I just came across Podzinger, a podcast search engine developed by BBN Technologies. BBN Technologies is the same company that developed the first Internet router, sent the first e-mail, pioneered voice over IP, and the TCP, to name a few. They have been working in acoustics and voice recognition for decades, funded by the government’s need to listen to broadcasts from all over the world and transcribe the results.
Podzinger is powered BBN’s AVOKE STX speech recognition technology which adds speaker recognition and confidence level detection to its speech-to-text technologies. The site’s database currently contains a little over 10,000 podcasts but should continue to grow as the site leaves beta. Each search result includes links to the RSS URL, iTunes subscription, and Yahoo! podcast directory entry. You can restrict your search to just one podcast, or view all the information Podzinger knows about the podcast. They even show you the Creative Commons license for the content right in the search results.
It seems like the site is definitely indexing show notes since the excerpts don’t appear to be conversational in nature. Looks like a pretty good start, and BBN definitely has the technical chops to make this thing work.