Re: XSI Triangulation to CDH

[André Adam <a_adam(at)49games.de>]

Thanks François, Wessam, really appreciated, but we don't have any 
problems to write a triangulation, the problems is that we need to 
precisely mimic the XSI algorithm in the CDH. This is about the 
generation of normal maps, where distortions in the Texture Projection 
are compensated relative to the XSI triangulation.

We already have an OGL Realtime Shader which does a, say, 90 percent job 
of showing even rather complex materials. Where it fails are render 
passes for transparencies, since XSI does not allow for defining the 
point in time when something is supposed to be drawn in the viewport. To 
circumnavigate this and other minor issues (environment mapping just 
does not work out between OGL and DX) we have our DX engine running in a 
CDH, basically allowing the artists to work in the OGL RT View and, from 
time to time, get a perfectly correct preview from the CDH - except for 
the triangulation issue...

Anyone from Soft who would like to give a hint at the actual 
triangulation algorithm used?

Cheers!

   -André


François Painchaud wrote:
> Some good links about triangulation algorithms and libraries to keep 
> you busy:
> www.cs.cmu.edu/~quake/triangle.html
> www.cs.berkeley.edu/~flab/cs294-5/project2/mesh.html
> http://gts.sourceforge.net
> www.andrew.cmu.edu/user/sowen/software/triangle.html
> www.cs.berkeley.edu/~jrs/mesh/
>
> François
>
> Le 2007-02-07 21:45, Wessam Bahnassi a écrit :
> > A pretty easy triangulation algorithm (it might be the one XSI is using, but
> > I'm just guessing) is the GLUtesselator. For the matter of CDH, I had to
> > code DSK|ShaderBass for a while, and that was basically another real-time
> > rendering engine for XSI. It used CDH though, and what we did was to attach
> > a custom RT shader node to the material, and in our plug-in, we ask the
> > GraphicsSequencer to render all objects using materials with that shader
> > attached. This way, you get your triangulation automatically done by XSI the
> > way it does it for all other RT shaders.
> >
> > I hope this helps,
> > Wessam Bahnassi
> > Microsoft DirectX MVP,
> > Programmer
> > Electronic Arts
> > --
> > 'Talk is cheap because supply exceeds demand'
> >   
> > > -----Original Message-----
> > > From: owner-xsi(at)Softimage.COM [mailto:owner-xsi(at)Softimage.COM] On
> > > Behalf Of Alan Jones
> > > Sent: Wednesday, February 07, 2007 6:47 AM
> > > To: XSI(at)Softimage.COM
> > > Subject: Re: XSI Triangulation to CDH
> > >
> > > Hi André,
> > >
> > > Unfortunately we can't access this - though I did request the ability
> > > to access the underlying representation of any set of components.
> > >
> > > i.e. from a selection of polies get the polygons which make the
> > > subdivided version, for edges get the edges which represent the
> > > subdivided version of those edges.
> > >
> > > It would allow things like what you're suggsted as well as building
> > > proper wireframe rendering solutions (pushing the sudivided mesh as
> > > curves to the renderer etc).
> > >
> > > Anyone else who wants to add their name to that box would be great for
> > > getting Soft to push it up the priority ladder :-)
> > >
> > > Cheers,
> > >
> > > Alan.
> > >
> > > On 2/7/07, André Adam <a_adam(at)49games.de> wrote:
> > >     
> > > > Hello alltogether,
> > > >
> > > > does anybody familiar with the C++ SDK know if it is possible to
> > > > reference the triangles resulting from XSI's internal triangulation
> > > >       
> > > of
> > >     
> > > > polygon meshes from a given facet?
> > > > We are currently trying to get the triangle structure of selected
> > > > objects into our Custom Display Host, and since we have to draw the
> > > > objects through their polygon clusters (holding the material), we
> > > >       
> > > can't
> > >     
> > > > simply access the triangles using GetTriangles() from the Geometry
> > > > class, since this is only available for the whole object.
> > > >
> > > > Another approach would be to mimic XSI's triangulation inside of the
> > > > CDH, any chance of gettung access to the internal algorithm?
> > > >
> > > > Thanks for any hint, cheers!
> > > >
> > > >     -André
> > > >       

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