A lot of commercial multitouch solutions is arising around the world from different companies. There’s a lot of different approaches, and I thought it could be cool to post our views and test-results, if others should try out commercial solutions. Right now I know of these companies offering some sort of multitouch technology - a few of which we’ve tested/are testing.
Nexio - http://www.inexio.co.kr (see below)
NextWindow - http://www.nextwindow.com (Not commercially available - have been shown in a demo with FingerTapps software on it)
ELO - http://www.elotouch.com (Microsoft have used this screen to demo Windows 7 multitouch features AND Surface on, rather than Surface itself)
Dahan T&S - http://www.dahantns.com/ (looks like a triangular optic technology)
MS Surface - http://www.microsoft.com/surface/index.html
Stantum - http://www.stantum.com (f.k.a. JazzMutant)
n-trig - http://www.n-trig.com
Perceptive Pixel - http://www.perceptivepixel.com (Jeff Hahn)
LGPhillips - http://www.lgphilips-lcd.com/homeContain/jsp/eng/inv/inv101_j_e.jsp?BOARD_IDX=1395&languageSec=E (article - does not seem available at this point, though)
We’ve tested out Nexio’s multipoint technology. It comes in the form of a simple overlay to a standard LCD-screen, is quite cheap too. It’s an optic IR-technology, with individual sensors for each axis. It creates an IR grid across the surface. This pose a problem right there, as fingers can make shadows when perfectly aligned on an axis, hence creating shadowpoints (x in the power of x, with x being the number of touches). They state in their API and documentation, that these ghosttouches can be removed using a certain status. This, however, does not work at all. We made our own algorithm, that work better than Nexio’s, but still no where near, what you would expect. It seems like Nexio is selling a singletouch-technology as if it were multitouch, and just made some (bad) adjustments to their existing C++ API. They insist this should be multitouch, but we can not see this technology work in a stable fashion - even in duotouch… NOT recommended! Don’t get to hyped about their demo-videos with google-earth. We have seen their source-code, and they “cheated” a bit here - they have static objects that just scans for any input, why it is not necessarily your fingers it uses, but could be the ghosttouches they generate. With static objects, you really can’t tell the difference, as precision is not a factor on an object that fit a quarter of the screen. On a sidenote, I can mention that the technology seems identical to NextWindow’s....
NextWindow have been very honest in saying that they are not there yet with multitouch, and deal with a lot of ghosttouches. They’re merely focussing on duotouch at the moment. I think this type of technology, where they split x and y in a grid really needs a third layer in the grid in a 45 degree angle, or something, to deal with ghosttouches properly.