|
PodZinger Search Engine Finds Spoken Keywords in Podcasts
TMCnet IP Communications Columnist
BBN Technologies has rolled out its new PodZinger podcast search engine, incorporating speech recognition research and development. PodZinger enables users to quickly locate specific information within podcasts by displaying text surrounding a search term. Users may skim the results just as they scan text search results, to quickly assess relevance.
Users may then listen to relevant sections of the search results by clicking on a specific word within the result for audio playback of that section. The PodZinger client scours the Web on a daily basis, adding several thousand podcasts per week as they become available.
PodZinger uses speech-to-text technology to create a text index of audio, enabling users to locate content from anywhere within a podcast and jump directly to the point where that keyword is spoken. This enables users to search audio by keyword, in the same way they currently find words using text search engines. The PodZinger site allows users to subscribe to podcasts and download or listen to them. The client will also automatically deliver podcasts on users’ topics of interest using standard RSS feeds.
Many podcasters have already submitted their podcasts to PodZinger for indexing, and have also placed the PodZinger search box on their sites for easy searching by their users. As more content becomes available via podcasts, particularly through use of RSS feeds, there is an increasing need for effective audio search technology.
"Relevant and content-rich podcast search is now a reality," said Alex Laats, president, Delta Division, BBN Technologies." At PodZinger, we've used three decades of speech recognition research and development to enable users to search multimedia content as easily as they search text. PodZinger delivers both visual and audio cues for users to gauge relevance in seconds, eliminating the need to listen to an entire segment to find the right nugget of information."
PodZinger is the culmination of 30 years of speech recognition research from BBN Technologies, designed to transform multimedia content into searchable words. BBN is known for the development of ARPANET and continues work in Internet and networking technologies through development of advanced protocols and quantum communications.
-----
Laura Stotler writes about IP Communications and related topics for TMCnet. She has covered VoIP and related technologies for seven years, contributing to Internet Telephony magazine and TMCnet, and as a freelance writer. To see more articles, please visit: Laura Stotler’s columnist page.
[ Back To TMCnet.com's Homepage ]
|
INDUSTRIES
Activate Email 
Find Solutions for Enterprises, SMBs & Service Providers at the INTERNET TELEPHONY Conference and EXPO East, January 20-22, 2010. Miami, FL.
