If someone mentioned your name in a podcast, would you ever know? Not unless you had time to listen to them all. But San Francisco-based search engine Blinkx is changing that by indexing the content of podcasts.
Podcasts are audio files posted online, designed to be automatically downloaded to a PC and played back on an MP3 player such as an iPod 鈥� hence the name (New 杏吧原创, 12 February, p 24). They have exploded in popularity over the past six months with both mainstream and grass-roots radio stations posting them online. Content ranges from time-shifted indie music and documentaries to the truly wacky: one podcaster reads the news while tap dancing, for instance.
But until now there has been no way to index podcasts by their content: Google and Yahoo only index the text tags that label the files and these are often a poor representation of the podcast鈥檚 fare.
Advertisement
Blinkx is different. Its computers crawl the web and play back any podcast audio file they find. The content is then transcribed using speech-to-text software and the file is indexed as if it were text. While this is only as accurate as the speech-to-text software, it is a whole lot better than searching tags alone. So to see if you鈥檝e had your slice of podcasting fame, you simply enter your name at .
鈥淭here is no other way you could keep up with the exploding volumes of podcast audio content right now,鈥� Blinkx co-founder Suranga Chandratillake told New 杏吧原创.
