David> As a fall-back, I am considering going the Arduino/ZigBee David> route. You can buy ZigBee doorbells, etc. However, this David> requires a fair amount of installation/wiring that needs to be David> done to the house. That was where I was going with my thoughts too. Not that I'm trying to kill your idea, I'm just trying to think how it would work in a reliable manner, since false positives are going to quickly get people ot NOT use your app. David> I lost my hearing during my middle age and I was given a box David> full of signalers by the Commonwealth of Mass. Hooking up these David> signalers is not the easiest of tasks and I was hoping to David> bypass this via having a wearable computer do the David> recognition. (BTW - I have a Cochlear Implant that obviated the David> further need for the signalers). Ouch, not fun! My eyes are terrible, but luckily correctable back to 20/20, though I'm getting old and going to need bifocals sooner or later. Ugh. Maybe instead of having an app which does this, would it be easier to have a small dedicated piece of hardware? When you got the signalers, did you have to wire them all up and run the wires everywhere? Moving to zigbee might be he answer to that side of the problem, though getting the sensors to work reliably is the hardest part I suspect. Good luck! David> On 12/12/2012 03:21 PM, John Stoffel wrote:
Doing DSP in your phone or other android device might become really difficult, really fast. In any case, I'd recommend that you get one of those development boards which runs Android and work with that to start. No need to sacrifice your phone ahead of time.
Also, instead of listening for sounds, maybe you can wire up a dedicated device into the home which would talk directly to the haptic device or the glasses?
My brother went to RIT (Rochester Institute of Technology) and they have a large deaf population there, to the extent that some dorms are fitted out with the bright stobe lights used to wake/alert deaf people of fire alarms, etc.
Listening for arbitrary sounds and trying to match them to a library of triggers sounds (sorry for the pun) like a very hard problem. Even apple is doing Siri by sending your speech to a remote datacenter for processing, since the phones don't have enough juice. Admittedly, that's a more difficult problem space, but it is similar and less open ended than your problem space.
Just hooking an Arduino into my doorbell/phone/other sensors and having the arduino talk to the alerting system might be a simpler way to do it.
John _______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
David> ---------------------------------------------------------------------- David> _______________________________________________ David> Wlug mailing list David> Wlug@mail.wlug.org David> http://mail.wlug.org/mailman/listinfo/wlug