Wednesday, December 19, 2007

Speech-to-Text Searching

A colleague just pointed me to the new search engine EVERYZING which searches digital media audio and video (YouTube, podcasts, etc.) for your search terms using a commercially available speech-to-text engine. I’m not in love with the results, but it’s a great application for a classic computational linguistics technology.

How does EVERYZING work?
EveryZing creates a text index of the audio data from audio and video files, using the industry's leading speech-to-text technology from BBN Technologies, to enable search within the spoken words of media, not just within the metadata.

In the interests of full disclosure, though I do not work for BBN, my company does have some customers in common with them and we have utilized BBN products in service of a couple contracts (but I have not personally had any contact with BBN personnel or products).

No comments:

TV Linguistics - Pronouncify.com and the fictional Princeton Linguistics department

 [reposted from 11/20/10] I spent Thursday night on a plane so I missed 30 Rock and the most linguistics oriented sit-com episode since ...