Hacker News new | past | comments | ask | show | jobs | submit login
YouTube-dl has received a DMCA takedown from RIAA (github.com/github)
4240 points by phantop 39 days ago | hide | past | favorite | 1443 comments



All: don't miss that there are multiple pages of comments in this thread. That's what the More link at the bottom points to. Or click here:

https://news.ycombinator.com/item?id=24872911&p=2

https://news.ycombinator.com/item?id=24872911&p=3

https://news.ycombinator.com/item?id=24872911&p=4

https://news.ycombinator.com/item?id=24872911&p=5


Note that RIAA is making this takedown because the software CAN be used to download copyrighted music and videos, and it uses examples in the ~~README~~(unit tests, see correction[1]) as an example of that:

> We also note that the source code prominently includes as sample uses of the source code the downloading of copies of our members’ copyrighted sound recordings and music videos, as noted in Exhibit A hereto. For example, as shown on Exhibit A, the source code expressly suggests its use to copy and/or distribute the following copyrighted works owned by our member companies:

They could, of course, have asked for the code to have been changed. Instead, they attacked the project itself. IANAL, but this seems outrageous the same way DMCA'ing a Bittorrent client would be. This doesn't circumvent DRM like Widevine. I don't understand what leg they have to stand on here.

This feels like DeCSS all over again.

P.S.: They also took down youtube-dlc, even though it's not listed.

[1]: It turns out I am wrong. It wasn't in the readme, but in the test cases. See extractor/youtube.py. To me this seems even more tenuous, but IANAL.


The fact that copyrighted works were included in the readme shows it was intended for that use, and the RIAA complaint will likely stand up to any legal scrutiny. Just because it can be used for legit purposes too won't matter in the slightest. I mean, Napster could have been used for legal means as well, and it got destroyed in court.

The only chance tools like this have legally is when infringement is an "unintended side effect."


Just to add a data point, but back when I still was working for a video distribution startup, we offered our customers the ability to directly import their video inventory from YouTube. They were the owners of the videos, it was just a convenient (and very popular!) feature for them to let us handle this import.

We used YouTube-dl for this, of course. No way we could have done this easily without it. We imported hundreds of thousands of videos like this.

I suspect there are many more legitimate uses of YouTube-dl than you would expect.


To add another item to the list of legal, legitimate uses of YouTube-dl: lots of public domain content is uploaded to YouTube, including a lot of media produced by the US government. For example, The White House has a YouTube channel, and my understanding of US law is that the vast majority of the content uploaded to that channel is public domain (produced by federal employees in the course of their job). Journalists or anyone else wanting to monitor the government would likely find YouTube-dl useful for archiving this public domain content from government channels.


I've gobbled hours of Cspan videos with youtube-dl. Turns out cspan is a hodgepodge of like 3-5 video hosting and codec schemes, but it is able to pull about a 3rd of them no problem.

Definitely forking it, I need this to continue my research.


Hope you can find a way to fork it. GitHub has disabled the repo and many forked repos.


The beauty of `git clone` is that there are copies of the repository on many thousands of hard drives (including mine -- last pulled a few days back). The main loss at this point is the metadata stored in GitHub's proprietary addons, like issues.

We can hope this inspires a) greater suspicion toward proprietary players that are sitting atop open-source work like GitHub (they're not the only one by any stretch); and b) greater adoption of systems that distribute the software's entire history together, including documentation and bug reports, e.g. Fossil.


I'm the Founder of BackHub and that's one of the reasons why we offer backups for GitHub repositories, incl. the metadata! Reach out for a discount.


If no-one else has anything more recent, would you be ok to put yours online somewhere for others to grab?


Turns out it's not needed. There are up-to-date ones already around.

eg: https://news.ycombinator.com/item?id=24873953


And there is GitCenter, which is git over zeronet (a p2p network similar to BitTorrent but for dynamic content)


The code still in pypi, you can still get it from there

https://pypi.org/project/youtube-dlc/#files


Is there alternative to github? Feels like there should be. One that isn't regulated, a wild west for lost code so to speak


Try GitLab.


Or Gitea!


Do either store issues as objects in the repo itself?


Not git, but I believe fossil[1] does this.

[1]: https://www.fossil-scm.org/


Do you know of any systems that do this? Seems like a good idea.


I really like https://github.com/gitbucket/gitbucket for this.


Google should add direct download links for public domain content, but that's another fight entirely.


Why should they ? YouTube is in the streaming business, not the archival business. If someone wants to make their videos available for download they should post a link on their website, served through their servers and bandwidth.


Youtube benefits from being everyone's go-to site for finding video content. Keeping that monopoly of mindshare is worth far more to them than any short sighted profit maximisation.


as recent legal action shows companies benefit from being monopolies up until the point that they monopolize too much.


Yeah that's not some kind of general rule. This is because the US legal system in particular lets this happen. And in fact, the US legal system works so badly that it often still benefits companies long after they've monopolized too much, causing all sorts of damage to consumers and other companies that actually did nothing wrong.


Presumably as a condition of being allowed to hold their monopoly status and treat the (publicly developed) internet as a source of annuities.


Isn't streaming live video? It seems to me that youtube is exactly in archival video and hardly in streaming. But it's a big site so maybe a lot of people use it in ways I don't know.


Streaming is not exclusively live video. Streaming is viewing the file as it downloads. As a stream. The alternative is what we did in the dark ages. Download the entire file before beginning playback.


Even in the dark ages (e.g. '97) we could watch the fragment of the avi videos as they downloaded.


‘97 is classic antiquity. The dark ages came later, after civilization fell. We obtained WMVs from Kazaa because that’s all that was left.


I myself preferred realplayer ("buffering...") and its more advanced cousin realplayer G2.


From what I remember WMVs were mostly fake so I would never download them


WMVs can embed instructions on where to grab codecs or authorization to play the content in the multimedia. Back in the day if you played WMV with windows media player this could be used to pwn your computer.


damn. then I've been living through those ages. I feel old.


Because it is the right thing to do.


Actually, you can host any amount of pirated video content on youtube, as long as its "visibility" set to "private"...


I'd love to be wrong here, but there are two separe things, content license and general YouTube terms of use, I think the terms prevent you from downloading any content, the fact that some content is licensed CC has to do more with being explicit, perhaps for displaying publicly or just informative, and I want to think they are leaving a bit of gray area for non abusive use(tons of downloads) , but you can see there is no download button, and if you build something that depends on it to certain scale they could go after you with terms of use, not license.

If you can prove me wrong with some link to where are we covered legally to download CC content from YouTube, it would make me very happy.


> I think the terms prevent you from downloading any content, the fact that some content is licensed CC has

Maybe you are right. If that is the case it is questionable whether contents produced with tax payers' money should exclusively hosted on Youtube to increase Google's income and restrict tax payers' rights.


I remember using youtube-dl back 8 years or so ago when I was an undergrad in India to download MIT OCW lectures (and others e.g. SICP vidoes, Coursera), because bandwidth was limited, download was capped, and 3G was too expensive. I could watch these videos anywhere, and whenever I wanted. And I'm pretty sure non of those videos had (or even have) ads.


Right, but that's not the heart of the matter.

It's one thing if a tool designed for legitimate archiving use results in incidental violation of copyright.

It's another thing entirely, legally, if a tool is designed for violating copyright. In such case, it doesn't matter if there are also legitimate fair use uses of the tool.

And that's what the README showed: that youtube-dl was built primarily to violate copyright.

EDIT: based on other comments, it turns out the offending evidence was in the unit testing code not the README file, which legally makes it a bit worse than if it had just been in the README file.


Downloading a Youtube video with youtube-dl IS NOT violating copyright in the USA.

Downloading videos may violate Youtube's TOS, but that's civil/contract law and not criminal and so the DMCA can't be used to enforce such TOS violations.

A person might share a video they don't own the copyright to, after downloading it from youtube with youtube-dl, but youtube-dl has no functionality in it to do such illegal sharing. It only has the time-shift functionality.

The only tenuous leg the RIAA has to stand on is claiming that allowing only streaming of content constitutes some form of digital rights management code, and by reverse engineering it, without imposing the same restrictions as Youtube's website and apps, do, maybe that's some form of circumvention which will be found not entirely legal.

They had to cite a German legal ruling which found YouTube’s “rolling cipher” was an anti-circumvention device. But German court rulings have no bearing on US law, and Germany doesn't have a DMCA statute, so it is a specious claim, and the RIAA is just hoping an open source project won't be able to fund a legal defense.

It's Betamax time-shifting plus macrovision/video stabilizers all over again.


> Downloading a Youtube video with youtube-dl IS NOT violating copyright in the USA.

Are you a lawyer? Ethically I agree with you: copyright law regulates distributing copies to other people, and YouTube is the only one doing that in this scenario.

But I've been told in other internet discussions that courts don't see it that way. Basically, that the point of the law is to allow big companies to stop people from accessing media without paying, and the interpretation is stretched as far as it needs to be in order to enforce that.

So I don't know what to believe here.


Yeah if you get realistic about it, courts in the USA simply see it the way of the party with the most money all the time. You can interpret laws until you're blue in the face, but they get to break the rules.

Clearly whenever we discuss laws in the USA as they pertain to a small party versus a giant corporation, we do so in the context of a hypothetical fantasy where these laws are actually followed equally.

It's like everybody's playing hide 'n seek, and you're like "of course a real SWAT team would surround and find you in moments".


> Yeah if you get realistic about it, courts in the USA simply see things in the way of the party with the most money all the time.

This might or might not be true, but even if it’s true it’s not (necessarily) a sign of bias or corruption on the part of the legal system, it might simply show bias (or rather strategic foresight) in what they bring to court.

In the spirt of Von Clausewitz’s On War, it is always easier to attack than to defend, because the defender has to maintain a solid performance throughout every inch of their line, while the attacker need only find a weakness and exploit it.

RIAA might’ve decided months or years ago that as part of their legal strategy they needed to curtail downloading an intact local copy of streamed content, since then they might’ve been seeking the perfect violation that, in the opinion of their lawyers, would show clear enough intent to break copyright law, be a core node in the ‘ecology’ of the downloading-streams strategic landscape, and have a hope of creating suitably broad and useful precedents to employ later. Maybe this case finally caught their eye. Maybe they’d been eying it for a while and only just recently something changed and opened the barn doors for attack (in the RIAA lawyers’ opinion — for example, was the example added to the test cases recently?)

A good question is: why didn’t they do this earlier? Another good question question is: could they do it to anybody else, and if they could, why haven’t they?

So you see it as the judges typically ruling in favour of the corporations. I see it as perhaps the corporations being very savvy in choosing what to attack and being willing to wait for ages for the perfect circumstances.


>I see it as perhaps the corporations being very savvy in choosing what to attack and being willing to wait for ages for the perfect circumstances.

Yes, RIAA has all the savvy of a deranged stalker.


> Yeah if you get realistic about it, courts in the USA simply see it the way of the party with the most money all the time. You can interpret laws until you're blue in the face, but they get to break the rules.

This is a result of money being a proxy for competence.


They may not be a lawyer, but they have basic grasp of copyright law. It is indeed distribution of copyrighted content without permission that is illegal, not obtaining it.


Wait, are you sure about this? All I've ever seen leads me to believe piracy in and of itself is illegal, sans distribution.


Piracy is something done at sea by pirates. Copyright infringement requires distribution, that is the core of the offense.


> Copyright infringement requires distribution, that is the core of the offense.

Infringement does not require distribution. The rights encompassed by copyright include reproduction, derivation, distribution, public performance, public display, and broadcast rights. Infringement of any of those rights is copyright infringement. For US law, see at least 17 USC 106(a) abd 501(a).


The crime of infringement requires distribution.

The tort of copyright infringement is as you describe it.

When people say copyright infringement requires distribution they mean you can't be tried in a criminal court for merely copying you can still be sued.


> The crime of infringement requires distribution.

At least in the US, criminal copyright infringement may be for any act of infringement. The statute 17 USC 506(a) [0] describes the necessary conditions for a criminal prosecution, which are not limited to reproduction or distribution. Here [1] is a nice article from the department of justice about the crime.

[0] https://www.law.cornell.edu/uscode/text/17/506

[1] https://www.justice.gov/archives/jm/criminal-resource-manual...


Courts take a very non literalist approach to technology. Even though YouTube is literally sending copies of the video to users’ computers, because they provide no convenient mechanism for saving the video then courts will tend to ignore that. Instead, it will be argued that a musician’s YouTube channel is acting like a venue and the users are attending an authorized public performance. youtube-dl would then be equivalent to someone recording a concert using a camera.

They might have a very strong case in court. One of the reasons youtube-dl is so active as a project is because they’re in an arms race with Google. Google regularly changes how videos are delivered in an effort to thwart downloading tools. The RIAA could seize on this and even have Google testify that they’re trying to protect against this unauthorized downloading. Given the revelations about the tests in youtube-dl’s own repo, it would be a very difficult legal battle.


In that context youtube-dl would be a camera which most assuredly isn't illegal.


In that context youtube-dl would be a camera which most assuredly isn't illegal.

In the criminal sense, you're correct.

In the civil law sense, you'd be wrong. Using a camera to record a performance is an act of copying and is subject to copyright law. Generally, if the performers and/or the owner of the copyright don't give you permission to record the show, you don't have the right to do so. (Note that taking pictures or recording short clips is usually but not always regarded as fair use. Recording an entire performance without permission is rarely treated as fair use.)

However, recording the performance is also an act of creation, resulting in a derivative work that has its own copyright, but one which is useless without a copyright license to the underlying performance.


I don’t understand your argument. No one is saying that recording a performance isn’t subject to copyright law. They are saying that youtube-dl is the camera, or the camera manufacturer. Canon or Nikon are not subject to copyright law because I use one of their cameras to illegally record a concert, I am. I am the one that did the illegal act and the fact that manufacturers make cameras, that gave me the ability to do it, is not the problem. The argument is youtube-dl shouldn’t be subject either, that it is the individuals using it to download copyrighted material that are breaking the law and a program having the ability to be used illegally is not the fault or responsibility of the program.


Downloading a Youtube video with youtube-dl IS NOT violating copyright in the USA. Downloading videos may violate Youtube's TOS, but that's civil/contract law and not criminal and so the DMCA can't be used to enforce such TOS violations.

Downloading a Youtube video is making a copy. Whether it is a copyright violation depends on whether Youtube's license includes letting visitors make permanent copies, as this is considered a separate type of license. The TOS has nothing to do with it.

A person might share a video they don't own the copyright to, after downloading it from youtube with youtube-dl, but youtube-dl has no functionality in it to do such illegal sharing. It only has the time-shift functionality.

This is false. Also, time-shifting is not a a valid fair-use defense for material available on-demand, because the justification for time-shifting (as in the Betamax case) generally no longer exists. There are other fair use arguments that could be made, but time-shifting isn't one of them.

They had to cite a German legal ruling which found YouTube’s “rolling cipher” was an anti-circumvention device. But German court rulings have no bearing on US law, and Germany doesn't have a DMCA statute, so it is a specious claim, and the RIAA is just hoping an open source project won't be able to fund a legal defense.

I don't know if you're joking or not, but they cited a variety of US cases. Kazaa, Limewire, Napster, etc, are all US cases that all came down the same way: a tool intended primarily for copyright violations violates the DMCA, and in certain cases can even expose its developers to statutory damages under the copyright code.

It's Betamax time-shifting plus macrovision/video stabilizers all over again.

Nope, not even close for so many reasons already explained in this and other comments.


> Downloading a Youtube video is making a copy. Whether it is a copyright violation depends on whether Youtube's license includes letting visitors make permanent copies, as this is considered a separate type of license. The TOS has nothing to do with it.

YouTube is not the copyright owner (in almost all cases). Downloading a video with youtube-dl may not be a copyright violation if you already separately hold a suitable license, or are in fact the copyright owner yourself (as in the case of one of the grandparent comments).


Yeah, I should have worded that better. I was referring to Youtube's license with the RIAA, since Youtube doesn't provide a copyright license to its users for content licensed from third parties.

If the license between the RIAA (or other licensor) and Youtube allowed visitors to make permanent copies of Youtube streams, then downloading a Youtube video would not be a copyright violation. But generally, having entered into streaming and downloading licensing agreements with the RIAA before, I would be extremely surprised if the RIAA would enter into such a license with Youtube, as it would kill sales of digital music downloads if people could just download permanent copies of Youtube videos instead for free.

If you own the copyright and upload a video to Youtube, then it wouldn't be a copyright violation to download that video from Youtube.


[flagged]


Really? Why don't you enlighten us instead of making ad hominem attacks...


I wouldn't be surprised if there's a legal violation when you're downloading a copy from YouTube even though you're the original copyright owner of the video. Youtube has a license to distribute the copy according to their TOS. That license will be subjected to specific terms (streaming + DRM protection). It might not include allowing creators to circumvent those terms for archival purposes.


"legal violation" is pretty vague. You might be disobeying their ToS, but it's not copyright infringement, since you own the copyright.


It's very possible to have a TOS violation but not a copyright violation if you're the copyright owner of the video being downloaded.

Owning the copyright means you don't violate the copyright in making a copy, but that doesn't mean you get to make that copy through any means; you still have to use legal means, and for a website, that means methods that are acceptable under the TOS of that website.


Technically, even streaming is making a copy.


I haven't seen anyone else making this argument yet, but yes! Chunk by chunk, streamers copy bits into their buffers and discard them as they are done watching. Is that discarding required, when the same companies can charge extra for additional bandwidth usage (not for making additional copies, but for clogging up their pipes)

I'm not sure if this is going to be considered similar to taking a full copy but I'm also not sure what substantial purpose there would be for taking that full copy except for the (probably valid fair use) purpose of shifting access to manage time, location, or network conditions. Given that you pay separately for your bandwidth, and there is an incentive to conserve it, I don't see any way this isn't considered fair use without a sharing component.

Taking a copy is incidental to lawfully consuming the content that was provided in an authorized setting, whether that copy is stable or not. Nobody is hosting copies of YouTube content, paying the bandwidth and hosting bills so anonymous co-parties to the infringement can enjoy the content separately from the authorized channel, (it is not even suggested that anyone would want to do this.)


Legally, even running a program is making a copy (from disk to RAM): https://en.wikipedia.org/wiki/MAI_Systems_Corp._v._Peak_Comp....

I think the question is whether the license YouTube implicitly provides to copy so one can stream also extends to downloading to a file. Probably not.

Why that should be youtube-dl's fault is yet another question, of course.


> Legally, even running a program is making a copy (from disk to RAM): https://en.wikipedia.org/wiki/MAI_Systems_Corp._v._Peak_Comp....

> I think the question is whether the license YouTube implicitly provides to copy so one can stream also extends to downloading to a file. Probably not.

> Why that should be youtube-dl's fault is yet another question, of course.

TV broadcasters do not give you a license to copy their content but it's legal to record TV. Why is youtube any different?


It was legal to record TV, using VCRs and DVRs, for time-shifted viewing, due to the alternative being that the content could not be consumed at all (at the time of the Betamax case, episodes were broadcast once, and generally not re-broadcast unless they were extremely popular, and were not available for purchase individually or even as part of a collection).

Today, there are a multitude of options for consuming TV content, and due to the (recent) creation of time-shifted licenses, the fair use justification for standalone-DVRs has mostly gone out the window (which is why they're not for sale anymore; for example, Tivo no longer offers standalone DVRs; Hulu, cable companies pay for time-shifted viewing).

As Youtube was always on-demand to begin with, the time-shifting justification for recording TV in the Betamax case never applied.


Are you saying that I can't still legally record TV on my VCR or DVR?


Legally, even running a program is making a copy (from disk to RAM):

Yes, that is why 17 USC § 117 provides that making a transitory copy necessary to run the program (such as a RAM copy), or an archival copy of the data needed to run the program, are both explicitly not violations of copyright.

I think the question is whether the license YouTube implicitly provides to copy so one can stream also extends to downloading to a file. Probably not.

It does not. Youtube does not explicitly, or implicitly, provide the user a license of any type. Youtube has licensed the right to provide a temporary copy of a video to a user.

Why that should be youtube-dl's fault is yet another question, of course.

It normally wouldn't be...except that some idiot decided to include unit tests to make sure that youtube-dl could be used to download music videos covered by RIAA licenses. That basically makes the RIAA's case, especially since youtube-dl must be regularly updated to handle Google's countermeasures.


Right, but 17 USC§117 only applies to programs, not video. Who provides me the license to copy the video from the buffer of the network card to main RAM, and then to GPU ram, and then to my screen?


Technically, even streaming is making a copy.

Yes, that is why streamed content is now generally covered by streaming-specific licenses, as noted above.


iTunes was designed [in part, probably a principal part of the original design] for CD ripping, I gather (I haven't personally used it more than twice, many years ago).

Format shifting was at the time not allowed in UK copyright law, and again is not allowed as of a couple of years ago (it was only allowed for about a year).

iTunes was designed for copyright violation. I'm sure the manual will show that.

Just for comparison.

It's my understanding that any media _can_ fall under Fair Use in the USA, meaning that there would have to be specification as to who was using the software and their circumstances within the unit tests in order to make it infringing use. You might fir example be archiving, or preparing educational materials, or compiling a review, or doing subtitling.

It's not sufficient to prove someone downloaded, you have to show it was infringing activity. How do you do that with a unit test? Software to perform legitimate, legal downloads would still have to do that same test.

In short you've found a smoking gun, now prove its principle purpose is murder.

This comment is provided solely for entertainment and in no way relates to my employment.


Tunes was designed [in part, probably a principal part of the original design] for CD ripping, I gather (I haven't personally used it more than twice, many years ago).... iTunes was designed for copyright violation. I'm sure the manual will show that.

No, Itunes was designed for organizing music. How you got that music was a separate function,and you could split out the CD ripping functionality and still have the primary function of iTunes. Additionally, making copies of physical media for personal archival purposes is recognized as a legitimate fair use (and was back then), so the the CD ripping functionality would have generally still have been okay. And yes, there is case law on when ripping CDs is considered fair use and when it is not.

It's my understanding that any media _can_ fall under Fair Use in the USA, meaning that there would have to be specification as to who was using the software and their circumstances within the unit tests in order to make it infringing use. You might fir example be archiving, or preparing educational materials, or compiling a review, or doing subtitling.

Yes, any media can be subject to a fair use defense. The second statement is not relevant to the copyright analysis. The last statement are examples of fair use, but note that acquiring the content for the purposes of preparing educational materials is not automatically "fair use". For example, preparing educational materials for a class provided at a for-profit institution, or in educational materials sold commercially, would not support a fair use defense, though otherwise the use of copyright content in educational materials has generally been deemed fair use.

It's not sufficient to prove someone downloaded, you have to show it was infringing activity. How do you do that with a unit test? Software to perform legitimate, legal downloads would still have to do that same test.

The unit-test is a part of a different analysis from a basic copyright infringement claim; That question is whether the tool is intended for violating copyright and for that, they don't need to show any infringing activity actually took place, only that the tool was intended for infringing activity. The DMCA does not require a creator to suffer harm before they sue, since it recognizes that in the computer age, once such harm occurs it may not be possible to put the genie back in the bottle.


No, iTunes was designed for organizing music. How you got that music was a separate function, and you could split out the CD ripping functionality and still have the primary function of iTunes.

Well, being able to rip music off your CDs was literally advertised as an iTunes feature; Apple promoted iTunes after its introduction in 2001 with the slogan "Rip, Mix, Burn." And Apple did get accused by various industry players of promoting copyright violation. I wouldn't say that iTunes was "designed for copyright violation" any more than I'd say youtube-dl was "designed" for it, but Apple definitely understood that iTunes had value to people who wanted to make digital copies of media in ways that did not strictly fall under fair use guidelines. :)


but Apple definitely understood that iTunes had value to people who wanted to make digital copies of media in ways that did not strictly fall under fair use guidelines.

Ripping CDs for personal archival copies had already been characterized as fair use by the courts by the time iTunes was first released. If it hadn't been, the industry players wouldn't have just "accused" Apple of promoting copyright violation, they would have done something about it. The music industry has always been very aggressive about protecting their IP rights.

It would have been very different if iTunes had the slogan "Rip, Mix, Burn, Share."


The kicker is that if their demo examples had been something like that, or e.g. students downloading copies of free lectures, or etc, the RIAA would have much less of a case. No actual code would have had to change.

If you have a tool which skirts (bad, overreaching, probably constitutionally invalid) laws, be smart about how you present them!


If this content is freely available, why can't we take a fair use copy? That is not a violation of copyright. (Even if it is entertainment / not educational.) Those are the terms of the deal, we made when we enshrined copyright into law. If you can show it to me, then I am allowed to take a copy. (Is it not? I thought that DMCA law itself was written to skirt around this fact of law.)


If this content is freely available, why can't we take a fair use copy? That is not a violation of copyright.

Making a copy on its own is not fair use. What matters is the reason for making the copy. If you are making an connectivity-shifted copy (aka, for viewing away from internet access) there could be a fair use argument to download content from a streaming site that is not otherwise available for offline consumption.

But the flipside of that: if there is a way to access the content in a way that fair use is not implicated (i.e., because the copyright owner makes the content available for that use case such as offline viewing, even if payment is required), then generally fair use does not apply.

Remember: fair use is a defense to violating copyright, and it must be reasonable under the circumstances. Not wanting to pay for a license for your intended use case is generally never considered reasonable.

If you can show it to me, then I am allowed to take a copy.

That has never been the law, at any point in the history of copyright, dating back to before copying works was trivial.


In Sony Corp. of America v. Universal City Studios, Inc. the majority opinion stated:

> [There must be] a balance between a copyright holder's legitimate demand for effective - not merely symbolic - protection of the statutory monopoly, and the rights of others freely to engage in substantially unrelated areas of commerce. Accordingly, the sale of copying equipment, like the sale of other articles of commerce, does not constitute contributory infringement if the product is widely used for legitimate, unobjectionable purposes. Indeed, it need merely be capable of substantial noninfringing uses.

Is there a subsequent opinion that reversed that? (How is youtube-dl substantially different from a VCR recording, if youtube is substantially similar to a broadcast medium?)

> [W]hen one considers the nature of a televised copyrighted audiovisual work ... and that time-shifting merely enables a viewer to see such a work which he had been invited to witness in its entirety free of charge, the fact ... that the entire work is reproduced ... does not have its ordinary effect of militating against a finding of fair use.


To play devil's advocate, I wonder if it can really be said that when you watch a youtube video you have been

> invited to witness in its entirety free of charge

when you consider advertising. It could be argued that when you watch a monetized youtube video you are being invited to view the video in exchange for also viewing advertisements as a form of payment.

I hope that you're right though, and that we get a ruling in favour of being able to make copies for fair-use.


> To play devil's advocate, I wonder if it can really be said that when you watch a youtube video you have been

>> invited to witness in its entirety free of charge

> when you consider advertising. It could be argued that when you watch a monetized youtube video you are being invited to view the video in exchange for also viewing advertisements as a form of payment.

The problem with this idea is that you're quoting a case about recording televised broadcasts. The quote you pulled your quote from begins "when one considers the nature of a televised copyrighted audiovisual work".

Televised works were shown with embedded advertising a lot more intrusive than what can appear on a YouTube video. There is no way for the presence of advertisements to affect how this argument applies to YouTube, when it was originally cast in terms of television.


Gets complicated, but watching ads to watch a video is a separate issue from the underlying copyright for the content.

The content holder licensed the content to Youtube (or Vimeo, etc), but does not necessarily control the monetization of that content by the licensee. (Generally, the license will say something like royalties shall be X% of revenue derived from the content, without limiting itself to specific methods of monetization.)

Youtube doesn't have a license to give to the viewer; the license they have is merely to show you the content.

I hope that you're right though, and that we get a ruling in favour of being able to make copies for fair-use.

This has been the law for several decades already. But the key thing to understand is "fair use" doesn't mean "to avoid paying." If digital content is available for purchase, the courts generally have not found making a permanent copy from a streaming source to be fair use. (Where it gets complicated: multiple copyrights applying to a single work, such as the dancing baby Prince video. The creator of the video was okay with people viewing it and downloading it freely, but Prince was not. In the music arena, this would be a non-issue due to compulsory licenses, but those licensing schemes don't exist for other types of media. Ultimately, the court said that the focus of the video was the toddler dancing, and that the music was incidental, so including it in the video was a legitimate fair use. But it took more than a decade for the case to be resolved.)


For sure, that is the argument they will have to make! As a counter-argument, what if I paid for my youtube premium membership though, to avoid the advertising, and really what I want to do is to location-shift within the bounds of the law, I want to take my laptop into the wilderness and watch the movies I paid for, on an Indian reservation in front of a mountain scene, ... where there is no broadband or cell radio tower service?

(Hypothetically of course! Youtube-dl is not only for Youtube.)


That's multiple different licensing scenarios. Paying for Youtube Premium doesn't give you a right to download whatever videos you want, since that right has to be granted by the copyright owners of the videos.

If you purchased a movie, presumably this included the rights to watch it offline (and most of the movie lockers like Vudu, etc, generally include such rights), then the selling website generally makes available copies for offline viewing (and many Blurays include codes to activate on movie locker service to get a license for a digital copy). In such case, you would need to use that service for downloading the offline copy, because the Youtube copy isn't part of that license to you.


There are fair use rights that are not able to be satisfied without a copy.

You can assert all of that, but unless you have some landmark precedent to cite that agrees with what you're arguing, I'd like to hear a court decide on that. Sony Betamax says what I'm arguing, and archival can be for fair use, too. There is 36 Cinema, which maintains copies of classic kung-fu movies that aren't available for purchase anymore, and invites Rza from Wu-Tang Clan on periodically for broadcast viewing with a value-add, Rza's commentary. Since the movies are no longer available for sale, there is no impact to marketability. Since commentary is added, transformative. That's all fair use, format-shifting. I say youtube-dl is a tool for format shifting, with substantial non-infringing uses that present a compelling value and don't impact marketability, exactly in line with the Betamax case. If format-shifting and copy for archival was strictly prohibited as you suggest, those movies could be lost forever whenever a format becomes obsolete. When was the last time you saw a VCR? (CD or DVD player?)

That's just one of the things that the fair-use rules and exemptions in copyright were written to help resolve positively.

You can record anything on TV with a VCR, even if it's available separately for purchase on video. There is no technical way I can see in which this scenario differs. You are not obligated to pay for content again and again in every medium, format-shifting is also an allowable kind of fair use. Everything you're saying makes sense, but I don't think it's as clearcut as you say, until it has been decided by a court (and appealed, and decided again by a higher court.)


You're literally just arguing against how the law works in the US. I'm not going to argue against the fictional legal system you've set up in your head.

The 36 Cinema stuff resulted in a new copyrighted derivative work: the commentary. And they can distribute that commentary all they want...but to distribute the underlying films as part of a commercial offering they need copyright licenses to those films. It's irrelevant that the underlying films are not for sale at retail, since they can still acquire a copyright from the copyright owner.

Format-shifting isn't fair use. It's literally an act of copying that is subject to copyright protections. Format-shifting might be fair use, but you have to do the full analysis, and format-shifting to avoid paying for a copy in the destination format is not fair use.

That's just one of the things that the fair-use rules and exemptions in copyright were written to help resolve positively.

Now you're just making stuff up. 17 USC 107 sets forth the basic rules for fair use, and they're quite limited.

When was the last time you saw a VCR?

Last week. Best Buy and Walmart still sell VCRs...Millions of people in this country still have old TVs with purely analog connections.

You are not obligated to pay for content again and again in every medium, format-shifting is also an allowable kind of fair use.

Actually, yes, you are, since each offering of the content in a different medium is a different copyrightable work subject to its own copyright. A VHS copy of a film is very different from the Bluray copy of the same film.


> but to distribute the underlying films as part of a commercial offering they need copyright licenses to those films

They sell tickets to these events. It's still fair use, (and it would not be possible to have these events without someone first acquiring a digital copy for archival.) You're arguing that they could broadcast the video with commentary but they can't make (or take) a copy? This is prerequisite to the activity! And many of these publishers may no longer exist, as not everyone agrees these movies are classics.

Also section 117 of Copyright Act explicitly calls out the making of a copy for archival as an allowable exception.


> It could be argued that when you watch a monetized youtube video ...

Not sure what one of those looks like. I hear people talk about them, but I've not seem an advert on a YouTube video.

That being said, I do use a bunch of anti-ad browser extensions... :)


Wherein did I sign such an agreement and wouldn't that also forbid getting up to use the bathroom during the commercial break?


Wouldn't adblockers violate this just as much as youtube-dl does, then?


That's the Betamax case. Congress generally overruled the Betamax case with the DMCA, which included specific provisions about copying tools.


Copyright included fair use and the courts ruled that a balance for fair use must be maintained for the monopoly rights of copyright to be enforced, but congress wrote the DMCA and said "nah"?

Yeah I'm gonna need to hear that from the courts, my assertion here is that the DMCA unfairly skirts around this provision of copyright, not that it superceded it outright.

Those fair use rights are not an optional provision of copyright, (although many opinions have stated that copyright provides the author may refuse to allow copies if they technically can, unless you have relevant case law that superseded Sony Betamax, in that case I believe it was the minority opinion), I am not aware of any decisions that say anything other than "copies for fair use may be required for there to be fair use."

But I am not a lawyer, and just because I haven't heard of the case doesn't mean it didn't happen...


Copyright included fair use and the courts ruled that a balance for fair use must be maintained for the monopoly rights of copyright to be enforced, but congress wrote the DMCA and said "nah"?

The Constitution gave Congress the power to right the laws on copyrights, not the courts. The courts don't write copyright law, they just interpret it.

Fair use, for example, is explicitly written into the copyright statutes by Congress (17 USC § 107), but with sufficient breadth and ambiguity in the language that the courts have added in uses that were not explicitly written into the law but could be reasonably read as being within the scope of the text of the law. That is indeed why the court in the Betamax case was able to characterize "time shifting" as fair use in the first place.

Yeah I'm gonna need to hear that from the courts, my assertion here is that the DMCA unfairly skirts around this provision of copyright, not that it superceded it outright.

Again, Congress gets to write copyright law... They get to change it how they want, so long as it does not violate the Constitution. (Note that the DMCA is an "act" of Congress, which is the method by which Congress actually writes and changes the laws of the US. When we refer to the DMCA, we are referring to the changes to US copyright law embodied by the DMCA.) Courts have upheld the constitutionality of the DMCA.

Those fair use rights are not an optional provision of copyright,

Yes, they are. The Constitution makes no provision for "fair use." In fact, by its bare text, it would appear that the Constitution would not support fair use because it provides for "exclusive rights" to go to creators and inventors (for the periods covered by copyright and patent). "Fair use" is entirely a legislative creation of Congress, and theoretically Congress could take it away.

unless you have relevant case law that superseded Sony Betamax

I don't know why you're hung up on Betamax. That case was based on the copyright law at the time of the case, i.e., the 1980s. Congress changed the copyright laws after that case (in the DMCA and other legislative acts), so Betamax is no longer relevant except as persuasive authority, and the underlying facts supporting the time-shifting as fair use ruling generally do not apply to digital content available on-demand.


> "Fair use" is entirely a legislative creation of Congress, and theoretically Congress could take it away.

If you have a court decision that shows a court interpreted Congress' actions as revoking fair use by the enactment of DMCA, then I'll concede the point, (but you won't have one as the DMCA does not explicitly revoke fair use.) DMCA laid out protections for copyright owners who sought to protect their copyright with eg. DRM schemes.

That doesn't revoke fair use. It just makes it practically difficult to utilize, since you might have to (illegally) circumvent a copyright protection device in order to access those fair use rights. They are still there, copyright owners just have a few more tools in their toolbox to prevent you from accessing them lawfully.

Is Youtube's "rolling cipher" such a mechanism? Debatable. Is the proper remedy a DMCA takedown of the entire youtube-dl source, or something else? That's all something for a court to decide. Only certain uses of youtube-dl are potentially foiling anti-circumvention devices like "rolling cipher", it's a utility that works on many video streaming sites, (and substantially many of the works on those sites are not protected by "rolling cipher" or similar, possibly any, DRM.)

> I don't know why you're hung up on Betamax.

Has there been a landmark ruling since, that reversed Betamax? It was decided by the Supreme Court, so unless you have one, I don't think I can agree that it is no longer relevant.


> But the flipside of that: if there is a way to access the content in a way that fair use is not implicated (i.e., because the copyright owner makes the content available for that use case such as offline viewing, even if payment is required), then generally fair use does not apply.

Can you cite a source for this? I do not believe that is generally the case. You are allowed to rip CDs even if the same music is sold digitally and you are allowed to DVR a TV show even if they sell DVDs of the same show.


You are allowed to rip CDs even if the same music is sold digitally and you are allowed to DVR a TV show even if they sell DVDs of the same show.

Fair use is a balancing act based on analysis of various factors. In the case of ripping CDs for personal archival purposes, the courts treated that as fair use because at the time there was otherwise no way to make archival copies. It's very possible that today a court could rule that making archival copies of CDs for personal use is no longer fair use, because the digital version of the music now exists, as a separate article from the CD, and can be legally acquired. (Note that libraries and other archives still have a statutory archival use exception. Also note that because a CD is a physical good, there are certain rights associated with it that would not apply to content acquired digitally.)

For TV shows, it's not the same thing, since a DVD of an entire season is not the same thing as being able to view just a single episode. That being said, with the rise of digital, on demand availability of individual episodes, the original fair use justification for VCRs and DVRs has basically gone out the window.

The original Tivo could not exist today, and indeed...it does not: Tivo no longer offers standalone DVRs. Hulu, Youtube TV, your cable DVR, all of those services license time-shifted viewing rights from the copyright owners. (Yes, the studios created a new type of right just for this...)


Do you have an actual precedent to cite or are you just speculating?


>an entire season is not the same thing as being able to view just a single episode.

Even from your own archive?

Originally many people started out with plain paper terminals rather than the spiffy video terminals that would later become common.

Well, sprocket-feed terminals more precisely, where you sit at a floor-standing wide-carriage _printer_ having a full QWERTY/ASCII keyboard, connected to the mainframe using a RS-232 serial COM port cable.

Or remotely dial-up over regular phone lines using external modems compatible with the kind first used for internet dial-up. Modems later found internally as standard equipment on PC's, made to accept a common RJ-11 telephone connector, eliminating the need for the RS-232 cable.

Not unlike a space-age teletype.

Either way, you type to the computer and it types back to you.

When available, as we all know it's been a while and computers still can never be expected to have 100 percent uptime, so time shifting has always been the norm in some way or another. At this point with a dumb terminal you just come back later when the mainframe is not too busy for input, or for output just wait for the printout until it's good and ready.

More effective session management would have to be accomodated by storage of some kind not unlike the punched paper tapes for sending and receciving on some teletypes.

Underneath the terminal you have a big box of the fan-feed computer paper so you fundamentally get an endless record of the communication in its entirety. Otherwise there is nothing. This is the default. Out-of-paper meant no communication and no data.

Pallets and pallets of boxes and boxes of printouts, excessive amounts of trees giving their lives and paper mill pollution up the wazoo (you ever smell that stuff?). But it's worth it, these are not copies, this is the original data, as you received it coming in live over the wire.

Depending on the institution or individual, and the risk of losing this unique output, archival handling procedures may apply.

Upgrading to a VT-102 type video terminal is actually analogous to a desktop PC when it comes to form factor, but the command line is still not from a local processor, and no local disk storage.

Naturally you still get your continuous printout as the screen display scrolls it on by, now possible from a plain serial printer (or from the same old paper terminal) connected to the second COM port on the VT-102 for pass-through printing.

You still use the terminal to operate the command line & display the output from a remote CPU, and with scrolling ability, can roll back to redisplay some recent earlier content. This was not a copy either, it was the same original live data as printed, just redisplayed. Not every terminal had that kind of memory though, and if present, not much.

Then you get _intelligent_ terminals with lots more memory plus local floppy storage, having more than two COM ports, and a simple local OS in ROM to handle these peripherals.

It finally became posible to judiciously save paper like never before, from that point on there has always been a local SAVE command of some kind. That's one of the only main purposes of any mainstream desktop workstation ever since.

You end up with a stack of floppies instead of paper containing the original data from that hardware session.

Interestingly, even today it is sometimes still faster to look something up in your paper records than find it on a disk though.

Anyway the disks simply have digital representation specifically crafted for the storage medium, often in appropriately treated text files.

From that point you could always play back data from a disk to your console screen, and/or one or more of the COM ports which may be connected to other terminals, computers, storage, or printing hardware. When the time is right.

The purpose of putting data into a computer file format to begin with is precisely so this type of communication can be achieved electronically.

And also, so the file or disk can be archived using one of the only non-communication commands in such a pre-DOS desktop environment, the ominous-sounding COPY command.

Well, it's electronic and you need to keep the original completely intact, so no telling how many times you need to run the COPY command to get a reliable archive.

But when one of the things that came across your wires is something like the text of The Godfather, that's an author's original work and it would be most questionable to manufacture copies for people as if it were your own creation, perhaps unfair even if proper attribution is given.

Each user has always been responsible for appropriately respecting the rights owners of any material in their archive, and questioning their own operation enough to avoid things like selling duplicates to just anyone. It's part of the responsibility of having such powerful equipment at their disposal.

Not just for those with dual floppy drives. Takes lots of floppy disks to boot sometimes too. Twenty years ago most PC's already came equipped with two CD drives in addition to a floppy drive, one CD drive was writable for archiving since floppies were not too popular any more.

MP3 sites earlier than Napster arose without question as experimenters compressed and shared their archives using the emerging codec. This was nothing like selling pirated copies of CD's, and was not made available to the general public by any means, just the few technology enthusiasts who had PC's with internet capability.

With progress in hardware and internet performance, there should be more advantage for archival use as time goes by.

Now exponentially more good material is at risk of being lost due to electronic or legislative failures, beyond the pre-existing threats from ordinary disasters like floods or acts of war. There's also the limitations of things like MP3 compression.

There should continuously be improving apps so users can work more effectively than ever with the desired portions out of their own complete digital history.

Meanwhile there is some consensus that operators without your permission are clandestinely eavesdropping to make truly illegal copies of your original complete digital history (along with many others') and extract portions that they my find desirable, sometimes against your own best interest.

You surely need to be entitled to better than this, and in favor of your own preferences instead.

In the USA according to copyright law archiving is not supposed to be impossible.

Nobody has a continuous paper record to fall back on any more.

Any threat to the usefulness of your digital equivalent would seem to be a form of cyberattack now that we have a better idea of how a cyberattack can do damage.

It's always been the same tool with possible uses for good or evil, the desktop workstation.

What you primarily do with it can only be determined by examination of its complete digital history.

From an entertainment law perspective it would seem like the only sensible violation could be if illegal entertainment had actually ocurred, not whether the file was being stored locally or remotely.

If most of the files just sit there almost all the time, with almost no dancing to the computer or anything like that, that's an archival device not an entertainment device.

Even more so when the primary user handling of the files is to maintain them intact against all odds, without ever intentionally using the file contents to achieve the files' own particular underlying purpose, whether entertainment files or not.


By that logic one would only be allowed to record with a vcr shows that weren't available on tape. Furthermore according to parents logic VCRs shouldn't be allowed to exist because they can record from both categories.


It depends. If the shows are available with each episode available on a separate tape (and each separate tape could be individually purchased), then yes, the VCR fair use justification evaporates. But generally VHS tapes include multiple episodes and individual episodes can not be individually purchased (other than a limited number of very special episodes), so there is still an argument to be made for fair use on the basis of each individual episode.

Furthermore according to parents logic VCRs shouldn't be allowed to exist because they can record from both categories.

No, that's not at all what I said. VCRs are just tools. Tools are subject to a different analysis post-DMCA: does the tool have a substantial non-infringing use or is it deliberately designed to violate copyright?

In determining whether a tool was deliberately designed to violate copyright, they look beyond just the mere function of the tool and examine why that functionality is present, and how the tool and that potentially-violating functionality is marketed.

On that note: DVRs generally no longer exist today as standalone goods (see, for example Tivo, etc). This is because the copyright owners introduced new time-shifting licenses a few years ago, and your cable company, Hulu, etc. pay the copyright owners for the right to let their viewers view content on a time-shifted basis. A standalone DVR would generally have the primary purpose of violating those (relatively new) rights, and thus wouldn't pass muster today.

What about VCRs? They're still okay. They make degraded, low-quality copies of broadcast transmissions for archival/time-shifted uses by people who have TVs that still connect to VCRs. As those people generally wouldn't be able to access the equivalent digital content on their TV, it's clear that the primary use of VCRs is for fair uses purposes.


Do you want to provide some cites for this line of reasoning.


I think gamblor is conflating the commercial publisher's responsibility to pay for each format commercially offered separately (they can only make commercial distributions that are appropriately licensed), with the consumer's fair use right to create a backup copy of their legally obtained copy of any copyrighted media, for backup or archival.

You absolutely do not have to buy the MP3s rather than rip the CD that you own to MP3, just because they also sell MP3s and you want your archival copy to be in MP3 format. You can make them as a backup copy. (If there's no anti-circumvention device like the famous CSS encryption in your way, that is.) The publisher cannot pay once for CDs and also sell MP3s, they need a separate license for that (if that is how the author's licensing is written, granted, that's a fact.) The consumer is not bound in this way, they "paid" for their copy (presumably, if payment was needed to receive it) and they can format shift if their use passes the 4 factors balance test for fair use, (and if it is technically possible to do so, eg. without bypassing an anti-circumvention device, (thanks DMCA.))

Moreover, we are talking about youtube-dl, which is not owned by the RIAA and they have no right to take it down in this way. They can seek relief in the form of an injunction, the scope of the injunction to be determined by the courts, who would have to consider the substantial non-infringing uses of the tool; they would be unlikely to decide that vaporizing youtube-dl from orbit is the appropriate remedy.


Why can't we take a fair use copy? That is not a violation of copyright.

That's not covered under "Fair Use": see [1] for fair-use criteria. It is definitely a violation of copyright, yea.

[1] http://fairusetube.org/guide-to-youtube-removals/3-deciding-...


Taking a copy does not affect the potential market of the copyrighted work in any way. Sharing a copy is another matter. Youtube-dl facilitates taking a copy only, it is not a tool for sharing. This case is substantially different than Napster, I disagree. In some circumstances, you must be able to take a copy in order to exercise your allowed "transformative" fair use rights. You cannot sample a record without a way to take a copy.


The article you cited is about fair use when distributing a video. It does not address making a copy for personal use.


A unit test seems like fair use.

I don't know where tranformation would land as they are just streaming data and not actually viewing it.

It's (potentially) the whole content, but a case that was literally just uploading an entire video as-is with a different title was ruled fair use not too long ago, so this isn't itself damning.

The nature of the work seems irrelevant here as it's not viewed. It could easily be replaced with any other video -- even one owned by the repo creators.

Likewise, since they are unit testing rather than viewing the content, I doubt there is any commercial harm and the market overlap between video viewers and unit testers is zero.


No, the unit test isn't fair use, because it wasn't necessary to violate the license holder's copyrights in order to test the code.

It's (potentially) the whole content, but a case that was literally just uploading an entire video as-is with a different title was ruled fair use not too long ago, so this isn't itself damning.

I'd like to see a cite for that. Generally, uploading whole content owned by someone else is not fair use absent something like the content being newsworthy in itself (such as video of a police beating). Otherwise, fair use would generally require some transformative effort.

Likewise, since they are unit testing rather than viewing the content, I doubt there is any commercial harm and the market overlap between video viewers and unit testers is zero.

Copyright violations don't require the copyright holder to prove commercial harm. That's the entire point of statutory damages.


"Not necessary" doesn't mean it's not still fair use. Lots of things aren't necessary but one definition or another. The argument here is that it's lack of necessity makes it incidental.

Here's the Carl Benjamin case covered in bloomberg (you can look up the specific case if you'd like) https://news.bloomberglaw.com/ip-law/youtuber-wins-attorneys...

> Copyright violations don't require the copyright holder to prove commercial harm. That's the entire point of statutory damages.

This is explicitly one of the four points of fair use as stated by the US government, so it's very important.

https://www.copyright.gov/fair-use/more-info.html


> And that's what the README showed: that youtube-dl was built primarily to violate copyright.

That's not what "primarily" means.


>it doesn't matter if there are also legitimate fair use uses of the tool.

What are you basing your legal theory on?


The text of the DMCA and cases based on it.

Once the infringing use has been established, it doesn't matter that there is also the potential for legitimate fair use. If it did, then Napster, Kazaa, and Limewire would still be around because they all had the potential for legitimate fair use.


Not having read any legal literature around or really understanding the DMCA, how can they not make a distinction between things with legitimate uses and not? How do torrent clients stand up to that?


Torrent clients don’t name themselves “Game of Torrents” and have test cases for downloading TV shows.

The problem YouTube-dl ran into is they literally named themselves after a prominent streaming site that is copyright holder friendly and also literally tested, in their codebase, the ability to do a thing that arguably violates the DMCA.


Not sure the "literally" things are anything to do with this. Only the "arguably violates DMCA" is relevant, and we already knew it was arguable.


The "literally" parts help the RIAA make their case.


Sounds like Google are the ones infringing, not youtube-dl.


why does that make it worse?


I don't think it does. Just because there are copyrighted works in the test cases, doesn't mean they are willfully in violation of copyright. If I am granted a license by Youtube to view this material, then is it not also fair use if I take a copy for later review, or otherwise maintaining the fair use rights innately afforded to me by copyright?

What other facilities could I use to exercise my fair use rights in this case, besides tools like youtube-dl? If the video is already available freely for viewing (with advertising), are there any greater damages that aggrieved parties could calculate than the loss of advertising revenue from my repeated viewing of that file?

What incentive is there for me to share this file in violation of copyright if anyone can download a fair use copy for themselves, with the tools provided by youtube-dl?


It makes it worse because otherwise they could have argued that the README was separate from the development of the tool itself, so the tool itself wasn't intended for violating copyright.

By putting the infringing act into the code of the tool itself, they lose the ability to make that argument, and moreover show that the tool was built to infringe copyrights, because clearly they thought it was important enough to be able to infringe copyrights that they created multiple unit tests for it.

are there any greater damages that aggrieved parties could calculate than the loss of advertising revenue from my repeated viewing of that file?

The loss of advertising revenue is largely irrelevant unless the RIAA chooses to demonstrate the amount. Copyright law provides statutory damages starting at $750, and for that they just need to show a single infringing act (and for each additional act they prove, they get another $750).

What incentive is there for me to share this file in violation of copyright if anyone can download a fair use copy for themselves, with the tools provided by youtube-dl?

You're assuming a fair use case for using youtube-dl to download a copy of a licensed music video. But fair use is defense to a claim of violating copyright, so you need to show why your use is a fair use, you can't just say "fair use" as if it were a magic spell that makes your legal problems go away. And generally, it would be difficult to prove fair use for any of the content covered by an RIAA license because they make that content available for use offline, time-shifted, etc...they simply require that you pay for that.


The mention of the copyrighted work eg. "Taylor Swift" does not lay out a case of infringement. Merely possessing a copy does not indicate infringement, you could be in pursuit of your fair use rights, with no intention to share. Given the cost of bandwidth and predilection of media conglomerates to meter and limit it, I would argue that there is actually no likelihood of facilitating any substantially infringing use.

The person who took a copy has no substantial reason to share it with anyone, as it is freely available through the content providers. Anyone with a copy of youtube-dl can get it, (does youtube-dl enable the download of any paywalled content? If so, that might be a fact that changes the game. But if so, does the RIAA claimant have standing to make that argument?)


Merely possessing a copy does not indicate infringement, you could be in pursuit of your fair use rights, with no intention to share.

Right, but you're looking at the wrong thing. The use of youtube-dl by you to download Taylor Swift is not what the courts are looking at (and note, that would generally not be fair use if your intent was just to watch it later as the time-shifting defense from the Betamax case is generally not applicable to content available on-demand, though using the video in a derivative work like non-profit educational content would probably still be fine).

It's the fact that youtube-dl holds itself out as being the tool to use so that you can download Taylor Swift that is at issue.

The person who took a copy has no substantial reason to share it with anyone, as it is freely available through the content providers. Anyone with a copy of youtube-dl can get it, (does youtube-dl enable the download of any paywalled content? If so, that might be a fact that changes the game. But if so, does the RIAA claimant have standing to make that argument?)

It doesn't matter if they intend to share the video further. The copyright violation act is the downloading of a permanent copy of video content provided on a streaming/on-demand basis. The downloader sharing the video with others would be a separate copyright violation.


This case is going to be really interesting if it goes to court, but right now it's not in court.

The letter makes two claims, "copyright infringement" which might be substantiated by the readme, (but not against youtube-dl authors, perhaps against their users) and "anticircumvention [sic]" which is the meat of their real issue.

The making of a copy is not outside of fair use unless it fails the balance test, aka "four factors."

Sony Betamax disagrees with you, if we can agree that youtube is similar to a broadcast medium and that youtube-dl is similar to a "VTR" from that case, aka VCR, unless there are other substantial differences that I'm missing.

Making a copy is only infringing if it isn't for fair use, and Betamax ruled that time shifting could be fair use.

One of the four factors is market impact, so how is the market for this content impacted by this taking a permanent copy? It would be impacted if the purpose and character of the copy was inconsistent with fair use, but remember youtube-dl hasn't taken or shared any permanent copies.

So unless you think that youtube-dl has made a copyright violation (which I think we've established they haven't) then youtube-dl is Sony Betamax, permitted to sell VTRs as long as there are substantial non-infringing uses, and the copyright claim will have to be brought against the infringing users.

That does not mean the courts won't find this is an "anticircumvention device" or will find that the takedown is improperly executed. They may very well rule it is an illegal anticircumvention device, Betamax happened before DMCA, and didn't decide anticircumvention.

But to my knowledge, in Sony one of the opinions spoke about "jamming" with hypothetical language, stating that it could be possible for Sony to build a box that jams unauthorized copying and I'm not aware of landmark cases that would have solidified those concepts.

If RIAA has accurately characterized the key rotation mechanism in YouTube then youtube-dl may well be ruled a circumvention device and that could be the end of it, fair use or no.


edit: I misread.


Yes, which makes this very sad that a simple slip-up can make it very difficult to come back from this.


I would hardly call this a simple slip-up. It is known that the software could be used to infringe copyrights and other open source projects make it clear that they do not support copyright infringement even when it is clear that their software can be used to do so. (If I recall correctly, Kodi and Calibre are two examples.) Meanwhile, youtube-dl has a test for that specific use case.


I feel that is rather thin legal grounds, if the examples are what dictate legality. The fact that I can change some text that does not change the function of code, or what can be done with it, to suddenly make it more legal does not sound like very great legal basis.


I don't know whether I agree that it should be thin legal grounds for a couple of reasons, even though I do agree that changing a few lines of text that doesn't alter the function of the code makes the argument sound sketchy.

Here's why I am not sure whether I agree:

The text in question is described as being part of a test. It is highly suggestive that at least one developer wanted to consider cases where the rights holder may be more assertive (e.g. with potential protection mechanisms), which implies intent. Granted, I don't know what the actual intent was. Only the people who added that text and accepted the commit can know that.

Intent is important when laws are broken. In the minor instances, it may determine whether the law is even enforced. In more significant violations, it may be considered when evaluating legal recourse or damages awarded. While I am uncomfortable with the law being fuzzy, I am also uncomfortable with it being absolute. The latter is especially true when you consider that it is far too common to create laws that protect singular interests.


As another data point, I used youtube-dl to download thousands of videos that were part of a cipher puzzle. It would have been infeasible without it.


not to mention its a console cmd and works well in scripts. i used it for a lot of sites that have their videos free but i used a console browser and used this in a script that i could call through the browser. i never looked at how it worked or what i could do with it in a non legitimate way but i wonder if i did unintentionally. i wonder if its the download part that is the issue though for some sites can have free to watch but might have issue with downloading funnily enough screen recording does exist.


I'm not a lawyer, but I am playing devil's advocate here.

Did your customers own those video files or the rights to the video? If they are the authors and owners of the content surely they uploaded something to YouTube to begin with. Why is that not what they uploaded to your service? If they preferred to import YouTube libraries clearly YouTube was adding some value there.


> Did your customers own those video files or the rights to the video?

Not sure why that's relevant?

I often use(d) Youtube-dl to grab 1080p videos to watch once, not to keep them. eg as an alternative to Firefox

That's because YouTube itself recently seemed to change something about their streaming, such that some 1080p video's now stutter and don't work correctly.

So, I download them to a temp directory then watch with VLC instead. Then batch clean up the temp directory when it's getting full.

eg Youtube-dl is just another Youtube client, no different from a browser. If you're saying that watching media via Youtube-dl is different than watching it via Firefox... to me that sounds incredibly bizarre.


The point is YouTube is a video streaming service. Not a file host.


Errr, yes?

A video streaming service is literally a host for video content + UI's for accessing it.

Whether YouTube internally stores those video's as single files, as multiple versions of a file, as chunked pieces, or whatever... who cares? They'll obviously optimise their backend storage for their particular needs.

Whether you start viewing in the middle of a video, or from the start... are you're trying to say it makes a difference?

Because VLC can seek around a video too.

With YouTube-dl, I download the video, and watch it. At some point, it gets deleted.

With Firefox, I download the video and watch it. At some point, it gets deleted.

Are you saying it makes a difference whether or not it gets assembled into "a file" on the receiving side in order to watch it in an appropriate player?


With firefox, you can see the ads that pay for hosting that content you are interested before, during and after you see it. I don't know how enforceable is that you should watch some content exclusively with the provided frontend, but it really makes a difference for them how you watch it.


Ahhh. I don't see ads using Firefox anyway, as I use uBlock Origin.


And you see that in the downloaded version as well whenever it is part of the video. The fact that youtube implemented this differently for some of their ads is their problem, not mine. That is like complaining that you aren't allowed to view a webpage on a kindle because the ads are supposed to be viewed in color.


Interesting. I never saw ads in video's downloaded via youtube-dl (nor Firefox either, due to uBlock Origin).


A copyright is for a creative work, not for a set of bits.

It's unlikely transcoding a video added anything creative to the process.


Transcoding a video is making a copy of a creative work. Ergo, by definition, copyright applies.

In fact, if transcoding a video added anything creative to the process, you would have two copyrights: one for the original creative work, and a second for the derivative (transcoded) work.


You missed it. OP has rights to the work. The question was whether downloading from Youtube breaks any laws. Youtube's add is transcoding, which is not a creative transformation. Ergo, OP probably has a legal right to download from Youtube, at least as far as copyright law is concerned.


No, I saw that. I was responding to the immediate parent, not the OP. The comment I was responding to was whether transcoding is an act covered by copyright, and it is.

The OP's comment was different. He was using youtube-dl to download videos uploaded by his own clients, which was fine because they're the copyright owners. Transcoding doesn't generally create a separate copyrighted work because at least for these types of media, the underlying representation of the bits doesn't matter.

Note: this isn't fair use. Fair use is a defense to violating copyright, and a copyright owner can't violate their own copyright ...but if they licensed one or more copyrights to a third party, they could violate the third party's copyrights depending on the terms of that license.


Yes I am aware that the bits are not what is copywritten and I am not suggesting Youtube has any right to the content itself. But using youtube-dl to download content from a competitors site is taking advantage of the functionality on that competitors site to add value to your own product.

It could still be in Youtube's interest to prevent this from happening.

I'm asking why you would expect a competitor to make your own product more compelling for your customers.

This is a further example that Youtube-dl is there to circumvent the intended functionality of Youtube.


> It could still be in Youtube's interest to prevent this from happening.

And in fact it is illegal for YouTube to lock your data within their service in that manner in the EU. You have a right to export your data and they have the responsibility to make it make possible to move that data to a competitor where technically feasible. And it clearly is technically feasible.


Do they have an obligation to export your data in a format you choose with a method of your choice or can they make that determination for themselves? Does the EU mandate the ability for YouTube-dl to exist?

There’s a lot of talking past each other going on in this thread. Nobody is defending the RIAA or DMCA. But they exist and this outcome seems painfully obvious.

RIAA being mean, DMCA being bad and copyright being outdated doesn’t matter here. What matters in that the law exists and YouTube-dl made it really easy for the RIAA to make a compelling case against them.


I think the law made it really easy for Google to make that case. I don't think the RIAA has legal standing here under DMCA.

Google isn't making that case because Youtube is YOUtube. The whole point is that it's a platform for family videos, amateurs, and all sorts of other things.

If the RIAA didn't want their content Youtube-dl'able, they should have used one of their locked-down platforms. They went to Youtube because the locked-down corporate platforms don't have the eyeballs. Now, they're trying to convert Youtube into one of those platforms.

My expectation is the RIAA will prevail, but because of deeper pockets and the potential for a multi-million dollar legal battle, not because they're on solid legal ground.


But YouTube is not the one taking legal action here, The RIAA is.


I understand that as well. I am saying it is possible to imagine a scenario in which someone looks at Youtube-dl as something other than a totally harmless aid to users.

Youtube has legitimate reasons to try and prevent people from downloading content from their site. Youtube-dl is clearly there to enable people to do that. Rightsholders can use this as another argument that Youtube-dl violates (or is designed to violate) the DMCA.

I'm not defending the RIAA or DMCA here but I can understand their argument and why they are making it.


I'm not sure what the line of thought is supposed to be here. Youtube also has legitimate reasons to make sure people watch their videos rather than their competitors, but that doesn't mean they can sue me for watching Vimeo.


Of course not. But what if Vimeo is using YouTube to host the videos? The point I am making has nothing to do with copyright. It has to do with leveraging a third party tool to make a competitors product an advantage for yourself.


And in your example they should be suing Vimeo. If we were competing hot dog stands and I use a hammer to break into your cart and steal your family mustard recipe, you don't deliver a notice to Home Depot.


Your analogy makes no sense. RIAA is arguing YouTube-dl is a tool designed to download copywritten material from YouTube because YouTube-dl included test cases in their own code for downloading copywritten material from YouTube. The existence of other streaming sites or other functionality of youtube-dl is completely irrelevant. YouTube does not need to be a party to this, the RIAA is just connecting the dots to make their own case. Their claim is that YouTube-dl is a tool for downloading copywritten material from YouTube because the creators of YouTube-dl literally tested for that in the codebase.

It’s a pretty easy case to make. Any other use of the tool is again, completely and totally irrelevant.

To be even more clear I’m sure YouTube-dl is a useful tool and I think such things should exist. However, this is an important lesson when making and owning these tools. Don’t do the RIAAs work for them.


That's all well and good, but I don't see where your earlier points about business competition fit into all that.


Not PC but,

Some people use YouTube as their repository. Or if they lost the original files it would be a way to recover something. I once had to download mp3's of my own music from MySpace because I had an HD crash and lost them.


"Some people use YouTube as their repository."

I wondered when this would happen. I'm. not referring yotube-dl specifically but the third-party repository problem such as Github and YouTube where material exists principally on one site.

The fact that the RIAA is involved makes it high profiled case and should warn people that distributed systems/repositories are a much better (safer) idea.


I too offload my personal vidos (family/kids) to youtube from my iphone. I used to back them up on a Hd but it has since crashed so whats on youtube is basically my sole backup. If youtube dl goes down I’ll have a hard time retreiving all that


I stopped sending my data to YouTube years ago for privacy reasons, so I have multiple HD backups. External drives are cheap now. If you use three drives you've redundancy and two of the drives can be second-stage/downstream backups so the drives can be configured so as they're used intermittently, thus they get very little wear and remain reliable.


Better get cracking then.


Sure that is perfectly reasonable. And the fact that YouTube is a resource for this is one of its competitive advantages. So it makes sense that YouTube might make it hard for other parties to take advantage of its position as a repository. Thus Youtube-dl is created to circumvent this design choice on the part of YouTube.


YouTube's design choice is to serve all their content via an open api, which is why youtube-dl works. It's not circumventing anything.


youtube-dl scrapes the website, from what I can tell. The API has rate limits that youtube-dl would bypass in a few minutes if it were to use the official API.

That's primarily what's got the RIAA's panties in a twist, is that youtube-dl is 'bypassing' youtube's protection measures to prevent their site from being effectively scraped.


Still can't understand the logic, the website is open! If the information can get to my eyes and ears I can put a recording device in-between. At the end of the day the video is being played on my computer. The only logical endgame, if stuff like this allowed to stand, leads to general purpose computing being outlawed.


If it was as easy as putting a recording device in the middle we would all use our cell phones and YouTube-dl wouldn’t exist.


It doesn't matter how much value YouTube adds. It neither owns, nor does it assert that it owns, the copyright to these videos. Banks provide a value-add service for your money. They don't own your money. Value-add is not a valid legal test for copyright ownership.


The question here isn't whether youtube owns the videos, the question is whether youtube's use of anti-scraping measures constitutes DRM that, by circumventing, violates the DMCA.


Right, and again, I'm not saying Youtube has any copyright ownership here. I'm saying they provide a service which adds value and as a business they may want to protect that competitive advantage.


"Napster could have been used for legal means as well, and it got destroyed in court."

But Napster was a service run by an organization.

youtube-dl is neither a service nor an organization. It's just some source code floating around the internet, to which volunteers contribute to.

A very different situation.

If anything, it's Youtube itself that's like Napster, and it's telling that the RIAA has gone after youtube-dl rather than Google.


It's not even clear that youtube-dl vs watching in the browser is any different from watching on tv vs using your vcr to record which has long been legal.


In this case, it "steals" the ad revenue that the copyright tries to protect, no? Or am I thinking about this incorrectly?


This is an argument copyright owners tried to argue with VCR and later TiVo, as it lets them skip out on the ads.


Nobody can steal their own attention


RIAA definitely does go after Google. They file countless DMCA takedowns on videos that re-upload content from official accounts.

Unlike Napster, Google just complies and hence doesn't have much of a case against it like Napster did.


RIAA already went after YouTube. That’s why Google added the ability to scan for copywritten material. Google wanted YouTube to be a friendly place for people to view content. That meant playing ball with RIAA.


Napster wasn't merely a tool it was a centralized service that provided an index of infringing files and where to get them. It both directly assisted in the act of infringement it profited from same.

You used the words "unintended side effect" in quotes but I am unaware of where that comes from while I am familiar with the term "substantial non-infringing use" and "contributory copyright infringement"

Youtube-dl neither assists nor profits from copyright infringement and it possesses arguably non infringing uses.

- You can download non infringing videos. Youtube-dl works on a variety of sites beyond youtube

- If you feed mpv a url it will transparently use youtube-dl to handle fetching the media for playback. Unless you specifically tell it to this does not permanently save the video and is not materially different from watching the video on firefox.

Mpv incidentally uses far less resources for decoding video even with newly added hardware decoding in linux.

- You can consume the same video you are allowed to access on a different device or in a situation wherein you lack network access. For example one could watch and then delete a youtube video one wanted to watch on the train where you lacked internet access. This is analogous to format or time shifting which have been deemed fair use.


Can't it be trivially reinstated then by removing any references to copyrighted content? youtube-dl has plenty of legitimate uses beyond just copyright infringement.


They can't fully un-ring that bell. The RIAA will always be able to argue infringement was the purpose since they used copyrighted examples.


Plus they've advertised themselves as having that capability. I suspect that even after the removal the RIAA will argue that the fact that the ability to download copyrighted material, which is something that the project itself said it could do in the past and has not been modified, makes it continue to be illegal.


I don't think that's quite how it works. It's not (generally) capability that matters but rather intent.

The RIAA would of course argue that examples of infringement in the test cases or readme demonstrate intent. A reasonable response to that might (or might not, depending on the context) be that infringement in those specific cases was never intentional but instead purely by accident.

If the infringement in this case ends up appearing to be intentional, it would probably make for a very uphill battle to argue that the tool itself was only intended for legitimate use cases.

Edit: Of course, it's reasonable to ask - if it's legal for YouTube to distribute the content, does using youtube-dl suddenly make it illegal? Is it a violation of copyright to record a pay per view stream? What about a publicly available stream paid for by ad revenue?


I think we're in agreement, it's just that I'm saying that the capability for that remains and the intent was shown to exist already.


> the intent was shown to exist already

Well that's the question, isn't it? Is downloading those particular URLs infringement if the content is legally allowed to be hosted on YouTube in the first place? If so, was the use of those URLs specifically intended to demonstrate or verify the ability of the tool to infringe copyright? Alternatively, can the tool be said to circumvent a protection device under DMCA section 1201?

(For the record, my answer to all of the above is an emphatic "no".)


And would a "derivative work" that happens to use "some" open source code from youtube-dl, but that never included a URL to a Taylor Swift song anywhere in the code, be considered to have a different intent?


I think this is actually an interesting question. Someone could fork youtube-dl to create "public-domain-dl" which has multiple references in the README to the fact that it is only intended for downloading public domain videos from YouTube (and elsewhere).

Instead of a unit test which successfully downloads a copyright-protected video, it would have a unit test which attempts to download such videos but returns an error message to the user based on a hardcoded blacklist of videos IDs, stored in a separate config file. (It would also have a unit test for successfully downloading a public domain video from the developers' own channel, of course).

The developers could also make clear that they are happy to receive DMCA requests to add specific video IDs to their blacklist. As for what happens if the user deletes the blacklist config file, maybe that won't have a unit test.

If such a fork was successful in fending off court cases, though, it would start to raise interesting First Amendment questions. The difference between the two pieces of software, in practice, is just the text in the README telling people not to break the law (plus a single "Delete" button press after installation, which the user would have to learn about from another source). This would mean that the difference between developers complying with the law and breaking the law is including that text in the README, which effectively seems like compelled speech.


> would a "derivative work" ... be considered to have a different intent?

Probably not. If intent were to be demonstrated in court, a derivative work would almost certainly be tainted in turn unless those authors had a _really_ good justification for their actions. (IANAL though; I'm just guessing based on other copyright cases that came up in the media before.)

For example, imagine forking Napster back in the day, redoing the UI, and rewriting half of the internals. Do you really think a court would let that fly? Judges are hardly idiots.

That being said, why not just mirror the original to a host running as a tor hidden service and continue all development efforts there?


My guess is that the answers to the questions is likely "no", "it'd be a hard case for youtube-dl since their lawyers are probably not that great" and "no", even though I wish the answers were "yes", "irrelevant", "irrelevant".


Isn’t the problem that this stuff is on youtube in the first place? The RIAA can’t win against google, so now they’re going against smaller players?


Youtube has a licensing agreement with the RIAA. If they didn't, Youtube would be bankrupt, since statutory damages for copyright violations are $750/violation and there's no legal defense once they show the violation occurred.

(And as someone who has dealt with the music licensing agencies, you always just pay for the license. Paying the big 3 licensing agencies annually for millions of streams cost less than the statutory damages for a single violation.)


No the DMCA says that so long as they respond to take downs they aren't liable. The agreement prevents litigation that COULD result in a judgement that might modify this deal which is clearly a risk in addition to the cost of ongoing litigation.


No, you're talking about something else.

Youtube has the right to stream those videos, because they pay the RIAA for a streaming license, which is a thing that music licensing agencies have provided for many years now. In many cases, the videos were uploaded by the recording studio itself, and in those cases presumably Youtube isn't paying a licensing fee because the license is implicit in the upload (but note that many videos uploaded by the artists themselves are not technically theirs to upload because the actual rights were held by the recording company, hence the reason Youtube needs the agreement with the RIAA).

The DMCA take-down exemption you are thinking of is a separate protection, that applies to Youtube taking down videos that aren't licensed by them containing music that isn't licensed by the creator of the music.

The statutory damages would apply in the former case, such as if an artist uploaded their own music video and Youtube didn't have the broad RIAA license. This happens a lot more frequently than you would think.


You seem to have a lot of legal theories that are based on your own understanding with no citations in cite. I suggest you do some research and either modify your opinions or add citations that show everyone else is wrong and you are right.


Didn't RIAA already win against Google? A lot of the messed up stuff with Content ID comes from Google's fight with record labels.


I think the case here is songs that are legitimately on YT that they don't want people to download (and avoid ads, etc)


What if another person will recreate this repository without history and offending links?


Might help a bit, but it is still pretty relevant.


Argue in front of whom exactly? Because thats ONE of the big problems with DMCA.


This is a proper DMCA claim, not YouTube garbage, so the process goes claim->counterclaim->lawsuit.


Federal court in the USA.


The DMCA only applies to the USA. In my opinion it'll only take days if not hours for the files to appear on a website in Eastern Europe, somewhere that the RIAA can't get at.

If that happens, the difference is that these are public domain non-copyrghted materials -- not the copyrighted videos, thus stopping access to the files will be more problematic.


Which is why RIAA is trying to kill the code itself.


Cure the disease not the symptom


The very premise of copyright itself then.


The issue with the readme infects the rest of the code, even if it's removed. It's still evidence that the repository was intended to facilitate copyright infringement. The problem isn't the words of the readme but what those words imply about what the authors of the code intended.

If they instead had released it and then said in a public forum "check out my cool code for copyright infringement" (but had a totally blameless README) that would be used as evidence in the same way.


Equally, then, couldn't another author fork the code and distribute it under a new name, with the offending test removed and a blameless README that says "please don't use my cool code for copyright infringement"?

It sounds like you're saying the problem is with the original authors, not the code itself, so taking down the code seems like the wrong approach for the RIAA.


Time to fork.


Should it stand up to legal scrutiny though?

What does it matter if I play the beatles on youtube by watching in a web browser or watching an .mp4 file that I just downloaded? It's functionally the same.


The intent matters. Browsers don't make it easy to get ahold of a copy of the file that you can share, youtube-dl does. Add that to a bad README and that's something that a court might treat very differently from a web browser.

This is not far off from how it's legal to carry lockpicks but it's often not legal to carry lockpicks around with the intention of using them to commit a crime. Either way you're carrying the same lockpicks, but if the court decides that you were carrying them with bad intentions, you're in trouble.


But it's perfectly legal to screencapture a video platform in full-screen, and previous courts have ruled that recording a stream you otherwise have access to is allowed (the VCR/TiVo lawsuits). The question here is whether youtube-dl is a tool that allows for the bypassing of DRM, as the RIAA is asserting that youtube has DRM on their videos that prevent automated tools from scraping the website.


> The intent matters. Browsers don't make it easy to get ahold of a copy of the file that you can share, youtube-dl does. Add that to a bad README and that's something that a court might treat very differently from a web browser.

I can't say for YouTube specifically (I haven't tried, I suspect they have the talent to make it actually difficult), but browsers do make it remarkably easy to extract most forms of media from a website. Even websites that attempt to hide it have to embed it somewhere. Whether you feel like it's easier to install Firefox and figure out how to extract the media, or figure out how to install Python and pip and then run youtube-dl is up to you.

There are likely even tutorials up by Mozilla somewhere about how to use the dev tools to pull an image (though I assume they're not stupid enough to do it on material they don't have rights to).

There are many reasons one might want to use youtube-dl. Furthermore, there are many creators that either don't care or are happy that you're downloading their stuff.

If we want to make analogies, every gun shop in America sells guns and human shaped targets. Do they have the intent to encourage people to murder other people? It is a tool that can be used for illegal purposes. They might even have examples of how you can do such a thing (as do gun manufacturers, lockpick manufacturers, security exploit announcements, etc, etc). Why would that make the tool illegal? The argument might hold if the only possible use of it is illegal; that puts it among the ranks of things we deem illegal to merely possess, like fully automatic weapons. I'm not buying the RIAA's argument that youtube-dl is so dangerous to society that we can't trust anyone to use it legally.

> This is not far off from how it's legal to carry lockpicks but it's often not legal to carry lockpicks around with the intention of using them to commit a crime.

No one has demonstrated intent. The publishers have demonstrated that it could potentially be used in an illegal manner (depending on your jurisdiction). They aren't going after someone who downloaded youtube-dl and is trying to use it to pirate things. They're going after the lockpick manufacturer because they have a tutorial on using their lockpicks on their site, and people could use that to commit crimes.


YouTube uses things like DASH, where they download pieces of the video from the server a bit at a time and feed into the browser's video pipeline. There's no single link to the entire video, necessarily; you have to piece together the bits. This is one of the things that allows changing quality mid-video, btw.

Furthermore, the video and audio are served separately so that codecs can be mixed and matched on the client side, and the same audio can be used for different quality video streams, etc. One of the things youtube-dl deals with is grabbing both and muxing into a single container.

So yes, extracting things from YouTube is not trivial at all.


> This is not far off from how it's legal to carry lockpicks but it's often not legal to carry lockpicks around with the intention of using them to commit a crime.

This is pretty far from lock picking tools. The crime commits with lock picking tools involves physically breaking into someone private property to either steal physical items or do physical harm to people. Breaking and entering and copyright law are not comparable. A closer analogy is smuggling a camera into a concert to break copyright law by recording a concert. Canon or Nikon are not responsible for what I did, just because they made the technology that allows users to record any content, which may or may not include copyrighted material. Even if Canon included how well their camera performed at recording concerts, that doesn’t suddenly make them responsible. Youtube-dl is much closer to a digital version of a camera than to lock picking tools.


The former gives money to the original content owner for every listen while the latter does not. Which is why you're able to listen to it on youtube without having to personally pay someone.


  Sony Corp. of America v. Universal City Studios, Inc., 464 
  U.S. 417 (1984), also known as the “Betamax case”, is a 
  decision by the Supreme Court of the United States which 
  ruled that the making of individual copies of complete 
  television shows for purposes of time shifting does not 
  constitute copyright infringement, but is fair use.[1][2] 
  The Court also ruled that the manufacturers of home video 
  recording devices, such as Betamax or other VCRs (referred 
  to as VTRs in the case), cannot be liable for infringement.
How is downloading something from youtube for later offline viewing ("time-shifting") in any way different from recording a tv show?


The Betamax case was pre-DMCA. The law has changed. I'm not saying it's invalid, just that there's a different argument to be made.


The Betamax case was about stuff broadcast over public airwaves, whereas this is a medium that the RIAA barely tolerates, rather than the (ideal) main distribution channel. Seems like a big enough difference to me.


The ruling there doesn't seem to rule out making copies of shows from a cable network, which are not exactly public airwaves in the same sense I think you mean.


I don’t see the fact that the RIAA dislikes internet based streaming as relevant to the analysis. Downloading a publicly available video transmitted over a nearly global communications network is directly analogous to recording a show to tape from a TV signal.


The signal is broadcast (ie not under your control) and so the sole purpose being permitted there is time shifting.

YouTube is an on demand stream. It would be comparable to recording a pay per view movie that you purchased. Is that legal? (The question isn't rhetorical, but I seriously doubt it.)


What if your intent is to shift to a time when Youtube doesn't exist?


It when you don't have access to it (plane flight, etc). That sounds like time shifting to me.


If, for example, you bought a movie and downloaded it to watch on a plane when you can't have access to streaming, why wouldn't it be legal?


That would be an interesting argument. I suspect it would be illegal because a pay per view stream is typically time limited. But if the stream itself were still valid while you were on the plane ... ?


What about using youtube-dl as the only technical way to play it on a low power device, like with mpv on an Rpi or old laptop.

That's a kind of tech adaption of the streaming/playback mode (and I've done it a ton on previous laptops)


Then RIAA would argue for a DRMd version of downloader that is prevented from doing it in high power deviced.


"...this is a medium that the RIAA barely tolerates, rather than the (ideal) main distribution channel."

Right, this is the precise problem, and it has to be solved - not by a free-for-all but with proper equitabl copywrite laws.

As matters stand, it's only going to get worse. Say we assume that if somehow both YouTube downloads are made much more difficult to crack AND no downloaders are available, then viewers and downloaders will then use newer tech that's even easier to use to record the videos. With new hi-resonution 4k cameras becoming commonly available, they'll be able to record stuff with still-excellent resolution by just photographing the monitor.

Of course, that still won't deter the RIAA, I can see them trying to ban cardboard covers and jigs that hold cameras and monitors accurately in alignment.


Probably because YouTube premium let’s you do that.


It could but probably doesn't

One could argue that YouTube success is mostly about unlimited access to illegal copyrighted material uploaded by some unknown person in some unknown part of the world

After all the thing it's already on YouTube and on my HD after I watched it


If the videos mentioned in the notice are hosted on Youtube without authorization, then the takedown notice should be sent to YouTube.

If the videos were uploaded to Youtube with authorization, then accessing them through youtube-dl is not an example of the program being used for infringement.

Either way, the purpose of the software is to download anything publically available from YouTube. Its purpose is not copyright infringement unless the purpose of YouTube is copyright infringement.


That's exactly the point.

The content is already available

The monetisation of the content depends on a lot of different factors, many have monetized unauthorized content over the years, many received money from the same unauthorized content by faking clicks and views and many avoid paying YouTube in the form of ads by using ad blockers

I believe youtube-dl users amount to a maximum of a single digit percent of the above (with the digit being between 1 and 2 with 2 excluded)


> YouTube success is mostly about unlimited access to illegal copyrighted material uploaded by some unknown person in some unknown part of the world

Good point. I quite like that argument.

Not that my opinion actually matters in this case.


Someone who is capable of downloading youtube videos can easily block ads and many do, not because ads are bad but because they are both excessive in any form possible and a malware spreading channel. I do too but don’t bother to download videos because it is still more convent to watch’em directly on youtube. But it seems those days are coming to an end soon, lets see


I watch youtube videos with mpv sometimes. That's not even downloading it. But it's using youtube-dl to stream the video :)


It stands up to legal scrutiny, and Betamax case is not applicable.

They have these things called "streaming licenses" these days, that are very different from "download licenses" because streamed data is not the same thing as a downloaded file even if they both use the same codec and are derived from the same source.


This is a wholly imaginary distinction based on a triviality. Streaming IS downloading. Also how is the betamax case not applicable. The user of a VCR isn't "licensed" to record whats on their tv


No, it's a distinction that exists based on the differences in how the video content is consumed.

It is irrelevant, legally, that one implementation of a streaming software may choose to download the streamed data to a temporary disc file, or that some content downloaders can structure the assembly of a partially downloaded file so that (some) content can be accessed before the download is fully completed. What matters is that a streamed file is intended to be consumed contemporaneously with the acquisition of the downloaded data however it is stored on the computer, and that caching for efficiency aside, each separate act of consumption involves a separate act of acquiring the content, whereas a downloaded file is intended to be downloaded once and consumed multiple times.

Also how is the betamax case not applicable. The user of a VCR isn't "licensed" to record whats on their tv

No, they are not licensed to record what's on TV. But the point of the Betamax case was that "time-shifted viewing" by a private viewer was deemed to be fair use because at the time broadcast TV content was transmitted on a fixed schedule and absent the private recording, was not otherwise consumable because many of these shows were not rerun or sold on cassette tape. Thus the court found that the Betamax had substantial non-infringing use (and indeed, that it's primary intended use was thus non-infringing).

However, digital content is provided on demand so there is no need for "time-shifted" viewing, and so the reasoning from Betamax no longer applies. Moreover, many content owners now make the content available for "connectivity-shifted" viewing (aka, away from internet connectivity) via a downloaded content license. Thus, downloading that content from a streaming source is generally no longer protected by fair use.

Consequently, using youtube-dl to download your favorite streamer to watch on a plane trip would probably be fair use because they probably don't make the content available for access-shifted viewing otherwise. But using it to download a music video would not, because the music and video owners make those available for download with a purchasable license. And the problem with youtube-dl is that they demonstrated that the primary intent of the tool was to download music videos, the evidence being that the README and parts of the code explicitly refer to downloading music videos subject to RIAA licensing. Under the Betamax ruling, this would be fine, because there are substantial legitimate uses...but the DMCA changed that and so it doesn't matter if there are substantial legitimate uses if the primary intended use was to violate copyright.


"README and parts of the code explicitly refer to downloading music videos subject to RIAA licensing."

It beggars belief that people can be this stupid. Why the hell would one mention it here? Are they truly that stupid to believe that the RIAA wouldn't see it? Then they made matters even worste by putting it on Github instead of some obscure website in Eastern Europe?

Talk about shooting oneself in the foot!


But it's not legally the same.


I followed the Napster trial closely. Napster’s argument that it had legitimate use was demolished because it had an ability to monitor the specific use of its service on an ongoing basis. That is, it could see queries from users as they happened. In fact, this ability was at the core of the service since Napster relied on a centralized search index maintained by the company. The court ruled this gave the company a more active role in contributing to infringement than the precedent Napster was trying to use (which involved VCRs - the Betamax case).

The courts remedy (at least at the preliminary injunction phase) was not to shut down Napster but to force it to comply with search blacklists provided by the record companies.

youtube-dl obviously is not a service so had no ability to monitor ongoing use. IMO it would have a much stronger case for legitimate use.

IANAL


> The fact that copyrighted works were included in the readme shows it was intended for that use

How could that be true? if the YouTube link in the test case was a video of farmer on his truck would that show it is intended for that use? No. YouTube dl is made to download any YouTube videos it has not been made to download specifically Taylor Swift music videos, it just has just been picked as a test case as any other YouTube videos could have been


While I agree with the point, a strong case can be made that the music video in the README and the test case were a poor choice.

A silly home video, preferably uploaded by the maintainers themselves would've just given RIAA's claim no teeth.


Or content from the public domain.


It was however a very poor choice, given that they could have picked something better.


We might as well ban screen shots because it can illegally reproduce photography stills. /s


Android allows applications to disable screenshots while they are running, as a form of DRM.


It's not a problem on a rooted phone, irrespective of what app developer decides or wants. And as I've mentioned, at a pinch, just photograph the screen with another device.

These days, just about anyone can copy anything that the DMCA covers despite any attempts to protect it. The answer is elsewhere, that is copyright laws have to be updated to be equitable to everyone. Unfortunately, it's dinosaurs like the RIAA who are standing in the way stopping it.


Distributing something that downloads copyrighted content isn't the problem though - the problem is "circumventing a technological measure that effectively controls access".

If the media provided by YouTube were DRM'ed and youtube-dl would be e.g. extracting the decryption key, this would make sense. Is that what youtube-dl is doing though?


From my perspective it kind of seems like YouTube-dl is like a knife, it can cut your food, but it can also be used to kill someone. In the case of YouTube-dl the knife was included with instructions on how to kill someone, so it was deemed a potential "murder weapon".

  What if someone sold a car and included instructions on how to run over pedestrians?  Should the car no longer be sold, or should the instructions be removed?
My point here is that banning all of something based on the fact that the something could be used to commit illegal acts, when that same something also has legitimate legal uses, is a slippery slope. These days the RIAA pretty much lives on that slippery slope.


Curiously, what do scrapers (aka "readers") have to worry about with this topic?

Eg, i'm making an archiver and reader combination that, for personal use, archives stuff in a Git-like store. Yet, Git (and Git-likes) can also be used to distribute.. so hypothetically i could use this software to scrape and distribute content.

My intention is primarily to make news articles/etc searchable, archived, etc. Yet i'm sure NYT would have something to say about my test cases scraping their site.

.. the world is interesting.


I think the recent LinkedIn case established that it is legal to scrape anything which is publicly available (ie no login required). Redistribution would be a copyright violation, but the scraping itself is legal.

I'm not sure what that means for youtube-dl though.


The same argument can be made for video that's also publicly available.

Not to mention, youtube has it's own Download function in the mobile apps, even for copyrighted content. So there's that.


> youtube has it's own Download function in the mobile apps, even for copyrighted content

Wait, seriously? I had no idea. I would think the presence of such a feature would severely undermine the RIAA's case.


It does, but it's only available for those who have a subscription to youtube premium.

As such, I assume the RIAA will try to argue that it doesn't fall into the same bucket that youtube-dl does, since the method of accessing the site is different (scraping the site, vs a sanctioned API)


I assume that YouTube's licenses with the RIAA covers the Youtube Premium features on their videos. If they didn't, they'd be in violation of RIAA's copyright (streaming licenses are not the same as downloading ones).


You know... I read the Youtube license and I see this:

Licence to Other Users

You also grant each other user of the Service a worldwide, non-exclusive, royalty-free licence to access your Content through the Service, and to use that Content (including to reproduce, distribute, modify, display, and perform it) only as enabled by a feature of the Service.

I am a user of the Youtube service. I read and agreed with the terms, I have watched videos. Thu I am allowed to reproduce its content. The RIAA explicitly agreed I can distribute it.


I'm not a youtube user, and I didn't agree to their terms of service. I can however plug their URL's into a terminal without actually accessing their web site


only as enabled by a feature of the Service seems like the operative clause.


1) the Service makes downloads available (Youtube on Android)

2) the Service makes downloads available on any browser, it is up to the browser to show it.

3) "If it turns out that a particular term of this Agreement is not enforceable for any reason, this will not affect any other terms."

Physics clearly make this unenforcable.


"Spitting out an mp4 file that you can copy to another device" is not a service that YouTube provides. That's why youtube-dl exists. If it was a service that YouTube already provided then there would not be as much reason for youtube-dl to exist.

"Unenforceable" doesn't mean practically unenforceable, it means legally unenforceable. The laws of physics permit most cars to exceed most speed limits. Those laws are still enforceable.

Unenforceable clauses in contracts are things that are either contrary to law or otherwise don't fall under contract law. So a clause that says "if you sign this you must give up your firstborn son" is unenforceable. That sentence just says that "if we screwed up drafting this contract, just because you were able to challenge one clause due to being unenforceable doesn't mean the whole contract is void." It's a severability clause.


Yeah, the last one was tongue-in-cheek.

However, Safari (not Firefox) is not affiliated with Google. Google has no say on how the content provided by the Service is presented. The Service provides a HTTPS connection. Its influences ends there.


> it's only available for those who have a subscription to youtube premium.

No. It's available for free users as well AFAIK.


> The fact that copyrighted works were included in the readme shows it was intended for that use, and the RIAA complaint will likely stand up to any legal scrutiny.

This. FWIW Grokster is the relevant case law here, not Napster:

https://en.wikipedia.org/wiki/MGM_Studios,_Inc._v._Grokster,....


Just an anecdata point but I was one of the people who used Napster for legitimate purposes (I used it to distribute music I created, gaining small recognition that wasn’t something I could’ve achieved otherwise with my music and my means and other technologies available). The thing that was disappointing in the fallout was that it caused many years of evasion tactics that made finding a comparable distribution multiplier very difficult. For my own use, I stopped caring about making music for others and trying to find a way to get an audience, and I happily play and write music now for my own joy and amusement, but I hate to think how many artists really struggled to find an audience the same way and came up short


> The fact that copyrighted works were included in the readme shows it was intended for that use

It's an alternative web browser for videos, of course it was intended for copyrighted works, just like Chrome is. I don't understand your argument, is creating alternative web browsers illegal in the US somehow?


This is exactly the point. All they had to was use their own videos for the tests and readme and the dcma takedown would probably never would have happen. Plausible deniability.


> The fact that copyrighted works were included in the readme shows it was intended for that use

Just because copyrighted works are on youtube and are seen as an example of youtube content does not mean the software was built with the express intent to violate copyright. That just doesn't follow.


It beats me why they'd use those videos as examples knowing how finicky copyright law is.


Lack of legal awareness / advice?


Perhaps the SFLC[0] should be actively seeking out popular Free Software projects and pre-emptively auditing them for legal issues such as these. That must be a more cost-effective (and crowd-sourceable) process than actually fighting a legal case against the RIAA.

[0] https://softwarefreedom.org/


Intention is a component of the law. Youtube-dlc apparently intended to assist in infringing the copyright.

We're still fighting over the color of bits. It's all a bit technically bozo, but, sigh.


A knife can be used to cut bread also.


No it doesn't follow at all, any more than had the repo owners included a couple of public domain recordings in the repo then the conclusion that the tool was clearly intended to download public domain recordings.


It's hard for a lot of techies to grapple with this, but courts and law are often decided by the intent of the offense. This shows intent, whereas public domain works would not. The difference, in court, is massive.


It's not that it's hard to grasp, we just think it's irrelevant, since plenty of people also have legitimate uses for this tool, and any other tool that allows them to do the same things is just as capable of downloading copyrighted youtube videos.


> since plenty of people also have legitimate uses for this tool, and any other tool that allows them to do the same things is just as capable of downloading copyrighted youtube videos

True, but _other_ people and _other_ tools aren't relevant to the intent of _these_ authors and _this_ tool. _Your_ intent is relevant in establishing whether _you_ violated the law (and thus whether an associated DMCA takedown is valid).


Intent does matter only if there's some action taken, or attempted.

The tool just sits there on github. Only the tool's users actually do download stuff or attempt to.

Also you'd be hard pressed to prove such specific intent from some random testsuite url, added by one or two authors, when the tool has very many authors/contributors and the README/manual for users just has non-copyrighted material sample urls, anyway.

RIAA just tries to make it harder to preserve regularly disappearing youtube content, like war crimes evidence, just to make some people richer on some stupid music. I can prove their intent, because that's what this tool is used for, and they're trying to remove the tool from the internet. /s This is the same stupid argument they're using. They will fail, but why not try to make other people's lives harder, right?


I do hope they fail, but regardless I think your analysis here is completely off base.

The intent or action taken in this case is the creation, maintenance, and publication of the tool itself. At issue isn't the intent of any particular end user, or even the majority of end users, but rather of the authors themselves.

To the best of my knowledge, it simply does not matter how the tool is used in practice. AFAIK, all that matters is the intent of the authors in creating and distributing it.


And that intent only shows that authors like to write code and contribute to such a tool. You really can't show much more from the act itself.

For example I authored and maintain megatools, a tool to download/upload from mega.nz that many people probably use to download the same kind of content that's being the issue with youtube-dl. I have no way of knowing.

I haven't used megatools myself in about 4 years, aside from quick testing prior to release/update. And I never really used mega.nz that much even before. I just wanted to learn a bit about cryptography, and return back to C programming after a few years of just doing PHP/JS, and writing a first mega.nz third-party client was an interesting opportunity at the time. Yet the tool is somewhat popular, and distributed as one of mega clients in various Linux distros.

It's ridiculous to assume intent from some test cases. The only intent I can extrapolate from the actions of authors is that they like to code, fix bugs in other people's code, and want to ease their maintenance burden by having tests. The rest is just speculation.


Are there other tools outside of software that we apply this sort of thinking too?


I don't believe the general principle is limited to either software or tools. It is my understanding that the courts generally operate based on intent, particularly when it comes to criminal law.

(For an arbitrary example, consider the difference between a gas station selling to the typical customer versus someone who asks the clerk for help filling some gas cans and in the process openly admits an intent to use them to commit arson later that night.)


You may feel it doesn't matter, but it does in fact matter to the entities where decisions are made on matters like this: the courts.


I guess that means those decisions are then based more on emotion than facts?


From a practical perspective, it's hard to wish for or imagine a legal system that doesn't rely upon judicial interpretation of intent. An example would be causing the death of another human being. Would you want all four of these incidents to be treated the same from a legal perspective?

- A surgery, which is known to have a 50/50 chance of success even at the hands of an excellent surgeon, goes poorly and the patient dies.

- A serial killer spends six months plotting the murder of his victim before executing them.

- A shoving match breaks out at a bar. Somebody slips, falls, hits their head on the ground, and dies.

- You accidentally frighten somebody by sneezing in a quiet library. They have a heart attack and die.

From a purely logical perspective, these are literally the same thing. A person dies! Realistically speaking, any legal system needs to consider intent.


If the law reached a point where the defendant in all of those scenarios would be found not guilty provided they were wearing a t-shirt at the time saying "My intent is to not kill someone", then we would probably agree that the law is not fit for purpose.

Similarly, if the difference between illegal software and legal software is the value of a random ID in a unit test, or the presence of a boilerplate "Please don't use this for copyright infringement" message in the README, then the law isn't really "promoting the Progress of Science and useful Arts", it is just a gotcha trap for developers who forget to copy-paste a piece of text.

The equivalent from the Prohibition Era would be the warnings on "grape bricks" that told buyers "not to leave that jug in the cool cupboard for 21 days, or it would turn into wine".[0]

[0] https://www.thedrinksbusiness.com/2017/07/rare-prohibition-g...


If you would put aside your emotions for a second, you would see that this is clearly just an issue of different priorities and values -- on both sides. And the side that has the power has priorities that presumably, given your phrasing, disagree with yours. I understand that can be frustrating and that that frustration might lead you to questionable rhetorical devices, but it won't change the fact of the situation.


No, the court looks to facts to establish intent. It's not possible to look into someone's brain, but courts are very, very comfortable with looking at facts to determine something about what's going on in there.

Crimes like first-degree murder have intent as a requirement. People are convicted of it all the time, and it's based on facts and evidence from before and after the killing.


> first-degree murder

Willful homicide, for non Americans.


Yes, it does mean that.


I think you are misunderstanding the legal argument. The DMCA Section 1201 specifically prohibits (among other things) technology that "is marketed ... for use in circumventing a technological measure that effectively controls access to a work protected under this title." The example in the README is evidence of this.

The argument is that youtube-dl is primarily used for breaking DRM not just that it could be used for doing so.


> The example in the README is evidence of this.

it wasn't in the readme, it was in the test suite.

I don't think it would be difficult to argue that downloading the world for the sole purpose of making sure the downloading worked was not an infringement (or, alternatively, was fair use).


I'm in no favour for the RIAA but just imagine. It would be a 300IQ move if that specific pull request was engineered by RIAA themself.

Please pull that test case... Fooled! Now there is a legal reason. Here is the DCMA takedown notice.


Calling the thing youtube-dl in the first place rather than GenericDownloadHelper or something like that could, I guess, be seen as marketing it for a specific purpose.


Sadly yes. Legally this was actually rather stupid.

Giving the project an innocuous name - think of a cute animal that isn't already being used as an open source mascot - and not explicitly mentioning popular artists in the README would have made the RIAA's case harder to argue.

As it stands now, whatever the ethics or politics, legally there isn't much of a defence.


so, let's fork it and call it something like FURIAA


Yeah I'm sure the RIAA won't notice their name hidden in there and argue the tool is designed to obtain their content.


Too late; the code is already tainted. It carries the colour of "intended, marketed and used for copyright violation", and you can't just wash it away by reuploading the repo under a different name.


Why not?

If youtube-dl incorporates some generic HTTP library (which it presumably does), is that now "tainted" too? If it used the one from Chrome, does everybody have to stop using Chrome? That seems problematic. Also, if that's the case I foresee some epic trolling ahead as people incorporate "interesting" code into their overt piracy tools. I bet some of them even incorporate code built into Windows or macOS.

If not, what stops somebody from taking all of youtube-dl, changing the name and three lines of code, and saying that the removed lines were the ones promoted for infringement?


The law isn't some computer program that you can trick if you try hard enough. If someone can reasonably tell that you just took a program and rejiggled it a bit to pretend it's something else, they are going to come after you regardless and likely be successful when they do so.


So lets have someone outside the influence of the RIAA do so? Not everyone lives under their thumb.


"The law isn't some computer program that you can trick if you try hard enough."

I don't know what you're talking about. It happens all the time.

Miscarriages of justice are common. As just one of endless examples, read about how much of forensic science is a joke, yet it passes muster in the courts.


Those things happen because powerful people are involved either on the side of the state or a plaintiff or defendant with lots of money. As a weaker party, you can't make the system work for you so easily.


Generally to do that you have to be willing to aggressively pursue your position and be somewhat ok with being unethical, which I doubt will be youtube-dl's position.


See my response here for summary, and the article linked there for details: https://news.ycombinator.com/item?id=24874225.

> If not, what stops somebody from taking all of youtube-dl, changing the name and three lines of code, and saying that the removed lines were the ones promoted for infringement?

What that somebody says is immaterial. If they did clone youtube-dl, their clone carries the "meant for copyright infringement" colour, by the virtue of being a clone of a project with that colour, and not a completely unrelated and independent project. It's the provenance and intent that matters. Web browsers and HTTP libraries do not have the "bad" colour, and being general-purpose tools, they likely never will.


> If they did clone youtube-dl, their clone carries the "meant for copyright infringement" colour.

Which part of it? Not the parts that consist of HTTP libraries, apparently? But you can break any given program into arbitrarily many components that are each independently useful as a component of a different program.

In this case the "meant for copyright infringement" part seems to be some of the unit tests. Does that mean the rest of it is fine? Or that the HTTP library part of it isn't?

You need a better way of distinguishing them than just claiming sorcery.

I mean here's a direct quote from your article:

> Most importantly, you cannot look at bits and observe what Colour they are.

So if the same code appears somewhere else, why would you expect to still have the same "Colour"?


> Which part of it? Not the parts that consist of HTTP libraries, apparently? But you can break any given program into arbitrarily many components that are each independently useful as a component of a different program.

Not parts, but the entire thing. Colour propagates through causality. The article I refer to explains pretty clearly about what it means. It's the intent and provenance, not the bits, that are important. If youtube-dl gets classified as illegal, then any trivial modification to it will get the same treatment. Even if you end up slowly replacing every bit of code, if you forked off youtube-dl and didn't change the functionality, that's still essentially youtube-dl.

(You can argue that after enough work done, the ship of Theseus isn't the same ship that sailed into the dock. But the important part is that it's still the ship of Theseus, no matter how many parts you iteratively replace.)

Consider cases like going after someone who took GPL code and republished it as proprietary, or plagiarism, or copyright infringement itself. For the law, it doesn't matter whether or not the bits you have are identical to those of the protected work; what matters is how did you get them. It's the same principle at work here.

> In this case the "meant for copyright infringement" part seems to be some of the unit tests.

In this case, unit tests are evidence that the whole project is meant for copyright infringement. The offending entries serve to establish intent.


> unit tests are evidence that the whole project is meant for copyright infringement

Minor quibble, it's not so much infringement (Wouldn't that require redistribution? Or that the original upload itself be a violation?) as circumvention of a mechanism as described under DMCA section 1201.

(But does _generally_ circumventing a mechanism that is only _sometimes_ used to protect copyright actually run afoul of this?)


People seem to continually be misinterpreting this as copyright infringement, and you are right that it isn't. The fact that you bypassed the viewing of ads changes nothing in this matter.

Otherwise, if we would allow this interpretation, then the entirety of modern internet technology is potentially infringing. This is ludicrous.

Nothing substantially different happens when you view the video in your web browser than when you download it using youtube-dl (mpv, the media player, even wraps youtube-dl so you can directly stream a youtube video). In both cases HTTP requests are made, followed by some processing of the data. Does that mean adblocking software is infringing too?

I don't know, maybe the US has totally lost its mind and this kind of thing would fly there, but certainly not in Europe.


It is copyright infringement (via the argument that ytdl facilitates and encourages others to commit the infringement), which happens to bypass ads.

The rights holders have effectively said "you can watch this for free as long as we get the ad revenue". Theoretically, if there was some way of preserving that ad revenue while committing that infringement, I would assume they wouldn't care about the infringement. They're just using the infringement as a tool to preserve the ad revenue.

What you're worried about is the converse, if there is a way to avoid infringing on copywrite while bypassing ads (which there is, ad-blocking), then there's no infringement to use as a tool, then they're going to want to make ad-blocking illegal. It's not hard to predict that's coming, especially in Europe given Article 13 last year.


I'm actually arguing that it is wrong to claim that a copy is even being made by the mere act of streaming (i.e. transferring the bytes of) a media file and hence you cannot violate copyright by merely downloading something. You only infringe once you distribute the content further.

In fact, as far as I'm aware of, this is also how the situation plays out de facto in most (all?) European countries.

> The rights holders have effectively said "you can watch this for free as long as we get the ad revenue".

The key lies in the word "effectively" here. They may have said and even meant that, but I am unaware of a legal mechanism that would actually allow them to do this (in Europe). The only way I can think of accomplishing this would be to make the content available under a dedicated, bespoke licence, akin to how GPL works, and even then I'm unsure whether it would be enforceable. TOSes, generally, are not.

> It's not hard to predict that's coming, especially in Europe given Article 13 last year.

This is wrong. Article 13 regulates content hosts (in this case, Youtube), not end users. Hence, Article 13 has no bearing on ad-blocking software.


> cannot violate copyright by merely downloading something. You only infringe once you distribute the content further.

Wrong. Copying something, without distribution, is still violating copyright. It's just taken less seriously than distribution. It's the same difference between drug dealers and drug users.

>> It's not hard to predict that's coming, especially in Europe given Article 13 last year.

> This is wrong. Article 13 regulates content hosts (in this case, Youtube), not end users. Hence, Article 13 has no bearing on ad-blocking software.

No, I did not say anything at all about Article 13. Please re-read what I said.


> No, I did not say anything at all about Article 13. Please re-read what I said.

What? You very clearly stated that the content of Article 13 makes it easy to predict that the EU will attempt to make ad blocking illegal in the foreseeable future. (Granted, the response to that didn't make much sense - the current law doesn't have to directly affect end users for us to make a reasonable prediction based on it that laws proposed in the future would attempt to do just that.)

> Wrong. Copying something, without distribution, is still violating copyright.

That really depends on the jurisdiction and context. For example, in the US, making backup copies of materials licensed in perpetuity (ex a movie on VHS) is (always, AFAIK) permitted. Another example is recording broadcast TV for the purpose of time shifting it (ie watching it later), which has been explicitly permitted by the courts here. Yet another example is ripping CDs for personal use, which falls under fair use in the US and is therefore not a violation.


> That really depends on the jurisdiction and context.

No it doesn't. If you haven't legitimately obtained Taylor Swift's Shake it Off, you can't argue the copy you made with youtube-dl is a permitted personal copy, in the US or anywhere with copyright laws.

If you did pay for it, then yes you can argue it's a permitted copy (just not in the UK, which has stricter copyright laws).

(this is mostly a copy paste of this comment https://news.ycombinator.com/item?id=24882280).


>> That really depends on the jurisdiction and context.

> If you haven't legitimately obtained ...

I'm well aware and never claimed otherwise. I very clearly stated that it depends on the context. (It's right there in the text you quoted!)

I was responding specifically to your previous claim that "Copying something, without distribution, is still violating copyright." which was overly broad for the reasons I specified.


Yeah, sorry I knew it was overly broad and should have been more specific. But it was in response to your completely wrong assertion that violating it requires redistribution. I think we're done here.


> your completely wrong assertion

Oops, apologies Reelin I wasn't paying attention to who said what, that was feanaro who was wrong.


> If you haven't legitimately obtained Taylor Swift's Shake it Off

Obtaining Taylor Swift's Shake it Off from a public YouTube video without watching ads is a legitimate way of obtaining it. That is exactly the point.


You are just wrong. Please go learn about copyright. Then you might understand why this takedown happened.


I like think you are wrong and I have presented my case previously, to which you haven't responded. Nothing in copyright law itself can force you to watch ads while downloading a public video. I am completely sure this is the case in my own jurisdiction. If you want, I would be curious to hear your argument about why you think it does in yours.

This takedown happened because DMCA is a broken, frequently exploited law which allows for illegitimate takedown requests, which you must abide by or risk exposing yourself to legal damage for no reason at all.


> to which you haven't responded.

Where? I responded here: https://news.ycombinator.com/item?id=24881656

I can only repeat what's there.


> Granted, the response to that didn't make much sense - the current law doesn't have to directly affect end users for us to make a reasonable prediction based on it that laws proposed in the future would attempt to do just that.

You're right. I assumed the OP was arguing that Article 13 itself somehow regulates ad blocking software, which is the only way I could interpret it so as to not be a non sequitur, but in doing so I made a non sequitur myself.


I never said Article 13 bans ad blocking. What I wrote was too difficult to parse and I should have realised that at the time and made it more clear. E.g.:

"It's not hard to predict they (EU and media corps) will want to make ad blocking illegal in future with a new law following on from Article 13." I could have said something more vague like "not hard to predict it given the direction the world is headed at the moment" to make the same point.


> Or that the original upload itself be a violation?

The original uploads are on the artists' channels and are monetised (i.e. when I accessed 2 out of 3 of them, an ad played, so I assume some of the cash the adverstisers paid Youtube somehow wound up at the relevant copyright owners).


> Wouldn't that require redistribution?

(sorry meant to reply to this earlier)

No, copyright infringement doesn't require redistribution. In the UK if I buy a physical CD and rip it to a PC, purely for personal listening, that's still infringement (there have been attempts to change the law but they failed: https://www.gov.uk/government/news/quashing-of-private-copyi...)


I was unaware that copyright in the UK was so extreme. That's unfortunate.

The topic at hand, however, is a DMCA action between two US entities (the RIAA and GitHub) and so is purely a matter of US law AFAIK. My understanding (possibly mistaken) was that the courts here had nearly always permitted making personal copies of otherwise legitimately obtained media. In fact, my understanding is that reversing this status quo was one of the primary motivations behind the DMCA; by disallowing circumvention of protection schemes, in many instances it effectively outlawed the tools needed to make otherwise permitted copies.


> otherwise legitimately obtained media

That's the issue. If you haven't legitimately obtained Taylor Swift's Shake it Off, you can't argue the copy you made with youtube-dl is a permitted personal copy, in the US or anywhere with copyright laws.

If you did pay for it, then yes you can argue it's a permitted copy (just not in the UK).


> Not parts, but the entire thing.

Which would include the HTTP library and the system call implementations it makes into the operating system etc., and you're back to an intractable mess.

> If youtube-dl gets classified as illegal, then any trivial modification to it will get the same treatment. Even if you end up slowly replacing every bit of code, if you forked off youtube-dl and didn't change the functionality, that's still essentially youtube-dl.

Play it backwards and see if it still works. Somebody forks Chrome and creates Chrome, Pirate Edition. The only difference is the logo and that the third party distributors overtly promote it for copyright infringement. It gets declared "illegal" but it's still 99% Chrome, and Chrome is still 99% it. Is Chrome now illegal? Is the shared 99% of it? If not, how are you drawing the distinction?

> Consider cases like going after someone who took GPL code and republished it as proprietary, or plagiarism, or copyright infringement itself. For the law, it doesn't matter whether or not the bits you have are identical to those of the protected work; what matters is how did you get them. It's the same principle at work here.

Except that copyright actually has some well defined mechanisms for determining which aspect of a work is infringement if you copy it. If you publish a book of poems and one of them is in the public domain, you can still copy that one. And the author of each of the other poems still owns the copyright to that poem, even if they're all published together in the same book.

So how does that apply in this case to an open source project with multiple contributors and components?

> In this case, unit tests are evidence that the whole project is meant for copyright infringement. The offending entries serve to establish intent.

Except that the project has multiple independent contributors each with their own intentions, so "the project" doesn't have any singular intent.


There's nothing fancy in youtube-dl. It's just a bunch of ad-hoc scripts maintained by folks which make it convenient to route the api responses to files on disks. Literally things like pulling html and regexing for certain snippets. In fact it supports many video services, cspan for one. It just dispatches to the appropriate scraper based on the endpoint.


The code is not tainted. It's code not some religious relic.


So rebase it and convert to rust.


Agreed, though the RIAA may well have sent a letter anyway. There's not really any downside to them sending a takedown that is overly broad.


No reason it can’t be hosted elsewhere with that name after a fork.


It definitely was very sloppy of them to show examples on actual copyrighted material. Couldn't they have simply put examples of public domain content?


JFYI i was wrong. There were no such examples in the README or “marketing material,” there was URLs and metadata in a few unit tests.


Not that big of a difference, since I see unit tests as examples, still sloppy either ways.


It's not "marketing," though, which is a key difference.

Programmers put weird things in unit tests all the time. Maybe the contributor was making a joke.


It doesn't have to be marketing. You're showing by example how your app can be used, and one of your examples is to download copyrighted content.


It would be interesting to see _who_ added this unit test. I wouldn't be surprised if it was some RIAA lackey.


That's quite a paranoid conspiracy theory you have there. The first mention of one of the urls (Icona Pop) in the git log is this:

commit f7ab6cbe160afbba60537c7a830a4c65c6f0b3ea Author: Filippo Valsorda <filippo.valsorda [at] gmail.com> Date: Tue Jun 25 14:38:00 2013 +0200

    add tests for use_cipher_signature videos (#897) and the ability to test multiple videos per IE
diff --git a/test/test_download.py b/test/test_download.py index 9af626dca..067bde4bb 100644 --- a/test/test_download.py +++ b/test/test_download.py @@ -153,9 +153,11 @@ def generator(test_case): return test_template

### And add them to TestDownload -for test_case in defs: +for n, test_case in enumerate(defs): test_method = generator(test_case) test_method.__name__ = "test_{0}".format(test_case["name"]) + if getattr(TestDownload, test_method.__name__, False): + test_method.__name__ = "test_{0}_{1}".format(test_case["name"], n) setattr(TestDownload, test_method.__name__, test_method) del test_method

diff --git a/test/tests.json b/test/tests.json index fd037d818..f6a70f153 100644 --- a/test/tests.json +++ b/test/tests.json @@ -11,6 +11,32 @@ "description": "test chars: \"'/\\ä�<86>��<9D><95><90>\n\nThis is a test video for youtube-dl.\n\nFor more information, contact phihag@phihag.de ." } }, + { + "name": "Youtube", + "url": "http://www.youtube.com/watch?v=1ltcDfZMA3U", + "file": "1ltcDfZMA3U.flv", + "note": "Test VEVO video (#897)", + "info_dict": { + "upload_date": "20070518", + "title": "Maps - It Will Find You", + "description": "Music video by Maps performing It Will Find You.", + "uploader": "MuteUSA", + "uploader_id": "MuteUSA" + } + }, + { + "name": "Youtube", + "url": "http://www.youtube.com/watch?v=UxxajLWwzqY", + "file": "UxxajLWwzqY.mp4", + "note": "Test generic use_cipher_signature video (#897)", + "info_dict": { + "upload_date": "20120506", + "title": "Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]", + "description": "md5:b085c9804f5ab69f4adea963a2dceb3c", + "uploader": "IconaPop", + "uploader_id": "IconaPop" + } + }, { "name": "Dailymotion", "md5": "392c4b85a60a90dc4792da41ce3144eb",


> That's quite a paranoid conspiracy theory you have there.

RIAA people have been known to do it, you know.


[citation needed]


One could argue that such a test is used only out of necessity - music videos by publishers are unlikely to have their title or metadata changed, which makes for a good static test.


He could have just uploaded a copy of Big Buck Bunny to an empty YouTube account just for tests...


I agree that they probably shouldn't have done this legally but programming wise, those videos are special (and that's why they are in tests) and looks like only VEVO can upload special videos like that and not a random pleb.


youtube-dl going out of its way to support VEVO's videos is kinda evidence against the project, seems to me...


What did the readme say that has no legitimate use? Because even downloading Disney can be legitimate.

LockPickingLawyer teaches you how to do things that could be torts, but there are circumstances where those same actions are entirely legitimate. Tort generally needs the context to show that something is not legitimate; so the same applies here. AIUI copyright is not a situation of strict liability.


youtube-dl aside, the LockPickingLawyer is an amazing channel.


I watched a horror movie once, the protagonist had hidden in a fortified escape room and was holding their breath.

Then, the scariest thing ever could be heard

"Click out of 3, nothing on 4."


Hm, but youtube-dl doesn't break DRM. You can just get the video link from the dev tools panel when you open a youtube page. It doesn't let you download anything you wouldn't otherwise be able to.


DRM is ultimately mostly obfuscation. And access to the video fragments in YouTube is heavily obfuscated.

There's very little case law on what the exact level of protection required to qualify, but I wouldn't want to go to court defending myself offer it. It's very risky.


Have you actually tried that lately? Particularly with a DASH stream? (Not that DASH itself is intended as DRM, but just that it's no longer anywhere near as simple as querying a single URL.)


Nonetheless, the underlying point holds. If youtube didn't want people downloading it's videos, it's trivial from them to close that up.


It isn't trivial for them to break platforms that won't get new YouTube apps.


YouTube have only been permitting YouTube apps on platforms that support Widevine DLM for a long time now.

I suspect they are just waiting for the percentage of older clients in use to fall a little bit further, but the plan is pretty clear.


I keep refusing to enable DRM on Firefox. Hopefully there are enough people like me to matter.


They did do that. That's why youtube-dl exists..


youtube-dl doesn't "hack" into Youtube. The process isn't anything that you can't do manually using chrome dev tools.


Congratulations on being the millionth person to say that, not one has given instructions on how to do it.

How would you go about putting a DASH stream together in dev tools?

Side note I didn't say hack, nor did the parent. Who are you quoting?


Except it’s not in the README, it’s in an obscure unit test. That’s not marketing.


But what is being circumvented? As far as I'm aware there are no controls put in place that are being circumvented.

Edit: I'm also curious why the riaa would have standing at all? Even assuming there is some protection being circumvented, it's not the riaa's control that is being circumvented.


A german court ruled youtube's obfuscation of video request signatures is an effective technological protection measure. YoutubeIE._decrypt_signature in youtube_dl/extractor/youtube.py bypasses this.


As a philosophical issue, since the youtube-dl code is running the cipher off of the code provided by youtube, just as my browser is, why is my browser not similarly circumventing this protection?

At least with CSS a key needed to be found by means separate from actually watching the DVD and wasn't shipped with the user in the same package as the video itself.


> why is my browser not similarly circumventing this protection?

Because your browser also runs the javascript that displays the ads (unless you've also blocked them, which I think could also be considered a "circumvention measure", IANAL).


If ads we're required, they could certainly not stream a video until the ads have been watched (or at least downloaded and the correct amount of time has passed), they could also insert ads as segments of video without having to transcode all videos for every user, but they don't.


> they could also insert ads as segments of video without having to transcode all videos for every user

That would work in a livestream very well, but not in a video. Imagine everyone getting adverts in the voice of live streamer in the different times of the livestream, and live chat getting time-dilated to compensate, without the streamer itself noticing.


Sorry this isn't really an argument. How can someone not make the argument regarding the browser in court and not have it align with the German court's prior finding?


Saving videos


Youtube-dl doesn't do anything your browser doesn't do already do when accessing said video.

Also, I asked what is being circumvented, not what the tool does.


It boils down to the intention. The intention is to download videos to your computer (as viewable mp4 files for end users) which isn't legal, as they are intended to be streamed (or possibly removed in the future) with ads, stats, analytics, recommended etc.

I definitely don't defend this and I know that this whole copyright and DRM thing is fundamentally flawed by design, but it unfortunately works this way legally.


If it boils down to user intention then you can't also argue the tool is circumventing anything. Either mimicking a browser and working the way google expects any client not written or authorized by them, such as a web browser, to is ok or it's not.


I don't think saving things is in general a copyright violation?

Knowingly and intentionally obtaining unlicensed content is a violation. Unauthorized redistribution is a violation. Saving and copying things isn't (AFAIK) a violation.


Saving things is creating an unauthorized copy, which is fundamentally the basis of copyright. Some territories have explicit allowances for personal copies but many don't.


There are a variety of reasons that you can have a copy of a copyrighted work it a way the copyright holder didn't intend, such as time shifting or media shifting.


And in many juristictions they are explicitly defined, and don't apply to copyrighted work generally, only when acquired under specific circumstances.

Format shifting is explicitly illegal in the UK for example, and timeshifting is restricted to works obtained from a broadcast source by the person performing the copy, and not to an on-demand one, in many territories.


Although the OS might do that as well when you watch a video and put the computer into hibernation mode.


isn’t DMCA a US law? so why not move youtube-dl outside the US github to a country that doesn’t have DMCA/copyright?


The US decided their laws, specifically the DMCA, applied globally back in 2001 [0]

[0] https://en.wikipedia.org/wiki/United_States_v._Elcom_Ltd.


I've always struggled with the idea of laws that seemingly seem to overstep their jurisdiction.


When it's corporations, they have special rights.


the US is free to decide whatever they want. that doesn’t mean it will also be accepted by other states.


If you don't comply they then threat to start trade warz and sanctions, simple deal IMHO


Then host it in a sanctioned state like Cuba or Crimea.


This makes me wonder: if I fork the project into another product without such marketing it in this manner, and BFG away all troublesome uses (e.g. in the tests) the git history, it would then be legal, yes?


I was thinking the same thing: The next step following this line of thinking would be trying to ban all torrent clients, because they CAN be used to download copyrighted material.

This is crazy.


Why stop with torrent clients? They should ban all browsers, because they CAN be used to download copyrighted materials!


Browsers?

We need to ban radios with tape decks and TVs with VCRs! The audio cassette and the VHS are bringing the downfall of the music industry with how easily they can be used to create unauthorized copies!

Oh wait.


i love how dead kennedys literally left one of their cassette tapes sides blank and printed "Home taping is killing record industry profits! We left this side blank so you can help." on it.

https://en.wikipedia.org/wiki/In_God_We_Trust,_Inc.#Artwork_...


Thanks for the throwback


I realize you're being sarcastic but with a line of logic like that, you have a bright career in Congress, should you want it.


Well, following their logic behind DMCA compliant, I guess fiddling long enough with developer's console on Chrome or Firefox would allow you to download potentially copyrighted material from youtube the same way youtube-dl does.

sed s/youtube-dl/firefox/g and voilà, DMCA for Firefox ready to submit...

Let's go further! Let's DMCA the Linux kernel because it runs Firefox/youtube-dl/curl/wget!


The logic behind the DMCA complaint is that youtube-dl is “marketed … for use in circumventing a technological measure that effectively controls access to a [copyrighted] work” [1], not that it “allows you to download potentially copyrighted material from youtube.”

[1] https://www.law.cornell.edu/uscode/text/17/1201


But it wasn't. This was just a unit test.


A unit test is an example of how the program is intended to run. It's the intention.


The hell it isn't. A unit test is a bench test for one particular functional unit. It's a check function. If you accept that tests as carrying intent, then I demand every weapons researcher or biologist to be arrested this instant. Clearly they are only out to murder people. That is patently ludicrous.


> If you accept that tests as carrying intent, then I demand every weapons researcher or biologist to be arrested this instant. Clearly they are only out to murder people.

Well, what do you think weapons are for? I’m not saying arrest them but that isn’t really a great example, considering the sole purpose of weapons is to kill or destroy. Using weapons for anything other than killing or destroying is basically a secondary usage.


One of the tests in question extracts the m4a audio file of a Taylor Swift song [0] that has nearly 3 billion views and shows me 2 ads before playing. That's a lot of revenue for UMG, who I guess are a bit unhappy with this mechanism that youtube-dl provides of listening to it without having to watch the ads.

[0] https://www.youtube.com/watch?v=nfWlot6h_JM


So it's only a problem if the video has a lot of views and ads? If that's the case, we should ban all Pi-holes, too, as they _can_ be used to block ads. What if I have YouTube Premium, then I don't see ads, so am I allowed to download via youtube-dl?


> So it's only a problem if the video has a lot of views and ads?

Pretty much, yes. No one is going to hire expensive lawyers to enforce copyright on something not generating revenue, although legally they can.

> If that's the case, we should ban all Pi-holes, too

If Pi-hole is anywhere saying or implying "use us to watch $vendor's content without watching their ads" then yes pi-hole is waving a red flag at $vendor who is now motivated to try to ban pi-holes. If there is a legal action they can afford to take they will.

> What if I have YouTube Premium, then I don't see ads, so am I allowed to download via youtube-dl?

No, downloading is a separate "right" from the right to watching without ads. (Downloading is the first step before uploading elsewhere and literally stealing the ad revenue.) But I think there's a legal grey area if you're watching (streaming) with youtube-dl, without saving the file. I'm sure it's against Youtube's TOS though.


> literally stealing the ad revenue.

Are you sure you want to say it's literally stealing the ad revenue? I would imagine literally stealing the ad revenue would require robbing the bank or safe where the ad revenue money is stored.


Yes. You have reduced the amount of cash they have and increased the amount you have by the same amount. If that's not stealing I don't know what is.

Trying to say that's not stealing is like trying to say "Someone hacked into to my bank account and "obtained" my money" is not stealing either just because there was no physical cash involved.


How does it increase the amount of cash I have? I watch the video, without ads, and I have exactly the same amount of cash as before.

It also does not decrease the amount of cash owned by the content owner or hosting company (apart from hosting costs, but that's the nature of the internet and they are willfully participating in it). The only cost involved is opportunity cost.



Yes, you are indeed talking about the opportunity cost. Re-uploading the video as your own (with or without ads) would indeed constitute copyright infringement, but without it, it is not.


I'm not sure this fits the definition of opportunity cost either? I understand why you think it does (because it was money not gained that could have been) but the content producer doesn't have multiple options to choose from here. They expect to have X dollars in their account, instead they have X-Y dollars and there's Y dollars in thief's account, because of the actions of the thief, not because of any decision they made.


The crux is in the expectation: they expect that they will make X in a given day, but an expectation is not a guarantee or a right. They also could have made X in a given day, but sometimes that just doesn't happen due to various factors.

I'm not sure to whom you are referring when you say "thief", but I was under the impression that we were talking about a person watching, but not re-uploading, a Youtube video without viewing ads. This person does not have Y dollars in his account as a result of copyright infringement since they simply watched the video and did not put it up for further distribution.


I really just shouldn't have mentioned ad revenue theft at all in the first place, it wasn't that relevant.

I wasn't saying "people who download are thiefs", I was saying "thiefs start by downloading". Then I had to explain what a thief is, and why they really have Y dollars in their account that should have been in producers account. In my other comment the thief is a separate party from the content consumers (so actually there are 6 parties, content consumers is split into those who watch ads and those who download without re-uploading).


How does watching an ad decrease the amount of money I have?


> How does watching an ad decrease the amount of money I have?

There are at least 5 parties here:

    · Content producer (spending time and effort making original content that people watch)
    · Platform
    · Advertiser
    · Content consumers
    · Thief
In your comment you're Content producer. The advertiser pays the platform who pays the content producer, and the amount they get is based on the number of views (there are other factors, but I think views is the main one). Let's say they're getting X views per day. Then the theif takes the content and uploads it to their own account with ads enabled, and gets Y views per day, and gets paid for that. No one is going to watch the content on the producer's channel a 2nd time just to make up their view stats, so now the producer is only getting X-Y views per day and hence is paid less.

This is the reason Youtube's Content-ID exists.


The illegal act here would be re-uploading the video. I'm asking how downloading the video by itself increases the amount of money I have.


It doesn't, and I never said it does. You asked for an explanation of how ad revenue theft happens, and I gave that.


No, it's not.

A unit test is there to allow the developers to test whether the program is producing expected output from certain inputs.

It doesn't demonstrate that it's intended for circumventing copyright.

They just picked examples to use for in test cases.


The Linux kernel does not have "youtube" in its name–it isn't advertised as a way to download videos.


So if youtube-dl had a different name it'd be ok?


I'm not a lawyer, and even if it was I could never tell you that anything would be "ok" in this kind of situation. But I suspect that it would be better off.


> Why stop with torrent clients? They should ban all browsers, because they CAN be used to download copyrighted materials!

We're talking about an industry that normalized rootkit DRM and persuaded pretty much all hardware manufacturers to implement expensive encrypted pipelines to lock you out from trying to even take a glimpse of content playing.

After a few years of working near this industry I really cannot overestimate how rentseeking and scummy every single part of it is. They will abuse you without a second thought if there's a cent to be made.


All computers with the option to install applications or unlocked/unlockable bootloaders can be used to make copies of copyrighted materials. And it certainly happens a lot.

But then, when secure boot came, Microsoft was forced to grant every owner of an Intel compatible computer the right to unlock it for antitrust reasons (luckily).

So were is the borderline between a legal tool and an illegal tool? Well, lawyers will find out...

In the meantime youtube-dl needs to be distributed some other channel. Which of course might raise the risk of back-doored or otherwise poisoned version floating around.


Why stop with browsers? With enough creativity you can use any network protocol to download copyrighted material, so they should just ban computer networks.


Because the law doesn't take kindly to "well, technically" arguments.


So what does it mean when someone says "he got off on a technicality"?

There have been plenty of cases where exactly that happened.

There's the spirit of the law and the letter of the law, and many are persuaded that when someone hasn't violated the letter of the law, they haven't violated the law.

Judges, lawyers, and juries will vary on this, but in many cases the letter of the law will carry the day, and people will in fact "get off on a technicality".


Because browsers and network protocols were not created with the primary purpose of violating copyrights...

Intent matters in the law.

After after all, two guns can be physically and functionally identical, but if one is purchased to be used in a conspiracy to commit murder then it's evidence of a crime, and the other gun is still just a gun.


Any input or output device


No joke -- if they could get the legislation signed into law they would be thrilled.

They could care less about side-effects and collateral consequences for adjacent technologies and use cases.


They would prefer it if you only used DRM input and output devices that only run approved code rather than general purpose devices that can run any code you can conceive of or download [1]

[1] "The Coming War on General Purpose Computation" https://boingboing.net/2011/12/27/the-coming-war-on-general-...


Another abuse of DMCA on the frontpage earlier: https://news.ycombinator.com/item?id=24867212

Indian startup taking down dissent, whistle blowers and leaks using take down requests to youtube and social media.


Same for any kind of information exchange, like speech or the written word. Common sense says their argument doesn't hold water at all. Should courts decide in favor of the takedown that'd just go to show how broken our legal system is. Let's hope not.


And you don't even need a browser for that. Your operating system comes with TCP support and that in itself is enough. (OK, some of them may probably be willing to build in all kinds of restrictions, so maybe they can get a RIAA stamp.)


Let's just ban internet protocol, while at it. :)

On a bit more serious note I wonder how much internet traffic is used for "illegal" purposes?


Why stop with browsers? They should ban all computers, because they CAN be used to download and play copyrighted materials!


We need to ban oxygen. It CAN be used to download copyrighted materials!


>we need to ban boron/phosphorus

FTFY


I think we should just replace all humans with an Ethereum smart contract. No humans, violation of law.


The next step following this line of thinking would be trying to ban all torrent clients, because they CAN be used to download copyrighted material

Didn't the RIAA put up a considerable effort against torrenting in the early 2000's? I remember quite vividly them going after just about anyone who made music available, even if you owned your own songs and were simply uploading them to a web UI to listen to in the browser (Remember Muxtape?)


Hush, don't leak their master plan! Of course they'd love to just ban about anything that can potentially be used to commit copyright infringement. Of course that includes your web browser, too. Unless it can only visit a list of websites pre-approved by RIAA ;)


Why not just ban computers or even why not jail people who even think about using ytdl? These organisations like RIAA need reality check. Sadly there is no body to stand against their bullying and stiffling the freedom of speech.


EFF takes up these causes sometimes, right? This notice is hollow sabre-rattling, so I wouldn’t worry too much. What can RIAA do if somebody forks the project and hosts it in the Lithuania? Not much. Yell into a pillow, maybe.


youtube-dl needs to be regularly updated as streaming sites add new countermeasures to block it. If those updates can no longer be crowdsourced from a popular, reputable site like GitHub, the tool has a good chance of dying.


This is the crux of it; finding something that is both DMCA-proof and trustworthy/reputable enough to allow auto-updating will be a huge barrier :(


The RIAA would absolutely ban all torrent clients if they could. If one advertises itself as a way to download copyrighted material, they would try to pull the exact same thing.


What about chromium? They should take that down as well, imagine what children can see using that thing. `curl`, `wget` as well as gitlab - massive offender, random number generators - one of the biggest offenders, any worst nightmare you can imagine can be produced by it!


And don't forget the evil math people can use to hide their illegal content!


Well, we (Lithuania) have a law that works along these lines. When you buy any storage device a part of the price is a tax that goes to IP protection agency. The reasoning is exactly that: because you may store some pirated music or whatever on it. How was this bs passed I have no idea.


This happens in Spain as well. About how did it pass, I assume it's because lobbying.


Now that 4chan is behind bars, it's time to sue the Internet!


Nope, just to sue "the hacker known as 4 Chan".


Yes, would be messed up given the state of play upon gun laws.


Does it really matter what they said?

This is like the reverse of those common disclaimers on hacking tools and tutorials which claim:

  This is for educational use only.
I've always wondered, are courts fooled by such disclaimers? Are their authors untouchable just because they put in some boilerplate disclaimer like that?

As for what was written in some code somewhere in the repo, it could have been written by anyone, even an RIAA plant who contributed that trojan horse to the project.

What makes anyone think what was said there is endorsed by or even representative of the views of the rest of the authors of youtube-dl or is what youtube-dl is for?

God, it pisses me off to no end to think that I'm going to be forced to use youtube's piece of shit, slow, ad-infested, tacker-infested, feature-poor, non-automatable browser interface.

But, after taking a few deep breaths, I think within a year there'll be multiple alternatives to youtube-dl which will all bear disclaimers that they are "for educational purposes only" and that they in no way endorse copyright infringement.


>I've always wondered, are courts fooled by such disclaimers? Are their authors untouchable just because they put in some boilerplate disclaimer like that?

I think the answer is that this doesn't always sufficiently protect them, but these small bits add up. It's useful to be able to point at words written beforehand that could help you in a legal situation. This is what corporate speak is about too and enough people engage in it that it likely does something.


> I think within a year there'll be multiple alternatives to youtube-dl which will all bear disclaimers that they are "for educational purposes only" and that they in no way endorse copyright infringement.

i dont think you have to wait that long. its client side, its open source, people have the code, it will most likely be maintained further, just not on github.

this is just meant as a backup, but here you go:

https://git.svarun.dev/ytdl-org/youtube-dl

or get the source with

    pip download --no-binary=:all: youtube-dl


> I've always wondered, are courts fooled by such disclaimers?

No. And there's no educational exception in the DMCA anyway.


The use youtubeDL for music would probably fit the fair use in Finland. The fair use allows personal copies of legally licensed material but not redistribution of said copies to third parties. It has been used to defend a service where ISP records the specific shows from TV like a VCR and streams them to customers at different time. The ISP:s and representation of copyright holders made a compromise to avoid legal battle, where ISP:s limited the storage duration of said recordings to two years. ISP:s still had service they could sell, and avoid unnecessary legal battle. I'm not a lawyer. So not expert on matter, but just educated local.


Why not host youtube-dl in Finland?


Seems like putting that in the readme was pretty stupid. Intent matters in legal things, making the example infringing undermines the argument that the tool is good and some people are just bad.

> This feels like DeCSS all over again.

I think napster would be the better comparision (and especially napster compared to vcrs)


They did not put that in the readme. The README contains only references to test videos, videos that don't actually exist, and one small-time video (in spanish?) that seems to be an old test video, but I have a hard time seeing for sure.

The source code does contain references to copyrighted videos in the tests, tests intended to make sure that youtube-dl can download the data from videos using the extremely "token" signature scrambling that youtube employs for certain videos. You can see the test cases here:

https://gitlab.com/HacktorIT/youtube-dl/-/blob/master/youtub...


There's stuff like Justin Timberlake in there and Taylor Swift. This was immensely sloppy considering how the RIAA has won cases before.


But from what I can tell, these unit tests don't even download the video - just pull metadata.


I don’t see how Napster is more applicable than vcrs. YouTube in this case is the broadcaster, putting content out to a general viewing public, ytdl is a recording device. For Napster to work, YouTube has to be imputes with illegally providing the copy and ytdl has to be the means to facilitate the copies transport.


Napster and VCRs were both accused of being tools to infringe copyright, VCRs won the lawsuit, napster lost. A big part of that related to how napster positioned itself in terms of marketing compared to vcrs.


> it uses examples in the README as an example of that:

>> We also note that the source code prominently includes as sample uses

as far as I can tell these videos are not referenced in the README, but instead in the youtube.py extractor file, which would go against the accusation that they were featured 'prominently'


Indeed. Those videos are three out of 94 test cases scattered among youtube.py.


It makes it worse lol. Using links to copyrighted materials in the tests meant to verify correct functionality?

It's "prominent" in the source code. They literally wrote out "Test VEVO video with video age protection" then linked to a Justin Timberlake song

https://gitlab.com/HacktorIT/youtube-dl/-/blob/master/youtub...

I clicked on this ready to see some stupidity from the copyright industry but literally downloading from links to copyrighted music in your unit tests?

That's how you give the RIAA a way to take down your project. -

Actually seeing this made the comments made in the worst faith about how "they took it down because you can use it to download bad stuff, what's next?!?!" seem pretty dumb.

If that was the case Bittorrent... which makes youtube-dl look like a guy slinging tapes out of a van next to Naspter HQ...

wouldn't have been able to send them packing with a simple "we don't endorse piracy": https://venturebeat.com/2015/08/07/bittorrent-to-riaa-youre-...


>It makes it worse lol. Using links to copyrighted materials in the tests meant to verify correct functionality?

Using copyrighted materials in test code is fair use and a well established practice:

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

The creators of youtube-dl are not distributing the material or inducing infringement by using a link to it in a unit test. I am typically on the RIAA's side,

https://news.ycombinator.com/item?id=23079092

but they're going to lose this one. If RIAA should be mad at anyone, it is youtube for distributing it without DRM. Youtube really really REALLY should have had their asses taken to the cleaners in viacom vs youtube. I don't know what the settlement was, but it wasn't enough, because youtube still exists.

https://en.wikipedia.org/wiki/Viacom_International_Inc._v._Y....

This is a python program doing exactly the same steps a browser takes to download the file. It's just written in python instead of js.


Here, though, Youtube (Google) have contractual agreements to license the use of copyrighted media on their platform, and which is uploaded directly from the record labels themselves.

The RIAA is claiming that YouTube does have DRM on their videos, which if the court agrees, may be the end of youtube support in youtube-dl.


It doesn't have DRM because youtube download only extracts the direct download link and downloads them. There is no DRM on youtube. Example from the alleged offending link,

youtube-dl --skip-download --print-json https://www.youtube.com/watch?v=UxxajLWwzqY

You can easily copy the links from the formats into your browser and they will play, you can wget them, anything. There is no DRM there, and if anyone convinced you otherwise, they lied.


> Using copyrighted materials in test code is fair use and a well established practice: https://en.wikipedia.org/wiki/Lenna

What kind of nonsense is that? Playboy did fight people for using it, the eventually let it go because it was one image that wasn't particularly important them. It's literally right there in the article you linked...

And I assume you forgot to link to this other program but again, not exactly a real argument.

RIAA went after them on the use of that link, if other projects (of which dozens exist by the way) didn't slip up like that and are still here, it only makes it clearer what the problem was


They can argue that downloading a publically available video for testing purposes is perfectly legal, because it is.


"You also grant each other user of the Service a worldwide, non-exclusive, royalty-free licence to access your Content through the Service, and to use that Content (including to reproduce, distribute, modify, display, and perform it) only as enabled by a feature of the Service." https://www.youtube.com/static?template=terms

It is enabled by a feature of the Youtube service.

They explicitly disallow downloading, but it is not enforcable (though that word probably doesn't mean what I think it means... the law probably overrules physics)

Severance

"If it turns out that a particular term of this Agreement is not enforceable for any reason, this will not affect any other terms."


But in that wording they do allow downloading. Features of the Youtube service accessed thought the Chrome browser don't have to be exactly the same as accessed though other browsers and tools, other tools are not prohibited to provide features they can extract and interpret from the received source code and data, they are not obligated to run the code at all or as is.


Ugh, I don't want to get into this Stallmanist "if the server returns 200 it's ok!" tarpit. If you believe that, it's fine.

You probably won't ever actually need to use that defense outside of internet arguments honestly... for all their noise the most repercussions the average person might ever see from infringing copyrights is a scary letter from an ISP.

Good luck if the legal system decides to test your hypothesis though.

This isn't about YouTube's TOS at this point, you're making a copy of a copyrighted video without the consent of the copyright holder

> downloading works protected by copyright without the authority of the copyright owner is an infringement of the copyright owner's exclusive rights of reproduction and/or distribution


First, it's not actually illegal, at least in my jurisdiction and in a great many, to make a copy of a copyrighted work.

Secondly, the author somehow gave permission to YouTube to make the video publicly available and downloadable, and YouTube is perfectly happy to provide to me the download URL.


It's illegal in the jurisdiction that Github and the RIAA operate in?

Laws don't exist in a vacuum, so it doesn't seem like a stretch for my comment to refer to the laws actually applicable here right?

Your second point is just remixing the same Stallmanist POV I call out above.

YouTube doesn't just provide you with anything. You go to YouTube, and you find a link, and you stick it in your unit tests.

The fact YouTube's server replies is neither here nor there in the eyes of the law (ironically Youtube does actually return 402 "payment required" and 429 "too many requests" of they figure out what you're doing)

You're free to make whatever esoteric argument you want, but after realizing they did something as silly as put copyrighted videos in the unit tests, there's not much of a leg to stand on.

Especially when there are literally dozens of YT downloader type projects of Github still up right now that didn't do that...


The DMCA is not intended for taking down such tools.

The proper procedure to do so would be to sue the authors, and from what I can gather none of the core authors are in the US.


You saying it's not for exactly the thing it is for... doesn't make it so. I don't know why you think it does.


> this Stallmanist "if the server returns 200 it's ok!" tarpit

That would be an Elbakyanist stance, not a Stallmanist one.


I can't tell if this is an attempt at a crappy joke or you actually think a) Stallman is not exactly the kind of person to make the argument and b) the overwhelming majority of people who prescribe to that line of though aren't aligned with Stallman in other thoughts on technology

Elbakyan doesn't hide behind these kinds of tired non-arguments and instead admits SciHub is illegal and pushes people to take action so it's not: https://engineuring.wordpress.com/2018/07/07/why-sci-hub-is-...


I mean supporters of Elbakyan, when I say Elbakyanism, not Elbakyan herself. Indeed, Stallman also asks for people to take action to end copyright entirely, but he uses a legal tool that twists the law to the opposite of intended purpose, instead of, say, creating a cracked software repository (like Sci-Hub does for scientific articles). Two opposite approaches that have the same intended result.


Ah so you've made up a term for supporters for a person, given them a collective philosophy that doesn't align with their "leader's" much less silly ideas.

My apologies for not divining the tea leaves on your made up term that I guess was not meant as a bad joke, but ends up being one anyways.


> this seems outrageous the same way DMCA'ing a Bittorrent client would be

popcorn time?

https://news.ycombinator.com/item?id=23075484


I wonder what commit added those three listed examples (from three different music companies) to the README, and when. They're the worrying component as they show intent.

In a dark universe timeline somewhere, the pull request for adding them to the README came from an RIAA employee.


A comment somewhere else pointed out that the listed examples aren't listed in the README, but rather as part of the unit tests.

It's confusing, because the letter calls it "the source code", which is not what anyone who knew what they were talking about would usually call it.


That letter is written by lawyers–trying to explain the specifics of "test cases" versus "source code" is unlikely to be a useful use of space. (Plus, if you can get the entire project taken down, why not go for that?)


Do you know of anything that might be applicable here for arguing a distinction between test cases and actual source code? It's a silly example but bringing a replica weapon through airport security is legal, I imagine, if done to test the security system. From a legal layman's perspective, testing on real data doesn't show intent but is instead good engineering to show that the code is verified for correctness.


The point being made in this takedown is that youtube-dl can be used for copyright infringement and is being used for copyright infringement, and the test cases pretty much do exactly that. (In fact, if you're a lawyer with a bit of an idea of software engineering, you might make the argument that the code is testing the correctness of whether it can infringe copyright!)

As an (hastily made, non-expert) example consider a particular gun called the "PandaHunter 9000". Now, pandas being endangered species and such, and having reports that people use buy the "PandaHunter 9000" to kill pandas, you might consider banning the purchase of such a gun. Of course, the manufacturer is going to say that you can use the gun for something else, and its intended purpose is not to kill pandas, but then imagine you find that they've actually been shooting pandas with the gun to test it. Do you see why youtube-dl has put itself in a somewhat unfortunate spot?


I do and thank you for taking the time to reply.

I know my gut reaction to all things IP and copyright (and US) law causes me to lose all rationality, but I'd imagine there's a case to be made that test code is exempt through fair use. I don't think it could be made in this case for the reasons you highlight, but in general as long as the offending tests are a small part of the whole they should be considered separately to the actual code. Like if all my test cases are cracking some DRM then a takedown is probably fair enough, but if I'm using 3 out of 70 plus tests to make sure I handle some YouTube specific algorithm in my tool to download YouTube content then the actual videos I use don't seem relevant. But I get why legally its going to be a huge issue. I just wonder if there might be some case in future where test vs library code is a factor.

Edit: for example in the situation where one of their other tests just downloading a normal video used a copyrighted example instead of a CC example that shouldn't be any different, the content is irrelevant due to it being a test case, it could be any licence, it's just a test, the intent of a test should count.


I had the same thought so I checked the git history and it seems, if I’ve checked the right lines, they were added somewhere between 4-7 years ago.


Without the existence of youtube-dl, youtube itself is necessarily commuting copyright infringement over the a good portion of the CC-By-SA works uploaded to youtube.

(not all of them, since the uploader grants youtube a license... but if the work has copyright holders who aren't the uploader, it applies)


I think in that case the copyright infringement would be the uploader rather than YouTube, but IANAL.


There are CC-By-SA works uploaded to youtube by google employees while at work.


Only if the employee is doing the upload as a function of their job can google have any liability. In any other circumstance google s protected by the safe harbor provisions of the law.


I think you have a good case, if the copyright holder/author is also someone else than the uploader.


That's some prime rate bullshit. Like saying since I could use Linux to download copyrighted materials, and there are guides around which tell how to do so, Linux should be removed from the Internet.

Also a good reminder - never trust a third party to host your files, always have a backup. And hosting your infrastructure on Github means anybody who wants it can take it down with a single letter.


> Like

That's the operative word, because it isn't "like" at all to a nontechinical person. If you explain it as "yeah so I have one tool that literally has 'youtube' in its name, and it downloads videos, also some of the test cases to make sure it works show it work on copyrighted videos" and the other one is "people can use it to watch videos…sometimes? Maybe you can even run the first tool on it? But like a billion people use it for completely different things" you can probably see how this works.


Conversely, the RIAA is making its media available on a known, at-risk platform to attain distribution. They're not protecting their copyright sufficiently, and the copyrights should be revoked.

I can't wait for ML to lower the barrier to entry for music to near nil. Make anyone a vocalist or instrumentalist and hose these assholes.


As much as I would like this to be true, copyright doesn't have to be defended in order to be kept. That's a very different law, trademarks.

You can have your copyright infringed for decades before you try to prosecute the infringers, and the courts will still rule in your favour.


I'm suggesting that ML will lower the barrier to entry so dramatically that the value of individual songs and musicians will plummet. The back catalog of copyright that the RIAA holds will become a fraction of its value today.

I could be wrong, but nascent technology in this field looks incredibly powerful.

I think this will happen with all media.


"ML will lower the barrier to entry so dramatically that the value of individual songs and musicians will plummet"

It's hard for it to plummet down much further.

There's already a glut of content, and the vast majority of musicians can't make a living from their music.

The ones who do well tend to have giant marketing machines behind them to promote their music, and ML will do nothing to change that.


This... doesn't make any sense to me. Are you arguing that songs written by machine learning will become sufficiently good at their "job" that there will be no value in actual humans writing music? And that this will be a good thing?


Yes and yes.

People will still be driving and making money. But there will be more people doing it and catering to a much wider audience.

It'll flatten the curve. All long tail.

Patreon is the first hint at this.


You wouldn't even need ML. Iirc there was this guy who coded a script that played every permutation of 12 notes in a 5 minute period and then made all the Melodys public domain. Ianal so I don't know if this would hold up in court, but technically that should invalidate every copyright on Melodys after that point in time.

E: https://www.youtube.com/watch?v=sJtm0MoOgiU&t=135 found the ted talk


Since there is little to no creative input in merely enumerating permutations, I don't think he has a valid copyright to waive. But the fact that the data set exists might sway a judge or jury away from upholding copyright claims on trivial excerpts of other works.


It seems to me that they are making the claim that this tool is expressly for doing something that is forbidden by the YouTube TOS and breaking the music licenses provided.

Of course YouTube has other videos that are under different licenses so it isn't clear that this is the only use case of the software.


Wait a minute, there's a difference between distributed copyrighted works, and merely copying it for yourself. In France for instance, we have this notion of "copie privée" (private copy), that says we are allowed to copy anything as long as we don't distribute it back. We even pay taxes on persistent memory for the privilege.

Up thread there's also a citation of the Betamax case, which says that it's okay for people to record shows so they can watch them later, and it's okay to sell video recording devices.

This is different from Bittorent, which automatically distribute any content you download. With this too, when you are downloading, you are also distributing, which is a much clearer case of infringement in most jurisdictions.


That's a very common misconception of private copy; the existence of the private copy tax is not excusing copyright violation.

What private copy wanted to cover is more like "I bought a physical CD and I make a cassette copy because my car doesn't have a CD player" or because I want to listen to a mix of my favorite music (so you paid a tax on cassettes and later on CD-R and CD-RW media). This was later extended to "I copy all my collection of music to a hard drive and stop flipping CDs in and out of the player" (so you pay a tax on hard disk drives).


It seems kind of the same to me if the result isn't being distributed again: "I want to view this YouTube video on the airplane when I don't have internet access." (Let's assume no international travel is involved to avoid questions of regional licensing rights.)

YouTube's paid offering in some countries includes this feature on Android and iOS, but for computer users there's no built-in way to do it. Wouldn't it fall within the spirit of the private copying rule for a subscriber to YouTube's paid service to use youtube-dl in this way?


> That's a very common misconception of private copy; the existence of the private copy tax is not excusing copyright violation.

You are wrongly assuming that a copyright violation has taken place. If a country's laws say it's legal to make a copy under specific circumstances, it's not a copyright violation to make a copy under those circumstances.


The issue here isn’t that youtube-dl infringes copyright, it’s that (according to the RIAA claim) its purpose is to “circumvent the technological protection measures used by authorized streaming services.” Even if youtube-dl was tested exclusively on public domain videos, so that the developers never infringe copyright, it could violate 17 U.S.C. 1201 [1] if the RIAA succeeds in proving that it was “primarily designed” or “marketed” for circumventing technological protection measures.

[1] https://www.law.cornell.edu/uscode/text/17/1201


Yup. I said as much somewhere else already:

https://news.ycombinator.com/item?id=24873950


I'm also NAL, but familiar with IP law.

My analysis is at https://news.ycombinator.com/item?id=24874277 and I'm broadly skeptical of the legal viability under US law.


What I haven’t seen pointed out is that this is really Githubs problem. They shouldn’t automatically respond to DCMA. The process should really be passed on to the owner of the repo and make them liable. Basically, defaulting to rolling over to DMCA requests had led to an environment where there’s no way to fight them. Even if I hosted something like YouTube-dl myself, my internet provider could cut me off.

It’s the same issue with deplatforming people, do we want a world where literally you lose the ability to share knowledge others don’t agree with Or in this case, a tool that could be used maliciously


> Basically, defaulting to rolling over to DMCA requests had led to an environment where there’s no way to fight them

That's the DMCA works I thought? When the host receives the notice they have to take down the content then forward the notice to the user, who can then file a counter-claim and then the host restores the content (if they don't take it down they they're liable).


Github and YouTube and most other places immediately take down content. It’s also extremely difficult to reinstate content, if possible at all.


GitHub's DMCA policy[0] is very clear on how to reinstate content. It's not extremely difficult, but you do have to send GitHub a properly-formed counter notice, and you do have to be prepared that the sender of the original DMCA notice may sue you to prevent the content from being reinstated.

[0] https://docs.github.com/en/free-pro-team@latest/github/site-...


actually that it is in the test cases instead of the readme makes it seem worse to me, after all the readme is not addressed to anyone in particular, they needed examples of things someone might do and they used popular videos (that happened to be copyrighted) as those examples.

But anyone running the test cases that does not have rights to those videos will have infringed, and then of course youtube-dl must also run their own test cases and we know they don't have rights to the videos.


IANAL, but regardless, according to RIAA, youtube-dl is circumventing the "DRM" built into YouTube, and therefore they're violating anti-circumvention parts of the DMCA. Therefore a DMCA claim can suffice to take it down.

Perhaps if youtube-dl used different examples (i.e. videos not protected by anti-circumvention) perhaps they could have avoided this.


Incorrect. DRMed YouTube-content cannot be downloaded using YouTube-dl. Only regular content.


Videos with obfuscated request signatures can be downloaded by youtube-dl. A german court ruled this is an effective technological protection measure.


A German court sets no precedent in US law.


> according to RIAA, youtube-dl is circumventing the "DRM"

I didn't find any mention of "DRM" or "Digital rights management" on the page. Can you be more specific?


This is so annoying... this means they can go after the OS we are using because some kids use their OS to download pirated stuff


Linux is already often banned in online videogames with claims that it's done to prevent cheating


It's pretty ridiculous. They might argue that a browser is a technological measure to protect a copyrighted work which youtube-dl circumvents since the DMCA is written so vaguely. Cases like this demonstrate why anti-circumvention litigation really has to go.


Except the licensing under which the copyrighted works are uploaded explicitly allow for access via a browser, so no you couldn't. That's like saying that a DVD player "circumvents" DVD encryption, or a key "circumvents" a lock. It's the intended use case.


You misread or misunderstood my comment. I'm saying according to the RIAA, the browser is the technological measure which the command line program is circumventing.


Meh, pretty good claim of fair use. They are just picking a smaller file that unlikely to disappear or move. The fact the file is downloaded and then deleted via a script isn't a substitute for actually viewing the file and has zero commercial impact.

And the take down could have been targeted against the particular files they found problematic rather than the whole project.

And the whole thing is ceded to the public domain. Anyone can remove the examples and unit test, call the code webvid-dl and be golden.


Fair use doesn't apply to DRM circumvention unfortunately. You can't even circumvent DRM for reasons completely unrelated to duplicating a copyrighted work.


Test cases using copyrighted works probably not a good idea... Since one could argue that part of the code is run frequently.


so should we take down web browsers as well? because they CAN be used to download copyrighted music and videos.


From the RIAA perspective it seems clear this tool was created to download copywritten material. You can disagree with the law or think the tool has other valid uses but the intent of the tool author seems clear here.

The author of the tool should have chosen a better example in the tests to at least maintain plausible deniability.


People use it for many purposes. Researches downloading public domain material etc.

Does any camera manufacturer showing a camera filming a live concert in some promotional video needs to worry?


Yes, obviously there are legitimate uses. That’s not the point. The point is the actual codebase of YouTube-dl features copyright infringement as a test case. RIAA is using this to make the case that the tool is designed to aid in copyright infringement.

That seems like a reasonable argument given the evidence and the DMCA. I’m not defending the DMCA or RIAA here.


It’s probably a case of fair use but it’ll have to go to court to decide that


IANAL but I don't even think DMCA applies. It's not like YouTube-dl is breaking copy protection schemes, at most it is handling anti-abuse schemes. After all, it's not like you can go and bypass Widevine with youtube-dl.


Unfortunately, it is making a copy. These guys also fucked up by "marketing" it for use in circumventing such controls. They should have showed how to pull OCW lectures.

(a)(2)(C) is the part. IANAL but that's my take.

[1] https://www.law.cornell.edu/uscode/text/17/1201


I can understand why it would seem like circumvention, but one usage of YouTube-dl even with the README invocations is so that you can listen to or watch the media in a personal capacity. Not technically any different from how you might using the actual YouTube client, albeit probably in violation of YouTube's ToS. "Copying" in a legal sense might need a little more rigor. (IANAL either, though, so hell if I know.)


What if I use youtube-dl like

  mpv <(youtube-dl youtube.com/video-url)
or similar. Then no permanent copy is made. I agree that they could have been cleverer in their marketing.


aka

  mpv https://youtube.com/video-url


Why in the world did they include copyright works in the README? That's contributory infringement!


Why? youtube-dl is a violation of YouTube's ToS most probably, but does using it to download a copyrighted video necessarily constitute copyright infringement? I have literally used youtube-dl to watch YouTube, in the past; it's handy to be able to use whatever media player you want, including ones that have better scrubbing.


> Umm, yes the law makes it clear downloading copyrighted material you do not own is not legal.

It most certainly does not, as almost nobody "owns" "copyrighted material," instead most users are granted a license in this context.


You replied to the wrong comment, I think.


Maybe not Copyright, but perhaps CFAA, for accessing the site in ways that was not authorized by the site owner.


So would using an adblocker, or any extension for that matter, also be a violation?

Another point: If you put your content up for public viewing, you're implicitly allowing people to download the material as that's what browsers do. You are downloading a video every time you go to youtube to watch a video.


>So would using an adblocker, or any extension for that matter, also be a violation?

This hasn't been tested, but yes. This is a plausible argument. And it's why the CFAA is complete garbage.


Umm, yes the law makes it clear downloading copyrighted material you do not own is not legal.

The RIAA and MPAA were suing users of Kazaa, BitTorrent, etc individually because of this law you may or may not like, but that's again the law!


I don't think you understand. Yeah, you can sue people for illegally redistributing copyrighted works. However, I am skeptical that you can do the same for people using YouTube-dl to consume copyrighted works. You are downloading the file from someone who has a license to redistribute it, you are just using a CLI utility instead of the page itself. You can't view something without downloading it in some capacity.

I'm not saying there's no possibility this is "illicit," I am saying this is not the same as illegally downloading music off of Kazaa.


I was thinking the same thing. Instead of using a browser to access a website, parse the files given to it, and download the data so that the user can view it, you are using youtube-dl to access a website, parse the files given to it, and download the data so the user can view it.

But I'm not a lawyer.


[Edit: I'm not completely sure how section 1201 interfaces with the fair use rules, but at the very least, a quick glance at the statute (as well as awareness of the First Amendment constitutional underpinnings of fair use) makes me believe the interaction does not entirely eliminate fair use in a context where section 1201 applies. It's clearly complicated.]

The law makes it clear downloading copyright material without holding copyright or a suitable license is usually not legal, but exceptions exist in many countries including the US (where the exception is called fair use).

Some uses of youtube-dl with RIAA-member-copyrighted content very probably fall under that fair use exception, even if other common uses do not. For example, if a music critic wants to use youtube-dl to import a music video into video editing software to produce a review of it, interspersing only a small excerpt of the original video with commentary on it, that's classic fair use and would be an easy win in US court if accused of copyright infringement.

If the actual and as-marketed purpose of covering those copyrighted videos by 3 of 94 unit tests is to make sure that such noninfringing fair uses of youtube-dl are possible with the technologies involved in those videos, I expect that the DMCA claim is invalid. If the actual or as-marketed purpose is to ensure that it's possible to infringe those videos, then the DMCA claim is valid.

That's a question of fact for a judge (in a bench trial) or a jury (in a jury trial). If the more favorable answer is true and credibly defensible in court, youtube-dl might be able to survive a motion for summary judgment, gain support in the court of public opinion, and lead the RIAA to consider settling or dropping the claim.

But if they were indeed trying to facilitating illegal circumvention or infringement, or if they can't credibly prove otherwise in court, yeah this won't end well for their ability to host youtube-dl in the many countries that now have a DMCA-like law.

(Disclaimer: I'm not a lawyer, just a former law student layman who remains a legal geek interested in and relatively informed in these areas. But I don't claim to be authoritative and this is definitely not legal advice. youtube-dl should definitely involve a suitably expert and qualified lawyer if they're going to fight this.)


Pardon i pointed to the wrong case example rather the better one is the RIAA suing Kazaa, Limewire, etc out of existence as they were tools to everyone used to download copyrighted material... 95% used for or more used for such.

Those companies mentioned and or others promoted and or were well aware their users were downloading copyrightable material which is contributory infringement.

This YouTube DL is used for the same thing and probably the same percentage are using it to download copyrighted content.

Not that I pro RIAA ... rather trying to detail the best case example for the RIAA potentially having a case against this tool.


The RIAA and MPAA were suing users of Kazaa who were alleged to have redistributed copyrighted works. These lawsuits were not because someone possessed or redistributed free software source code.


Using a copyrighted music track as the example in the README is really stupid or just shows a total lack of respect for copyright.


FWIW, they did not. This was misinformation from me, misreading the DMCA. I am not glad that my misunderstanding will now spread far and wide, but I at least caught it within the edit window.

The DMCA is complaining about test cases. See extractor/youtube.py in your local copy.


Copyright law is illegitimate and should not be respected.

The stupid thing is hosting youtube-dl in the Untied States.


"The stupid thing is hosting youtube-dl in the Untied States. "

It's not like a million people don't have youtube-dl's source code.

It can be hosted anywhere in the world in a heartbeat, and by now probably has been a thousand times over.


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

Search: