Many people have asked us, directly or indirectly, the question: ”Why would I want this?”.
This question has a short as well as a long answer.
The short answer is that it’s the next logical step in user interfaces, for the long answer we have to dive into the history of user interaction. When we first started communicating with machines, we did this using plain, generic buttons. After a while, buttons gave us insufficient freedom, as they provide limited input possibilities and full alphabet keyboards quickly took over. As time passed by, text based commands and responses didn’t satisfy us anymore either, and graphical interfaces were born, using a mouse as it’s main input. The graphical user interfaces got better and better and a slow transition to touch interfaces began, our first step into natural interfaces.
The key difference between the previous means of interfacing and touch interfaces lays in the fact we can touch objects in real life as well, unlike mouse and keyboard interfaces, which have no real life counterparts. After single-touch we progressed to multi-touch and gestures. Both of these techniques try to join virtual objects into our real-life world using simulated physics to make them react more realistic and dynamic to our input.
For us it was very important to find out exactly why these techniques make interaction more natural. we did this by comparing the virtual gestures to real life manipulation, like for example the “Pinch-to-Zoom” gesture known from several mobile phone OSes. The “Pinch-to-Zoom” gesture resembles how we would interact with something stretchable like a sock. By pulling each “side” of the sock outward we stretch up the opening to put the foot in, which resembles the zooming effect. Another, psychological, aspect is that by stretching the sock we reduce the effort needed to accomplish the desired result. This can be translated into the fact that zooming into an image enables you to view a particular part of the image enlarged, thus reducing the effort needed to see a certain detail of that image.
With this new knowledge in mind we can add some key requirements for wide-spread adoption of motion-tracking enabled interfaces to our list:
- Gestures must resemble natural manipulation techniques.
- Simulated physics are a must. They enhance the dynamic and raise the veracity.
As a result of these findings we’re currently experimenting with the capabilities of the Kinect sensor to calculate, or rather, approach the speed of a movement, given the fact that the Kinect’s frame rate of 30fps combined with the (inconsistent) USB-transport and image processing overhead make it difficult to do an accurate calculation.