Dave, are you talking about darkened edges around separate textured polygons/suraface or where two textures join on one surface? I tried to find out from Sun engineers at siggraph 2000 if the first was an issue in Java3D. I was encountering it in VRML and wanted to know if Java3D would have that problem. Wasn't able to get a difinitive answer apart from yes, you have control over clamping etc.
"Seemed to me the solution was to add throwaway pixel border around the texture and use texture coordinates (1/width) .. (width-1)/width instead of [0,1]." I have heard of that solution too but not tried it as I was encountering this problem in VRML and had no control over it except by turing off smooth texturing in the VRML browser (which I assume to be billinear filtering). This gets rid of the dark edges around textured tiles but the texture looks worse (pixelated). >From a discussion by C/OpenGL programmers I once saw I think they concluded this was >not a complete solution - I'll try and dig up the info. cheers, Chris :) The correspondence I had was a bit confusing to me. David Yazel wrote: > I have determined that turning bilinear filtering off helps, but does not > seem to solve the problem. So setting the min and mag filter to > BASE_LEVEL_POINT does make it better. CLAMP_TO_EDGE also does not solve the > problem. Surely someone has successfully clamped direct x textures? > > Dave Yazel > > ----- Original Message ----- > From: David Yazel <[EMAIL PROTECTED]> > To: <[EMAIL PROTECTED]> > Sent: Saturday, February 09, 2002 11:02 PM > Subject: [JAVA3D] Edge clamping in direct-x > > I seem to be having a difference between edge clamping on direct-x and > opengl. If I have a 64x64 texture with an image of a box on it, and I use > texture coordinates [0,1] in opengl with mode CLAMP it works perfectly, > but in directx it loses the last pixel. I seem to dimly remember this issue > way back in my mind. Seemed to me the solution was to add throwaway pixel > border around the texture and use texture coordinates (1/width) .. > (width-1)/width instead of [0,1]. > > Is this right? Is there any way around? > > Dave Yazel > > ---------------------------------------------------------------------------- ---------------------------------------------------------------------------- This Email may contain confidential and/or privileged information and is intended solely for the addressee(s) named. If you have received this information in error, or are advised that you have been posted this Email by accident, please notify the sender by return Email, do not redistribute it, delete the Email and keep no copies. =========================================================================== To unsubscribe, send email to [EMAIL PROTECTED] and include in the body of the message "signoff JAVA3D-INTEREST". For general help, send email to [EMAIL PROTECTED] and include in the body of the message "help".
