Re: Non Destructive Character Workflow

["Tim Leydecker" <BauerOink(at)gmx.de>]

Hi Frank,

I haven´t touched the Animation mixer in XSI, nor animated at all
so far but currently try to setup a PassSystem that should allow to
move the Lighting stage infront of the Texturing.

Basically, I do the lighting in greys plus a colored environment bounce
and later add a flat textures pass plus a selfoccluding reflections pass
from with the same colors.

Here´s a breakdown of passes I´m playing with, I´m 75% there:

Incident angle (camera position), no shadows/spec/refl/lights.
Environment sampled occlusion, as above
Ambient occlusion
Fake shadow pass using a white Lambert material (not just shadows)
(env) Reflections, via an amb_occ spread=0
Optionally, a global broad specular pass to determine bounce area.

For setting things up:

The Incidence controls the amount of reflections with the viewing angle,
facing get´s less reflections...

The environmentsampled occlusion is used as the main filler, replacing
a FG pass (to prevent flicker). You can try incidence and brad specular
as the mask and take whatever you prefer.

The Shadowpass is multiplied using a mask of the env sampled occlusion
(multed over the incidence if you like).

The stanadard occlusion is used to shade down caveats where neccessary.

Boiling things down, the incidence brings the most *filmic* effect and makes
it a joy to play around with front/backlights, e.g. the amount of fill and bounce.

It´s a bit difficult for me currently to squeeze the colors in correctly without
flattening out the tonalrange but as I said, I´m just ~75% there.

Cheers

tim

P.S: You could also check out a SSS pass (for the textures) looks quite good.











----- Original Message ----- 
From: "Daniel Bachmann" <dbachmann(at)mediagenerator.ch>
To: <XSI(at)Softimage.COM>
Sent: Monday, May 15, 2006 12:12 AM
Subject: Re: Non Destructive Character Workflow


> hi franky,
> actually i think this is a good idea to compile a workflow fact sheet.
>
> as newbie on xsi i cannot tell anything about that at this time but since i have 
> started to work on my new short that involves a lot of parallel working i can state 
> about my thoughts and workflow from time to time.
>
> so any pro hints are welcome about the non destructive workflow.
>
> daniel
>
>
> Frank Lenhard wrote:
> > Hello XSIlers
> >
> > i finally got a bit of time between projects and try out what i
> > actually bought xsi for :) the ability to have
> > modelling/texturing/rigging/animation in an almost paralle workflow.
> >
> > now my steps are rather small and slow at the moment but i would like
> > to hear if ppl out there really using the promised flexibility and if
> > there are traps and downfalls i should be aware of.
> >
> > i checked the xsi4 production series, which has some good hints here
> > and there. i read the manuals on the topic. but i didnt found a
> > "ready" workflow somewhere. like its advertised here
> > http://www.softimage.com/about_us/pdf/getting_into_character.pdf
> >
> > i imagine the following scenario and would be happy to hear if its
> > possible.
> >
> > naturally we would start with roughing out the character based on the
> > designs. that would give us a model which already got the right
> > proportions. but it might missing details like finders, face, toes
> > whatever.
> >
> > this model gets passed on to the rigger, who builds a basic rig into
> > it so it can be posed for animatics, layouts.
> >
> > thats what the animator is getting and starts on the preproduction.
> >
> > while the layouts are in works the modeller doesnt sleep but gets the
> > model refined, also based upon client response, which might come late
> > in the process. he applies textures+uvw's
> >
> > while the modeller is busy the rigger is building the actual rig for
> > animation. with new revisions ready from the modeller he just takes
> > the new model and has to reenvelope the newly added vertices. the rest
> > of the skin stays intact.
> >
> > the first rigged char for animation gets to the animator who starts
> > animating it.
> >
> > refinements are done to the model+textures and to the rig, as it
> > appears that it needs to support moves there where not anticipated and
> > where the rig currently breaks.
> >
> > the animator requests a quaternation spine instead of the fk spine.
> > but he doesnt wanna lose the already made animations.
> >
> > the face rig becomes available and the animator takes his ready body
> > animation and applies with the new face rig the facial
> > animation+lipsynch.
> >
> > finally everything is finished and goes of for rendering.
> >
> > in a perfect world the modeller would finish shortly before the rigger
> > who shortly finishes work before the animator.
> >
> > is THAT possible? at least that how the softimage advertisement sounds
> > to me :)
> >
> > maybe if there is interest and several ppl know about different
> > aspects about the pipeline, i can compile a little "how to" and "what
> > to be aware of" list, so there is something we can safely go along in
> > production.
> >
> > ciao
> > franky
> >
> > ---
> > 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


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