[brokentone]

FB also seems to be a bit of the former. Here (https://www.facebook.com/GoodMorningAmerica/videos/101533728...) is an anecdotal clip that ABC/GMA stole from this guy (https://www.youtube.com/watch?v=EU13ByIYkew) with no credit, they flipped, and re-watermarked his video. They got 22mm views on something the author only has 0.5mm views. Terrible ABC, terrible.

[deathhand]

There needs to be CRC/SHA for video. Once that is figured out identifying these sorts of things will be trivial.

I would in fact pay someone to develop this as I think it is something humanity needs. Or I could pivot it into a startup and offer content creators a "safe harbor" to let them know whenever their stuff has been pirated.

[nekopa]

Isn't this what YouTube calls content ID? They alert you that your content is up somewhere, then you have the option of monetizing it, or having it taken down.

[deathhand]

Uh interesting but sadly it doesn't look like it is open for all:

https://support.google.com/youtube/answer/1311402

[roywiggins]

Not quite trivial: people can crop, skew, vignette, mirror, timestretch and scale video to try to defeat it. YouTube's Content ID seems to be vulnerable to this sort of thing still.

[x0x0]

I think that's why many of these systems focus on audio. It's harder to change in a way that beats fingerprinting but doesn't ruin the viewing experience.

[izelpii]

I think the majority of the cases are uploaders that basically post the same content.

[izelpii]

ContentID is limited to few companies. It requires training, is not multi platform, and doesn't stop freebooting.

[r721]

https://en.wikipedia.org/wiki/Digital_video_fingerprinting

[thatoneguy]

YouTube already has Content ID and it's by no means perfect.

[obituary_latte]

It's possible ABC licensed it from him. There is a link in the description to license the vid.