YouTube as a streaming client

Copyright algorithm scans private listings.
>>>/g/sqt next time.

>>60698120
It's possible to scramble video image with a filter, upload the video and then on streaming back use a filter in reverse order. Same goes for audio. The scrambling mechanic needs to be somewhat compression proof.

>>60698233
devious.

>>60698233
no
google can recognize images with noise and shit added
you would probably have to render the material unviewable for that to work

>>60698233
It would look like trash due to the compression (even if somewhat compression proof), and if you're going to all this trouble to just stream some movies you might as well pay a VPN/seedbox/whatever.

>>60698339
Adding noise wouldn't be a sound scrambling mechanic as that would destroy data.

>>60698339
He didn't say "add noise and shit"
He said "scramble"
Are you blind?

>>60698355
Circumventing Google's own system is more adventurous.

>>60698097
just use google drive you dumb fuck, you can upload anything

File: A_For_Effort.jpg (94KB, 1280x720px) Image search: [Google]

94KB, 1280x720px

>>60698500
To be honest it feels good to live somewhere where 0 fucks are given about pirating. Well, at least if you don't go around selling it, you're fine.

At least one good thing from living in this shithole.

>>60698233
what if you slow it down,speed it up by 2x or reverse the video thats less likely to be spoted i belive

>>60698551
>you can upload anything

No you can't.

>>60698551
15GBs youtube has unlimited

>>60698589
Then how the fuck can mkbhd upload 8K videos? these surely exceed 16GB

>>60698097
I've been thinking about this myself, but with google/outlook drive.

would they be able to detect shit if I uploaded .rars with passwords? plus I'd be keeping it just for myself.

>>60698339
No, no they cannot.
There's MILLIONS of copyrighted videos on the site right now because they have weird borders around them. That's all it takes to defeat their scans.
The only other method they have to scan is audio.
Scrambling the audio just enough to be heard but still noisy enough to not be detected.
Equally dynamically compressing the audio at noisy sections, or stretching, it can throw their filters off even more so.
Not detectable to humans as much, but it will shit on average-shape based analysis that gets rid of noise to create those audio-hash segments.
A simple comparison would be turning a V in to a U by smoothing the waveform out over a few frames. Doing that in noisy sections is very easy to do without being very audible.

Google scaled back their auto-flagging shit years back because it was throwing false-positives out the ass.
A fucking birds background video was copyrighted as a film for fucks sake.
Sure, they might get removed with time using new analysis powered by stuff like Tensorflow, but as-is that's an absolute negatory.