Background audio can reveal where a video was recorded
cybernews.com
100 meters may be close enough for this background audio privacy risk, as aircraft sounds can help locate videos. See what the evidence shows

cross-posted from: https://lemmy.world/post/54070272

When uploading a real recording, there is a good chance your location & therefore possibly your identity can be revealed from the footage itself.

  1. You can triangulate low volume background sounds such as air planes: https://cybernews.com/privacy/aircraft-doppler-audio-geolocation-privacy/

  2. You can check the position of the sun/moon/stars, shadows, weather…: https://www.bellingcat.com/resources/2020/12/03/using-the-sun-and-the-shadows-for-geolocation/

If Streamers & Journalists CAN do it, government agencies ARE doing it.

If you need privacy be aware of this when recording / editing friends <3

  • grahamja@reddthat.comEnglish
    4·
    6 hours ago

    Typically your first step in avoiding things like this is don’t be important. Your second is don’t record everything and share it. If you do have to share it, wait until long after you leave the place you recorded it.

  • Imgonnatrythis@sh.itjust.worksEnglish
    6·
    18 hours ago

    We need user friendly video scrubbing software to remove or at least identify tracking risks as the tools to find them are going to keep getting better.

    • Psi9@kbin.earth
      71·
      7 hours ago

      watch the YouTube video. its really very very good. and it pints out that attempting to scrub the audio of the types of indicator he uses would be very difficult. as airplane noise is both in the same frequency range as our voices, so it can’t be taken out without affecting voice quality, and penetrates walls so will always be present.

      • meaansel@lemmy.worldEnglish
        3·
        4 hours ago

        Video only demonstrated that high-pass filtering doesn’t work. If you (or that hypothetical software) know where you are and at what time you recorded, you’ll know exactly where those ミ lines are are gonna be, so you’ll only need to erase them and fill them with noise indistinguishable from rest of background.

        And just in case your noise model is not good or seams aren’t that well hidden as you think, make it look like you erased a bunch more plane sounds in other paces on recording, so that there are thousands of places you could be in (picking should be weighted with population density)