Yes. I want to play back the user’s own voice or any voice that around the user, just like a hearing aid.
OK, I think you could use SaveThatWave for this. It saves a WAV file of recognized speech that you can then play back using an AVAudioPlayer.
Another question I have is normally how close to the device for an voice to get accurate recognition? If the user talks to another person in one meter distance, can the engine get pretty accurate recognition for the other person?
For open speech into the built-in mic, I usually test from about a meter away, so I would expect that to work pretty well. A bigger issue is going to be cross-talk between two different speakers.