    Setting this option to "Always" has the side effect that for rays reflected from this material, the secondary GI engine will be used to evaluate GI. In your case this is the light cache. This means that the option is similar to a local "Use light cache for glossy rays" option - which accounts for the faster render times.

    The drawbacks of this option come also from this fact - as with "Use light cache for glossy rays", for very specular reflections you may be able to see the light cache samples in the reflections.

    Best regards,
    Vlado
    I only act like I know everything, Rogers.

    Comment

    Working...
    X