Thread: [Algorithms] Bicubic normals for a bicubic world
Brought to you by:
vexxed72
Menu
▾
▴
gdalgorithms-list
|
From: John S. <jse...@ho...> - 2000-08-04 00:03:59
|
Hi! I'm new to this list, but already I'm asking questions... ;-) Does anyone here know how I can approximate the normals across a bicubic surface (i.e.: a Bezier patch) using another bicubic? I want to use normals for backface removal, and a simpler interpolation doesn't help much. On the other hand, an exact biquintic solution would be too expensive to compute. Thanks. John Sensebe jse...@ho... Quantum mechanics is God's way of ensuring that we never really know what's going on. Check out http://members.home.com/jsensebe to see prophecies for the coming Millennium! |
|
From: Akbar A. <sye...@ea...> - 2000-08-04 01:24:06
|
hello john, expensive? calculate the tangent vectors, compute cross products at each corner of patch. lamens; calculate normals to the parametric surface at the patch corners and assign these values to the relevant surfaces >interpolation doesn't help much you where probably just assigning the plane normals to the triangles, instead of doing the method above. this should improve your results. doesn't dx8 have some high level code that does this ?? watt and watt, "advanced animation and rendering techniques" has a wonderful section on this topic. page 83 hope this helps, if you need more details please ask. peace. akbar A. "We want technology for the sake of the story, not for its own sake. When you look back, say 10 years from now, current technology will seem quaint" Pixars' Edwin Catmull. -----Original Message----- From: gda...@li... [mailto:gda...@li...]On Behalf Of John Sensebe Sent: Thursday, August 03, 2000 7:02 PM To: gda...@li... Subject: [Algorithms] Bicubic normals for a bicubic world Hi! I'm new to this list, but already I'm asking questions... ;-) Does anyone here know how I can approximate the normals across a bicubic surface (i.e.: a Bezier patch) using another bicubic? I want to use normals for backface removal, and a simpler interpolation doesn't help much. On the other hand, an exact biquintic solution would be too expensive to compute. Thanks. John Sensebe jse...@ho... Quantum mechanics is God's way of ensuring that we never really know what's going on. Check out http://members.home.com/jsensebe to see prophecies for the coming Millennium! _______________________________________________ GDAlgorithms-list mailing list GDA...@li... http://lists.sourceforge.net/mailman/listinfo/gdalgorithms-list |
|
From: John S. <jse...@ho...> - 2000-08-04 02:35:11
|
Well, I'm trying to avoid five cross products (one per point generated - forty multiplies and twenty subtracts total) at every patch subdivision, not to mention that I don't have the Bezier control points once I start subdividing (using central diffrencing). I just want an approximation better than linear, so the threshold value the normals are compared against to determine backfaces can be as small as possible. A technique to approximate the normals with a bicubic is supposedly described the the Proceedings of SIGGRAPH '82, but since that's so old, I can't seem to get it anywhere. Somehow, I think finding the Holy Grail would be easier... ;-) Thanks again. John Sensebe jse...@ho... Quantum mechanics is God's way of ensuring that we never really know what's going on. Check out http://members.home.com/jsensebe to see prophecies for the coming Millennium! ----- Original Message ----- From: "Akbar A." <sye...@ea...> To: <gda...@li...> Sent: Thursday, August 03, 2000 8:21 PM Subject: RE: [Algorithms] Bicubic normals for a bicubic world > hello john, > expensive? > > calculate the tangent vectors, compute cross products at each corner of > patch. > |
|
From: Akbar A. <sye...@ea...> - 2000-08-04 03:19:01
|
hmm, if you know what your going to be subdividing too, what about a precalculate lookup method for the normal data? then once you hit a certain subdivision level, you just look up the normal in the precalculate array? i guess you have to make one of those trade off decisions. lots of memory versus cpu cycles in the real time sector. isn't it acm's job to keep track of the siggraph papers? peace. akbar A. "We want technology for the sake of the story, not for its own sake. When you look back, say 10 years from now, current technology will seem quaint" Pixars' Edwin Catmull. -----Original Message----- From: gda...@li... [mailto:gda...@li...]On Behalf Of John Sensebe Sent: Thursday, August 03, 2000 9:09 PM To: gda...@li... Subject: Re: [Algorithms] Bicubic normals for a bicubic world Well, I'm trying to avoid five cross products (one per point generated - forty multiplies and twenty subtracts total) at every patch subdivision, not to mention that I don't have the Bezier control points once I start subdividing (using central diffrencing). I just want an approximation better than linear, so the threshold value the normals are compared against to determine backfaces can be as small as possible. A technique to approximate the normals with a bicubic is supposedly described the the Proceedings of SIGGRAPH '82, but since that's so old, I can't seem to get it anywhere. Somehow, I think finding the Holy Grail would be easier... ;-) Thanks again. John Sensebe jse...@ho... Quantum mechanics is God's way of ensuring that we never really know what's going on. Check out http://members.home.com/jsensebe to see prophecies for the coming Millennium! ----- Original Message ----- From: "Akbar A." <sye...@ea...> To: <gda...@li...> Sent: Thursday, August 03, 2000 8:21 PM Subject: RE: [Algorithms] Bicubic normals for a bicubic world > hello john, > expensive? > > calculate the tangent vectors, compute cross products at each corner of > patch. > _______________________________________________ GDAlgorithms-list mailing list GDA...@li... http://lists.sourceforge.net/mailman/listinfo/gdalgorithms-list |
|
From: John S. <jse...@ho...> - 2000-08-04 03:37:08
|
I'm doing an adaptive subdivision, so I have no idea how far I'm going to be subdividing going in. John Sensebe jse...@ho... Quantum mechanics is God's way of ensuring that we never really know what's going on. Check out http://members.home.com/jsensebe to see prophecies for the coming Millennium! ----- Original Message ----- From: "Akbar A." <sye...@ea...> To: <gda...@li...> Sent: Thursday, August 03, 2000 10:16 PM Subject: RE: [Algorithms] Bicubic normals for a bicubic world > hmm, > if you know what your going to be subdividing too, what about a precalculate > lookup method for the normal data? then once you hit a certain subdivision > level, you just look up the normal in the precalculate array? > > i guess you have to make one of those trade off decisions. lots of memory > versus cpu cycles in the real time sector. > > isn't it acm's job to keep track of the siggraph papers? > > peace. > akbar A. > > "We want technology for the sake of the story, not for its own sake. When > you look back, say 10 years from now, current technology will seem quaint" > Pixars' Edwin Catmull. > |
|
From: Angel P. <ju...@bi...> - 2000-08-04 07:48:36
|
> Hi! I'm new to this list, but already I'm asking questions... ;-) > > Does anyone here know how I can approximate the normals across a bicubic > surface (i.e.: a Bezier patch) using another bicubic? I want to use normals > for backface removal, and a simpler interpolation doesn't help much. On the > other hand, an exact biquintic solution would be too expensive to compute. What about creating another patch with control points displaced one unit along the normal ( Easy to do with the control points at the corners, but what about the other control points? This is very easy with DX8 N-Patches). Then the normal can be computed like this: Normal( U,V ) = P2( U,V ) - P1( U,V ) |
|
From: John S. <jse...@ho...> - 2000-08-04 13:35:06
|
I've experimented with this, but as you've indicated, the closeness of the approximation hinges on the two approximated (middle) points. If someone could give me some help determining where those points should be, I'd certainly appreciate it. ;-) John Sensebe jse...@ho... Quantum mechanics is God's way of ensuring that we never really know what's going on. Check out http://members.home.com/jsensebe to see prophecies for the coming Millennium! ----- Original Message ----- From: "Angel Popov" <ju...@bi...> To: <gda...@li...> Sent: Friday, August 04, 2000 4:19 AM Subject: Re: [Algorithms] Bicubic normals for a bicubic world > > Hi! I'm new to this list, but already I'm asking questions... ;-) > > > > Does anyone here know how I can approximate the normals across a bicubic > > surface (i.e.: a Bezier patch) using another bicubic? I want to use normals > > for backface removal, and a simpler interpolation doesn't help much. On the > > other hand, an exact biquintic solution would be too expensive to compute. > > What about creating another patch with control points displaced one unit > along the normal ( Easy to do with the control points at the corners, but what > about the other control points? This is very easy with DX8 N-Patches). > Then the normal can be computed like this: > Normal( U,V ) = P2( U,V ) - P1( U,V ) > |
×
Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.
Thanks for helping keep SourceForge clean.
X