7/14/08

Radiohead & 3D Scanning


Thom York underwent a scanning session to capture his performance of House of Cards. The video was created using two different scanning methods, laser and structured light. It's encouraging to see fine artists using technology instead of abhorring it. In an exciting move, Radiohead released the raw data from the session as a CSV sequence which I was able to process into a polygonal mesh to reveal what Mr. York looked like during the capture session. There is some surface warping due to the deliberate distortion of the process by shooting through a piece of glass or Lucite, but the results were interesting.

7/13/08

The Story So Far...

The core image processing script is functional, now it's time to get it functional for the user. I've decided that the Pane Layout will be the best to hold the different program elements because of the resizable areas. In the future I may make the image area it's own window but only if it doesn't impact the performance of the script by handling the cross-window chatter.

The trickiest part so far has been the file-handling. MEL has about 4 ways to handle file dialogs and each seems to have their strengths. For now I only need to worry about storing directory paths.

FIRST!

Master of Fine Arts Thesis Project : Creating an Optical Motion-Capture System from "off-the-shelf" Hardware and Software.

What is Motion-Capture? Motion-capture (commonly known as "mocap") is a method of recording the movements of a real-world performance, usually a human. The movements are then applied to a computer model which mimics the actions of the original performance. The performer usually wears a special suit which has markers or sensors which are used to turn the movements into a numeric representation, easily understood by a computer.

Who uses Mocap? Both the entertainment industry and research entities use mocap. Many of today's video games and movies use mocap to create performances where realistic movement is desired. Researchers use mocap to study the movement of the body in detail. Since the performance is captured as data, it is infinitely repeatable and allows for greater scrutiny with less variables.

Why write your own? Traditional drawbacks to a commercial system are its prohibitive cost and complexity, thus limiting motion capture to primarily larger studios and research entities. These barriers are on the verge of falling away due to the dropping cost of near-professional electronics and new methods of motion data capture by said electronics. Independent animators and small studios will be able to utilize motion capture as they have with other technologies which previously were unavailable to them such as 3D animation software itself and high-end workstations.

How does it work? The program is actually a script that runs within a 3D package called Maya. The user will record the performance using off-the-shelf video cameras and bring it into the script (which I've dubbed rotoCap). The video is analyzed frame by frame to synchronize the multiple data streams and to locate the markers placed on the performer at the time of the capture session. The markers are isolated in 3D space and used to drive the performance of a computer model.