Multitouch prototype: First video

I didn’t post in a time, but here is the first multitouch video. It lacks of projector yet, but works.

Multi touch interface (part 5): First FTIR test

Here my first FTIR test. It works really well. I didn’t used a visible light filter yet, but you can see it works.

Multi touch interface (part 4): Almost done

Finally some pics of my almost finished prototipe. The electronics are feeded by 12 Volts, and it works fine.

I was doing some quick tests with my mobile phone to check if it generates blobs, and yes, it does blobs barely touching the screen. Them are soft, but it’s due to my mobile cam cause it have IR filter. I couldn’t do tests with the PS Eye without filter, but judging by the images that I have taken with it, I think it will capture the blobs very clear. 

Now I only need a projector, but if I can’t get one soon I will try with an standard monitor and a mirror to do some real tests.

Playstation Eye

I bought recently a Playstation 3 (probably I will write about it in a near future, because of it’s amazing Power Cell processor), then I bought too it’s camera, the PS Eye, cause it’s specs are great for multitouch, and it’s price very affordable (another cam for MT with a similar specs it’s about 300€, and not so easy to obtain).

Sony PS Eye:

• 4 audio input channels: 16 bits/channel, 48kHz, SNR 90db 
•  FOV: 56º o 75º
•  640×480 at 60 fps or 320×240 at 120 fps
•  Uncompressed video, optionally JPEG compression.

60 fps is just what we need to a smooth blob tracking, but we can reach 120 to get a real smooth tracking.

The problem with this camera is to run it on a PC. I was starting to write some drivers for windows, but finally I found that the great AlexP at NUI Group was quicker an already published the first working driver for the PS Eye (and consecuently I saved my time writting my own). The AlexP driver it’s a direct access driver, and he still have to write a wrapper to use it with touchlib. I think he will bring us the wrapper soon.

For the moment I already removed the IR filter from the cam, and I will use some exposed photographic film, or a floppy internal disk to act as a visible light filter. Then I will capture only the IR light with the camera, not the “image”, to proccess it with touchlib (but probable I will write my own libraries based on OpenCV to adapt it to may needs).

About my prototype, I already solded all the circuit and soon I will show some pics.

Multi touch interface (part 3): The LED frame

Today I was starting with the LED frame to illuminate the acrylic panel. As I wrote before, for the prototype I’m using a couple of shape “L” aluminium profiles, 90cm each (the plexyglass widh is 70cm, but I toke 90cm to use as a bracket over the trestles). I will put 25 IR LEDs on each profile.

I still didn’t any test, but the firts thing I worried about is the aluminium I bought. I have natural aluminium, but it’s possible that I will need polished aluminium to reflect better the possible light leaks throught the panel. Anyway this is only for the prototype, with the finished project I will use shape “U” polished aluminium profiles to wrap the LEDs completely.

I have a 3cm distance between each LED, and I will mount them in a manner that when we have an “space” between 2 LEDs, we have a LED on the other profile, covering all non illuminated parts of the frame:

0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0

   0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0 

I printed a template of that to put it on the profiles and have the exactly marks where I have to drill the holes.

The LEDs thickness is 5mm, so I will use, obviously, a metal drill bit to do the holes. I will put then the LEDs as in the next pic (casually my rules has a 5mm hole on it).

The assembly will be as in the pic underneath. I haven’t drilled yet, so the pic doesn’t show the LEDs, but you can see how it will look like:

Tomorrow I will start the drilling, I’m waiting for the cables (I will use non-soldering cables while I don’t adjust all the system, but it’s recommendable to solder the electronics at the end).

Multi touch interface (part 2): Prototype scheme

At last, I received the first components to build the FTIR interface an do the first tests.

I have 100 LED (I bought 100, but I will use only 50, but just in case). I choosed the Oshram SFH485, because them have the best results with FTIR configurations: A wavelengh of 880nm and an emission angle of 40º. With superior waveleghts it’s harder to filter the light for FTIR, and the 40º angle it’s perfect for my plexiglass.

 In the other hand, I received at last the acrylic (I was waiting too long for it). It is 8mm thickness and is 76×56cm (panoramic view). I could use thiner one, but due it dimensions, I prefered to don’t risk and ensure it will not bend with the pressure strokes. In any case, the LEDs aer 5mm thick, so an acrylic of 5-6mm will be fine.

I ordered a couple of aluminium profiles in L shape, 90cm each. I will use them as “rails” to put the acrylic in and to mount the LED arrays. I choosed the L shape for the prototype, but of course when I reach an stable prototype I will change them for a good aluminium frame. 

The first thing to do at this point was to decide the LED scheme for the project. I will use 25 LED, with 2cm between them on each side of the acrylic, to illuminate the entire panel.

The LEDs operate at 100mA, with a forward voltage of 1.5V  I will use a source of 12V, so I will do a 10×5 array, with 47ohm, 1W resistor on each row. Perhaps I need a less power resistor to get the LEDs brighter, but I will try first with that, to don’t risk burning the LEDs (they are not cheap).

Note: The resistor formule (following the Ohm’s Law) is that:

R = (Source Voltage - Forward Voltage) / LED intensity

 And then we choose the nearest standard value wich is greater.

I still have to decide how to make the silicone complaint surface, I don’t have neither projector neither diffuser yet, but I have enough to start testing and generate the first blobs.

This week I will start this, so I will post some pics and explain the obstacles I could encounter in the process.

Multi touch interface (part 1): Planning

When I watched the Jeff Han’s multi touch video a year ago, I got impressed with that technology and I was researching about it. In fact, this is a not so new technology, and is really simple.

There are tons of geeks all over the world building their own multi touch displays, and researching how to improve them, and nowadays these home made prototypes are as good as any kind of comercial device.

I’m planning to build my own prototype. As a coder, I have an special interest about this technology and how to develop applications for it.

When I decided to start this projects, first of all I had to choose the technology to use on my interface. There are two main approachs to develop a home made multi touch interface. Both of them are similar, but the construction are totally diferent:This first method is the most used, perhaps because is the method used by Jeff Han, and it did famous this technology.

With this method, we use an acrylic panel and flood it with IR light from the borders. We use acrylic and not glass, because acrylic has a good transmittance for IR, while glass will not work.