The state of audio search
Following on the heels of his post about the state of video search, Niall Kennedy outlines the current state of audio search:
I expect even more speech-to-text work in the future as the CPUs, memory, and disk space available continues to become computationally and monetarily cheaper. Perhaps we might even see client-side analysis of content similar to analysis work being conducted on images.
While automated transcription still isn’t something that we’re going to see anytime soon – it’s that sort of technology that will make audio search fulfill the dream of what we all want it to be…
My wife and I run a website focusing on live theatre in the Washington DC region, http://www.showbizradio.net. It started out as audio for our Live365 Internet radio station, echoed on our web site as a podcast. After Live365 complained our reviews were unauthorized advertising on our station, we are now only a podcast, with the radio station only playing Broadway showtunes. After about 2 months we decided to post a transcript of every podcast we post so that our site would be better indexed by the search engines. I believe that the transcripts are paying off, our traffic from search engines is up considerably, and our page rank has increased two points to six.
We use Transcriva from Bartas Technologies. It takes a fair amount of time to publish the transcript, but it is worth it.