RE: Reference model limitations?

["Andre DeAngelis" <andre.deangelis(at)ubisoft.com>]

I see your point Vincent, but referencing after all is mainly for the benefit of the modelling/animation stage. 

 

There are inherent limitations to referring which will remain in place until Deltas expose the operations you mentioned.  Still, if you use your imagination and stick to some rules, there is much you can do already.  For example, if you want a cloth item to be a reference model, there is nothing to stop you making it it’s own reference model and connecting it to your main character ref model with constrains (though you are then struck with the fact that all the instances of the cloth ref model will be pointing to the same Syflex cache directory).  See Helges examples of what can be done on XSIblog here (http://www.xsi-blog.com/archives/174) and here (http://www.xsi-blog.com/archives/169).

 

When it comes to sims like cloth and particles etc. there is a lot of poly extraction, merging, retopologizing (is a that a word?), and other operations that you would never want to be reflected in the reference model.  Nor would the FX guys enjoy being forced to use more than what they need to crunch computationally intensive collisions etc.

 

So if you want to preserve a non-linear workflow from animation to FX, there is nothing stopping you from passing deltas/actions sources/mixers or point caches from animation to the FX crew.  After all, the more granular you can make your pipeline, the lest likely it is to break.

 

I would be curious to find out to what extent Mayas reference scenes allow simulations to be performed of references assets without committing them to local.

 

Finally, when you get to lighting, you will usually want to make the Ref models local anyway, especially if you want to optimize scenes and remove unnecessary detail.

 

---

From: owner-xsi(at)Softimage.COM [mailto:owner-xsi(at)Softimage.COM] On Behalf Of Vincent Fortin
Sent: July 25, 2007 4:44 PM
To: XSI(at)Softimage.COM
Subject: RE: Reference model limitations?

 

Hey Andre, yes that’s one way to do it but making references local tends to ruin the concept of referencing.

I think what Delta is aiming at (and I have to salute the clever work done by Soft in recent versions) is to create a non-linear collaborative pipeline from modeling to rendering. I have seen this working smoothly with Maya on big pipelines and Delta now simply seems to make it possible in XSI with the added advantage of having all the animation-sharing goodies built right into it. Again, I want to express my –almost- entire satisfaction, this stuff is very powerful and will hopefully change the way companies work in XSI.

 

However I just feel we’re still missing the point by denying access to new operators, poly hiding, point tweaking and the works.

The way software often work is there are connections linking the referenced model to the -later added- operators although I’m not sure how XSI works internally and that’s what I meant by “XSI’s object structure” in my previous post.

 

---

From: owner-xsi(at)Softimage.COM [mailto:owner-xsi(at)Softimage.COM] On Behalf Of Andre DeAngelis
Sent: Wednesday, July 25, 2007 3:27 PM
To: XSI(at)Softimage.COM
Subject: RE: Reference model limitations?

 

Simulation and FX are normally applied here after the animation has been complete.  The FX and sim and done on the models I local mode and the sim backed out to point cahce which is applied to the final model.

 

The models and made local anyway in lighting.

 

---

From: owner-xsi(at)Softimage.COM [mailto:owner-xsi(at)Softimage.COM] On Behalf Of Vincent Fortin
Sent: July 25, 2007 2:09 PM
To: XSI(at)Softimage.COM
Subject: Reference model limitations?

 

Hi there,

 

How come one can’t add operators and clusters to reference models? I’d have expected to be able to stack stuff on top of them while keeping the original file’s integrity.

I can only guess this is somehow due to XSI’s object structure that doesn’t have any special mechanism for handling this. Otherwise, from a user point of view, it just doesn’t make sense.

So assuming I didn’t overlook some basic info in the doc, how would you guys go about implementing a character finaling workflow in which an animator could do something as simple as connecting a ground collision to a referenced character that has Syflex on it?

 

Thanks

 

Vincent