We're fleshing out the technical feasibility of a project that would involve remote sites recording voice responses from various stimuli and sending that captured stream to our server.
At our sites we can guarantee:
1) consistent hardware ("gamer-style" plug and play usb headset)
2) java vm (current)
3) flash >= 9
4) stations running >= xp pro
5) Internet Explorer >= 7.0
We would need full duplex recording... in other words, a stimuli played with a timer going while recording only the voice response (think continuous translation services). The recorded stream can be anywhere between 5 -60 seconds. The playback stream would be the same length. Quality is important as a human will need to understand what was recorded at a later time.
What are my options? Could an actixe-x control be used? The main problem here I'd imagine would be that the stream would have to saved locally, then uploaded after the fact, which could be a problem.
VOIP server might be the way to go?? Any thoughts/experiences would be greatly appreciated.