Urban Zen
January 18th, 2006, 19:33
Hello !
This topic follows hard tries I gave otpimizing Final Gather in a special way, as I'm now familiar with mrformaya basically.
I'm using an envSky as IBL texture for fg and it works great with very low accuracy settings, because color info is spreaded from everywhere homogeneously. Renders are lighting fast.
I figured out also that one could use the actual sun in order to light a scene by crancking up "overall" and "sun" (+halo) brightness. The problem here is that VERY HIGH accuracy is needed because the very little angular size of our sun.
That's where the mess started for me.
First I've tried to feed miDefaultOptions.finalGaterRays with the truncated alpha of an alternate envSky via script node, setting sun color up to 8192 and sky color set to 64 (remember these were intented to feed fgr attribute). This worked fine except the value was computed once by frame, thus giving a nice +-2200 average. I've found in the doc attributes evaluated by sample but could'nt figure out how to implement them... There is some limitation here ... :(
Finally I understand everything happens at fgmap creation time.
The basic idea is:
- When a fg ray exceeds a certain color value in an initially low accuracy fgmap,
- then feedback mr so he computes same ray with higher accuracy,
- then goes back to low accuracy.
Kind of an adaptative final gathering I guess !
Finally, here's my question :D
Is this implementable ? via mel (I would prefer so) or by .mi editing (that's where I'm a no0ob) ? Or is the whole stuff a myth ?
Thank You !
This topic follows hard tries I gave otpimizing Final Gather in a special way, as I'm now familiar with mrformaya basically.
I'm using an envSky as IBL texture for fg and it works great with very low accuracy settings, because color info is spreaded from everywhere homogeneously. Renders are lighting fast.
I figured out also that one could use the actual sun in order to light a scene by crancking up "overall" and "sun" (+halo) brightness. The problem here is that VERY HIGH accuracy is needed because the very little angular size of our sun.
That's where the mess started for me.
First I've tried to feed miDefaultOptions.finalGaterRays with the truncated alpha of an alternate envSky via script node, setting sun color up to 8192 and sky color set to 64 (remember these were intented to feed fgr attribute). This worked fine except the value was computed once by frame, thus giving a nice +-2200 average. I've found in the doc attributes evaluated by sample but could'nt figure out how to implement them... There is some limitation here ... :(
Finally I understand everything happens at fgmap creation time.
The basic idea is:
- When a fg ray exceeds a certain color value in an initially low accuracy fgmap,
- then feedback mr so he computes same ray with higher accuracy,
- then goes back to low accuracy.
Kind of an adaptative final gathering I guess !
Finally, here's my question :D
Is this implementable ? via mel (I would prefer so) or by .mi editing (that's where I'm a no0ob) ? Or is the whole stuff a myth ?
Thank You !