mollyrocket.com - Old Tech E-mails

Old Tech E-mails

In the distant past, some of the things I wrote about code were posted only on private discussion lists. I've tried to collect some of my postings and reprint them, for anyone who's interested. I've learned a lot since many of these posts, so if you asked me about the topics today, I'd probably have a lot more to tell you, and a lot better information to give. But since I just don't have that much time to do re-writeups, this is as good as it gets for now. Where possible, the old images and e-mails are listed here. Note that the [CENSORED] stamp has been placed over addresses and stuff that would give away the name of the private forums where the original posts were made, or the people on those forums.

2003-06-06 15:25:00
by casey

Discontinuous b-spline curve fitting

I did some interesting tricks to make my b-spline solver handle multi-order discontinuities. This lets it do stuff like correctly fit the motions of things that teleport, bounce, etc., without having to break the curve up into sections (you still get one nice big b-spline, that you can evaluate normally). I wrote up some notes on how I did it.

Text of e-mails

2002-10-24 01:45:00
by casey

Distance from plane to axis-aligned ellipsoid

At one point, for some unknown reason, there seemed to be some confusion as to how to find the closest point to an ellipsoid from a plane. So, I tried to dispel that confusion. Hopefully, I succeeded.

Text of post

2002-10-22 02:39:35
by casey

High-res to low-res model map projection

I wrote an extremely fast and accurate high-res to low-res mesh projection tool that I included in the exporters for Granny. It actually knew the exact worldspace quadrilateral mapping for each texel in the projecting texture maps, and would cast rays out of the resulting shape.

Text of e-mails

2002-02-01 00:00:00
by casey
6 items

Pose-space deformation

I wrote a Pose Space Deformation solver. It was cool. There's no art tool that lets you author them, but I was hoping someday there would be, so I was ready to add it to Granny if they ever did. They still haven't. Anyway I had a solver in MathVis that let you author poses and play back deformations. You can see from these simple screenshots how you can totally control the creasing, unlike the rigid example where it just interpenetrates and sucks. I don't have screenshots of matrix blending, but that sucks only slightly less ass than the rigid case, as we all know.

2002-01-18 12:26:46
by casey
3 items

Periodic loop decomposition

I developed a technique for looping motion extraction called Periodic Loop Decomposition. While it works very well for single motions, it breaks down for transitions between motions mid-loop, and so I don't think it's quite where it needs to be to be a standard tool. But someday it might be worth looking at again.

Text of e-mails

2001-06-29 03:12:05
by casey
3 items

Quaternions

Quaternions for real-time animation were very poorly understood when I started working with them. The conventional wisdom of the day was that to interpolate quaternions, you had to use something expensive like SLERP, which is what I did at first. But I later discovered with some experimentation that you have many more powerful options. My initial e-mails on the subject are here, along with the original screenshots from my cute little MathVis test harness.

Text of initial e-mails

Text of later e-mails