Are Automatic Captions On YouTube Getting Better?
Dare I say that Google/YouTube’s automatic captions might have gotten better? You don’t have to rub your eyes or clean your glasses. You’ve read my words correctly.
On June 6, 2015, I uploaded a video talking about how the automatic captions aren’t captions. I even called them “craptions”, which is a term a man named Michael told me about back then. Most of the words were incorrect. There was no grammar. (For the record, I’m no expert when it comes to grammar, but the lack of punctuation and capitalization sure was something.) Everything was essentially one long run-on sentence. Captions would stack up on each other and move at a slow pace.
All of this is why I hated seeing that this was an option (and still kind of do). Every time somebody responded to, “Please caption your videos!” with, “Click on the CC button. There’s automatic captions!”, I would cringe. Every single time, I would reply saying that they’re not the same thing, not even close. This is why I made videos stating that automatic captions were not captions and why #NoMoreCRAPtions was born.
Now fast forward to sometime around 2017. What are the automatic captions like now?
Last summer, a news company of some sort or brand emailed me and asked to do a little interview in regards to automatic captions. I said yes and the article was posted. Not too long later, I received an email from one of my contacts at Google. She told me that she read the article and she wanted to tell me about what Google was doing to make the automatic captions better. I can’t remember the details of the conversation but her point was that they were improving and they were continuing to work on improving them.
So, have they improved? The short answer is, in fact, yes. The longer answer is that while they have improved, they are not perfect.
For the most part, I’ve found that a lot more words are correct. As long as the audio is clear, the system doesn’t have a huge issue translating voice to text. Even my own automatic captions are a lot better than they used to be and I’m a person with a weird combination of a deaf, American southern, and German accent.
My only real issue at this point of time is the lack of proper grammar/format. The format matches television format in the sense that they stack up on each other (this format isn’t on all networks, by the way), but it all just looks like one run-on sentence. Without punctuation and line breaks to break everything up, it becomes difficult to read. Hopefully, this can be improved on down the road.
Overall, I’m pretty pleased with how it’s going so far. It’s clearly improved since 2015. The best thing about this improvement is that it will make it so much easier for creators to caption their content, whether it be those who currently caption now or those who have never captioned because they thought it would be too tedious. There has always been an option to simply edit the automatic file and now with the improvements, it should take less time to fix.
With all of that said, do I think that we should choose to use only automatic captions? No, I don’t suggest that. I will always suggest manually written or edited captions because it will be the most accurate. Automatic captions are not 100% accessible and that is what captions should be. However, when push comes to shove, I’m likely going to use (and I already have) automatic captions on some videos if I really want to watch it and have issues understanding what’s being said.