20 April 2010 ~ Comments

Embarrassing YouTube Captions

Technology is a beautiful thing. It has brought us space travel, the artificial heart and instant access to information. Still, it has its drawbacks. Hard as we might try, technology just isn’t the same as a human touch.

Let’s say you work for years and your company invest lots of money to take a spark of an idea and make it reality. You create a video to tell the world about your wonderful product, and you hope the video will go viral and that everyone in the world will want to buy your product, right now.

You post it, and once your baby has left your hands and is out in the wide world of the internet, voice recognition technology puts its two cents in. It takes your dream, labels it with a bunch of inaccurate captions, and the only attention you’re getting is side-splitting laughter from the millions of people who receive your video as a funny forward.

Last year, YouTube brought voice recognition online, and used the technology to automatically caption millions of videos on the hugely popular internet channel. That’s great for viewers, especially those with hearing loss, right? Yeah, not in all cases.

We applaud the fact that YouTube when out on a limb to provide viewers with new voice-recognition captions, however, it is impossible for software to be accurate all the time. As a result, examples of seriously incorrect captions have flooded the internet.

Viewers watching an episode of the PBS series “Sid the Science Kid” heard the alphabet song. But those watching the subtitles saw “a B C D E F G a giant day an enemy and he laughed and he made me now the US black and scene.” Hello, how am I supposed to learn to spell my name using that alphabet?

An interview segment from a trade show caught a spokesman talking about Windows 7. The auto-captioning had him saying, “The good news is that they want a more reliable of racism they want.” OK, am I supposed to believe that my computer has been making fun of my ancestry behind my back? I hope not.

Macintosh is known for technical innovation, but the video introducing the company’s new operating system, Snow Leopard, read “you know this is is more monkeys and transplant patients have mental Clinton’s aides to Obama.” Of course, I can’t be positive, but I’m guessing that’s not the message Mac wanted to send to its customers.

What about you? When your company places a sales, promotional, or training video online, what message do you want to send? And how can you be sure that when potential customers search for what you’re selling, they’ll be directed to your doorstep?

Accurate captions keep your work (and your product) from turning into a joke, but they also make a big difference when it comes to the effectiveness of internet searches. A client searching for the hot new Mac OS probably won’t be Googling Clinton or Obama, or with any luck, mental patients.

Accurate text in the video means your search engine optimization factor, or SEO, shoots up like a rocket. Better information for the search engines means a higher placement and better ranking, which will lead your audience right to you. With accurate text in the video, people searching will be more likely to land on your video.

A professional captioning company can create accurate and useful captions that will bring respect instead of giggles, and help boost your worth when it comes to internet searches.

YouTube is a cultural phenomenon that isn’t going away any time soon, and there’s no reason every business person shouldn’t consider using it – to create a buzz, to create an image, to provide demo videos and tech support. But leaving the captions up to a technology that’s still in its infancy might not be the best choice. When it comes to your livelihood, best to do what your clients do – go to the experts and let them do their thing.

  • Thom Lohman

    Great post, as always!

    What are your thoughts about the "auto-sync" feature, which relies on a human-generated text transcript (which is likely to be miles ahead of a machine transcription in terms of accuracy)? Since the launch, I've felt that auto-sync was the feature that should be out front (in terms of marketing YouTube user friendliness vis-à-vis accessibility) because it gets users involved in the process. Sure, the auto-sync will "bork" on speaker IDs, sound effects, and the like (which is why professionally-generated captions are always preferable), but those can always be added after the syncing is completed. Seems to me that synchronization is one of the most time-consuming parts of the captioning process, and--if some improvements are made to the speech recognition engine--auto-sync can save the casual YouTube channel owner a great deal of time if he/she takes the leap into making his/her content accessible to viewers who (for whatever reason) can't or don't want to access audio content.

blog comments powered by Disqus