[DrBoring]

Tangential

I've imagined some sort of video watermark that would be dynamically injected into video streams from services like Netflix.

The watermark would ID the account to which the stream was sent. The purpose would be to catch whoever is making the web-rips the make it into the wild.

I don't actually expect anything like this to be practical. It's a lot of technology to develop and support just for the slight chance of catching video pirates. But it's a fun mental exercise.

Questions I ask myself include:

1. What would the watermark look like visually? I image a slightly off-colored pixel appearing at various x,y locations at various time intervals. Sort of like the yellow microdots on laser printers that ID an individual printer.

2. How would the visual data be injected into the existing cached video byte chunks residing on CDN (I'm not sure of the correct technical term for the packets). I suppose the altered packets would have to be created on the fly and sent to the CDN host serving the individual streamer.

[wildzzz]

Since the only way to watch most streaming services is through their dedicated player that can already overlay text and graphics, why not just take advantage of that? It would need to be something that can survive video transcoding. Something like very slight dimming and brightening of specific segments of the video could encode an ID. The video would already have a list of random segments identified for encoding and each bit in the ID would tell the player to dim or brighten that segment. If you based the segments off of keyframes, the dimming or brightening might be imperceptible.

[zuminator]

Watermarking the streams doesn't do anything to prevent Blu-ray rips from being propagated, just as watermarking streaming audio doesn't prevent someone from sharing their CD rips. All it takes is one pristine source rip for the watermarking effort to be moot.

As far as catching web rip culprits, that becomes a cat and mouse game when easily available stolen credit card numbers are a thing.

In the end, i think these type of schemes end up protecting information security division income more than they protect the artists' income.