On 1/20/07, Daniel Rind <daniel.rind(at)chello.at> wrote:
General Rendering Plugin Question:
Well - is anyone even interested in using a different renderer than mental
ray in XSI? If yes, then why? What are the features you are looking for in
other renderers?
I don't have in-depth experience with other renderers, but there are
some features here and there that can make other renderers more
desirable than mental ray. I will refer to the 3ds max scanline
renderer and the Pixar implementaion of RenderMan.
- Shadow maps. Area lights are generally very prohibitive, so shadow
maps are an interesting alternative to get soft shadow edges. Here,
both the 3ds max scanline render and PRMan shine. Imho they are much
faster than the mr implementation (although I have not spent
considerable time with deep shadow maps). In 3ds max in particular I
could throw in several dozen of shadow mapped point lights and still
render in record time. On the other hand, PRMan gives you full access
to shadow map buffers, which means you can reused them through a
library, and can make troubleshooting a lot easier.
- Motion blur in PRMan is, well, *amazingly* fast compared to mental
ray. And I'm talking about deformed 3D motion blur. In fact, it's so
fast that you can reasonably budget a production with it, instead of
relying on 3rd party plugins that use motion vectors and the likes.
- Delayed read archives in PRMan. Several times a year *I pray* that
XSI would implement such a feature with mental ray. You can do sort of
delayed read archive already, however with a big limitation: the
referenced mi2 files are not renderable in XSI and XSIBATCH, only
through mi2 rendering (at least last time I checked). I was told by
reliable minds that this limitation is due to a name space issue in
XSI. Anyway, delayed read archive makes rendering unbelievably
scalable without hurting the user experience.
That's all I can think for now, I've been in mr world for a while :-)
Cheers
Bernard
---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
unsubscribe xsi
