Topic: Closed captions

Posted under Tag Alias and Implication Suggestions

The bulk update request #6451 is pending approval.

change category soft_subtitles (0) -> meta
create alias soft_subtitles (0) -> closed_captions (6)
change category closed_captions (6) -> meta

Reason: Closed captions, also known as soft subtitles, are text or other visual display to provide additional or interpretive information that can be turned on or off by the user. This contrasts with open captions, also known as burned-in captions, which cannot be disabled because they are part of the video stream.

A common closed captions format used on the Internet is WebVTT. An example of a website that uses WebVTT is YouTube.

Are they even displayed to the user when embedded into the video? As in, if you go to a video with WebVTT embedded into the video, will they be displayed? If not, there's absolutely zero reason to tag it.

donovan_dmc said:
Are they even displayed to the user when embedded into the video? As in, if you go to a video with WebVTT embedded into the video, will they be displayed? If not, there's absolutely zero reason to tag it.

presumably as long as the video player your browser is using supports CCs, yeah.

sipothac said:
e6 uses the browser's video player, so it'd depend.

We don't provide the required track element, so it doesn't really depend on the browser.
The implementation also requires it to be an entirely separate file, making it being embedded into the video itself irrelevant for us.

Given this started over it being added to a video via the metadata, I'd assume there's a way to attach them via that method.

https://www.webmproject.org/docs/container/

It is hard to tell from the way it is written, I don't read technobabble (English technobabble, anyway), so I can't tell if it's already implemented or not, but it suggests this is an intended functionality of how webms work. Unfortunately, I don't know of any examples on here where that metadata is present to test it.

Please! This would be so useful! Even if they aren't displayed on the website, being able to find files that have them would be a blessing. It's an incredibly important aspect that too often gets overlooked in terms of archival. Even if it's buried in the metadata, having a transcript means that the file can be searchable amongst a large library, or be translated with ease. On top of the barebones basics of making the video accessible to hard of hearing people.

Aren't they some kind of text format with markup, like HTML or JSON?
I seem to remember that. Typically, I'd post text file link or story_at_source with archive for stuff like stories. I guess you could do something like that. Putting links to external resources (like videos over the 100MB limit!) in the description is not exactly unheard of. Putting subtitles directly in the description or comments would be brain dead, but I figure that I still had to state the seemingly obvious.

At least it's not a hard sub situation. "Oh, better translation and much higher resolution or uncensored! But oops, we forgot that subtitle files exist and only provide it with embedded watermarkssubtitles. Our bad!"

oopsitripped said:
Please! This would be so useful! Even if they aren't displayed on the website, being able to find files that have them would be a blessing. It's an incredibly important aspect that too often gets overlooked in terms of archival. Even if it's buried in the metadata, having a transcript means that the file can be searchable amongst a large library, or be translated with ease. On top of the barebones basics of making the video accessible to hard of hearing people.

Yeah, there are people that put in as much effort for the subtitles as the actual video files.

As a side note: I was comparing PNG files with different hashes that compressed to the exact same pixels, and there WAS some data that was technically erased from the (much) bigger files. As in, not just different compression of the same data, but with PNG you can optimize away invisible but still present data. Namely, stuff the alpha channel hides. IOW, it's a layered format, and good compressors will actually wipe away redundancy between those layers if they're invisible in the final image. Usually though, I saw where some sites ran PNG images through a more effective compression algorithm, to save space/bandwidth, at the cost of CPU time. Files that were within 20% of sizes were very very likely identical (even including the layers).

Updated

I don't see a reason not to include the tag, even if e621 doesn't display them - it is embedded meta information after all. If I download post #4562753 and open it in VLC the closed captions show with no problem, I imagine it wouldn't be impossible for a browser extension to make them appear in-browser.

If being visible on e621 without any browser extensions or downloading a file is a requirement, we may as well go nuke all the tags on all the SWF files ๐Ÿ™ƒ

faucet said:
I don't see a reason not to include the tag, even if e621 doesn't display them - it is embedded meta information after all. If I download post #4562753 and open it in VLC the closed captions show with no problem, I imagine it wouldn't be impossible for a browser extension to make them appear in-browser.

If being visible on e621 without any browser extensions or downloading a file is a requirement, we may as well go nuke all the tags on all the SWF files ๐Ÿ™ƒ

๐Ÿคก

Pup has a mirror of every single file, ever, right? I assume because that's kind of required, to do stuff like checking resolution, (alpha) channel presence, etc. If we're extremely lucky, the metadata was already extracted in a form where the original files don't have to be rechecked, but I doubt it. Not sure how the bot works, so this'll be a TIL day.

faucet said:
I don't see a reason not to include the tag, even if e621 doesn't display them - it is embedded meta information after all. If I download post #4562753 and open it in VLC the closed captions show with no problem, I imagine it wouldn't be impossible for a browser extension to make them appear in-browser.

If being visible on e621 without any browser extensions or downloading a file is a requirement, we may as well go nuke all the tags on all the SWF files ๐Ÿ™ƒ

Comparing to flash isn't really fair, considering most tags added to flash would have been added when it was widely viewable in browsers
They aren't visible now yes, but we don't allow uploads of flash anymore

cloudpie said:
Yes please. Accessibility is important and there's a significant community of deaf/hoh furries.

also one that just don't want to have porn sounds playing through their speakers.

cloudpie said:
Yes please. Accessibility is important and there's a significant community of deaf/hoh furries.

And those to whom English is not their native language, so having a text version of what is said can be very helpful. You can even pause and read the subtitles - you can't pause and listen to the spoken dialogue.

Pup

Privileged

Only just saw this thread, and it's definitely something I could check for.

alphamule said:
Pup has a mirror of every single file, ever, right? [..]

I don't, it just overwrites one file over and over to save space, but ffprobe <url> gives the tracks and metadata of the file without needing to download the full thing, so it'd be really easy to go through every video and check for embeded subtitles.

Given it also includes the track name I could possibly make <language>_closed_captions as well. It'd be great to let the hard of hearing or non-English speakers be able to search for them, even if they have to download them to see them.

alphamule

Privileged

pup said:
Only just saw this thread, and it's definitely something I could check for.

I don't, it just overwrites one file over and over to save space, but ffprobe <url> gives the tracks and metadata of the file without needing to download the full thing, so it'd be really easy to go through every video and check for embeded subtitles.

Given it also includes the track name I could possibly make <language>_closed_captions as well. It'd be great to let the hard of hearing or non-English speakers be able to search for them, even if they have to download them to see them.

Ah, given that HDDs are cheap I just assumed. One time effort and a lot faster to run locally. Then again, I have 4G ISP speeds. XD

Mentioning local mirrors, it seems that a lot of these never got seeded: topic #37724

  • 1