YouTube has come out with its voice-to-text technology to power up its search functionality while providing rich metadata for the selected videos. The feature makes it possible that the search of any keyword results in it being found within the entire audio/video file instead.
The new feature uses unique automated spiders that whizzes across he audio file of every video to transform the spoken words to text. The result is then applied to the video file, where the text is placed within it making it very searchable.
It has been reported at Beet.tv that the technology ha previously been applied to the videos of McCain and Obama for their election campaigns and enabled users to search for a few topics like ‘gas prices.
YouTube has been up with quite a lot features lately, it introduced annotations in June and the present feature is more likely to gain widespread popularity. The concept isn’t new as companies like EveryZing, Delve, etc have been providing voice-to-text transcriptions in the form of metadata. Plus there is news of Adobe readying itself up to provide the same technology for its Flash videos.
Speech-to-text can turn out to be a very effective tool for the search across online e videos as the user doesn’t need to be confined to getting results locked by the headlines. The search would take you directly to the point in video where the keyword has been spoken. To me it points at finding every possible result for my search as headlines often miss out the word that forms the crux of quite a lot of news.