Re: Some questions regarding shaders and SPDLs

Date : Thu, 14 Jul 2005 16:36:48 +0100
To : XSI(at)Softimage.COM
From : Alan Jones <skyphyr(at)gmail.com>
Subject : Re: Some questions regarding shaders and SPDLs
Hi Thomas,

Yeah - unfortunately that's it. FCurve would be much better, but we've
not got it :-(

Though writing lightwrapping shaders in mental ray isn't as simple as
you might hope. You'll have a few issues to contend with - though none
of them killers. Happy to detail it if you like.

Cheers,

Alan.

On 7/14/05, Thomas Helzle <xsi(at)screendream.de> wrote:
> Sorry, I'm unsure if I understand you correctly. Are you basically suggesting using a gradient instead of a FCurve?
> And am I right - the other problem is that I can't just call the FCurve/Gradient with some value to be evaluated by the spline function without the back and forth you describe? Autsch ;-)
> 
> Hm - funny that splines are so rarely available for shaders. It is so much more powerful defining for instance your specular or diffuse with a spline and it is much easier to visualize the result of a spline than that of a gradient.
> Especially cool is, if you can work on the full 180° light incidence, not just 90°...
> 
> In messiah I was able to "misuse" the animation graph editor for that by sampling a time other than the current:
> http://www.cgtalk.com/showthread.php?t=187319
> Scroll down to the 16th post to see it in action with a rather simple example.
> Can I access an animated parameter of a shader at a different time than the current in XSI from within the shader?
> 
> If anyone has found a workaround that uses a spline interface for shaders, I am all ears ;-)
> 
> Thanks for the feedback!
> 
> Cheers,
> 
> Thomas Helzle
> 
> 
> On Thu, 14 Jul 2005 16:05:22 +0200, Alan Jones <skyphyr(at)gmail.com> wrote:
> 
> > Unfortunately we can't do anything to use them yet. If you're after a
> > workaround I can suggest a very ugly one that will be a bit
> > uncomfortable for users.
> >
> > You can save some information with mi_shaderstate_set then have
> > another shader that reads that information in (mi_shaderstate_get) and
> > outputs it (as a color, scalar whatever).
> >
> > So to get something fcurve like you could save a scalar off, read it
> > in, pump it through a gradient mixer. Then plug the gradient mixer
> > into the parameter that you'd like to have the evaluated fcurve. So in
> > your shader code you'd save the state, then evaluate the parameter. If
> > you needed to do multiple evaluations remember to flush your cache
> > between them.
> >
> > Cheers,
> >
> > Alan.
> >
> > On 7/14/05, Wayne Williams <wayne(at)realtimegaming.com> wrote:
> >> I believe I remember someone else asking about the Fcurves for shaders
> >> and think the final conclusion was that at this time it isn't possible.
> >> Someone please correct me if I am wrong though! I'd love to have that
> >> kind of functionality for sure!
> 
> ---
> Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
> unsubscribe xsi
>

---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
unsubscribe xsi


Search the XSI List archives here or use the advanced search form to search across mailing lists. Searching help is available.
This site supposedly brought to you by Benjamin Grosser and the Imaging Technology Group.