23 votes

Swearing and automatic captions

13 comments

  1. [6]
    sparksbet
    Link
    This is really interesting! I'm someone who generally likes to always have captions on when I'm watching something (probably due to a combination of my own ADHD and having lived with a...

    This is really interesting! I'm someone who generally likes to always have captions on when I'm watching something (probably due to a combination of my own ADHD and having lived with a hard-of-hearing roommate back in college) and I've noticed and been frustrated by Google's censorship of swear words like this in their Youtube auto-captions. I can still hear them, so I luckily don't suffer from actually losing that information the way a d/Deaf user would. But I do share the frustration the author has with the infantilisation of it -- in the context of captioning content like YouTube videos, there isn't even the excuse that it's supposed to be a business environment!

    21 votes
    1. [2]
      lou
      (edited )
      Link Parent
      If I fire up Sopranos on HBO right now, the Portuguese subs will make it appropriate for 10-year-olds. And it's not just curse words they censor, it's the content as well. The phrase "I'm gonna...

      If I fire up Sopranos on HBO right now, the Portuguese subs will make it appropriate for 10-year-olds. And it's not just curse words they censor, it's the content as well. The phrase "I'm gonna fuck that whore's brains out" becomes "I'm gonna marry and have three kids with that girl!". I'm only half joking!

      5 votes
      1. sparksbet
        Link Parent
        lmao that's so egregious that it's kinda hilarious. Hope no one was trying to use those subs to improve their English!

        lmao that's so egregious that it's kinda hilarious. Hope no one was trying to use those subs to improve their English!

        4 votes
    2. [3]
      Weldawadyathink
      Link Parent
      What does d/Deaf mean? Does it convert more than just deaf? This article is the first time I have seen it.

      What does d/Deaf mean? Does it convert more than just deaf? This article is the first time I have seen it.

      2 votes
      1. [2]
        phoenixrises
        Link Parent
        deaf refers to the condition while Deaf refers to the community!

        deaf refers to the condition while Deaf refers to the community!

        5 votes
  2. [6]
    Wolf_359
    Link
    Interesting! I'll probably go down a rabbit hole on this subject now. I always use captions. Honestly I use them mainly because I like them, but I find they're helpful with media where the...

    Interesting! I'll probably go down a rabbit hole on this subject now.

    I always use captions. Honestly I use them mainly because I like them, but I find they're helpful with media where the dialogue is inexplicably mumbled and made quieter than the music.

    If anyone wants to see a lot of incorrect captions, watch The Wire. Anyone even semi-familiar with AAVE will catch a ton of inaccuracies and even inconsistencies across episodes. Definitely an old white woman behind those captions somewhere, but she did her best.

    13 votes
    1. [5]
      redwall_hp
      Link Parent
      As someone else who always uses captions, both to enhance understandability in English media and because I watch subtitled anime, I have a tangential and kind of amusing anecdote about automated...

      As someone else who always uses captions, both to enhance understandability in English media and because I watch subtitled anime, I have a tangential and kind of amusing anecdote about automated captioning...

      CrunchyRoll tends to offer both subtitled and dubbed versions of anime in their catalogue. When watching a show in its original language, the subtitling will be provided by a human translator attempting to be fairly faithful to the original script. When a show is dubbed into another language, the company handling the dub will create a new script that is often more localized as well as translated, so the dialogue will seem fairly natural in the new language. This means one cannot simply use the subtitle text from the non-dubbed track for captioning the dub...

      So, CrunchyRoll cheaps out on this and uses automated captions for dubs. Which does not work well at all, with often bizarre and amusing results. Not only does the system have no concept of different people speaking, so dialogue going back and forth results in weird sentences based on wherever it decided to end them, but it completely butchers names. Watching the Re:Zero dub, you'll be treated to such gems as "mabeasts" (creatures in the show) being referred to as "lobbyists," Subaru occasionally being "super," and an inability to distinguish between the characters Rem, Ram and Rom. When it doesn't concoct more bizarre sentences.

      12 votes
      1. Weldawadyathink
        Link Parent
        There is a good Tom Scott video about this. It doesn’t help your CrunchyRoll situation at all, but does provide some interesting looks into the industry.

        There is a good Tom Scott video about this. It doesn’t help your CrunchyRoll situation at all, but does provide some interesting looks into the industry.

        3 votes
      2. [2]
        williams_482
        Link Parent
        I'm reminded of a classic in the garbled subtitle genre, The Backstroke of the West. Which, counter to the examples in the OP, definitely was not crafted to conform to an obscenity filter.

        So, CrunchyRoll cheaps out on this and uses automated captions for dubs. Which does not work well at all, with often bizarre and amusing results. Not only does the system have no concept of different people speaking, so dialogue going back and forth results in weird sentences based on wherever it decided to end them, but it completely butchers names. Watching the Re:Zero dub, you'll be treated to such gems as "mabeasts" (creatures in the show) being referred to as "lobbyists," Subaru occasionally being "super," and an inability to distinguish between the characters Rem, Ram and Rom. When it doesn't concoct more bizarre sentences.

        I'm reminded of a classic in the garbled subtitle genre, The Backstroke of the West. Which, counter to the examples in the OP, definitely was not crafted to conform to an obscenity filter.

        2 votes
        1. sparksbet
          Link Parent
          A lot of older, low-quality Chinese to English translations have variants of "fucking" in them in inappropriate places, and it's almost always based on the following issue: the word/character 干...

          A lot of older, low-quality Chinese to English translations have variants of "fucking" in them in inappropriate places, and it's almost always based on the following issue: the word/character 干 means a lot of things. With one pronunciation (it's one of the handful of characters with multiple different tones in Mandarin) it means "dry" and is thus frequently present in descriptions/names of dried foods. With another, it's pretty frequently used as a verb that means "to do" or "to work". Unfortunately, that same verb is also sometimes used to mean "to fuck". To a human, even one with pretty poor Mandarin skills like me, it's generally pretty obvious from context what meaning is intended with this particular character.

          Modern machine translation is pretty good at figuring out 干 from context, but back in the halcyon days of Babelfish that definitely wasn't the case. And when you need an English translation but there aren't any proficient English speakers around... you end up with a lot of "fucking" problems that don't get spotted until they're on social media. "Backstroke of the West" (which I remember watching years back and enjoying) almost certainly was made using this older low-quality machine translation.

          1 vote
      3. sparksbet
        Link Parent
        I'm honestly impressed Crunchyroll even bothers -- most services just don't offer "dubtitles" at all for dubbed media. I've wanted to use German dubs of some favorites to help improve my listening...

        I'm honestly impressed Crunchyroll even bothers -- most services just don't offer "dubtitles" at all for dubbed media. I've wanted to use German dubs of some favorites to help improve my listening comprehension skills, but I can't use the German subtitles to help me when I struggle unless I'm watching something that was originally made in German. It's a huge pain in the ass.

        1 vote
  3. ShroudedScribe
    Link
    As someone who is not d/Deaf, I still prefer to read compared to listening to audio when it comes to educational content. For the majority of my online courses, subtitles exist. But they are of...

    As someone who is not d/Deaf, I still prefer to read compared to listening to audio when it comes to educational content. For the majority of my online courses, subtitles exist. But they are of varying quality.

    There was one time someone in a video said a word that could have been misinterpreted as "ass" and was partially censored in the subtitles. So that leads me to believe they're auto-captioned.

    Even though I do not have an auditory impairment, being able to search through a transcript is incredibly helpful. And I completely agree that censorship of captions is absurd when the audio content contains such language.

    2 votes