RE: Problems with SSS split and other render-wrangling questions....

Date : Tue, 17 Jul 2007 16:24:35 -0400
To : <XSI(at)Softimage.COM>
From : "Halfdan Ingvarsson" <hingvars(at)Softimage.COM>
Subject : RE: Problems with SSS split and other render-wrangling questions....
Title: Re: Problems with SSS split and other render-wrangling questions....
The problem is that in XSI you specify SRT as a local->global transform. Mental ray uses the inverse (most of the time) and you can't invert a scaling of zero (i.e. 1/0 == infinity). So in XSI we try to clamp the scaling at around +/-0.0001 (depending on which side of the zero you're at). This creates gigantic transform matrix values for the inverse which can muck things up (those cases depending largely on what happens on a internally within mental ray).
 
One approach would be to simply push those zero scale objects as dummies, but it that only works if two axes are zero. When flattening an object by scaling only one axis, this obviously won't work and you're liable to end up with the same problem again.
 
 - ½


From: owner-xsi(at)Softimage.COM on behalf of Andy Jones
Sent: Tue 17-Jul-07 07:47
To: XSI(at)Softimage.COM
Subject: Re: Problems with SSS split and other render-wrangling questions....

You can definitely hose MR pretty easily by scaling an object's center
to 0.  The geometry stays where it is, but the center has a zero scale. 
When you go to render, everything will be way slower than normal and
some things will just break entirely.  Scaling to 0 in general is not a
great idea, since the software would have to think a lot harder to do
the same optimizations it can do if the geometry is just hidden.  I
don't think it currently does that extra thinking, so the result is just
slower software.

I can see where a zero scale could present some problems for lightmaps,
but not so much for raytracing and scanline, since in those cases, the
odds of ever having to do anything with the geometry (aside from missing
it in a ray cast) are effectively 0.

-Andy

Chris Marshall wrote:

> It's just something I've never thought was a good idea. Probably from
> years back when the software wasn't so sophisticated! ;-)
> I've always thought that if the software has to do any calculations
> based on a value of zero, it could screw things up. I would tend to
> scale to a very small number then turn the object off.
> If it works for you though, that's fine.
>
> Dan Yargici wrote:
>
>> Hi Chris.
>>
>> I'm not sure why, I've been doing it for years with no problems at
>> all!  There are even other things in the same scene that are also
>> doing it!
>>
>> Are there any other know caveats with scaling to zero?
>>
>> DAN
>>
>>
>> On 7/17/07, *Chris Marshall* <chris(at)eclipsecreative.co.uk
>> <mailto:chris(at)eclipsecreative.co.uk>> wrote:
>>
>>     That's a good general point. Never ever scale anything to zero!!
>>
>>     Dan Yargici wrote:
>>
>>>     OK, I solved it last night, the issue was with an object being
>>>     scaled to zero.  As soon as I changed the animation to 0.01
>>>     instead of 0 - normal service was resumed... Interesting indeed...
>>>
>>>     Cheers for all the suggestions.
>>>
>>>     DAN
>>>
>>>     On 7/17/07, *Frank Lenhard* <franky(at)ixdream.com
>>>     <mailto:franky(at)ixdream.com>> wrote:
>>>
>>>         ah, i got this when i used volumics once.
>>>         i switched volumics off (render manager) and everything was
>>>         fine.
>>>         we had lots of area shadow casting lights and these
>>>         volumics(shadow
>>>         mapped) as well....
>>>
>>>
>>>         ciao
>>>         franky
>>>
>>
>>     --- Unsubscribe? Mail Majordomo(at)Softimage.COM
>>     <mailto:Majordomo(at)Softimage.COM> with the following text in body:
>>     unsubscribe xsi
>>
>>
>
>
> --
>
>
> *Eclipse (Creative) Limited*
> Pascoe House
> 54 Bute Street
> Cardiff
> CF10 5AF
>
> Cardiff Tel: +44 (0) 29 2047 0070
> Cardiff Fax: +44 (0) 29 2047 0071
>
> http://www.eclipsecreative.co.uk & http://www.eclipsecreative.tv
>
> Please Note:
>
> The information, concepts and ideas contained in this email are
> privileged and confidential, intended only for the use of the
> individual or entity to whom they are addressed.
> If the reader of this email is not the intended recipient, you are
> hereby notified that any dissemination, distribution or copying of
> this email is strictly prohibited and may be unlawful.
>
> If you have received this communication in error please return it to
> the sender then delete the email and destroy any copies of it.
> Eclipse Creative cannot accept any liability for any loss or damage
> sustained as a result of software viruses.
> It is your responsibility to carry out such virus checking as is
> necessary before opening any attachment which may be included with
> this message.
>
> Any views expressed in this email are those of the individual sender
> unless the sender is acting under instruction by an authorised officer
> of Eclipse Creative Limited.
>
> Copyright theft is illegal.
> © Copyright 2007 Eclipse (Creative) Limited
>


---
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.