Does anyone have any ideas on how to incorporate Weight/Pressure sensitivity into a multi-touch setup? By this I mean in terms of hardware scales or something, not by computer vision… It would be quite cool if you could work out how much force is being applied to the surface in general or how much something sitting on top of it weighs, e.g. food or drinks. You could then use that info to automatically notify a waiter in a bar for a drink refill or for a multi-touch kitchen work surface, or perhaps for some novel interactive games…

Great question. Pressure is something I’m also interested in and a experimental pressure view is added to a new tracker that will be released.

The only technique that really shows true pressure is FTIR (and possibly DSI). DI and LLP don’t show real pressure since the light intensity doesn’t change with greater pressure. The intensity of light in FTIR WILL change based on the amount of pressure if you have a good setup. In FTIR if you calculate how bright pixels are in a given touch/object you can say that it has more pressure at this particular part.

What some people do is infer pressure with height and width of an object. Unfortunately, that won’t work with objects that can’t change size. Like a cup, or fiducial. So using LLP or DI with static objects that can’t change size means you can’t infer pressure. Also, a finger getting larger (bigger height/width) doesn’t automatically mean more pressure was used.

Imagine having sets of buttons in an application, but each one has to be pressed with a certain firmness to work. This would give some vactile feedback to how the buttons work. Some would feel easy to push while some would require more pressure.

This is a really interesting idea.  It would certainly increase the cost of the table and the engineering complexity, but if the surface rested on four fairly sensitive scales at each corner you could record overall pressure.  I’m just thinking out loud, but it seems like this way if you press near a corner, you’ll get one of the four sensors with the bulk of the weight, and if you press in the middle the weight will be evenly spread.  But you can record the sum of all four, and it seems like you’d have a fairly accurate record of at least the overall pressure.  This would have a limited application, though, since you can’t determine very accurate pressures for each blob.
It would be useful if you’re pressing with only one finger.  With having four different weight recordings, you may even be able to infer some interesting information...I’ll leave that to someone more creative to figure out.

Memphis - 24 July 2008 06:45 AM

This is a really interesting idea.  It would certainly increase the cost of the table and the engineering complexity, but if the surface rested on four fairly sensitive scales at each corner you could record overall pressure.  I’m just thinking out loud, but it seems like this way if you press near a corner, you’ll get one of the four sensors with the bulk of the weight, and if you press in the middle the weight will be evenly spread.  But you can record the sum of all four, and it seems like you’d have a fairly accurate record of at least the overall pressure.  This would have a limited application, though, since you can’t determine very accurate pressures for each blob.
It would be useful if you’re pressing with only one finger.  With having four different weight recordings, you may even be able to infer some interesting information...I’ll leave that to someone more creative to figure out.

That reminds me of the wii balance board