Digital Media Lab logo

Research

Digital Media Lab logo

Research in this lab/studio consists of a variety of activities. We do publish papers in journals and conferences based on what we learn in the lab. We construct artistic objects and evaluate them, and allow audiences to assess them. We create.
What are we workin' on now?? A lot of our work concerns the use of games and other interactive tools for use in teaching and training. However, the sexier work (IE that has cool images and stuff) can be found below:

Sound and Music as Input to Games Focusing on music, the underlying problem is how to recognize a musical note. Pitch is not just a matter of frequency, but has psychoacoustic elements. We have a system that recognizes pitch pretty accurately and in real time. It has been used in a simple game, and we expect to develop more games, especially for the teaching of music performance.

Short video:





I'Powahsin Project This is an attempt to create a computer game within which players speak an aboriginal language. It is intended as a means to teach such languages generally, and is part of a larger virtual world project called Turtle Island.
It should have been easy to fund this project, but efforts to do so have failed, and so progress is slow but work continues. See a short article about I'Powahsin

See the Turtle Island web page.


Virtual Theatre Live theatre in virtual spaces, or 'Puppetry of the Pixel.'

See a short article here.


Exercise Revolution
We have built a dance game that uses heart rate input from the payer. This game can maintain the palyer's heart rate is in the aerobic range; if the rate is too slow, the game speeds up the music, and if too fast then the music will be slowed down.

We hope to add in-shoe sensors to eliminate the dance pad, and will be experimenting with positional audio.

Booze Cruise
An older project that got a lot of attention - a game that teaches the perils of impaired driving. Here's a short b-roll video.






Cutting the Strings
Multiplayer games require a great deal of bandwidth on the Internet because the current position of all of the actors in the game must be sent back and forth to the server on a regular and frequent basis. We propose that avatars should be able to learn how their player reacts to specific situations and be able to behave with some autonomy. So long as the avatar is behaving as the player wishes, the player need not intervene, and the network traffic needed to support this is very small. Thus, the avatars are behaving more like actors than puppets.
A former student, Nathan Sorenson, has built a prototype of this system for the basic case as a proof of concept. We can see a better than 10 fold decrease in network traffic so far.


Send Us Mail!