PDA

View Full Version : Core Bucket Rendering?



Paul Brunson
02-10-2009, 05:15 PM
I know this is pure speculation but I'm really hoping to see bucket rendering in LW core. I'm excited for instancing which allows us really high polygon counts if we're duplicating the same geometry all over the place.

But I have to say one thing I really like about modo is that you can pump tons of extra geometry into everything and subdivide the daylights out of things.

Even in 64 bit Lightwave I find myself running out of ram (8 gigs) when dealing with high detail models.

Sarford
02-10-2009, 05:31 PM
Maybe we'll have bucket rendering, maybe not. I'm more excited that since the SDK is now fully open, Sensei can make a new bucket renderer without the limitations of the old one. :-)

Sensei
02-11-2009, 12:02 AM
We'll see when they will release the first public SDK..
Open SDK (if you're understanding it as the same SDK for 3rd party plug-ins as built-in tools), doesn't automatically mean that limits are gone..

creacon
02-11-2009, 01:55 AM
I don't know how seriously you have been using Modo, but as far as polygon count goes, we had to switch back to LW Modeler, because Modo couldn't handle it anymore.

But that might change in the future.

creacon



I know this is pure speculation but I'm really hoping to see bucket rendering in LW core. I'm excited for instancing which allows us really high polygon counts if we're duplicating the same geometry all over the place.

But I have to say one thing I really like about modo is that you can pump tons of extra geometry into everything and subdivide the daylights out of things.

Even in 64 bit Lightwave I find myself running out of ram (8 gigs) when dealing with high detail models.

Sensei
02-11-2009, 02:03 AM
I don't know how seriously you have been using Modo, but as far as polygon count goes, we had to switch back to LW Modeler, because Modo couldn't handle it anymore.

He was talking about rendering, rendering in bucket renderer in Modo.. Modeling requires completely different strategy to optimization than rendering..



But that might change in the future.


Yeah. Modifier and history stacks will decrease chances for smart speedups and optimizations.. Everything will have to be generic and final..

Lightwolf
02-11-2009, 02:22 AM
I don't know how seriously you have been using Modo, but as far as polygon count goes, we had to switch back to LW Modeler, because Modo couldn't handle it anymore.
It depends a lot on what polygons you mean ;)
Raw, direct editable meshes -> LW wins hand down.
Instances and micro-polygon displacement -> a no brainer for modo

Cheers,
Mike

JBT27
02-11-2009, 02:22 AM
Or is it that the rewritten renderer for 9.x will simply be 'plugged' into Core?

I'm curious if two major new LW renderers have been in parallel development, given that the new 9.x renderer is one of the major accomplishments of that cycle.

Julian.

Paul Brunson
02-11-2009, 01:08 PM
Yes, I was thinking micro-polygon displacement.

For high polygon count models that aren't using sub-d surfaces, like an STL import or something. I usually rely heavily on qemloss3 and high resolution normal maps loaded with infinimap to keep the memory use down while getting as much detail as possible.

I'm curious how much of 9.x's render engine is in core as well, or maybe its more of a question of how much of core's rendering engine did they put into 9.x :)