Friday, April 07, 2006

RE: Lag in the Scrolling Gestures

A question from Rob "Xemu" Fermier regarding the lag in scrolling gestures.

[Rob] Very cool technology and a nice approach for demonstrating it. Using contemporary examples like Google Earth and Warcraft 3 is an excellent way of taking relatively abstract concepts and making them real for people. The gap between academia and "real world" software development is often pretty huge and it's great to see more approaches like this that can bridge that gap.

[Ed] Thank you. One of the things that I like most about my research is that computer games such as Warcraft III allow it to appeal to a larger audience. That is, my research is not only of interest to academics, but to the general public.

[Rob] I noticed the frame rate for scrolling, etc hitching a bit ... I'm curious if that was an artifact of the technology used to do gesture detection, or just the machine playing the game?

[Ed] Diamond Touch is a special type of input device for tables that can detect the gestures and movements of up to four people simultaneously. This input currently runs at a frame rate of 30 frames per second which does not seem like a lot but they are more than sufficient for gross gestures such as using a whole hand to pan a map. Also, since four people can interact simultaneously the effective frame rate is really 120 frames per second. Modern windows applications will often respond to the mouse at a rate of around 120 per second. This means that there is a bit of jerkiness in the Warcraft III panning gesture. This could be resolved by using interpolation of mouse events between frames. This is done in the Google Earth demo, thus the jerkiness is almost non visible.

Certainly the limitations of today's tabletop technologies would make it difficult to play Warcraft III as well as you can with a mouse and keyboard. But eventually, these limitations will be overcome and we will be able to interact with computer games in ways that were previously not possible. I detail some of the possibilities in a recent paper published at Pervasive Games 2006. It's important to realize that tabletop games are not replacements for mouse and keyboards over Warcraft III. Rather, tabletop games represent a new genre of gaming where people can interact face to face rather than having to look away from each other as we do with current console games. Being able to interact with rich hand gestures and speech provides an engaging experience that normally can only be found when manipulating physical objects such as a gun in an arcade.

The goal of this research is to understand the capabilities and limitations of speech and gesture tabletop interaction. This will hopefully inform the design of future multimodal tabletop games.


Rainer Schuller said...

Dear Mr. Tse,

I´m a huge fan of RTS Games and just saw ur W3 Video of the Control
Panel. First of all: Wonderfull Idea! Really love it!

I would be very interested what your opinion is about the release date
and the aprox. cost of such a Command Panel?

Thanks for the Reply and keep up the good work!


Casey J. Sillito said...

I’d just like to say that I think your work is pretty impressive. I’m really interested in new ways of using computers, especially ways that make it more social. In my own small way, I like to experiment with these things, but certainly nothing as innovative as what you do.

Zagary said...

Great job Edward,

So when will the real product be available in the market?

Kent said...

Very impressive work...

Would it be possible to do this with an projecter and a high res camera bounced off a mirror on the ceiling w/maybe a reflective glove to ease tracking...
Con: not as visually clean
Pro: less expensive(for now), portable