Closed Captioning with CEA-608 (or anything else!)

[Canaan Whitston]

To start, I work in the production department as an intern Editor for a larger business. We are the only production department that our company has and there aren't many of us. I have been put in charge of our captioning needs as we're switching from Open Captions (burnt in) to Closed Captions -also, we only post content to YouTube. That said, I've done as much research as possible in pertinence to Closed Captioning capabilities but am still new to this. Please excuse me as I'm sure that some of my terminology will be wrong, but I'm trying my best.

My boss has set 3 requirements for our captioning needs:

1. Captions can only be one line of text tall (as opposed to 2 or 3 lines)

2. The ability to caption ourselves, without waiting on other human factors. (no services or anything, we can't wait for that because we work on a really tight schedule.) That said, I've been downloading .srt files from Temi (semi-instantaneous transcription service) and tweaking my import settings in order to convert my file in an attempt to execute the following requirement stated below. (#3)

Also, I'm importing .srt's into premiere so that I can "export captions" as a different file type that will embed positioning information. I then upload my exported captioning file directly on the YouTube video.

3. Dynamic Positioning. This has shown as the hardest factor. We need to be able to vary positioning on captions at different points in the video. This is because we have a "lower third" with a phone number (at the bottom of the screen, where the captions default to) that is introduced about half way into our videos. We can not cover this number. I've done research into embedding positioning information in caption files and have found 3 file types to be compatible with YouTube. (Link to YouTube's supported file types )

These file types are:

.scc (optimal for YT)

.vtt (WebVTT)

.stl (british standard)

Anyway, I've tried the Target Stream Format "Open Captions" and that carried some of the positioning information from my Premiere exported .stl file (I chose .stl because it was the only viable option available with "Open Captions") when I uploaded to YouTube but it kept setting random X axis values to my captions. Then I set all my X values to the same number (0) and they got even more random!

After, I tried importing my .srt into Premiere and converting to Target Stream Format "CEA-608" (708 wouldn't let me export .scc's) and this worked. The positioning carries over and this stream format allows me to export .scc files. My problem with this is: It only allows my caption lines to be 32 characters of text long which looks small when uploaded to YouTube. It also displays differently than it does when exported and uploaded to YouTube but I don't care, I just need my lines to be longer in length. Is there any way to do this? Is there a better route? Any feedback is appreciated. 

[Stan Jones]

Sorry that I haven't responded sooner. I don't have a good answer, but I'll tell you what I know.

#1 should not be a problem.

#2 is not a big issue, ONCE YOU have the transcription. But for your purpose (sidecar file; Youtube in their preferred format; but with an unusual positioning), I see no advantage to using PR for the captioning. I'll comment further below.

#3 is a problem. See this exchange that is recent. My limited experience/experimentation has not produced a reliable positioning. And I don't see a help topic on Youtube re this. https://support.google.com/youtube/thread/2995175?hl=en

Instead of PR, I would simply work in SubtitleEdit or one of the other free apps. Just add your final product from PR as the video, import the srt into SE, use it's positioning, and export stl or whatever works. SE (and others) will also do amazing conversions and adjustments. A great tool.

Re positioning, I'm not going to delay this further by looking for more info. SE will add code that should work. It uses this type format: {\an7} (for upper left). Left to right, top to bottom, the codes are:

{\an7} {\an8} {\an9}

{\an4} {\an5} {\an6}

{\an1} {\an2} {\an3}

You want more control, and I do not see/recall how SE does more precise location. I believe that it does.

Let us know what you find.

Stan