audio to text

[tandi]

is there any software or tool that can convert from audio to text ?

[unskinnyboy]

By audio, I hope you mean speech or spoken voices. Music would be nearly impossible to recognize and translate accurately. What you are looking for, is a Speech Recognition software (http://en.wikipedia.org/wiki/List_of_speech_recognition_software).

[tandi]

yeah i mean like speech, voices , not song
thanks unskinnyboy for your help

[fibbingbear]

If you don't mind blowing some money, you could try Dragon NaturallySpeaking, but I am not positive that has a function to give it a wave file and output text for you. It is supposed to be one of the best, but it often makes translation mistakes, even after you train it. I believe Carnegie Mellon has open-source software called Sphinx4 that takes in wave files and outputs text. I tried to use that, but it is a nightmare to set up, so I gave up on that.

Just realize that if you want the translation to be accurate, you will have to go through and proofread the text, pretty much of any recognition software.

[CWR03]

Dragon NaturallySpeaking won't work with recorded audio - it requires that you "train" the program by reading a very large amount of text and correcting the program as needed. It won't accept just any random input and work properly.

[Ajax_Undone]

I own a copy of NaturallySpeaking It does suck I just transcribe from audio to text manually

[saint-francis]

I use DNS (Dragon Naturally Speaking) all the time. I am dictating this right now! DNS does allow you to feed it an audio file but if you haven't trained it for that particular voice the accuracy will not be as good as it could be. That said I have given DNS several audio files to transcribe that weren't in my voice and it worked out alright. Though I certainly did edit the text it gave me as it had some errors. All in all DNS is probably about as good as you can get with the current technology. Just like with OCR (optical character recognition, which I am also a great fan of) the technology just isn't as good as it could be or will be in the future.
I must say that nothing beats sitting back and dictating posts on a forum or writing a paper with speech recognition.

[DigitAl56K]

Dear aunt, let's set so double the killer delete select all (http://www.youtube.com/watch?v=IkeC7HpsHxo)