YouTube has been spotted testing server-side ads, which could pose a problem to ad blockers.

    • linearchaos@lemmy.world
      link
      fedilink
      English
      arrow-up
      0
      ·
      4 months ago

      A bit longer if they play their cards right. If we have to move to a detect, buffer and skip model, it’ll be quite a while.

      • CheeseNoodle@lemmy.world
        link
        fedilink
        English
        arrow-up
        0
        ·
        4 months ago

        There are already addons to skip in-stream adds for when people do those sponsor segments. It just crowd sources the most consistently skipped parts and does it automatically.

        • SkaveRat@discuss.tchncs.de
          link
          fedilink
          English
          arrow-up
          0
          ·
          4 months ago

          If you’re referring to sponsorblock: no, those timings are manually submitted, not automatically generated by Skip time

        • Strayce@lemmy.sdf.org
          link
          fedilink
          English
          arrow-up
          0
          ·
          4 months ago

          Timing and length of ads won’t be consistent between users so this approach doesn’t work.

          • CheeseNoodle@lemmy.world
            link
            fedilink
            English
            arrow-up
            0
            ·
            4 months ago

            Depends how wide the range of ads is, if its a set amount/month then you could use the same method to pretty quickly identify which segments of data are ads and automatically cut them out wherever they occur.

        • linearchaos@lemmy.world
          link
          fedilink
          English
          arrow-up
          0
          ·
          4 months ago

          Sponsor block works so well because the content to skip is exactly in the same place for everyone’s video.

          They don’t have that kind of limitation.

          Sb also requires you to initiate a fast forward during a commercial. They could ignore that command while commercials are running.

          At that point we need to store the buffer as it’s coming in. Detect the commercials based on sound or scene changes.

          Since they do have a limited number of of commercials, if we did the crowd effort to identify and catalog /fingerprint commercials we may be able to do a more accurate method for detection and skip. At least until they are modifying the commercials with AI to avoid detection.