5 ms·
A "in a hurry" hack might be to run it through a bad phone line? Still understandable, but you can probably guess that the phone provider is going to crush the
by mhh__ 6y ago
A "in a hurry" hack might be to run it through a bad phone line? Still understandable, but you can probably guess that the phone provider is going to crush the spectrum and significantly reduce the bit depth.
- adamjb 6y agoSimilarly my first thought would be to reencode it as <96kbit/s MP3.
- megous 6y agoWhy do such risky unpredictable things if you can simply filter out certain parts of the spectrum?
- j4yav 6y agoHow do you know what bands they are watermarking? And isn't it likely they use some within the normal speech band?
- megous 6y agoWhy would re-encoding into mp3 help? The audio is probably already in some low bitrate similar codec.
- semi-extrinsic 6y agoIf you pick a bad enough codec, it should mask watermarks by distorting the sound, while still allowing speech to be audible. Maybe something like mu-law, which only has 3 kHz bandwidth centered on human speech (as opposed to mp3 with usually 44.1 kHz) would perform best.
- ddingus 6y agoAt those rates, the audio loses credence.
- quenix 6y agoI think the idea is that if they use audible frequencies, the user will, well, _hear_ them.
- eru 6y agoPeople are used to artifacts in zoom calls, the sound quality ain't always the best. I also noticed that zoom corrects for lag in my connection by speeding up and slowing down speech. (I noticed that when my counterpart was running a metronome.) Zoom could artificially introduce micro-lags to encode data. Eg as a human I can't really tell whether some sound starts on an even or an odd milli-second.
- ddingus 6y agohttps://blogalytics.typepad.com/files/a-technical-look-at-arbitrons-ppm-system.pdf https://blogalytics.typepad.com/files/a-technical-look-at-ar... Not really. And if a user does, they will hear rather mundane codec artifacts, and or noise that is basically like other noise. Given a redundant mark, and the system I linked has multiple, concurrent redundancy, the chance of the mark surviving grows with the amount of audio. They can literally drop marks during speech the entire time.
- j4yav 6y agoIt's not like it's a voiceover that explains who is watching the video. It can be a series of quiet pops.
- ddingus 6y agoExtremely likely. That is probably the best place to put it.
- m463 6y agoI doubt that would work. I would imagine there are trivial ways to let a watermark survive transcoding. Sort of like the content recognition systems in use: https://en.wikipedia.org/wiki/Automatic_content_recognition https://en.wikipedia.org/wiki/Automatic_content_recognition on the other hand, I found this interesting: In January 2018, a YouTube uploader who created a white noise generator received copyright notices about a video he uploaded which was created using this tool and therefore contained only white noise. from https://en.wikipedia.org/wiki/Content_ID_(system) https://en.wikipedia.org/wiki/Content_ID_(system)
- luch 6y agoI think the better hack would be to run it into a speech-to-text transcoder, and then use a vocoder to play it back. You lose the "voice" in the process, but in some situations it might actually be fine
- nucleardog 6y agoAn earlier comment[0] says there's a reliable method for including ids and things in AM audio at 5KHz bandwidth. POTS is a few KHz. So less, but without looking into Nielsen's fingerprinting methods I wouldn't say you could assume that the fingerprint wouldn't be preserved running it through a phone line first. [0] https://news.ycombinator.com/item?id=25831303 https://news.ycombinator.com/item?id=25831303