Hacker News new | past | comments | ask | show | jobs | submit login

If the RIAA notice is to be believed, you've still admitted to a crime as you've bypassed YouTube's DRM, which is their hosting of the video in other public links.



IIRC, YouTube actually does apply extra DRM to some videos (ones "owned" by the RIAA, maybe others?)


Nope, they argued that their URL obfuscation is DRM.

edit I think YouTube-Red's successor has DRM on it's videos, I don't think youtube-dl ever worked on them though.


I better throw out my web browser then. It is circumventing YouTube DRM by opening those URLs.


By that logic, playing Netflix on your browser is breaking DRM. That’s not how that works.


Netflix sends the videos to your computer with a form of DRM and then uses a key from either your browser or hardware to unlock the content. That is nothing like YouTube, which sends both the links and the media to you unencrypted.


The video torstenvl downloaded wouldn't have had the rolling cipher on it.


From my understanding, every video has the rolling cipher as it is just what YouTube calls the links the media files are at.


My understanding is in general a fixed signature is used, only certain content has the rolling cipher.

If all videos used the same system then youtube-dl would not have had any reason to make major label music videos part of their unit tests.


Youtube-dl's counter claim states that though those lines of code did not violate the DMCA, they have replaced them with videos without copyright music.

That sounds like they didn't really have any reason to make major label music videos part of the tests, it was just a developers personal preference. Though, it doesn't prove this is the case.


It is not the case, and all that has happened is a removal: https://github.com/ytdl-org/youtube-dl/commit/1fb034d029c8b7...


That commit details the replacement of the music videos with a generic test video, exactly what I said. I'm unsure how it is supposed to show it is not the case.


Incorrect. There's a single green line that has an alteration to replace a music video ID "UxxajLWwzqY" in a test case that actually only makes use of ID "BaW_jenozKc" ("Use the first video ID in the URL").

The removed "Test generic use_cipher_signature video (#897)" case did make use of ID UxxajLWwzqY.


I see what you mean, but the extractor file still features how to deal with videos containing the cipher, all that was removed was the tests. Testing the generic cipher video may have been ruled unnecessary as it's universal.


That's the partial win here - repo restored with tests referencing the RIAA related videos removed but the code dealing with the rolling cipher itself still intact.

And next time youtube makes one of their frequent changes to their website the extractor will break in some way. Somebody will work to fix it and make use of the same tests, only now some of them won't be in the public codebase.


Except we haven't established whether there was anything special about those videos at all


But on the other hand there is always an omelette to whisk.


Youtube-dl works on way more sites than just YouTube.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: