Hi all
 
Just wondering if there is any speech to text software that will actually extract text from the audio of an MP3?
So I could just load up an MP3, press play and it can turn the speech into text.
 
I've offered to help transcribe an interview and didn't realise how much it would take.
Would be great if such software existed and if it doesn't then I'm sure it must be possible?
 
Cheers
 
Steve

Steve I am not familiar with any but I googled/bing speach to text software and found a number of them.  Maybe one or more of these will work for you.
 
Good luck.

Thanks Mario
 
I did google some and although they said speech to text when I actually read more into them they couldn't load MP3's and since I'm only doing this as a one off I'll probably just knuckle down and get on with it as the stuff I think I need is expensive.
 
Cheers
 
Steve

I think the Dragon Naturally Speaking stuff will convert a WAV file to speech--if you first convert the WAV to MP3.  Quality might be questionable, though...

When I did electronic discovery work there were definitely products that did this as companies would extract voice mail memos and producer text that could be analyzed. Also I would think that medical transcription products would need to have a decent accuracy rate.

I will see what I can dig up.

this looks promising
 
http://audiototxt.com/
 
 

Thanks for your help guys, the link you posted looks interesting if a little time consuming gmon72.
I don't have my mic set up and wouldn't I get feedback if I'm pointing at at my speakers?
 
I decided I'd have to do this bit by as it's an interview between two people and there's lots of laughing and interjections which I don't think any software is gonna get as I had to listen to lots of sections 4 or 5 times to be clear as to who's saying what.
 
I'll get there in the end
 
Steve

Good luck. I thought you installed software to create a virtual link between the speakers and the mic so it was more or less seamless. Sometimes though it is easier to bite the bullet.

Windows has speech to text built in, but it trains on your voice. It works fairly well. Dragon dictation is a big one too.

The difficulty I see with this is that speech recognition software needs to be "trained" by the user's voice. It will print on screen various words that the user repeats into the microphone for analysis.  Universal speech recognition would be a really big deal. 
    The person who recorded the audio needs to train the software.  If so it may work pretty well- good enough to be completed with a bit of proof reading.