763
Sean Murray explaining what the No Man's Sky team has really been up to (lemmy.world)
submitted 1 year ago by to c/whitepeopletwitter@sh.itjust.works
56 comments fedilink hide all child comments
 
you are viewing a single comment's thread
view the rest of the comments
[–] 43 points 1 year ago (25 children)

it's crazy to me that for all the ai "advances" in the past few years nobody has thought to improve subtitling.

  • permalink
  • fedilink
  • source
  • hideshow 25 child comments
    •
  • [–] 7 points 1 year ago (8 children)

    Andrew Ng did a video when he gradually added noise to the training audio to improve the quality.

    But here we are dealing with homophones so it's not just turning speech to text, it also needs to be context aware.

    Possible but too expensive to implement automatically.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 8 child comments
    •
  • [–] 6 points 1 year ago (7 children)

    context awareness is the entire point of language models tho :(

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 7 child comments
    •
  • [–] 2 points 1 year ago (6 children)

    I'm highlighting that speech to text and context awareness are different skills.

    YouTube is unlikely to waste loads of compute power on subtitles that don't need it just to capture the occasional edge case.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 6 child comments
    •
  • [–] 1 point 1 year ago (5 children)

    i mean, it's a one-time-per-video thing. they already do tons of processing on every upload.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 5 child comments
    •
  • [–] 3 points 1 year ago (3 children)

    So if you can reduce compute there then you save money.

    There is no technical difficulty. It's a business decision.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 2 points 1 year ago (2 children)

    right now they're dynamically generating subtitles every time. that's way more compute.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 1 point 1 year ago (1 child)

    For real? That's incredibly dumb/expensive compared to one subtitle roll. Can you share where you saw that?

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 1 point 1 year ago*

    well, i have no evidence of this. however. looking at the way auto-generated subtitles are served at youtube right now, they are sent individually word-by-word from the server, pick up filler words like "uh", and sometimes pause for several seconds in the middle of sentences. and they're not sent by websocket, which means they go through multiple requests over the course of a video. more requests means the server works harder because it can't just stream the text like it does the video, and the only reason they'd do that other than incompetence (which would surely have been corrected by now, it's been like this for years) is if the web backend has to wait for the next word to be generated.

    i would love to actually know what's going on if anyone has any insight.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 6 points 1 year ago* (2 children)

    It's even worse for captions.

    Captions and subtitles aren't even the same thing.

    In fact, most DVD players don't even pass the code captioning through HDMI ports, so old captioned DVDs don't work anymore.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 4 points 1 year ago* (last edited 1 year ago)

    explain. edited with explanation. i've seen the technology connections video, thanks.

    my comment is still about the actual post above, and i was specifically thinking about auto-generated subs rather than, say, movies. apparently that's not obvious.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 6 points 1 year ago (2 children)

    My student friend tells me that the auto-generated captions for non-English MS Teams lecture recordings recently have improved significantly and have even become usable.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 1 point 1 year ago (8 children)

    Probably because a lot of countries either dub the content or it is already in their native laguage. You generally see a lot of subtitles on OpenSubtitles of countries like The Netherlands where that doesn't happen

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 8 child comments
    •
  • [–] 2 points 1 year ago (7 children)

    on youtube and twitter?!

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 7 child comments
    •
  • [–] 1 point 1 year ago (5 children)

    No, but on the fields where there is money to be made for subtitles, like movies and tv shows.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 5 child comments
    •
  • [–] 1 point 1 year ago (4 children)

    what does that have to do with the OP though?

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 4 child comments
    •
  • [–] 1 point 1 year ago (3 children)

    Where do you think money for development of new shit comes from? From places where money is made, like TV, movies etc. YouTube doesn't really make more money because the ads are there.

    And Twitter hahaha I can only laugh at that

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 1 point 1 year ago (2 children)

    ...i feel like you're having a different conversation than i am.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •