Multitouch interface: Preparations


Since I watched Jeff Han’s video and his multi touch interface a year ago, I was totally charmed with that technology and I was researching a lot about it. It actually is not a new technology, and is relatively simple.

There is a great community around the multi touch displays, and geeks all over the world are building their own devices of this kind, researching and improving them. The prototypes that you can find in that community don’t have anything to envy to commercial ones.

I’m a fulltime geek, and I’m the first to melt with this wonder so I was researching and doing my homewroks to start building my own device. I’m interested on it, because as a coder, I’ll have a new world where to experiment with new software paradigms.

When I first started this project, the first thing I needed was to choose what kind of technology I’ll be using. Currently there are two approaches to make a big multitouch table. They are based in the same concept, but are built in a different manner.

FTIR (Frustrated Total Internal Reflection):


This first approach is the most used to date, probably because jeff Han used it.

With this method we will use an acrylic panel, and then we will fill it with infrared light from LEDs. Using an acrylic instead regular glass does the trick, because it’s transmittance with infrared is top notch, while glass will not work with this physical phenomenon.

We fill the panel with IR light, and that light bounces inside it (total internal reflection). When we touch the panel with our bare fingers, we are frustrating the internal reflection, and refleting the light to the floor, where we have an infrared camera, or a normal camera with a filter that will detect these light “blobs”, so we can process the capture video with a computer to use them as controllers.

Over the acrylic panel we will put a diffusor fabric so we can project the image from the computer from the bottom.

Using the base fingers, the “frustration” effect works very well (even better with moistured fingers), because the skin of the fingers is semi transparent and deforms itself when applying pressure; but if we put the diffusor we need to put a lot more pressure that makes it unusable. To fix avoid that, we need an intermediate surface between the acrylic and the diffusor (known as complaint surface). It was extensively reseached, and the best result to date is making a thin transparent silicone (Sort A Clear Silicone Rubber). That will emulate our fingers skin, but makes the building process much more complicated (and expensive). But not as much as to stop an incurable geek

DI (Diffused Illumination)


The second method is the diffused illumination. In this case we don’t fill the panel with infrared (this tim eit can be acrylic, glass, or any other transparent or translucent material), but we light the bottom side using LED lamps, and trying to illuminate the entire surface. When we touch the glass, we refelct the light back to the floor, where our camera is.

This kind of display, can track fingers, but also fiducial markers (the famous Reactable is a clear example of this kind), but the generated “blobs” are not as clear as FTIR ones.

DI Panels are easier to build, because we don’t need to solder LED circuits, neither a complaint surface. But the device need to be isolated inside a box, and calibrating LED lamps is harder to achieve to obtain an optiaml result. Another disadvantage (but easily fixed), avoiding residual IR from ambient light (in special sunlight). If we don’t get rid of that, probably our display will detect inexistent “blobs” all the time without touching it.

I have to recognize that DI is very interesting because it’s capacity to track fiducial patterns, but I have chosen FTIR for my first project.

Right now I’m starting to order the first pieces, as the acrylic, the LEDs… and I’ll continue reporting my progress as soon a I receive them at my home.