PDA

View Full Version : compositing buffer, no joy with hypervoxels....



stylus10
05-19-2015, 06:27 PM
so how do i do it? get hypervoxels to show up with color and definition using the composite buffer? in 11.6.3?

Every4thPixel
05-20-2015, 03:25 AM
You don't. Not Possible.

prometheus
05-20-2015, 05:35 AM
select all other items in scene editor and make them unseen by camera, render out hypervoxels in seperate pass.
Would be nice if the lw team could make it work in thecomposite buffer though, but they need to improve hypervoxels first of course :)

lardbros
05-20-2015, 06:38 AM
As prometheus says... or could try with 'matte objects' in the object properties panel. So all objects render black.
Helps when objects pass infront of the HV, they'll obscure them when you put them together in post.

jwiede
05-20-2015, 08:15 PM
There are workarounds to generate something akin to an HV buffer/pass (as noted), but really, LW3DG needs to implement proper, distinct HV buffers for Lightwave's Composite Buffer output.

Sensei
05-20-2015, 11:45 PM
It should be possible to make 3rd party plugin for this and have generated just volumetrics in it. So it's just a matter of your's "determination" ;)
It can work parallel to regular render (not needed another pass).

3DGFXStudios
05-21-2015, 02:02 AM
I would render a color matte and use that to comp things but it's not really a solution to get HV render passes out of LW.

jwiede
05-25-2015, 02:57 PM
It should be possible to make 3rd party plugin for this and have generated just volumetrics in it. So it's just a matter of your's "determination" ;)
It can work parallel to regular render (not needed another pass).

Are you referring to the using the VolumetricEvaluationFuncs->* APIs to generate the volumetrics' contributions? If so, how would you "advertise" that output to be a compositing buffer output?

(curiosity only, any interest I had in writing LW plugins died long ago)

Seems like a lot of work (esp. considering custom cameras/sub-pixel/etc. for view frustum ray origination, AA, etc.) just to replicate a process they already have to do, and for which they already have everything easily available to produce distinct volumetric(s) output buffer(s). If people are willing to pay, though, sure, why not. I might buy such a plugin were it reasonably priced and available for Mac. ;)

If the processing really does involve re-raycasting the volumetrics chain (incl. HV shading, etc.), seems like generating the buffer would basically double the volumetrics' render (& AA) cost, correct? Depending on exactly when it ran in parallel, ofc, that might not add so much in reality.

Sensei
06-02-2015, 06:48 AM
If the processing really does involve re-raycasting the volumetrics chain (incl. HV shading, etc.), seems like generating the buffer would basically double the volumetrics' render (& AA) cost, correct? Depending on exactly when it ran in parallel, ofc, that might not add so much in reality.

Full HV render in 5.7 sec:
http://forums.newtek.com/attachment.php?attachmentid=128416&d=1433249137

while HV matte pass in 1.7 sec:
http://forums.newtek.com/attachment.php?attachmentid=128415&d=1433249137