YouTube Automatic Captions Vs. Hand-Made Captions - Which is Better?

OK, the answer to the above question is obviously hand-made captions, but it takes time (and therefore, if you're hiring someone to do it, money) to do.

But the real question is, can YouTube's automatic captions serve a purpose? Maybe they can for purposes of generating search terms, but as far as actually serving to help people who are hard of hearing or who otherwise want to follow along with the speech on your video reading captions, I've found that they're not worth using.

The nonsequitur errors make the text unreadable for any hard of hearing or deaf person, and more of a distraction than help reading along for others.

Can they be corrected to make them usable? Yes, you can download and edit YouTube's automatic captions and make corrections, but I have discovered that doing so, at least for someone with the typing skills of a transcriptionist, it takes significantly less time to actually do a transcript and convert it to captions than to correct YouTube's automatic captions, thus making them inefficient in terms of cost.

The one purpose they may serve is as a way for slow typists to caption videos, or for people who don't know how to make captions of their own. It might be faster to correct captions for them than it is to make them. But if you're hiring someone to do captions, YouTube's automatic captions take too long to correct to be worth using. You should just hire someone to do them from scratch.

I offer this service and many other multimedia and translation services at
http://www.direct-translation-services.com
directtranslation@gmail.com

Category:

Education

Tags:

• youtube
• captions
• automatic
• autocaps
• subtitles
• timing
• transcribe
• transcription
• translation
• tyt
• young turks
• wingnuts
• closed captions
• open captions
• cc
• john avlon

[funkalunatic]

It demonstrates that it's impossible to do accurate transcription without analyzing context. The guy actually does pronounce "rabid" as if it were "rabbit", and the smooshing of "the, and, uh"s and stuttering are like code that can only be deciphered in melodic context.