4

I don't want my laptop or mobile to just dictate/type down what I say, but instead write what the caller on the other end of the line says. Why doesn't Google or Nuance enable that functionality? Is it even possible? What would be needed for that to work?

2

2 Answers 2

1

When you are speaking your voice is being recorded through your computer's microphone and audio hardware. This audio signal is available to any applications on your computer. When you are speaking to someone over Skype the voice signal is encoded in a bunch of packets. You would need to reverse engineer Skype's protocol in order to decode the voice signal.

If you're talking on a cell phone in speakerphone mode the quality might not be good enough for voice recognition. Also the voice recognition software would need to be trained on both your voice and the other person's.

I'm not sure what relevance Google has here.

7
  • Well, there's some wrong information in that answer. The software could always just grab the audio signal from the line out for example.
    – Falcon
    Jul 29, 2011 at 13:17
  • true in theory it could pick up on the line out, But emddudley is true Dragon requires training on your voice before it will correctly recognise the words. Although this is just for Dragon the intelligence agency's have been doing this for years but Im guessing the reluctant at releasing their algorithms for the greater good!
    – squareborg
    Jul 29, 2011 at 13:39
  • Not that it belongs here but there is also some legal issues with recording phone calls.
    – Chad
    Jul 29, 2011 at 13:46
  • 2
    Google's relevance is in the Google Voice tool. Technically, Google Voice does transcribe incoming calls, but only when you send them to voicemail. Then it can text the voicemail message to you. Jul 29, 2011 at 14:26
  • I'm with Chad on this one. There are some serious serious legal consequences here. I'm not a lawyer, but I'm pretty sure you need consent to transcript a phone call. . .
    – surfasb
    Jul 30, 2011 at 3:05
0

Dragon NaturallySpeaking (unlike Google) needs to be trained before being able to transcribe speech into words.

Google does speech to text transcriptions for the caller on the other end of the line when someone leaves a message on your autoresponder on Google voice.

Note that the voice quality is often altered quite heavily during a call, which makes voice recognition even harder.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .