And they say you can’t improve on perfection. Just kidding. Yes, speech recognition levels can always improve (I am looking forward to speech technology translating the Indian telemarketers that call me from English to English). We all know that speech technologies that are extremely accurate are more expensive than others that have lower levels of accuracy. You get what you pay for I guess.
Perhaps the cost of recognition will go down. I was fascinated to see this excellent TMCnet article on speech recognition and how MIT has used their labs to work with 14,000 authors to improve the accuracy of speech recognition and they’re giving it away for free. Take a look at how MIT is improving the quality of speech recognition. Hope to see you at Speech-World next week.
Peter Maddern
January 14, 2007 at 7:59 amI haven’t tried out the windows Vista speech recognition yet. I am more familiar with Dragon NaturallySpeaking.
I think that speech recognition programs have some frustrations associated with them but once you’ve got used to using them, I can there’s no going back.
The world at the moment seems to be in a constant downsizing mode which seems to have a big impact on secretarial resource. Everyone seems to have a PC in the office and at home and in the workplace, because of reduced secretarial resource, everyone is expected to type up their own work.
So to my mind, because of the pressures and demands on everyone, Speech Recognition is a technology which is bound to take off.
I have some demos on my site if anyone is interested in taking a look. They give an idea of the results with the latest release of dragon NaturallySpeaking v9. They’re on
http://www.speechempoweredcomputing.co.uk
See “Online demonstrations” on the above page.
Peter Maddern, Wales, UK
Related Articles
News From Day 1 of Ribbon Perspectives19 #rbbnp19
ZaiLab Powers the Contact Center With Happier Agents
Voci Technologies Touts Advanced, Cost-Effective Speech-to-Text