Audio To Text Software

so if you do a quick google of audio to text, youll find a bunch of confused people looking for audio to text software, some techy types who kind of know what theyre talking about, and i even found what seems to be a forums blocking the information. it seems to be that most people think that it is required to "train" the software first, though i hardly beleive that this is true for all programs

i just need a wicked program, that i can toss in an mp3 or wav file, and have it output text

simple?

no, very very complicated.

It's complicated because of the vast variety of spoken accents, even within a single language (especially one as widely spoken in so many varieties as English.)

Processing speech within the human brain is one of the most intensive functions it performs. Separating speech from background noise, seperating one word from the next - and you have to learn to recognise differing accents and tones in totally different octaves as representing the same sound? What about homonyms whose meaning and spelling comes from context? Even a computer as powerful as the human brain takes about 2 to 3 years to learn this task!

Heck, several hundred people on this forum can't seem to decide if one word is "heat" or "night" - you expect a bit of freeware software to do a similiar task?

heat

You're looking for voice recognition yeah ? It's doable but I think it's still early days before it's actually of any use (I remember using an early version of Dragon Dictate about 10 years ago, god it was shite - it's probably changed a lot since then though http://en.wikipedia.org/wiki/Dragon_NaturallySpeaking):

  On 2/16/2010 at 2:58 PM, ET said:

it seems to be that most people think that it is required to "train" the software first, though i hardly beleive that this is true for all programs

Yeah pretty much, speech is a complex beast.

it's pretty long and boring to teach soft to ocr, I don't even want to imagine how it is with vr

invent an ingenious way of making a record of what you are saying by not actually having to speak out loud. perhaps, using your hands somehow, you could link directly to your brain and enter information straight into the computer.

Upload the audio file attached to a blank video file onto youtube, turn on Transcribe Audio, ?????, Profit!

http://www.youtube.com/watch?v=5m9pn6x-9Yo

Click the "CC" button at the bottom right-hand corner of the flash player.

Oh my God, this new feature is SO GREAT.

  On 3/4/2010 at 3:28 AM, Yegg said:

Upload the audio file attached to a blank video file onto youtube, turn on Transcribe Audio, ?????, Profit!

 

http://www.youtube.com/watch?v=5m9pn6x-9Yo

 

Click the "CC" button at the bottom right-hand corner of the flash player.

 

Oh my God, this new feature is SO GREAT.

oh lol

jesus

the google voice search for smartphones is really good i think. I even threw in a bit of Japanese and it got it, so considering youtube is google, that's pretty fucking weird.