There are some misconceptions out there about how accurate voice recognition technology is, and also about how captions get formatted.
MISCONCEPTION 1: Voice recognition technology automatically transcribes the words spoken on screen and “makes” the closed captions.
Uh, no. Ever watch the auto-generated captions on YouTube? The captions are often comically wrong. But on the plus side, the inaccurate computer-generated captions provide job security for real-life captioners. Also, I applaud Google for going to the expense and effort of providing closed captions when the content providers haven’t provided them.
I was taken aback when I told someone about my job, and they questioned the need for human captioners, pointing out that Siri on their iPhone was so accurate at interpreting their speech that it could be used for captioning, also.
Arrrgg…
Misconception 2: Voice recognition technology can format captions.
Uh, not by a long shot! Voice recognition technology doesn’t identify speaker changes. Voice recognition technology mixes up homophones. Voice recognition technology doesn’t know when one sentence ends and another begins. Voice recognition technology doesn’t always capitalize proper nouns. And don’t even get me started on voice recognition technology’s complete inability to format in AP Style or Chicago Manual of Style.
The good news here is that– for now, at least– humans can still provide more accurate and readable captions than Siri can.
Yours verbatim,
Caption Breaker
“Captions: so hard to make; so easy to break.”
Leave a Reply