this post was submitted on 15 Aug 2024
174 points (98.9% liked)

technology

23182 readers
461 users here now

On the road to fully automated luxury gay space communism.

Spreading Linux propaganda since 2020

Rules:

founded 4 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
[–] GaveUp@hexbear.net 7 points 1 month ago* (last edited 1 month ago) (1 children)

You're interpreting the term watermark too literally

It will be a small unique arrangement of just a few pixels to identify the user

It can even be distributed across the screen pixel by pixel to make it less noticeable

All they'd have to do is make each pixel 1 hex code lighter or darker or something

Assuming each pixel can have no change, 1 step lighter, or 1 step darker, it'd only take 22 pixels to cover 31B accounts = 3^22

I believe there's 25B Google accounts in total out there atm

[–] ashinadash@hexbear.net 4 points 1 month ago (2 children)

In every frame, easily identifiable by a shitty pinhole camera though?

[–] GaveUp@hexbear.net 4 points 1 month ago (1 children)

I updated my comment with more details

[–] ashinadash@hexbear.net 4 points 1 month ago* (last edited 1 month ago) (1 children)

It's plausible but unlikely I think, putting a lot of faith into shitty pinhole cameras to be able to see twenty two 4K pixels one hex value lighter or darker, when most cameras have atrocious definition/sharpness and get blown out by light, blinded by darkness. I dunno, this reminds me of the screaming around Microsoft Kinect in 2013. They had bad and shitty plans for Kinect but, cheap hardware everyone hated Idk.

[–] FloridaBoi@hexbear.net 2 points 1 month ago (1 children)

I feel like if you just slightly turn up the compression ratio then all that nuance is lost making the watermark nonexistent or unusable

[–] ashinadash@hexbear.net 3 points 1 month ago

Yes especially since Netflix in particular has atrocious compression.

[–] The_sleepy_woke_dialectic@hexbear.net 3 points 1 month ago* (last edited 1 month ago) (1 children)

There exists a technology that takes elements in a picture, like a bird in the background, a character, a glass of water, etc and moves them just a few pixels. You can encode a lot of data like that and it's undetectable given just one example. They can encode your unique user identifier 1000 times in even a short video. A camera is bound to pick up at least part of it each time.

[–] ashinadash@hexbear.net 4 points 1 month ago (1 children)

Quotin'

putting a lot of faith into shitty pinhole cameras to be able to see twenty two 4K pixels one hex value lighter or darker, when most cameras have atrocious definition/sharpness and get blown out by light, blinded by darkness.

I guess if the TV itself was doing the DRM recognition? Idk though, I've seen alarmist posting like this before... seems to me evil tech shit usually gets done in more mundane ways?

Its definitely possible and even trivial to do there are a thousand ways to encode just a few bytes of data undetectably in a video and nothing but motivation stopping them from using every one every where. I think it's plenty mundane and even trivial for what they get.