Re: [NTLK] Speech Recognition

From: James Fraser <wheresthatistanbul-newtontalk_at_yahoo.com>
Date: Wed May 20 2009 - 07:56:48 EDT

Hello,

--- On Tue, 5/19/09, Ryan Vetter <physicalconstants@yahoo.ca> wrote:

> The huge downside though is that a quiet environment is a
> must which many times makes it impractical... microphones
> need to advance a lot more for speech to be more
> practical...

Is there a words-per-minute limitation as well? I'm wondering how fast today's speech recognition compares to keying in text (I'm guessing that, under optimum conditions, it's roughly one-third to one-half as fast).

I know that a few years ago, using voice recognition meant that you had to pause.....between......words. Which made voice recognition still somewhat slow as compared to conversation-speed speech (let alone us Torchy Blane types). Is this still the case? Or has voice recognition software progressed to the point where someone using it now sounds as though they're conversing normally (albeit one-sidedly)?

The article you linked to talks about improvements in speech recognition, but I'm afraid the author discusses those improvements almost entirely in qualitative terms.

Best,

James Fraser

====================================================================
The NewtonTalk Mailing List - http://www.newtontalk.net/
The Official Newton FAQ - http://www.splorp.com/newton/faq/
The Newton Glossary - http://www.splorp.com/newton/glossary/
WikiWikiNewt - http://tools.unna.org/wikiwikinewt/
====================================================================
Received on Wed May 20 07:56:57 2009

This archive was generated by hypermail 2.1.8 : Wed May 20 2009 - 10:30:00 EDT