Hacker News new | comments | show | ask | jobs | submit login
“My ten hour white noise video now has five copyright claims” (twitter.com)
860 points by ryanlol 46 days ago | hide | past | web | favorite | 241 comments

I used a song, with permission from the artist, on a demo reel and got a copyright claim on YouTube. I was surprised to find that the claim was on behalf of a French band who covered the song and added lyrics to it. I tried to dispute the claim, but there was no option or possible way to indicate that their match wasn't correct, you could only dispute whether or not you had a license to the track they matched.

In the end I just gave up, and let some French band monetize my video that contained a song written and recorded by my friend's dad. Enjoy the 25 cents you earned!

That's no longer the case. I post a bunch of game playthrough vids and there are a lot of options to contest claims like that now.

I had a recent one where the sound effect from a game released in the 1990s was claimed as being part of a song released in 2015.

I had a similar thing happen with gameplay videos containing wind noise. The videos were uploaded in 2014, but I get copyright claims from a song that was released in 2017. This seems like a major oversight in Content ID, it should force the claimant to provide a release date for the song and exclude videos that were uploaded before that date.

> This seems like a major oversight in Content ID, it should force the claimant to provide a release date for the song and exclude videos that were uploaded before that date.

The major oversight rather lies in the fact that the law does not introduce huge penalties/indemnification for false copyright claims - if these existed, copyright holders would be a lot more cautious with potentially false claims.

The law does in the form of perjury charges for false DMCA claims; content ID is intentionally operates outside the DMCA legal framework.

The perjury coverage is weak, but they at least provide something.

EDIT: Perjury comes into play as part of a permanent takedown request, not the initial temporary takedown request. Just a clarification.

The perjury charges are in respect to the claim that you are authorized to enforce the copyrights on the work you claim is being infringed.

If you own the rights to a 2013 advertising jingle for yo-yo wax, and you assert that you own the rights to it while making a DMCA takedown claim against a birthday party video posted in 2010, there is no perjury. The claim that the work is infringing is not the part covered by the perjury threat.

This is why the DMCA is broken. There is no reason whatsoever for legitimate copyright holders to avoid making frivolous claims, and the burden lies entirely on their victims to prove their innocence.

(I am not a lawyer. I might be wrong about this. Don't mistake this comment for legal advice.)

The perjury claims come into play with the permanent takedown. There's no burden of proof on the initial takedown, as there is no burden of proof for the request to have it put back up. It's after those two steps that the takedown can become permanent, which is where the legalities and associated penalties come into play, and where the claim can go to court.

As I said, weak, but it's there.

Wouldn't even that go away if the claimant had a legitimate reason to believe the subject was infringing? ...maybe if an automated matching algorithm flagged it as such?

If the algorithm returns false positives more than 5% of the time, does the claimant truly have a legitimate reason for such a belief?

Yes. Which is why it only prevents false claims (the original concern), and why I label it as a weak protection. Of course, it's also only really going to come into play during the lawsuit phase of the notice and takedown process, a process unlikely to be automated.

IANAL, of course.

You should make a copyright claim against the song that was released in 2017 because they clearly sampled your work without permission and without payment.

This is exactly what I was going to say -- but it is worse! Since they (2017 song) already claim a violation, they are essentially admitting to a violation (except, on themselves!)

I would actually consider doing this if I were in the same position. You'd have to wonder what the consequences were, considering that by filing a claim against someone you are establishing that you believe there to be a violation. If the claimant then is shown to have the newer work by several years, I wonder how such a claim would be taken?

DMCA would be establishing that you believe there to be a violation, but ContentID is just automated matching YouTube does on the behalf of artists - so no such claim exists.

surely, by enrolling to have ContentID match a work which you claim is yours against others, you are initiating the chain of events which leads to the automated claim of infringement being made? This implies (to me) that you are the one making the claim?

That's exactly what I was thinking: by making the claim they've admitted that both the number and your video are using the same music. Only you have proof that your video existed first, ergo they stole your work :)

LOL! I thought that too, but I do not own the copyright on the game or the sound effect.

Is it valid for a game author to claim a copyright on something like the game's music in a playthrough though? Or is that fair use?

I think arguably, both options are correct.

It is probably valid for a game publisher to claim copyright on the game's music--assuming they created it for the game [as opposed to licensing it].

I also believe that a playthrough video probably counts as fair use in most cases.

Nintendo is known for having a 'bad' reputation with playthrough videos on YouTube.

You mean to claim infringement. Nobody has to claim copyright because you automatically get it when creating a work. Fair use relates to what is infringing and what isn't, it doesn't make a work any "less copyrighted".

Since there appears to be no consequences to making a false or misleading claim, and although I wouldn't call it a solution, one course of action would be to make a claim against their song.

Kick 'em where it hurts.

The cover band's claim that they were acting in good faith would likely hold more water. Not sure if you want to be playing games with the law like that.

Meh, it's not a major indictable: no one's going to gaol for misusing a copyright claim system.

Anyway, as others have noted, I believe you agree to an arbitration clause when you sign up. Lawyers aren't likely to get involved, you'll just get kicked off when a human works out you're abusing the terms and conditions.

(I didn't down vote you)

> The cover band's claim that they were acting in good faith would likely hold more water.

This claim of the cover band can only hold water if they made strong investigations to ensure to their best that the creater of the video did not have a permission of the original copright holder of the song that they covered.

> Not sure if you want to be playing games with the law like that.

There's no law involved here. "Infringement" in this context relates to a term defined in YouTube's TOS with regards to the ContentID take-down service provided to the big right-holders industry. Not the legal sense of "copyright infringement", because nobody is making a copyright infringement claim in this whole story.

The entire ContentID take-down service is designed so that YouTube will have to deal with as little copyright infringement claims (from the big right-holder industries) as possible.

From the perspective of the uploaders and viewers, ContentID is just this thing that helps YouTube take down videos, which they have the right to do (as per the TOS) "for whatever reason". The dispute procedure is just something to make that easier to swallow, but unsurprisingly it's not at all effective in getting a more just result.

That's what you get when you upload your work to a platform like YouTube instead of publishing it yourself. As a small player you can't really get YouTube to sign a contract detailing the license under which you allow them to publish it on their platform.

Wouldn't Content ID just be a civil matter if the revenue is under a certain amount, given that it operates outside of the DMCA?

It's in fact just a TOS matter. It's entirely YouTube's call.

IIRC there are penalties if you lie about who you're bringing the claim for.

I didn't even get that far on one video I tried to upload. I had the permissions to distribute the video (video game trailer direct from the publisher, so the entire purpose is to redistribute it), yet the upload was rejected outright with a "This video violates the YouTube ToS message" (no nudity/violence etc. either, so I can't think it should have tripped any other content filters).

US copyright law has very stiff penalties for false copyright claims. Just sue YouTube directly.

It was probably a YouTube internal content ID claim and not a DMCA takedown request per se.

No, it doesn't.

For example, in a DMCA takedown notice the only thing asserted under penalty of perjury is that you are, or are authorized to act on behalf of, a copyright holder. And YouTube doesn't even use the DMCA system; it uses its own separate terms-of-service-governed system.

Yes, it does.

17 U.S. Code § 512(f) states:

(f) Misrepresentations.—Any person who knowingly materially misrepresents under this section—

(1) that material or activity is infringing, or

(2) that material or activity was removed or disabled by mistake or misidentification,

shall be liable for any damages, including costs and attorneys’ fees, incurred by the alleged infringer, by any copyright owner or copyright owner’s authorized licensee, or by a service provider, who is injured by such misrepresentation, as the result of the service provider relying upon such misrepresentation in removing or disabling access to the material or activity claimed to be infringing, or in replacing the removed material or ceasing to disable access to it.


This has been tested in Online Policy Group v. Diebold, Inc. - and Diebold lost and had to pay $125,000 to Online Policy Group.


The Diebold case is basically the only time anyone's ever made that work, though, because the bar for it is so high. It's not enough to prove "they made a complaint and it turned out it wasn't their stuff". You have to prove that they knew, which is next to impossible unless you basically have a notarized statement from them saying "Yes, I was aware it was wrong and did it anyway", and the payout is a pittance.

Meanwhile, the idea that DMCA notices carry some sort of super-strict ruinous penalty for errors remains completely false. And YouTube -- as I said in the original comment -- uses a separate takedown system governed by their terms of service, so the whole thing is moot.

Well, that's not actually what you said. You actually said:

> in a DMCA takedown notice the only thing asserted under penalty of perjury is that you are, or are authorized to act on behalf of, a copyright holder

And the Diebold case didn't have anyone with a notarized statement that you speak about. I don't think what you are saying is necessarily supported by the facts of that case.

It is true that the only thing asserted under penalty of perjury in a DMCA notice is that you are or are authorized to act on behalf of a copyright holder. The rest of the notice is not made under penalty of perjury.

The notarized statement comment was hyperbole to make a point about how hard it is to prove, sufficiently for a court, that someone "knowingly" and "materially" misrepresented information in a DMCA notice in order to trigger 512(f). In the Diebold case you had the perfect storm, because Diebold's lawyers basically admitted to the court that they knew they were using the DMCA not for copyright enforcement but to take down something that embarrassed them. That is literally what it took to get a judge to nail someone for DMCA abuse.

And even then: the judge only found that they knowingly/materially misrepresented for a subset of the material, and they only paid out $125k, which is literally nothing to a company of Diebold's size.

Meanwhile, DMCA abuse remains routine and rampant today, in large part because it's next to impossible to get someone penalized for it.

"Knowingly" is probably too high a bar for this to be a significant deterrent. A mental state of negligence or recklessness would make it much more effective.

Knowingly misrepresenting that material is infringing to get it removed probably ought to be subject to criminal penalties.

You probably 'agree' to an arbitration agreement when you make a youtube account.

Sue them in any country that ignores arbitration clauses.

IANAL but seems like you cannot just sue people/companies in any country. I think you need to be at least resident of such country.

Each country has their own laws. The EU is mostly harmonized, and most other countries have similar laws. However if you want to do jurisdiction shopping you might find a country (in Africa?) that will allow you to sue.

> In the end I just gave up, and let some French band monetize my video that contained a song written and recorded by my friend's dad. Enjoy the 25 cents you earned!

I would have elected to not post the video. NEver let anyone profit from your efforts if they have done nothing to contribute. Don't cave to these bottom feeders.

> NEver let anyone profit from your efforts if they have done nothing to contribute.

Don't focus on the "bottom feeders". Focus on those to whom you have brought joy. Github would be an empty place otherwise...

Ah! But in the case of open source, a person willingly releases code, usually with a license, fully knowing someone will profit from it. That is their choice. Git hub is full of open source licensed code at the will of the authors.

In the case of the parent post I was addressing earlier; This is akin to a company claiming part ownership of your open source code because you used part of a public domain library your friend's dad wrote which they copied and incorporated into their closed source product.

This video of the ending of Star Wars without John Williams' music got a copyright claim from the rightsholder to Williams' score [edited from original post, not actually claimed by Williams himself]. The little bit of music at the beginning is from Gustav Holst and is public domain.


edit for more context: https://www.wired.com/story/the-star-wars-video-that-baffled...

John Williams stole (borrowed) a lot of music from Gustav Holst and many others. He's famous for it. Specifically parts of Holst's The Planets. Does this sound familiar?: https://youtu.be/Isic2Z2e2xs?t=4m24s

There's a long history - I suppose it's a tradition - of composers leaning on one another for inspiration. Sometimes this takes the form of them outright copying another's melody and elaborating upon it.

Beethoven, Mozart, Haydn, Bach, Mendelssohn, Brahms etc. all did this. Put in this context what Williams has done in borrowing from Holst and others isn't scandalous whatsoever and shouldn't be thought of as such.

Just let me fix a few words, for science:

> There's a long history - I suppose it's a tradition - of composers claiming copyright on someone else's work and sending DMCA takedowns automatically

That's how absurd it has become.

You're making an excellent case against copyright.

I always joke about how much money Europe should pay China for ideas like the compass or gunpowder.

Everything is a remix, https://vimeo.com/14912890

Beethoven, Mozart, Haydn, Bach, Mendelssohn, Brahms etc. probably didn't send each other DMCA takedown demands though.

Oh of course, Auralnauts replace so many of Williams' works with Holst in their various videos that you hardly even know it. Similarly look at the opening of the 4th movement of Dvorak's 9th (https://www.youtube.com/watch?v=p9OBf8f55tU) for Jaws' inspiration.

Or listen to any of Ralph Vaughan Williams' 9 symphonies and you'll hear John Williams' orchestration and voicing directly.

Listen at 1:45, whole piece is amazing though.


The overall structure is similar, many themes sound like they were modified from Star Wars (or rather, the other way around), and 6:55 sounds like it was literally copy pasted.

“...great poets imitate and improve, whereas small ones steal and spoil.”

Every so often, I come across an old blog post or other content with a YouTube video embedded in it. About half the time, it'll have been taken down by a false-positive copyright claim. It seems that YouTube videos can't be permanent, only ephemeral; as a long-term hosting service it simply doesn't function.

Yet another reason not to host on YouTube. Here are some more reasons not to host with YouTube:

- some YouTube videos use nonfree Javascript to implement DRM. https://notabug.org/GPast/avideo/ is a youtube-dl variant that avoids running that JS code. notabug.org seems to be down as I type this, perhaps it will return later. - should you wish to allow your video to be downloaded, YouTube sets up needless barriers to allow downloads. - hosting with YouTube means you and your users are being tracked by Google (which owns YouTube). - YouTube doesn't make it clear how to download the precise video file the submitter uploaded; users get rewritten videos instead. - YouTube only hosts videos, not arbitrary kinds of other media.

All of these problems are easily solved: use the Internet Archive (at https://archive.org/) and your own hosting instead of YouTube. IA offers a download URL which is static and will redirect to your uploaded data. You can use this URL in an HTML5 video or audio element to embed the data in a webpage (again, no JS needed). IA will (by default) rewrite certain kinds of files, but your original uploaded file is also available and machine-determinable (via an XML file that clearly indicates which is the source file and which are derivative files). IA can be used fully without JS turned on (no need to worry about executing code to get files). IA hosting is gratis (zero cost), as is YouTube.

I have no evidence which indicates IA tracks users akin to how Google does.

I encourage you to consider hosting your media with the Internet Archive and your own website instead of YouTube.

Mr. Stallman, is that you?


One more reason Google sucks:

Google recently deleted hundreds of videos of the war in Syria. Potential evidence of war crimes erased. Original uploaders may be dead.

I don't think some tech people realize the consequences of their decisions.

> Google recently deleted hundreds of videos of the war in Syria

Wow, that is truly evil. Do you have a source on this?

Edit: found article about it: https://www.nytimes.com/2017/08/22/world/middleeast/syria-yo...

> Wow, that is truly evil.

Arguably. Youtube is a video hosting site, it's not a political platform and has no obligation to host content it doesn't deem suitable for their platform.

Those videos should be sent to news agencies, (inter)national archives, the UN, war crime tribunals, etc instead (or in addition to) "entertainment" video hosting sites.

They are under no obligation to host the videos, but they can't just delete them. It's irresponsible, immoral and unethical.

By parallel, misprision of felony is a crime in the US:

> Whoever, having knowledge of the actual commission of a felony cognizable by a court of the United States, conceals and does not as soon as possible make known the same to some judge or other person in civil or military authority under the United States, shall be fined under this title or imprisoned not more than three years, or both.

There is a moral imperative created when someone else is giving you these records. It's like someone asking you to call 911. You don't have to go out looking for emergencies to call 911 about, but you do have to deal with it when one is dropped in your lap. In this case "dealing with it" just means retaining a copy and making it available on some terms.

> They are under no obligation to host the videos, but they can't just delete them.

One implies the other. If your media backup and recovery plan is Youtube, then you need another plan.

> Youtube is a video hosting site, it's not a political platform and has no obligation to host content it doesn't deem suitable for their platform.

For "not a political platform," Youtube has quite a lot of political content, and it's not at all clear to me why someone would think that the idea of a general interest video hosting site and a political platform are somehow exclusive at all.

And as far as obligations go... I'd imagine that there's no legal obligation, and maybe there shouldn't be, but that isn't the end of the question. As far as I know, I'd have no formal obligation to stop and render aid or even contact emergency services if I came across someone unconscious and bleeding on the sidewalk near my house, but that doesn't mean I might not be under some obligation.

And one problem is that organizations like Youtube often don't put themselves together in such a way that there's anyone interested in and capable of responding to these questions.

It's almost like they come across a turtle in the desert, and it's flipped over on its back, belly baking in the hot sun, beating its legs trying to turn itself over. But it can't. Not with out their help. But they're not helping. Why is that?

> why someone would think that the idea of a general interest video hosting site and a political platform are somehow exclusive at all.

A general interest here seems commercial in nature and is thus susceptible to whims of the owner(s) and their ToS.

One possible solution seems to be something like a Wikipedia for videos, a non-profit org, but the cost of video hosting seems to still be prohibitively high for a non-commercial entity in 2018.

The likely reason why they ended up on Youtube was because every last Android phone (and a fair few other phones) have Youtube bundled, complete with a one button upload.

In the spur of the moment, in particular in a warzone, the proper place to store something is the last thing on one's mind.

Regarding "Youtube is a video hosting site, it's not a political platform": politics enters any situation where 2 or more people are involved. Another poster has clearly explained how YouTube is quite fine hosting shows with what are clearly political views as well as choosing whose views are currently in favor. Calling any editing effort non-political is either woefully uninformed or an attempt to manipulate others (snowed or trying to snow others).

Regarding "has no obligation to host content it doesn't deem suitable": Yes, sadly, this is a problem not unique to YouTube. All hosters get to decide what to host and to change that decision at any time.

Some people choose to respond to hosting on sufferance by becoming sycophants (placate the hoster with speech one believes won't offend the censors) but that is ultimately detrimental to one's own message and free speech overall. We have no solution for this that is really easy for nontechnical computer users because we all depend on other parties for Internet access. If homes had Internet access that could not be taken away, we might be able to recommend self-hosting. For now, in the age of the ISPs, the best we can do is massively decentralized hosting which uses no DRM, running additional nonfree code (such as YouTube's DRM in JS), and somehow getting a lot of peers with lots of storage and bandwidth to keep available copies of everything uploaded (we ought not decide that one message is worth sharing and not another message because that is the very problem we're trying to solve).

This is a tough problem to solve. I don't know what a viable solution will look like quite yet. But I know what to avoid: we won't solve this problem by placating big business desires for more control over what we say and how we get copies of other people's works.

A correction: running no additional nonfree code.

YouTube is not just a video hosting site. They are a primary conduit through which global human culture is conducted. Their actions to actively control and manipulate human culture (as Eric Schmidt said in his book The New Digital Age it is Googles responsibility to do because they are rich and therefore know better than humanity how human culture should go) is something they bear total responsibility for. And, at least for me personally, "but we thought it would make more money this way" is not a legitimate reason to facilitate mass murder.

Sure, in theory... In the end you either educate the users if you can, or try to make the big platforms more accepting of this, or both. There is no perfect solution. Being an "entertainment site" isn't a good excuse.

For those that don't get the joke: https://gimletmedia.com/episode/106-is-that-you-kd/

I can't imagine listening to the audio, but reading the transcript is nauseating due to the frequent occurrence of the work like.

197 occurrences of the word like.

This particular section of the transcript annoyed me the most.


He’s fired. And becomes like basically a celebrity for certain parts of like the right. And they were kind of like… Like this is not what he was necessarily saying about himself, but for them, it’s like… He’s like this perfect symbol of like the qualified, brilliant, genius coder who’s the best person for the job, whatever his race or gender happen to be, who said one reasonable thing and like got fired because of political correctness.

That's the paragraph that led me to hitting the back button.

That couldn't possibly be the source of this joke, so why this particular link?

(Also, frankly, this interview doesn't exactly present its participants as the sharpest tools in the shed.)

The comments that followed my comment are the reason I wish it was possible to simply delete a ycombinator/hacker news profile.

It's an entertaining episode from a show I enjoy. Apparently, there are no other fans of the show here.

Very sorry.

You're entitled to your tastes; just because no other fans of the show chimed in, doesn't mean you need to feel sorry for posting it (though some explanation of relevance to the topic would be appreciated).

So don't worry; in a day nobody will remember this subthread anyway :).

It builds up over time with a slow drip drip drip. Just the constant soul destroying pointlessness of socializing on the internet though comment boards.

Anyways, I googled it and no it is not possible to delete a hacker news account. I successfully disabled reddit, facebook, twitter and linkedin though. So, this is the last outpost of an abysmal time on the web.

Thanks everyone. I am out.

Wtf is this?

> I encourage you to consider hosting your media with the Internet Archive and your own website instead of YouTube.

unfortunately this makes your content pretty hidden. You are not within youtube's search anymore, you're not in any sidebar as related video either

This is why I am increasingly of the hub and spoke method. You create content and host it on your own, but then reupload it to various sites that fill a need (youtube/vimeo/dailymotion for video, etc). Then not only is the content when removed from one (say youtube) still there in another place/form, but it can just be reuploaded.

Own your own content, but if you seek to spread a message you might want to use closed proprietary ecosystems as long as you aren't dependent on them.

For example, I was recently updating linkedin, and went to add Bruce Schneier, but he had a post that said he never checks or uses it and only links to his website from there. He still has a presence there though, and that really hit me how much these little walled gardens we let take over so much of our compute ecosystem are starting to look like prison, even for your standard copyleft hating SV hipster.

Of course Stallman would say don't even use them, and he makes some good points, but the problem we have now is even if you have a good message you can't reach the masses without going through some priestly gatekeeper of the government-corporate media power. So you have to use the system to beat the system.

> So you have to use the system to beat the system.

Which by the way is what RMS did back when the GNU system didn't exist: use proprietary UNIX to develop GNU, replace the system piecemeal.

The two cases (YouTube and using nonfree OSes) are not comparable as the grandparent poster and this suggest.

In Stallman's case there were no free software OSes. One could not use a computer in freedom. Therefore he used a nonfree OS to write free software until he could run a free software OS and thus had no more need for nonfree OSes. For years now he (and the FSF) don't run nonfree systems. They even endorse computers (and other computer hardware) that can be run completely in freedom (see https://www.fsf.org/ryf for details).

We already have file hosting options that don't use DRM, nonfree JS (or indeed any JS), and avoid the other problems of YouTube. There's no need for coming up with something new as we already have viable other means of sharing files.

We don't need YouTube, they need us. By buying into their limitations and restrictions uploaders endorse YouTube (which is undesirable and certainly not very thoughtful) and help foist YouTube's choices on others (which is unethical).

Still, the issue remains: if you have a message to spread, how would you go about spreading it? The brute facts are, not putting copies of the message in those evil centralised social networks means the message will be invisible to many people, in some cases the majority of the target audience.

I'm all for getting rid of YouTube, but I'm not sure how we should do it.

Well yeah, think of it like backups. You wouldn't keep just a single copy of an important document, would you? And certainly not on a cloud service with randomly-enforced ToS, right?

Upload to as many sites as possible, link between them. Keep offline copies if you can.

Do you contribute to the Internet Archive? Video service is ridiculously expensive, and using them to hotlink HTML5 video onto your Web site seems...kinda antithetical to the spirit of what you're saying.

There's nothing special about "video service" nor is that even a relevant concept to discuss and in no way contradicts anything I said. If you're going to claim otherwise, please be specific about this alleged contradiction so we can have a meaningful conversation. Serving data is serving data. Hosters have to put effort into disguising/obfuscating a URL to make it not be a static link (and thus completely useful for HTML5 video/audio use, or pointing others to download and keep, or something else).

If a hoster doesn't want one to distribute the files, that is the issue. Not whether one distributes them for use in an HTML5 video/audio element, or downloading for local storage, or use in some other way.

https://blog.archive.org/2009/07/01/archive-org-supports-the... would seem to address your claim.

Video service is special because it is expensive on account of video files being large. You’re pulling a lot of bandwidth off Archive when you embed HTML5 onto your site. On its face, that’s a pretty shitty thing to leech from a charitable organization; YouTube pays for that with advertising. Archive takes donations. Are you donating?

I’m confused that you think this isn’t relevant. You’re directly advocating hotlinking against Archive and you don’t even mention supporting their operation in the slightest. I have no idea what you think I said, because your reply makes little sense.

This is true but I think OP meant Internet Archive in addition to self hosting.

> use the Internet Archive (at https://archive.org/) and your own hosting instead of YouTube

Use IA to set up a canonical, permanent URL for the media, which resolves to a host under your control as long as it's available, then resolves to the IA-hosted archived copy if your site ever disappears? As long as the URL doesn't change, a CDN could do most of the heavy lifting anyway, right?

No, OP meant embedding Archive hosted assets onto your site, or any site.

> You can use [the Internet Archive URL] in an HTML5 video or audio element to embed the data in a webpage

>IA hosting is gratis

Not a scaleable solution though. If everyone were to do the same they'd quickly be consumed by bandwidth costs.

That's quite a claim made without any pointer to evidence. Do you have any evidence to back up this claim?

IA has been around for many years and not bumped into this problem. Nor are they making choices that even hint at this problem. In fact, they have long made choices (such as default derivatives generation) which suggest they have plenty of storage space and bandwidth.

Also, this isn't an evidenceless complaint specific to IA.

RasPi + Cloudflare should be a good solution. Even live streaming is possible with CDN as live streams are just conventions with playlists and metadata. I can't comment on costs at scale, but Twitter claims equal ingress and egress bandwidth for Persicope (ie. first egress hit caches), standard HLS will just have more latency and not more cache misses if I understand correctly. My stream is 1500-2500kbps 720p.

How about distributing one's videos through BitTorrent instead? Should scale pretty well. Granted, viewing the video becomes a hassle. But it works.

Or with PeerTube, which makes videos available in the browser AND uses a bittorrent-like protocol: https://github.com/Chocobozzz/PeerTube

IA (https://archive.org) offers a BitTorrent feed for each of its uploads. You should check out what IA offers.

Exactly, when I want to look my watch history to see what I was watching 5 years ago, most videos have been taken down or removed by the author. And Youtube does not even display the title of the video or its description, so I have no idea what it was.

No its human greed.Don't monetize your video and it will be fine for years.

YT never deletes anything, which is why they're losing money. Any archive ALWAYS loses money. Hosting cat videos for 10 years isn't a good business case.

I think automatic systems like Content ID should be illegal. Plus I doubt a robot can consider things like fair use. Well maybe they could do some machine learning of past cases where videos were complained about but ended up being fair use...

I think a real human should have to spend all day finding content and manually doing the work. Kinda like how some states have banned red light cams. Technology is not perfect. I heard of those cams going off even when it's legal to turn on right. A real human swearing under penalty of perjury or giving sworn testimony seems perfect for accountability.

At least the money is held until you do a dispute now with Youtube. I know if you license royalty free music - the system will even flag it. I know someone personally who had that problem. I guess the systems aren't linked so no idea you had a license but annoying. So say you made a video for a client and uploaded it, and embedded to their website. Even though you didn't monetize it, your client website now has ads for their competitor as they'll allow the video to stay up but with ads where the money goes to whoever claimed it! I guess the way to protect that is to upload as unpublished or unlisted then dispute it before publishing it when you already purchased a license.

I suspect Youtube's Content ID and flagging system is the way it is to placate companies like Disney and Viacom who couldn't care less about fair use, but do like arbitrary power.

It shouldn't be illegal, google can do what the want on their hardware. The problem is that everyone has centralized to googles hardware, and issues like this will happen for as long as that remains unchanged.

We need a path back to the distributed web where people upload videos to a server they own and control. Instead of IoT juicers we need "video broadcasting in a box" solutions.

Yeah... It seems like some search engine for some decentralized video might work.

Some search engine indexing, videos hosted on independent sites(regular http), favorites/playlists can be stored in something like Dropbox privately or shared.

Probably could be just a video tag with additional elements and builtin functionality.

Some advertising spec too... Sites can either implement third party ads or host their own ads for example. So instead of being told your content isn't advertiser friendly. You could run your own ad server that's in the video metadata or use a external one agreeing to their terms for video content. So total freedom to go out and get your own advertisers or signup for a service and plugin a URL and they get to handle that part.

So say you just want to make videos, not doing controversial or political stuff and not really into being a sales person just use this ad network... or maybe you are a very political channel, point to this other ad network that targets that audience. Maybe you are a non-profit, run your own ad server with your own in-house fundraising ads.

Not sure how you'd go about handling sensitive stuff some people might not want to see, maybe at the search engine part using AI/machine learning but also allow sites to indicate certain things within the spec itself... Decentralizing the search part even better.

But ideally you'd have a consistent UI sorta like YouTube but all content is fetched from servers like a web browser, you'd have subscriptions, sorta like RSS, etc...

But I guess subscribers counts and commenting/liking would need to be apart of it too. Maybe use somthing like OpenID. but sites that want to host videos apart of it would provide certain metadata and some callback urls to post to for like comments, getting commentings, likes, etc.

Just thinking out loud... One of the benefits of YouTube is discovery and organizing your watch later, subscribers, etc. So some sort of formal spec for a video app would make sense.

Searching random sites, signup and comment and have like 10 sites to check for new videos would get annoying.

You may be interested in PeerTube: https://github.com/Chocobozzz/PeerTube

I'm still stuck on the notion that the accused is innocent until proven guilty.

I know, I know, that makes me a pinko commie liberal Freedom Markets™ hating fruitcake.

A common theme during this late stage capitalism is the burden of proof being placed on the weaker party in all these transactions, in other words, we're presumed guilty until we can prove our innocence.

I think a real human should have to spend all day finding content and manually doing the work.

Writing as someone who has been that human, in a startup creating original content that someone was repeatedly ripping and uploading to YouTube after signing up with different credentials to bypass the paywall, unfortunately your approach doesn't work very well either.

It's true that excessively aggressive takedowns result in YouTube not hosting some legitimate content. But it's also true that sites like YouTube allow an infringement to reach thousands or millions of people in a very short space of time, potentially completely undermining the value of the original work.

If you're a Hollywood studio or big record label, you can factor piracy into the cost of doing business, and you can take steps to make it less attractive or more trouble than it's worth. If you're a small independent content creator, the economics and practicalities are entirely different.

Ultimately, once content is shared the genie is out of the bottle, while relatively little content is so time-sensitive that you can't allow for challenging it and keeping it offline while the dispute is resolved, so requiring major hosts to take down by default or lose their safe harbour protections isn't unreasonable. If someone doesn't like that, they're free to host the content on their own systems instead, but such a site is unlikely to have the resources to cause the same scale of damage as a major video hosting service like YouTube or Vimeo, and if it does then it's going to be worthwhile to bring a real copyright infringement action against the site owner, who in turn is likely to be identifiable.

Apparently this is the development of Artificial Intelligence. From a corporate point of view, one big fix cost is obviously better than ever growing (now rather small) variable cost.

Also, google is leading in this AI sector. Your feedback will help them to improve. The more hassle they introduce, the more you react upon it, the faster it can learn.

I've had a number of claims against tracks I've made where the basis of the claim is a sample I've used - not a sample of a commercial track, but one from a sample library that I've bought and used legitimately, but where said sample has been used and published on YouTube before the track I've done. There seemed to be no worthwhile way of contesting it as responses seemed to go nowhere. It's not a big issue for me as there's no income from YouTube for us, but I'm sure there are cases where it is a significant problem for a legitimate user.

I had a copyright claim once on songs my band wrote and performed. Original material. Our original material. None of which vaguely sounded like any one else's songs. The person I had to talk to to resolve the issue was very difficult to deal with and we went round in circles until I managed to prove we owned the copyright and publishing on our own material.

1. Assemble a botnet

2. Automatically download and reupload popular videos, with extremely subtle modifications.

3. File automated takedown notices against each copied (popular) video

4. Destroy youtube's ad ecosystem for cheap

Or, use an adversarial DNN to transform videos into ones that will get copyright claims. Extra points if you figure out to how to have it using reinforcement learning to improve itself.

Demonstrating that the system is broken is not an effective strategy. It is impossible to get someone to understand a fact when their salary depends on not understanding it.

That's like saying that grocery stores are broken because you can just print out bar codes and use them at the self-checkout. Stores have enough security for their needs and balance it with customer convenience. I don't think even Google would claim that their system is perfect, only that it's currently good enough to stave off ravenous lawyers and keep them from losing safe harbor.

The point isn't demonstrating anything; the point is making continuing to run the system as-is untenable. If every large corporate publisher on YouTube is suddenly having takedown claims filed against them, the system isn't going to stay the same for long.

I suppose not, but what do you expect will happen instead?

There's a good chance the answer will be "YouTube shuts down", and while that could be a good thing, you'd better have a replacement ready for deployment. Preferably something open source and distributed.

> If every large corporate publisher on YouTube is suddenly having takedown claims filed against them, the system isn't going to stay the same for long.

That's wishful thinking, they'll just be whitelisted.

Wouldn't that just result in the videos being taken down/demonetized?

Not sure how this would harm youtube, wouldn't you just end up silencing yourself?

Baiting fraudulent DMCA claims and then countersuing for lost revenue COULD work, but would likely require more legal fees than would be 'profitable'.

Making one copyright claims on youtube will basically just move the revenue of the ads to the copyright holder.

If you are able to trigger this to three separate claim holders you can actually jam the system into not paying out or any monetization.

I might be explaining this badly so if I am mixexplaining this but Jim Sterling does this.

Umm I wasn’t really being serious about this at first but due to overwhelming response I might try this.

5. Find out it doesn't work.

YouTube's enforcement system is not set up to treat every actor in it equally. It's set up to favor big rightsholders like record labels at the expense of everyone else. So your attempt to get all those videos demonetized, or get copyright "strikes" on the big labels' accounts, isn't going to succeed. Notices filed against those accounts probably go into a black hole (and you have no DMCA remedy, because YouTube doesn't use the DMCA system, requires you to use that system as part of its terms of service, and so isn't obligated to act DMCA-style on complaints filed through its separate non-DMCA system).

> YouTube doesn't use the DMCA system

Are you sure? Wouldn't they then lose the "safe habour" protection from the DMCA, and then be more liable?

I know they can, and do, go beyond the DMCA, and take down things more often. But I thought they would still want to keep the benefits (for them) of the DMCA

YouTube will react appropriately to a DMCA notice if you send them one. That's all they have to do to keep their safe harbor. Anyone can go look up their registered agent and do it.

But YouTube's own easy-to-use automated enforcement system is very different from the DMCA, and almost certainly will ignore or at least massively de-prioritize notices-within-that-system-which-are-not-DMCA-notices against a major company. So if you felt you had a real case you'd probably have to actually send a DMCA notice, which would just result in automatic counter-notice and YouTube saying "well, we followed the rules, if you want more to happen you have to go sue that gigantic company all by yourself".

OK, that's what I thought. I thought the DMCA required something to be down for a certain amount of days (1? 3?), like you cannot just auto put it back up, right?

2.1 embed Rick Astley "never gonna give you up" within the video.

That’s already been done. See here: https://youtu.be/dQw4w9WgXcQ

Uh uh. Nope. Not gonna fool me this time!

IIRC I have only been Rick Rolled 2 or 3 times. You had to have Navy SEAL level of discipline during that year or two.

...or be trained in gorilla warfare, so that it's just another (non-click) target to you.

Great, now I'm imagining some sort of Planet of the Apes crossover..

For the people who hadn't had their coffee yet:


> dQw4

Nope, not falling for that one.

Funny, same reaction, but I memorized XcQ :).

Haven't heard that song (on youtube) in years. Blocked in my country.

I would hope the older and more popular original video would ‘win’ in that scenario

True story: on soundcloud, I had a private track that I later (months) released on a label. Shortly after that label published the track, soundcloud reached out to me to say that I had illegally-uploaded content. I told them I owned the rights to the song and they confirmed this with the label, but it was still a weird experience.

> soundcloud reached out to me to say that I had illegally-uploaded content. I told them I owned the rights to the song and they confirmed this with the label

That sounds exactly like how the system should work.

Except, perhaps, it'd be nice if they asked politely if he had permission rather than claim he didn't.

5. Create an atmosphere of fog in the DMCA dispute system, slowing down operations and reducing the effectiveness of automation efforts

6. Package this in such a way that anyone can deploy it, making it eternal and diffuse. Otherwise it’s just going to be one or two people with the feds eating them alive to “make a point” about law and order.

What I don't get is how when I post a 15 second video that has some music in the background at a coffee shop and I get a YouTube-bot telling me I have violated copyright but then bazillions of others post full HD rips of Hollywood movies and commercial albums.

They might post them but the money goes to the rights holder, which I guess is why they'll tolerate it.

Youtube copyright claim system and ContentID shows us a glimpse of what happens when unchecked AI in many cases reports and handles the report by just removing it.

It can be cold and inhuman even if it is a needed service, very vogon-like.

YouTube copyright claim system needs to add penalties for false claims with the same banning system that they have in place for youtubers with too many take down claims. This is just insanse...

YouTube grants all monetization to the claimant instantly, for the entire period of the claim, even if it's bunk.

So take an algorithm that can flag a million YouTube videos, and maybe you steal $10,000 of monetization over cost in a few days. Nobody can afford to sue you hard on the false positives because you're mostly only hitting "nice guys" who will just change the sound and reupload. So you up the algorithm to be more profitable, wash, rinse, repeat.

That's why this is really happening.

I'm surprised there isn't a claim from John Cage's estate.

That would be silence, or, the absence of music; not white noise. Seriously this is insane that there is a claim against this.

It really depends on where you perform the piece.

If you amplify almost any recording enough, you can get some form of noise. Any white noise of a sufficient duration is effectively a recording of that one John Cage piece.

no, it isn't, because that piece is unconscious- it is meant to be what happens without any intention. A crank up to noise is a definite intention of the performer. In 4'33", the performer does nothing but open the score and the keyboard, mark movements, and bow at the end.

If you go through enough generations of analog copies of a recording of that performance, you get 4'33" of noise.

no, it would not be silence. I don't mean to be all snooty here-- but since we are talking about it-- it would be whatever sounds occur during that time. He was inspired by those Rauschenberg paintings ( that appear to be blank canvasses, and are but are functioning as shadow catchers- real-time spontaneous art framers, if you will)- same with 4'33". It's about whatever happens in that space in that particular moment. It is a kind of unconscious spontaneous group improv, sort of. Interestingly, improvised solos over tunes aren't copyrightable- bandleaders ripped them off players all the time.

I don’t think it’s snooty, but I did clarify as to the absence of (traditional) music.

I’ve preformed pieces like this several times, but not Cage specifically. I agree all performances of these types of compositions will be unique, which is something I admire about all performed pieces. Our canvas is not only one of tones but also time and space.

I hear what you are saying. But isn't it interesting... If I play a solo over "Ask Me Now", the Monk estate doesn't own it, but I don't think I "own" it either. But if I did it on a recording made by Monk, with his group, I think then he owns it, depending on the recording contract I sign (which is usually like a gig - I get paid for the recording once, like a gig, but I don't own it) But if I play in his band one night at a club and he hears my solo and digs some section of it and makes a tune out of that, note for note-- I can make no claim against him, because it was an improvised solo. But what I'm wondering, and it is really strange, is it different in John Cage's work because he's claiming to own the idea of framing what sounds randomly occur in that space in his composition as music? (I mean, if he made a recording of the concert)If this is true, then there's an interesting article to be written about you-know-what as it relates to classical and jazz and what each represents and the kinds of entitlements offered by our culture as a result of these pervasive race issues. Thank you! I'd never thought of this idea before as it relates to John Cage and jazz improv and ownership. It is certainly something to think about. Cheers!

It would be silly if actual human file these claims so the claims are very likely to be filed by some algorithm, which somehow found matching pattern in the noise.


1) Some copyrighted music has a period of noise that closely matched the segment of noise of the video.

2) There is adversarial pattern in the noise in the video that confused the hunting algorithm.

Three of the claims are for other white noise recordings. It's arguably not even an algorithmic mistake, but a result of someone claiming they have rights over a recording of white noise.

What's wrong with copyrighting a particular concrete example of white noise? A particular string of randomly generated bits? The fact that an algorithm just hears generic white noise is the algorithm's problem.

What's wrong is that copyright doesn't work that way:

"must have been created independently and contain a sufficient amount of creativity"

from : https://www.copyright.gov/circs/circ33.pdf

US copyright of course, ymmv but I doubt other countries are going to be more likely to allow white noise / PN sequences.

If we're talking about creative work, the state doesn't really have any authority in deciding what's adequately creative. A aiff file full of random or nearly random data sounds like a plausible form for a creative work to take to me.

Sure they do, by virtue of creating the laws that they then adjudicate and enforce, the state has the authority to legally determine what is creative, at least in the legal sense.

It's likely that your "creative work" will need to be similar to other forms of creative work in order to enjoy that status. Either that, or you need to earn a reputation within a creative community as a person how pushes the conventional boundaries.

In other words, randomly generated sounds are not widely considered to be creative in the conventional sense. So, unless you become the next deadmau5, it's unlikely that a court would rule that your random noise is considered art.

Those with no knowledge of the law constantly spouting off as if they do is easily one of the worst features of HN.

How do they define "creativity"? Because surely, a specific sample of properly random white noise would have maximum density of unique information (useful or not - but then is a song "useful"? Where's the line?)

It's not a precise standard, the exact boundaries of "creativity" would be left up to a judge to decide in a case where it was truly ambiguous. I'm pretty sure that this specific case would count as a minor modification to a work in the public domain, which does not qualify.


Yeah, that was kind of my point. Including non-precise words in a legal document is like saying "a wizard did it".

Using deliberately undefined or "non-precise" words is a core concept in legal documents. It provides leeway and "escape hatches" where humans can decide if a law actually makes sense in a specific case. This is why we are talking about "spirit" vs "letter" of the law.

The information does have to be useful or original. If there was something innovative or original about the way the random information was created, then that might be copyrightable, but if it's just arbitrary randomness, there's no reasonable claim.

You are confusing different things.

The issue is with making a particular concrete example of white noise and then claiming that your copyright over that particular example now extends to all white noise.

It'd be like John Cage claiming that, since he 'wrote' c4′33″ (consisting of 4 minutes 33 seconds of silence), any powered-off music-playing device is infringing on his copyright and must pay royalties.

> Credited to Batt/Cage

He was silly enough to explicitly give Cage a songwriting credit. That's the entire reason the case exists.

sigh That's what I get for assuming humanity is capable of being sensible about obvious things.

I agree this is pretty absurd but I could imagine that in this case the actual creative action (which I guess now can be copyrighted) might have been presenting silence as a musical score.

So my guess is your turned off hi-fi system wouldn't infringe their copyright - but if you produced an audio file containing nothing but silence and marketed it as a song, you would be infringing.

At least that's my try to find a spec of sanity in this.


If you just produced an audio file containing silence and marketed it as a song you would be fine. The only way you would be found infringing is if they could somehow prove that you copied Cage's silence (or produced your silence using the "score" to Cage's silence) instead of independently composing and producing your own silence. Basically the only way that can happen is if you are dumb enough to actually admit that you copied Cage's silence.

So you're saying it's the difference between 0's and NaNs. I guess that makes sense.

I have low expectations of human behavior when it comes to IP. They have a chance to be exceeded in that case, and when they are, that feels good.

> What's wrong with copyrighting a particular concrete example of white noise?

It would be functionally impossible for a start. Copyright does not cover ideas and information themselves, only the form or manner in which they are expressed.

Given white noise is a random signal having equal intensity at different frequencies it's basically like saying "what's wrong with copyrighting a particular number?" it is nonsensical.

When you're dealing with copyrighted files, every single one is just a very large number.

I'll see your large number and raise you an army of monkeys on typewriters, which if I follow the pseudo-intellectual logic going on in this thread correctly, means I own the copyright of everything that will ever be created in the future.

Or maybe we can all agree that just being able to describe, view, feel, smell or listen to something doesn't necessarily make it something that we can copyright or patent. We can walk back from the brink of madness towards sanity, it just takes the first few steps. :)

Right, but as - I repeat - Copyright does not cover ideas and information themselves but only the form or manner in which they are expressed then the computers information isn't the point - the expression of that information is.

Which is why it is not actually in violation to have a mp3 but only to upload or play it - as that is the expression. The digital data is not actually copyrighted.

Which is, again, why white noise makes no sense to copyright - it's random noise.

Actually, it's more like a function that maps a huge (maybe infinite) vector of very large numbers to a single concept in a human brain when interpreted in a very specific manner.

Think about how many ways you can encode a single song: there are dozens of formats, each with varying bitrates and other attributes. But no matter which "number" of that song you play, a human will hear interpret them all the same.

A copyrighted work has dozens of different encodings that would map to different files.

A copyrighted file is just a large number.

It's not the number that's copyrighted. It's the content that number represents that's copyrighted.

But when it's on a computer, the two are indistinguishable.

A cam recording of a movie is still copyright violation even though the cinema movie and the cam recording are different numbers.

> Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith.

Clearly I didn't mean that the copyrighted work is only associated with a single number. The "two" in the "two are indistinguishable" was in response the the comment above:

> It's not the number that's copyrighted. It's the content that number represents that's copyrighted.

This classic essay is ever relevant http://ansuz.sooke.bc.ca/entry/23

It's not possession that constitutes copyright violation, it's distribution. And that's up to a court of law to decide whether your actions constitute it.

pseudo random

if people think there is nothing wrong with it , then the concept of IP is dangerous to society

The concept of IP is that that there is no concept of IP. Copyright is one thing and that's all that's being discussed here. Everything else that you might call "IP" is different, has different origins, works differently, and is entirely unrelated to trying to determine if someone copied another video deliberately.

We're not saying that some people should be allowed to confuse consumers by trying to pass off their products as someone else's products (those are trademarks) or that people shouldn't disclose all of the details of their inventions in exchange for a short monopoly on those inventions (that's patents), or that anyone should be allowed to say that their sparkling wine was prepared in Champagne, France (that's regional designation).

"IP" is a broad, vague term that describes different laws, with different purposes. Laws that are so different that it hardly makes sense to discuss them in the same breath.

then the concept of copyright is dangerous to society

(could you really abolish copyright without bringing down the rest of IP protections with it?)

I think you could. It's completely different thing to have laws intended to not confuse consumers compared to laws intended to give authors exclusive reproductive rights over creative works (trademarks need not be creative or original at all, for example).

Ugh, this argument is tired and dumb. Thank you for bestowing upon us all the sacred knowledge that there are different kinds of IP, but we all already knew that, and choose to continue to refer to them, collectively, as IP, because they actually are similar in a lot of ways.

No, it's informed and clearheaded. We aren't well served by conflating trademarks (which, at their best, help with consumer protection via anti-fraudulence) with copyright infringement, or patent power. The similarities are far and few between, and mostly not relevant in any discussion. The term "intellectual property" is usually used by people trying to snow someone or someone who is confused about what's going on.

That there's confusion between the various concepts speaks to their similarities.

Copyright and patent are both intended to encourage innovation by providing the originator with a temporary monopoly on the use of works/processes/etc. Both are frequently discussed here in the context of large corporations/publishers using them to stifle the innovation they were designed to encourage.

Trademarks are often used in tandem with copyrights to protect a body of work in different ways - while "Steamboat Willy"s copyright will eventually expire (if it hasn't already), the trademark of the Mickey Mouse character will last indefinitely (so long as Disney protects it).

Distinctive trademarks are also commonly protected by design patents. Yes, they protect the work in different ways, but the two are related in their use and concept.

If you're so against collective terms, I hereby insist that you no longer refer to the "United States", or "European" culture. But I digress.

Patents are not intended to encourage innovation. Innovation happens all the time without patents. Patents are intended to abolish trade secrets! The bargain society offers with patents is tell us how you did it, because we know you're going to do it anyway. But if you tell us how, we promise to not compete with you for a short while.

If the trademark on Mickey Mouse were used forever, this doesn't mean we can't pass Steamboat Willie around. If its copyright expires, we can copy it, we can reproduce it, we can exhibit it, we can remix it; we just can't confuse the public into thinking we're Disney while we're doing any of this. If Disney thought it could prevent people from copying Steamboat Willie without copyrights, it wouldn't be lobbying for infinite copyright on it. The best Disney could do is make us state, in no ambiguous terms, "Mickey Mouse is a trademark of the Walt Disney Company" somewhere very prominent while we do whatever we want with Mickey.

A trademark doesn't mean that the trademarked thing is Lord Voldemort. Otherwise, every time we said "Microsoft" or "Apple" we'd be infringing. It just means that we might have trouble calling our OS "Hackintosh" or "Lindows", and even that's debatable in court (Microsoft was about to lose the Lindows case).

I like how you define the intention of patents but I am fairly sure that is not the intention in the law definition of patents. And while what you say is nice it simply doesn't work. Anything worth keeping as a trade secret and able to be kept secret will remain a trade secret regardless of the existence of the patent system.

It would be very interesting to live in a society where once a product is available on the market any trade secret on it is illegal and the only option is to have patents that you receive royalties for but you can not restrict their use.

> If you're so against collective terms, I hereby insist that you no longer refer to the "United States", or "European" culture.

that is being reductive. sufficiently so that you're basically straw-manning their point.

i would wager that, for example, extremely few people who view patents as negative (at least in this era), also view trademarks as negative.

IP = suitcase word (it contains any and all sort of meanings you can cram in).

wouldn't it have to be recognisable/distinctive to be copyrightable?

That's only in a world where copyright was used to encourage interesting work. Instead, it's just another property right, and there's a land grab for every possible permutation of bits, or genes, or whatever.

Isn't anything that can be converted into bits and compared recognizable and distinctive?

No. If that was the case you could buy a vinyl record, play it a few times (such that some slight crackle developed from wear) and declare it a derivative work. If the distinction between two slightly different works (eg two white noise recordings) can't be perceived by a reasonable person (or a jury with 6-12 of them) then it doesn't exist for legal purposes.

You could sit there in court with a hex editor and show how every byte of two noise files in different, but nobody procures or pays for audio recordings for the aesthetic pleasure of hex dumping them. Law isn't a competition to see whether you can figure out a way to articulate what the difference is.

There's nothing wrong in theory with someone claiming copyright over white noise. If someone else blatantly copies their white noise, they are perfectly within their rights to file a copyright claim.

However in this case it does seem to be youtube's content matching algorithm that is finding the match (as far as I can tell, but it's hard to know for sure). My guess is that their matching algorithm just doesn't work well for white noise.

Copyright, in USA, requires a creative component. Facts, randomness, and other stuff lacking any creative expression is not covered by copyright. The bar is ridiculously low as in you can claim copyright on particular formatting or page layout of say names and phone numbers.

I could imagine, it's a stretch, a specific "arrangement" of white noise being copyrightable. But in general it is wrong to think you can copyright white noise.

I'm pretty sure John Cage's works are under copyright, which is interesting since he made such a big show of erasing himself from the work and embracing the random. It never occured to me how obtuse that was. I'm thinking particularly of 4'33". So, I guess the 'him' part of that composition is the fact that someone sits at a piano in particular, rather than say... bagpipes or even pipe organ. I wonder if it would be considered satire or a derrivative work on ukelele choir...I wish he was alive to ask him his thoughts on this.


> I'm thinking particularly of 4'33".

I read that using Apple's Text to Speech (macOS High Sierra) like:

> I'm thinking particularly of 4 feet 33 inches.

LOL Apple. You haven't heard of Classical Music yet, have you?

As you said, I could imagine that maybe, maybe a specific white noise file is copyrightable. This would be similar to how typeface copyrights work. The specific font file for Helvetica Neue is copyrightable, but printing out the letters, tracing them, and digitizing that to make your own font is legal (in America at least).

There is always going to be some kind of creative component in actually generating the white noise. And even if you argue that white noise is theoretically uncopyrightable, you have to remember that it is also theoretically impossible to prove that any particular piece of sound is actually truly random.

Neural networks are easily fooled:


Almost as easily as humans.

I had the same problem happen years ago with video using a song that's way out of copyright and already adjudicated in courts as being in the commons. I think most use automated systems to attempt to claim copyrights over your video. At the time, it was very hard to contest claims.

While a lyrics or a melody might be public, it doesn't mean the song is.

While a song might be public, it doesn't mean that a specific recording of the song is.

The owners of the rights to the recording can only exercise rights over their recording, not on the song itself.

Sorry to nitpick, but if we are going to talk about legal issues let's, at least, be precise.

Yes, that's what I meant.

I recorded a video once and added ocean wave sounds to it. I googled "ocean wave sounds" and used part of someone's recording of the ocean. When I posted it to Facebook it got taken down for copyright infringement.

I appealed and they put it back, so there's that.

IMO that actually would be a legitimate case of copyright infringement (not that I'm a proponent of it) - I'm surprised you were able to get it put back.

There needs to be an element of creativity for a work to be copyrightable in the US. If it's just the same sound as you would hear from going to a beach on any windy day then it may be hard to argue that there is any creativity to it.

An audio professional would probably argue that there's a level of creativity required to create and edit such a recording for commercial use, but I'm no expert on the subject so you may be right.

Due to a combination of technical limitations and the coconut effect, many natural sounds don't sound very natural if you just naively record them. Often the artist has to edit it at least a little to make it sound good.

A common misunderstanding: There is no formal/legal claim of infringement being made anywhere in the ContentID process, it's just the automated process YouTube uses to decide whether or not to take down a video.

Any formal/legal claims of infringement should be made by the right-holders industry towards YouTube (which YouTube really wants to avoid) not the uploader. The DCMA was made to facilitate this, but for a huge platform such as YouTube it just falls short. Here's the important part, let it sink in very clearly:

ContentID is a deal that YouTube made, demanded by the larger players in the right-holders industry that goes (something) like: IF the big right-holders can easily make large amounts of claims, and YouTube provides some sort of ML detecting filter to scan for possible infringements, and when flagged the ContentID process will handle these cases such and such, THEN the big right-holder industry people won't file gigantic claims of infringements and take down requests to YouTube.

From the perspective of the little uploader, no courts, claims or legal things are involved. You have no legal relation with the right-holders when YouTube takes down your vid for infringement. You can't sue YouTube for taking down your vid since as a US corporation they have every right to take down any video for whatever reason (and TOS, but mostly because big corporations just get to do whatever). And that reason is mainly: Whatever keeps the big right-holders happy.

YouTube is first and foremost interested in keeping the big right-holders happy, and of course keeping their advertisers happy. That's it. They're not interested in being fair or even ethical (as the recent Logan Paul drama has shown us--considering that YouTube's trending videos are hand-picked by humans ...).

ContentID is not at ALL about YouTube dealing with copyright infringement in general. If you write, perform and record an original song with your band and upload it, ContentID is not going to protect you. It's just there to help the big players scan for their own stuff and then flag whatever. If they accidentally flag the wrong thing, original content or even the actual original sample (if it's not licensed to the big right-holders) that was used in a newer big right-holder licensed work, then it's going to get removed. A false positive is only a very tiny little "oops" for YouTube, they don't really care. The dispute thingy is just a pacifier.

They have literally zero reason to reduce their false positive rate. The software works; the big right holders are happy.

Just imagine if they all claim each others videos lol

Jim Sterling adds multiple instances of copyright infringing material to some his videos to take advantage of a similar effect[0].



All of these claims have been filed under penalty of perjury.

There has to be a reason for why not a single copyright troll has been prosecuted for it.

On my youtube channel I use Kimiko Ishizaka's open source bach recordings (under CC0) as background music. I got copyright claims every single time (from the same trolls -- forget their name). Eventually, when replying I got the idea to say, "You have made a claim on the same recordings for every single one of my videos. You have subsequently relinquished the claim when you discovered that I had a license to use this music. Given that you know that I have a valid license, you seem to be acting in bad faith. Please do not flag my videos again".

And, you know what? It seems to have worked. They never flagged me again.

Glad to hear this!

It's a complicated problem. In classical music, ContentID has a hard time differentiating between various recordings of say the Goldberg Variations. People were getting dozens of ContentID matches from Glenn Gould, Andras Schiff, and so forth, and each time they'd dispute the claim, the next one would come with a different recording. Nobody was doing anything nefarious, but you couldn't get any peace. Soe we (Kimiko and myself) decided to put her music into ContentID ourselves. Now it mostly identifies correctly as Kimiko's recordings, and people can ask us (robert@opengoldbergvariaitons.org) to de-monetize any video where the ads bother them. Some people don't mind, and as a result, Kimiko gets a small amount of revenue for her work as well.

Write a letter to your state DA asking them to pursue barratry charges against the lawfirm.

Unfortunately (?) neither I nor the company claiming infringement are in the US. I think the reason they actually gave up is not for legal reasons, but because they have to abide by Youtube's terms of service in order to make claims. If Youtube kicks them out, they lose their revenue stream, so potentially serious for them.

I have a video of my own noise generator audio plugin. It's repeatedly been copyright struck by somebody who's put white noise into YouTube's ContentID. Every time I notice this pop up again I dispute the claim and they give up.

They're not in the US. I think they were East German? This is not new, I'm sure it's literally the same people.

> Kimiko Ishizaka's open source bach recordings (under CC0)

Thanks for the pointer. I hadn't heard her recordings of the WTC and Art of Fugue, and they seem pretty good.

Unfortunately the only part that has to be under penalty of perjury is a declaration that "the complaining party is authorized to act on behalf of the owner of an exclusive right that is allegedly in-fringed." So even if these are bullshit, as long as they are acting behalf of the owner of the work they claim it to infringe, they haven't perjured themselves.

And that's for a standard DMCA takedown, which these may not have been. I believe YouTube has their own process separate from the DMCA which may have looser requirements. (They still have to accept standard DMCA takedowns to get the safe harbor protection, but it doesn't mean they can only act on takedowns that are DMCA compliant)

They likely weren't. ContentID enforcement doesn't require a DMCA takedown request. It's essentially a pre-get-laywers-involved level of enforcement. That makes YouTube the judge.

The industry was careful when they wrote the DMCA to make sure it couldn't be used against them. Theoretically our lawmakers should have taken the public interest to heart before passing the bill, but that appears to have been too much work.

Isn't youtube's suck up to industry style of process not at all subject to that sort of penalty?

I do not really understand what the problem is about (I am not a lawyer) : why not simply respond "sue me" and request 10 M dollars of prejudice for false accusations?

The next copyright claim should be more careful, I guess

Youtube is about taking culture and adding your own touch to it. Mostly fair use. This copyright thing is just as stupid as software patents.

To be fair, the copyrights claims are from other videos containing white noise. An algorithm that matches two white noise audio by claiming they contain the same information is doing its job pretty well.

There really should be penalties for filing an incorrect complaint.

Maybe the white noise messes up the AI trying to find songs in it...

It would be nice to see a timeseries correlation.

Have you guys listened to Justin Bieber samplings? I bet he is the one claiming copyrights for all that white noise.

The edge case to this classification problem is quite literally noise. If you filter out noise, you're kinda in a pickle ... a Pickle Rick! ... I'll let myself out..

Applications are open for YC Summer 2018

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