Mailinglist Archive: opensuse (2532 mails)

< Previous Next >
Re: [opensuse] Speech-to-Text
  • From: "David C. Rankin" <drankinatty@xxxxxxxxxxxxxxxxxx>
  • Date: Sat, 03 May 2008 15:23:00 -0500
  • Message-id: <481CC9A4.9060805@xxxxxxxxxxxxxxxxxx>
Donald D Henson wrote:
Greg Freemyer wrote:
On Sat, May 3, 2008 at 3:45 AM, Rikard Johnels <rikard.j@xxxxxxxxxx> wrote:
On Wednesday 30 April 2008 23:37, Donald D Henson wrote:
> Does anyone know of an Open Source application to accept continuous
> speech and convert it to text? I've found a couple of proprietary apps
> but you have to use Voice mail as an input. Any suggestions appreciated.
>
> Don Henson

A few years back there was something called ViaVioce from IBM (i think)
I never got that to work, and as far as i know its been discontinued for
ages...
I havent seen any speech-to-text utilities for linux, nor for windows too for
that matter.
The last time i saw it was on my brothers Amiga way back in the early days.
He actually had it responding to voicecommands.
And there was also a simple voicerecognition software bundled with a
SoundBlaster sound card at one point. But that vas a very long time ago.

I too would love to have speech-to-text, as i am LOUSY at speed typing, and as
i sometimes tell short stories to my friends, it would be heaven to be able
to get them into printed matter...


I think in Windows "Dragon NaturallySpeaking" is the leading solution.
I don't use it so no firsthand feedback here.

http://nuance.com/naturallyspeaking/

A little history of products (including some OSS ones) and whats
available is at:
http://en.wikipedia.org/wiki/List_of_speech_recognition_software

Greg

Thanks for the links. I'd seen most of the OSS stuff. All that I've seen so far are nowhere near production use. Dragon Naturally Speaking looks promising, though. Their web site convinced me to plunk down a hundred bucks for a copy of their cheapest product. After I've had a chance to use it for a few days, I'll post a review here. It has to be physically shipped so it'll be a couple of weeks.

Don Henson


It will take more than a few days before you can make an objective comparison. With all speech recognition software you have to "train" it which is a fairly time consuming process. Though from other I know who have actually done it, after the training is done, you can use it very reliably.

The "training" will create a "voice" file based upon your voice inflection, pause, annunciation, etc. The process involves beginning with a default male/female voice file and going though a correction process where the difference in your specific manner of speaking is captured in your voice file and that is used to augment the recognition of words, etc..

Since your voice file will take some time to create, tweak, get just right, etc., make sure your know what that file is and BACK IT UP regularly. Otherwise, after several weeks of use, if something bad should happen, then you are right back to step one.

I hope that helps.

--
David C. Rankin, J.D., P.E.
Rankin Law Firm, PLLC
510 Ochiltree Street
Nacogdoches, Texas 75961
Telephone: (936) 715-9333
Facsimile: (936) 715-9339
www.rankinlawfirm.com
--
To unsubscribe, e-mail: opensuse+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: opensuse+help@xxxxxxxxxxxx

< Previous Next >
Follow Ups