Understanding Motion Capture QuestWorld Criticism

From QuestFan
Jump to: navigation, search

The following is from Understanding Motion Capture for Computer Animation and Video Games by Alberto Menache. It was recovered online here.


The Real Adventures of Jonny Quest

The Real Adventures of Jonny Quest, produced by Hanna-Barbera, was a reproduction of an old cartoon series. Each episode consisted of at least 18 minutes of cel-animated cartoons and a computer-generated virtual reality segment of up to 3 minutes. THe CG segment for the first season was produced by Buzz FX. They decided to capture the body and facial motion of all the human-shaped characters because the budget was too low for keyframe animation. The motion data was provided by House of Moves, located in Venice, California.

Data for several episodes was captured at the House of Moves studio, and post-processing started as the rest of the production was being orchestrate.d As the motion data started to arrive, it became obvious that more work would be required to enhance the quality of the characters' motion; unfortunately, time and money constraints did not allow this to happen. The actual motion data wasn't the main problem -- it was the size difference between the performers and the actual characters. The characters were unsteady on the ground and seemed to be floating or skating. In addition, many shots that included interactions between characters or characters interacting with props did not match when applied to the digital models. The characters sometimes needed to be shown walking on uneven terrain or climbing rocks to other objects, which was impossible to match using motion data.

The Real Adventures of Jonny Quest is an example of a show for which motion capture data could have been a good idea, but that ended up as a failure because of the many mistakes that were made, starting with the budgeting. It is always a mistake to assume that all the characters' body motions in every shot can be captured. The data that can be captured has to be planned on a shot-by-shot basis. Furthermore, the captured body motion data is never plug-and-play. Always expect some kind of human involvement. You also cannot assume that the same character setup will work for all shots. In some cases, such as when there are interactions between characters or props, some areas of the character need to be animated by hand or by using the data combined with inverse kinematics. When a pipeline is set up for mass production, these cases have to be considered in advance. Some testing should be done before making ca commitment to such a large undertaking. Finally, when using a motion capture studio to collect data, one must remember that it is a motion capture studio and not a computer animation studio. It is up to the people who will use the data to decide if and how a certain shot will work.

Facial capture is a different story. As long as the performance is acceptable to the director at the time of capture and the character setup is well designed, facial motion capture can be a plug-and-play solution. In the case of Jonny Quest, such data was integrated into the pipeline much better than the body data. It produced a medium quality result, but still acceptable for the kind of television budget this project had.

From: Press, Interviews, and FAQs