Earlier this week my colleague Stuart McKee sent through an email highlighting a Microsoft Research project called MAVIS (Microsoft Audio Video Indexing System) which is designed to “index the spoken content of recorded conversations, whether they are from meetings, conference calls, voice mails, presentations, online lectures, or even Internet video”.
Stuart has been working with the Washington State Digital Archives people to build a system to archive tens of thousands of hours of audio records and now, via MAVIS, provide the public with the ability to search those audio archives quickly and accurately. The really cool thing is that this isn’t a vapourware announcement – the search capability is up and running right now on the WSDA website. Below is a of snippet from a Washington State press release on the project:
“The Washington Digital Archives is the first government program in the country to offer Microsoft Research’s Audio Search technology, which takes record keeping to the next level: it doesn’t just preserve audio recordings – it gives people an innovative way to search through them.
The partnership grew from an ambitious project undertaken three years ago by the Washington State House of Representatives and the Washington State Archives to save over 30,000 cassette tapes of committee hearings.”

Try it out for yourself. All you have to do is click on the “Search” link, select “audio recordings” from the Record Series dropdown list and enter the word(s) you’re looking for in the keywords field. I searched on “New Zealand” and found multiple references each of which was returned as a link which, when clicked, started playing the audio stream from the point at which the search term was found.
This is a very cool and extraordinarily useful piece of technology and it’s a great example of Microsoft Research making good on their goal of “turning ideas into reality”.