AWS Thinkbox Discussion Forums

Custom Geo Proxy Object Support

This seems not to work but it would be super cool if it did as nobody else can seem to get it right. fr-proxy to TP seems to be broken, attached 10k vray proxies to pflow particles via script on animated geo not very practical, plflow won’t handle proxy objects properly as shape instances.

Solution Frost! You can create particles in TP or PFlow, assign rotations, even better populate the shape integer channel and use ten different proxy shapes!, ect, ect.

Custom Geometry - vray-proxy or fr-proxy

As of now, understandably, it errors out with a can’t create trimesh.

Any thoughts on this?

The Proxies of VRay, fR and mr are custom loaders for the render engine and don’t really provide useful geometry data to generic Max plugins.
So the question is: Could FROST create a bunch of INSTANCES of these objects in the scene pre-render, set their properties based on the particle data and then destroy these instances post-render. So FROST itself would not render but will behave as an instance manager. It sounds dangerous, but it might be possible.
Will let Paul give his opinion…

Aye understood. So it really doesn’t then currently behave as one, in a manner of speaking?

It would be an awesome way to control LOD of say a forest.

I am sure someone said that about rendering ten gazillion particles too :laughing:

In the case of VRay, there might be a way to feed the proxy data to the renderer without creating thousands of scene nodes via the SDK, but I don’t know enough about that.
I suspect that fR supports something like that to allow TP to do billions of polygons renderings.
Not sure how this would work with mr though.

I could ask Edwin although I am sure it would come out in some programmatical dialect that my ears aren’t fully attuned too.

isn’t MR technically dead?

cb

Mental Images is. Mental Ray - probably not so much.

after gelato fiasco, i wouldnt hold my breath that nvidia will sell a software renderer for very long! :open_mouth:
it isn’t their core business, and releasing the people who would fight for it seems to me like a sure fire start to dissipating the interest.
it’s like apple and shake, final cut etc. when 98% of your revenue comes from something else [nvidia: selling chips/cards] the other 2% tends to get neglected.

cb

Thing is, they don’t sell it. Most of MR is just OEM licensing to companies like ADSK. The rest is services like Reality Server. Nvidia gets to sell the hardware, and someone else sells the software/service. I think they still come out ahead.

You have to keep that 2% around to develop for the other 98%

So I guess that means we can leave out the MR Proxy and just work with the FR and Vray proxies? :smiley:

We haven’t done anything specific to support mental ray so far, why start now… :wink:

I like this idea, and I’ve added it to our wish list. I’m not sure about the details but I think we could make something analogous.

Eww I shall await with much enthusiasm and eagerness :smiley:

You’d need to create data specific to the rendering engine to support this. You can’t change the material on an object after it’s been instanced by the renderer unless you area able to talk to the renderer. Way beyond just passing the triangle mesh. Not saying that cebas or chaos group or nvidia would have a problem with that, just saying it would be in their court to allow you to do that.

Privacy | Site terms | Cookie preferences