The Firefox browser now has a built-in page translator that works even without the Internet::Mozilla has announced the release of an update to its Firefox browser. In version number 118, users will find a significant innovation - a built-in translator

  • Marius@lemmy.mariusdavid.fr
    link
    fedilink
    English
    arrow-up
    39
    ·
    1 year ago

    I’ve been using for a few months. Here is my opinion:

    • Translation quality is still far from good, but is good enought to be understandable.
    • Can’t translate PDF files (hope it could do it in the future, even if that mean reflowing it)
    • The extension allowed to keep translating this tab. That’s a future that, in my opinion, would be highly appreciated in the built-in translator (instead of enabling the “always translate”).
    • The language choice doesn’t correspond with what I usually need (which is chinese. But I know chinese is notably hard to translate.)
    • It seems that translation into french first goes thought a first pass of english translation. While this still produce readable result, targeting english is for now probably the best option (even thought the cost of implementing a new language translation pair doesn’t seems too high, I understand they might prioritise adding more language, at least for now. Actually, I should probably contribute to this myself if I care as much about it)
    • stevedidWHAT@lemmy.world
      link
      fedilink
      English
      arrow-up
      3
      ·
      1 year ago

      Very interesting points, does OCR work with PDF? Might be a possible temporary work around to run the pdf through ocr and pipe back into Firefox (maybe by running it on a local version of the .html and just directly injecting the pdf back in? Not sure, could be a fun project!

      • Marius@lemmy.mariusdavid.fr
        link
        fedilink
        English
        arrow-up
        1
        ·
        1 year ago

        Well… I once tried to just copy the pdf into a .txt file that I then opened into firefox, but it seems to not translate .txt, thought it may be cause they are not HTML.

        • stevedidWHAT@lemmy.world
          link
          fedilink
          English
          arrow-up
          1
          ·
          1 year ago

          Yeah sadly PDF content is practically baked in like an image. If you want to learn more you can search for the breakdown of pdf structure at the infosec institute website, remember there being an article that was pretty informative. I’ll see if I can find it later today and update this