Week 7 // July 14 - July 20
Last week people in mocap lab were very busy running a two-day intensive motion capture session for the BDI company. They recorded a wide range of motions that will be used in their simulator to train the personnel. After the session was over Mo asked everyone in the lab if they could help clean-up the data. The process of cleaning up involves first retargeting the motion onto the human skeleton and then skinning the human model to the skeleton. After these steps are taken care of, the best part begins: the person has to playback a motion and every time there are motion artifacts such as self-penetration it is necessary to use Track editor to fix X, Y and Z motion graphs of the joints. After the entire motion looks natural there are just a few more steps left to finish up the process: center the model on the x-z axis to make sure that it is always in contact with the floor.
Since we had so many people working on cleaning up the data we acutely felt the problem I mentioned last week: the lack of stations and the inability to accommodate everybody. Following my prediction we got more stations with Maya installed on them! I was able to finish up cleaning my part of data right from the Graphics Lab :-)
Going back to my project: I spent quite some time sifting through numerous (200+) motion capture files looking for a jump-with-both-feet motion. For my previous bowing pencil animation I didn't include the motion of the second leg, which threw off the dynamics of the movement. That was on purpose but now it would be much better to try the techniques on a more natural-looking motion. I settled on the six files from which I chose a motion, which was the closest to what I was looking for and needed the less cleaning.
I also played around with the LaTeX some more. I found out how to fix the Adobe Acrobat crash (every time I converted my work to PDF and then closed a document, Acrobat would crash). For my final test document I adjusted the margins (the default one inch is too much) and inserted a picture (although a caption is not showing up).
 |
| A chaotically deformed mesh under lattice's influence |
On Wednesday and Thursday I continued working on my pencil animation. For some reason I kept having really bizarre chaotic deformations of the mesh. I try changing the lattice resolution and relative influence but I just couldn't get rid of the distortions. On Friday we had out regular meeting with Jessica. Laurel did a good job on her progress yet because of the setback I didn't have much to present. At least I was still able to participate in the discussion during which we decided to use more motions. My task was to meet with Mo and see if we can figure out together why I was having the problem. Jessica's gut feeling was that the issue is due to "the order of operations." Unlike the usual programming, where you may solve a problem in a variety of different ways, which all might be correct, sometimes with Maya, you have to repeat the exact sequence of steps in order to arrive at the solution.
After our meeting Mo met me at the lab and he tried to re-create the process of lattice skinning that I showed him last week. And what do you know! Somehow in my notes I didn't record a proper order of steps. Looking back at my last week's comments I see that I didn't denote that after "skinning the mesh to the skeleton and then binding the lattices to the specific joints" I unbound the mesh. The best way to do that was just to not bind the mesh in the first place! The way the mesh is positioned and the retargeting is done, it is possible to do lattice skinning right away.
Armed with this re-discovery I easily finished my pencil animations and started working with the flower model. I organized the folders to where I uploaded the necessary flower mesh, .v files with motions and the .vsk skeleton. I'm all excited about working on it next week!
| <<== Previous Week |