If you follow any tech blogs, you’ve probably been coming across a lot of things multi-touch interface related. I was introduced to the idea of a multi-touch interface when I saw a video of Jeff Han demonstrating his own display at a TED Conference. But, it wasn’t until I found this short description some time afterwards that I became enamored with the thought of building my own. It was all I could think about for awhile, so much so that I took the leap into the unfamiliar world of physical computing to build my own. It was an awesome learning experience.
This was almost a year ago now, but I wanted to write a little bit about the process. It took around three months to get a working prototype (being in DAAP full-time doesn’t allow much free time for extracurriculars). I had a friend who was in to the idea also so we worked together.
Our list of parts
- 18" × 20" × .4" piece of acrylic
- Logitech webcam
- 20 OSRAM SFH485 Infrared LED’s
- Power supply
- Wires
- Silicon rubber surface
- Projection surface
The Basic Idea
We shine infrared light into the edge of the acrylic screen. Most of the light bounces around inside the acrylic until something interrupts, or ‘frustrates’, the light. This phenomenon is called frustrated total internal reflection, or FTIR. To put it simply, Putting your finger down on the acrylic reflects light downward.
Image credit Jeff Han, http://cs.nyu.edu/~jhan/ftirsense/ftirschematic.gif
We modified our webcam to be able to see infrared light and placed it underneath our acrylic surface. Software processes images from the webcam and tracks finger movement. Finger movements and presses are sent as events to an application that is projected onto the acrylic screen.
Fine Tuning
After we had the basic setup working, I had my dad put together a more permanent setup to house all the parts. He built a wood frame to hold the acrylic and LED’s and attached legs to it. I bought a DDR dance pad off of Ebay and sacrificed it for a silicon rubber layer housed inside. I attached this directly on top of the acrylic surface, and then added a layer of mylar for a projection surface on top of that. The layer of silicon rubber helped to improve the touch sensitivity. We were able to borrow a projector from our professor which helped keep our overall cost down to under $200.
It was a fun project, but we didn’t document the process nearly as well as we should have. You can see some more photos from the project on our Flickr album. For anyone interested in trying to build one of their own, I have to mention NUI Group as a great resource. I was fortunate to find the NUI community early on during my research, the people there were a huge help. They are also home to the open source multi-touch software Touchlib, which I used to test out my prototype.
