Audio file to text

Status
Not open for further replies.

JoannaV

Puritan Board Sophomore
Does anyone have any good recommendations for software which will transcribe an audio file to text? We would like to be able to make transcripts of the sermons from our church in addition to having the audio available. Ideally I would like to be able to use mp3s. My father-in-law found this: CoolSoft LLC Launches SpeakToText™ 2.0 but the only thing I can find mention of is WAV to text. Ideally I would prefer to avoid having to convert my files to another format. :coffee:
 

VictorBravo

Administrator
Staff member
Joanna, you are asking for something that, to my knowledge, isn't there yet. The most mature software that converts audio to speech is Dragon NaturallySpeaking. It works pretty well as long as you have well controlled input: either a high quality mic in real time, or a high quality recorder and quiet background. Regular mp3s produce a lot of transcription errors (as in almost hopeless gibberish with the occasional good guess).

I don't know anything about SpeaktoText, but if it will actually do what you are asking it to do, converting to WAV is a minor hurdle. Audacity does it for free.
 

VictorBravo

Administrator
Staff member
Follow up. I looked at the website and note that it requires training and configuration. My guess is that it will not work very well with a regular converted mp3--mostly because of recording quality.
 

Skyler

Puritan Board Graduate
On the other hand, transcribing sermons by hand is great practice for boosting your typing speed. ;)

---------- Post added at 10:04 PM ---------- Previous post was at 10:03 PM ----------

"...state-of-the-art voice recognition technology from Microsoft Corporation" is not state-of-the-art voice recognition technology. Sorry. =)
 

Tim

Puritan Board Graduate
YouTube is trying to develop technology that will provide automatic generation of "closed captions" (i.e., text at the bottom of the screen). I tried it once, and there were many (sometimes humorous) mistakes. It seems that a major hurdle with this sort of thing is that you can't "make" the speaker on an audio file slow down and pronounce words more carefully, so that the program can discern them.
 
Status
Not open for further replies.
Top