View Full Version : Anti-aliased Buffers in Compositing Buffer Export

11-05-2013, 11:20 AM
LightWave 11.6 addendum - "The Compositing Buffer Export now supports anti-aliased buffers. This means that requests for Anti-aliased surface ID, Object ID or even Depth buffers can now be met."

Someone was able to generate antialiased Surface ID or Object ID buffer ? My buffers are still aliased as before.

12-03-2013, 02:27 PM
Same here. RAW buffer export is still aliased and I can't find anything that I'm doing wrong with the setup. A bug?

12-03-2013, 02:31 PM
I found the issue. It's adaptive sampling. Disable it (sadly) and the buffer is now getting anti-aliasing like you'd expect. You just take the render time hit I guess.

12-04-2013, 09:45 AM
I disabled adaptive sampling and buffer is still aliased.

08-18-2014, 04:55 AM
Hope you don't mind me bringing this back. Sanchon, did you find a fix? Is there a list of buffers that get anti-aliased somewhere or should we assume it's all of them?

I'm using Node Pixel Filter and DPs Amb Occlusion node replacing the Luminosity channel, then using exrTrader to export a png seq. I had hoped it would get anti-aliased but I'm getting a very strange mixture of the two.

08-20-2014, 09:37 AM
Forget Compositing Buffer Export, use DP's node pixel filter/image filters instead. If not for these plugins I would have dropped lightwave for something else.

08-20-2014, 03:33 PM
I am, I think. Using Pixel Node Editor with an Amb Occusion node overwriting the Luminosity channel. Then exrTrader for output. LW still needs to do the antialiasing of that channel, yes?

08-20-2014, 06:51 PM
in Node Pixel FIlter have you double clicked the Pixel Filter when in the node editor and checked on 'Multi Thread/Persp.AA and 'Raytrace/Glob 3D shad.?

08-20-2014, 10:01 PM
I still use "surface custom" buffers - they are always antialiased but you must manually set surfaces. I like this method because I can get only these buffers that I exactly need without messy aliased SurfaceID image buffer.

08-21-2014, 04:15 AM
That's an interesting approach, Sanchon.

My approach should work though, yes?