Pronunciation training for second language learning

ISLE (Interative Spoken Language Education) has a demo of software that uses speech recognition to help non-native speakers see how good their pronunciation is.  I haven't tried it (it looks to be targeted toward Italian and German learners of English) but it's an idea I've wanted to see productized for a long time.

Once your SR system takes in a spoken audio file and analyzes it for statistical properties related to the way the engine was trained, you can do lots of cool stuff.  Was the utterance spoken by a male? female? young person?  angry person?  How angry?  How female?

I think this would be a wonderful way to learn a foreign language.  Or maybe a midwestern English speaker like me could work on getting one of those pompouseloquent British accents that seem to be so appealing to the ladies. :-)

Published 27 February 07 11:47 by sprague

Comments

# Chuck Watson said on March 14, 2007 5:16 PM:

We have been working on computer-based pronunciation training for about 20 years.  Our software called HearSay teaches ESL students to avoid pronunciation errors at the single-word level.  Of course there is a lot more to correct pronunciation than saying single words correctly, but being able to do that does help.  Some published work on HearSay is listed on our site (comdistec.com).

# Richard Sprague WebLog said on April 13, 2007 1:22 AM:

I like the idea of using SR technology for purposes beyond simple recognition or synthesis, and I've

New Comments to this post are disabled

Search

This Blog

Syndication

Page view tracker