The purpose of this experiment was two fold. I was looking to do something more interesting with motion capture, so I wanted to try a type of puppetteering approach, and it would also be the springboard for another project that's grander in scope than this one.

The entire piece is motion capture. The motion of the vase is provided by markers on a rolling chair. The motion of the eyeball is from a flimsy wooden dowel rod with three markers on it. These were both controlled by Fran Kalal who's an undergrad at OSU and works at ACCAD.

I provided the eye blink and the eye motion. The left to right eye motion was provided by a rotating assembly that was a concept piece for the Jane project (thought up by Ryan English, an ACCCAD graduate student), and the eye blink was just a marker on a stick. Everytime I blinked, I'd use my thumb to flick the marker along the stick in order to get some type of motion.

The motion for the vase and the eye were mapped directly in Maya using aim and point constraints, but the eyeball and the blink were done differently. These were mapped by expressions.

The eyeblink was set up as an expression. When the Y translation value of the marker was above a set value, the nurbs sphere closed, when below that value, it was open.

The two markers that were used for the eyeball looking around where parented to gemoetry that then used aim constraints to make one "look" at the other. I just ripped the Y rotation values from the constraint and applied it to the geometry of the eye.

The next step is to smooth out the mocap data, add more definition to the character and the environment, and possibly replicate the character so there are several of these floating around but slightly different. I also need to keep the lense of the eye from poking out when the eye is closed.

This is an early attempt with the aim constraints and simple geometry set up so I could seee what was going on.

First render. very dark and still contains some of the elements of the aiming devices

Next to last effort. Full on ray tracing, but I haven't smoothed any of the paths, so everything is still 100% mocap driven. (large file)

After years of being involved in motion capture, I've started to get bored with human motion. There is a lot human motion is good for, but I felt that this limits motion capture. I think there is a lot more that can be done with it. I'm currently working with getting realtime feedback to work reliably, and hopefully after that's done, apply it to this digital puppetteering approach. I think this adds to what the technology is capable of as well as opening it up to different creative ideas and ways to look at mocap data.

After working on the Jiggly poject, I got an idea for more of a chinese dragon approach. Human motion capture is from the waist down and the performer is holding a pool noodle with markers on it. The motion is then copied five times and offset in time and location to provide the different segments of the dragon.

The larger problems really were in the dynamics on the blue "ropes" that are suppose to be following along. I'm thinking of turning these into smaller dragons instead of ropes. The lanterns at the back are another experiment in dynamics that went haywire, but I think they're a little interesting.

The last one is playing around with creating animated goemetry with he motion capture data. I'm using the same sets of data moved in time and space, but I've instanced and mirrored the geometry and added some rotating glowing lights within it to set it apart from some of the other renders. The images being difficult to see as human driven motion is part of the process I'm exploring these days. I'm very interested in driving non-human characters with motion capture data.

Maya Playblast

A different interpretation