AW: multi-touch?

Federico Lorenzi florenzi at gmail.com
Fri May 30 19:30:31 CEST 2008


On Fri, May 30, 2008 at 4:06 PM, Joerg Reisenweber <joerg at openmoko.org> wrote:
> Am Fr  30. Mai 2008 schrieb Fabian Off:
>> Hey!
>>
>> How does synaptics handle this? When I look at the output my touchpad does,
> I can see "X Y Fingers" Values... Maybe we could look into this code and see
> how they do detect the amount of fingers? Dunno whether they work nearly the
> same, but I believe this could maybe point us into the right direction :)
>>
>> Only my idea... 2-finger-scrolling is so great!
>
> To the best of my knowledge, at least MY synaptics in front of me right now is
> a capacitive type, and even this one doesn't support multitouch right now
> (though maybe a driver issue).
> Please google or wp for capacitive vs resistive ts! Our 4-wire resistive type
> ts is a device as dumb as bread, NO silicon inside. You simply can't do muto
> with such device in a reasonable straight way. How do you get info of
> X1,Y1,X2,Y2 out of a device with 4 ANALOG connectors (GND incl!)??? It may be
> feasible, but it's rather tricky and needs quite some special hw AROUND the
> silicon-free ts.
> There's a way to detect the "surrounding square" of _all_ touchpoints on a
> 4w-R-ts, at least with GTA02 i think. You may use this to detect there's more
> than one touchpoint. Still you have no correct data for the coords of the 2
> (or 3?) points.
That sounds interesting, if you can tell for a fact that two points
are being pressed down, then you could use some fancy maths and
determine their locations, it would probably solve the one of the
multi touch emulation problems AFAIK, which is not knowing if two
points are being touched on either side, or one is being touched in
the centre.

I probably make no sense :)

Federico




More information about the community mailing list