Hunter Williams Posted August 13, 2019 Share Posted August 13, 2019 Hey guys, I'm trying to deform a volume "contrail" by a spline with direct projection without the use of converting to points and back. I've tested this method on points and it seems to work, but am unable to get it working on volumes without major stepping. It relatively follows my curve, but I can't seem to figure out where the stepping is coming from - it works perfectly with points and conserves it's thickness, so I'm pretty stumped. I attached the project file and few screen grabs. Let me know if anyone has any ideas volumedeform_v03.hipnc Quote Link to comment Share on other sites More sharing options...
jonidunno Posted August 13, 2019 Share Posted August 13, 2019 Hey Hunter, Is this what your after? volumedeform_jl_v03.hipnc Quote Link to comment Share on other sites More sharing options...
Hunter Williams Posted August 14, 2019 Author Share Posted August 14, 2019 36 minutes ago, jonidunno said: Hey Hunter, Is this what your after? volumedeform_jl_v03.hipnc Hey Jon! Thanks for the reply. I should have specified- this is for post-sim deformation, and I was only using the tube as test geo- so doing to deformation on the geometry would not be possible in the theoretical setup. I think I got something working though - volumeDeformBySpline_HW20190813.mp4 missiles_v02.mp4 1 Quote Link to comment Share on other sites More sharing options...
fuat Posted August 14, 2019 Share Posted August 14, 2019 (edited) What was the solution? Edited August 14, 2019 by fuat Quote Link to comment Share on other sites More sharing options...
jonidunno Posted August 14, 2019 Share Posted August 14, 2019 17 hours ago, Hunter Williams said: Hey Jon! Thanks for the reply. I should have specified- this is for post-sim deformation, and I was only using the tube as test geo- so doing to deformation on the geometry would not be possible in the theoretical setup. I think I got something working though - volumeDeformBySpline_HW20190813.mp4 missiles_v02.mp4 No worries, glad you got it working. Looks GREAT! Quote Link to comment Share on other sites More sharing options...
Hunter Williams Posted August 14, 2019 Author Share Posted August 14, 2019 51 minutes ago, fuat said: What was the solution? Well for the stepping problem, i think it was caused by me accidentally sampling the interpolation data from scattered points as opposed to interpolating directly on the volume. The tool works well enough for my use-case, but the volume doesn't match 1:1 with the line, which is frustrating.. so maybe I'll have time to fix it in the future Feel free to improve /update and post it back up volumedeformBySpline_v05.hipnc 1 Quote Link to comment Share on other sites More sharing options...
fuat Posted August 14, 2019 Share Posted August 14, 2019 (edited) Thanks for the reply. Not in front of houdini right now. Will have a look and learn from it when back from vacation. Very interesting. Could not look into the scene file. I once read that trilinear interpolations work quite well with fields. But dont know if this applies to your scene. Edited August 14, 2019 by fuat Quote Link to comment Share on other sites More sharing options...
Hunter Williams Posted August 15, 2019 Author Share Posted August 15, 2019 19 hours ago, fuat said: Thanks for the reply. Not in front of houdini right now. Will have a look and learn from it when back from vacation. Very interesting. Could not look into the scene file. I once read that trilinear interpolations work quite well with fields. But dont know if this applies to your scene. Yeah I've heard that too! So this project is my intro into this type of math and whatnot so I'm still learning this stuff, but I think that trilinear interpolation would need a 3 dimensional deformed object to implement as you would need to sample linear interpolations from 2^3 or 8 points, which would mean i couldnt use a spline or I'd have to sweep it to make a volumetric representation based off of the deformed spline. I downloaded a few other peoples project files to try it seems much slower than my bilinear interpolation. Obviously they're going for deforming volumes by a 3d grid so they need the extra dimension of interpolation, but for what I'm doing I think it would be more efficient with bilinear interpolation. Someone at WETA used RBF interpolation which I would love to know what they did- I'd imagine its faster as its radially interpolating from the source point, but I have absolutely no idea how or why haha I think RBF is used for pose-morph / blend shapes? Kinda talking out my ass and have no idea what I'm talking about, so definitely could be wrong on multiple fronts Would love for someone who knows more to set me straight Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.