Mailing List
Home
Forum Home
Softimage
Carrara
trueSpace
Dir3d-l
Maya - a powerful 3D animation and visual effects software
Macromedia Flash Development
Subjects
Cameras
scaleDown command
black out solved
Aircraft Tutorial
Mathematical XYZ ?
Its done This vs That
Its done first week
recommendations for screen video captures?
3DExplorer "Oddity "
New Director
ProTeam renewals
Fuel 's new websites (X post)
Blue peter create a make toy
targeting groups question
XPost: Shockwave 3D game ( sort of )
RES: RES: RES: Fish Modeling
Emitting particles from object intersection
Fuel 's new websites (X post)
Texturing
Big Break Contest Videos
New Plugins
Models and Texture on my updated site
Error Installing Patch tS6 6
Plasma?
Looking for Inspiration
Weird EMail Q
It 's done first week ?
Cherry not cranberry
New game
Camera Animation Problem
Particle plugins?
 
faux bevels?

faux bevels?

2004-04-05       - By Daniel Bachler

 Back
Reply:     1     2     3     4     5     6  

> [mailto:owner-xsi@(protected)] On Behalf Of Sam Cuttriss
> ive been wondering if a shader could create the illusion of
> microbeveling an object?
>
> when it comes to animating, UVing, adjusting form overall proportions
> etc microbeveling is the bain of my life.
> edge crease values are great for objects that are subdivided, but its
> often preferable to leave geometry unsubdivided.
> in these cases i would really like a shader that can distort normals
> across specified edges constrained within a radius...
>
> that way geometry could be left nice and clean and the edges would
> produce nice speculars.
> is that a feasible idea?

Well if you know how to write MR shaders there might be a way to fake it.
(Note that all this is theory, I have never written a MR shader myself, so
there might be some obstacles that render this approach impossible or some
shortcuts that make it a lot easier). Basically look up the bayocentric
(sp?) coordinates of the triangle you are rendering, take a look to see if
it is within a specified threshold with one of the selected edges and if yes
twist the normal vector of the current intersection point so that lighting
calculations think that there is no hard edge there. It might be a bit
awkward to transfer the edge data to MR because AFAIK you cannot simply link
to parameters within XSI in a MR shader, but you could pass it in a text
field or text file or whatever. Also, one should probably transform the
bayocentric coordinates into world space before evaluating the edge
distance.

Feel free to correct me if I make no sense :)

Daniel

---
Unsubscribe? Mail Majordomo@(protected) with the following text in body:
unsubscribe xsi