Automatic Voice Recognition Caption Blunders

YouTube now recognizes the need for online videos to be closed captioned and for that we say, “Hooray!” However, it is becoming increasingly clear that while their new automatic voice recognition software allows their videos to be quickly captioned, the technology still has its bugs and its accuracy leaves a lot to be desired. In fact, the actual errors are so great that it has now spawned tons of new websites solely dedicated to posting screenshots of the ridiculous captions that have been found on certain YouTube videos. These are not just random user videos, either. Some are found on professional software training videos, including a rather embarrassing captioning error within a video clip by Macintosh for their Snow Leopard operating system.

Per our last post, we discussed the benefits of using voice recognition software. However, we understand that part of the process is to go back and correct the errors. Voice writing “on the fly” by a trained professional still only produces captions that are 95% accurate at very best, but usually under 90%. (And this is using software that has been trained to understand an individual captioner’s voice.) Closed captioning remains a field where a trained human professional’s product will surpass any computer program.

So while this new embrace of closed captioning is still a step in the right direction, we encourage companies to remember that accuracy is important especially when it comes to your message. (Also remember that your SEO will be enhanced if the captioning text of your online video is completely accurate.) Why risk having your professional video ridiculed for closed captioning blunders?

To get an idea of just how inaccurate voice recognition captions can be, check out: