Hacker News new | past | comments | ask | show | jobs | submit login
YouTube blocks recording of public domain Schubert/Liszt composition (youtube.com)
176 points by imiller 26 days ago | hide | past | web | favorite | 101 comments

The big problem with classical music is that the compositions are all in public domain, however most performances are not. Given the subtle differences between performances, algorithms are not smart enough to figure out the difference. In fact, even humans would have a hard time to distinguish between two performances of certain classical pieces.

I don't think it's so much that the algorithms aren't smart enough to tell, it's that Google have had to make the algorithm still match in the case of people altering music with filters/effects, time stretching, pitch shifting etc. - so it's been made deliberately a lot less sensitive to most of the factors that a musician brings into their interpretation.

That's fine - it's probably necessary for Google to not be constantly dragged into the courts by actual infringement slipping through with small alterations. The issue is that there is absolutely no recourse for an incorrect match. There's no option to say 'no, the algorithm is wrong' or 'the claimant has made a false claim' - the nearest option is to claim fair use, and that kicks it back to the claimant to be able to lie again if it was fraudulent, with no fear of consequences.

There needs to be some process where it can be appealed to a human. I think the best way I have seen suggested is to stake a small amount of money on it to have somebody review it - say, $15 from each party, and then you lose it if it's ruled against you, or get a refund if you win. That should easily pay for half an hour of somebody's time, so they could hire staff to do it.

If the claimant lost and still didn't agree, it could be escalated to DMCA, where there are potential penalties for fraudulent claims (even if it doesn't happen much).

Make it $50 and give half to the winner.

Or you could even turn it into a system where you can post a (prepaid) bond for a video. Ex: $50 gets automatic human review. $500 gets a phone call from a rep. The catch is you lose the bond if you're legit infringing. Make it channel wide and I bet there's a class of YouTubers that would be willing to post multi thousand dollar bonds to assert their legitimacy.

But why would you want anyone to have to pay anything? Google is one of the richest companies. This should be simply the cost of doing business. Google wants to make money by removing humans completely. They should have to pony up the expenses for when their system fails...

The cheaters and liars need to be penalized. If they're spending others' money they won't stop trying.

>The cheaters and liars need to be penalized.

You want Google to adjudicate on more things?! Now you're going to need an appeals process for when they get it wrong here as well.

They already are, this is the same process and it already has an (ineffective) form of appeals. The suggestion is only to add cost to it to disincentivize abuse.

We don't necessarily disagree, but what bothers me is when people blame Google for this, as opposed to focusing on the draconian laws that have forced their hand.

poor Google, making billions of dollars out of other people's work with an automated system where it's impossible to talk to a real person. feel real bad for them

Poor me for the results our idiotic laws have produced. I do in fact feel real bad for me.

These laws didn't appear out of the blue, they're created and lobbied for by rich companies making money off copyright. Blaming Google for facilitating further abuse of copyright laws is a fair thing to do, and frankly, the unlikely event in which Google uses its money to put pressure towards fixing these laws is still more likely than the laws getting fixed because "we the people" somehow coordinated enough on the issue.

Seems like a great solution if it doesn't get abused by claimants. Maybe the verification status should remain permanent, or further disputes should escalate the sum staked for the claimants.

Still sounds like a Youtube problem, not with copyright.

The amount of false positives Youtube's copyright system has generated is crazy. It's clearly broken and wide-open for abuse by bad actors.

Even with the complications of copyright there are so many examples of people having entirely fair-use videos taken down. It goes well beyond just the difficulties of two works one copyrighted/one not matching.

This is a side effect of having onerous copyright laws. It takes quite a bit of engineering effort that could have been spent on other projects. This means economic loss and frustration for consumers.

As an aside, without copyright, we could have both Disney and Sony Spiderman movies. The consumer wins, and there's more than enough money to go around for competing studios.

>> "As an aside, without copyright, we could have both Disney and Sony Spiderman movies. The consumer wins, and there's more than enough money to go around for competing studios."

Wouldn't that be trademark, not copyright?

The problem is fair use isn't something that can be algorthimically determined. All you can really do algorithmically is say 'this bit of content matches this bit of content' determining if that match is part of fair use requires determining the context in which it's used which will be a hard problem until we either figure out uploads (and some poor schmucks' uploads have to watch and filter YT videos all day) or general purpose human level AI.

There are some heuristics that could be used but so many of them break down in pretty common cases. Eg: If we say using a small portion of someone else's video in a larger video won't strike it leave out some types of videos like Destin from Smarter Everyday or the Slowmo Guys where the most interesting shot could be 3-5 seconds total or if we just measure by length of copyrighted material vs the whole video we'll wind up with bit 10 hour compilations of copyrighted stuff from different places maybe. It's a hard problem and unfortunately Google (pushed somewhat by the law and somewhat by their own internal decisions) take a flag and contest approach that favors the 'original' creator.

Yes! And Kubrick knew this: for "2001: A Space Odyssey" he bought the right of "Also Sprach Zarathustra" directed by Böhm and replaced it at the last minute by the one directed by Karajan hoping no one would notice.

Citation? That doesn't seem to match the story as given in https://en.wikipedia.org/wiki/2001:_A_Space_Odyssey_(soundtr..., for example.

There is some reference to this in French:

> Les droits d'enregistrement de Karajan n'étant pas disponibles, c'est la version de Karl Böhm, avec l'orchestre philharmonique de Vienne, qui fut créditée au générique. Mais au cours de la postproduction, Kubrick remplaça discrètement l'enregistrement de Böhm par celui de Karajan et personne ne le remarqua3.


I own many recordings of Liszt. Some performances are better than others.

It's ironic that a system of rules (DMCA) that was designed to help performers not lose revenue is actually causing a performer to lose revenue. Rousseau is one of the most recognised pianists on YouTube. He works tirelessly to perform a new piece about every week. This must be incredibly tiring, and the skill required to be able to do that week in, week out, is absolutely breathtaking.

If anyone deserves to be made the poster child for the damage the DMCA and YouTube content ID is causing, it is Rousseau.

And how is it that an hour after this is reported, the problem still persists!? A million people are missing scheduled content. If this happened to a television station as often as it happens to YouTube there would be widespread, vocal, public complaints. Their competitors would also be having a field day.

The DMCA is for publishers, not creators. If Rousseau would just conform to the system and give a publisher a cut everything would be fine. Independent creators have always been the enemy.

Edit: I am being completely sarcastic w/ respect to the second sentence. I believe that there are people (those who are enriched by the current copyright system) who do believe that, though.

I assume you are being facetious. Rousseau doesn't owe anything to any publisher. Just checking you understood this.

I am being facetious, but I daresay a lot of people who work in the publishing industry sincerely believe that viewpoint.

I believe we're living in a time when publishers are becoming increasingly irrelevant. They're clawing onto anything they can to retain their position as middlemen and as a tax on commerce.

Years ago read a music exec claim that 'republishing' content should re-extend copyright.

The publisher is not entitled to that cut. In fact this doesnt involve the DCMA at all but YouTube's own process.

I don't think publishers are entitled to anything, but I bet that publishers do. Content ID exists as a response to the DMCA, which was heavily lobbied-for by publishers. Content ID wouldn't exist if not for that lobbying.

Content ID is not the DMCA. Content ID got it wrong here, and that happens all the time. It's not a big deal for the claimer. Which is totally contrary to how DMCA works, if someone makes a wrong claim there it gets very expensive for them.

Youtube could easily copy that aspect of the DMCA and make it very expensive for mis-claimers, and easy to counteract automatic mis-classification. That they don't has nothing directly to do with laws, it's simply Youtube being a bad platform.

> Content ID is not the DMCA. Content ID got it wrong here, and that happens all the time. It's not a big deal for the claimer. Which is totally contrary to how DMCA works, if someone makes a wrong claim there it gets very expensive for them.

That simply is not true. DMCA misclaims are pretty much not punished at all in practice.

The provisions are there. If they are not used it's either because the violated party does not sue or because the US justice system again is not working properly. But at least that's how the system ought to work. With Content ID there is no recourse at all against misuse, it all relies on Youtube to care. Which it doesn't seem to do at all.

The provisions don't protect against a wrong claim but a purposefully false claim. Unless you can come up with proof that the claimant knew they were filing a false claim you are out of luck.

This is everything to do with the DMCA. YouTube claims it is required by law to take down content that receives a takedown notice. That law is the DMCA. Edit: I'm probably wrong. It seems a claimant can also issue a block. They do not have to issue a takedown notice. That seems to be the problem right there!!

(However) I find it really difficult to believe the claimant in this case has no control over what YouTube does, as they have previously claimed. It may have been Content ID that got the identification wrong, but the claimant should not issue a block if it is a false claim.

Anyhow, Rousseau found a way around it by uploading a new version that received a copyright claim from a different company, instead of a block. He'll still have to go through the dispute process, at least we can watch the video because YouTube is not obliged to (and does not) take down the video.

Youtube's content ID system is not the DMCA. It stands in front of that system, it never gets invoked. DMCA has a notice/counternotice system with a formal requirement for what a notice requires. Content ID and the monetization nonsense that entails can be whatever Youtube wants it to be.

> Which is totally contrary to how DMCA works, if someone makes a wrong claim there it gets very expensive for them.

No, it doesn't. There is only one part of a DMCA takedown notice that is made under penalty of perjury, and it's not the part that is usually at issue in a false notice.

If anyone is wondering, Schubert and Liszt didn't collaborate IRL on these. Schubert wrote the pieces, died, and then Liszt transcribed them for solo piano after they had posthumously been found in Schubert's belongings.

Thank you! The answer to this question is why I even clicked on the link. (A collaboration like that would be awesome, though.)

YouTube's filter is so sensitive it's scary. Music from cars driving past in vlogs, or even <2s of music is enough to trigger the filter. Dialling the sensitivity up so much is bound to have second order effects negatively impacting somewhere else.


From @hyperionrecords regarding a recent unrelated incident. Hyperion's content is the one making the claim against Rousseau's video.

They seem aware that ContentID often misidentifies claims, but there doesn't seem to be a public statement on what they proactively can or will do to help fix it.

The usual canard is:

Google is a private Corp and can do whatever the hell they want.

The real problem is that tech companies are monopolies and oligopolies and they hide behind "The people can just leave if they want".

The end result is that monopolies beget and work with other monopolies to strengthen each other. So we see YouTube working with RIAA and MPAA on non-DMCA underhandedness and what is clearly fair use (car driving by with 2s of a tune).

I think the real problem is no one (or not enough people) want to pay for a service that hosts user uploaded video, filters each and every one accurately for illegal/copyrighted content, and then is able to serve it worldwide 24/7 instantaneously.

That’s the reason I think no alternative service has sprouted up.

That's just another reason to consider advertising-funded services as fundamentally anticompetitive and get rid of them.

We had a policy for that built into the DMCA. It was a procedure that copyright holders could send an email and go through a process of attestation and rebuttal.

Google shortcutted that by working with music and movie studios and made a ML version that bypasses the DMCA. The harm is caused to everyone not represented by a major media company.

In the end, I have no due process afforded to me by the DMCA because Google bypassed it for cronyism.

And enough people aren’t willing to pay for the people receiving that email.

Nothing stopping anyone else from starting a similar service as YouTube that offers that kind of service, other than lack of money to burn.

I'm not sure what to fix.

Let's give the benefit of the doubt to Google and say they're constantly revising their algorithms; even then, due to the sheer amount of video uploaded to YouTube every day, there will be some of them that are misidentified, it's simply unavoidable IMHO.

I think the improvement can be made on the smoothness and speed of disputing/reviewing these claim, though. This process should be as short and painless as possible, for sure.

In OP's case, considering it only seems to be a couple of hours, I think everything is still OK-ish at this point.

I didn't notice the timeline of the OP's post, but they could pressure Youtube to implement something that requires manual review before action is taken, instead of automatic action. Especially with a genre that typically has lots of similar recordings of out-of-copyright compositions.

It's an entire genre that ContentID is incompatible with.

YouTube seems to be taking the brunt of the criticism when it's the repercussions of copyright infringement lawsuits that are the problem.

I'm sure YouTube would love to be more lenient. If they could say "anything goes" they'd be thrilled, but we know they cannot because of the LAWS that are protecting IP owners.

IP should be protected, but to what extent?

And the bigger problem is the amount, and rate, of which content being produced threatens everyone's ability to do anything!

Try naming a business. NOTHING is available. If you want to run a competing business in a saturated market, the naming options are almost non-existent, especially if you want to name your company something that embodies what your business does.

Music is having this problem too. When you can take a few stanzas and say you OWN that music, could someone not just use a computer to generate every possible sequence of notes, slap a copyright on it, and own all future music?

Photography, video, games, puzzles, speeches, books..... it's all about to hit a brick wall if we can't be more lenient with IP.

In my opinion, complete works should be awarded a copyright, but sub-sections of works should not.

And software patents..... dear god.

Meanwhile, book publishers are suing Amazon because they added a captioning feature to make audiobooks accessbile for copyright infringement: https://gizmodo.com/is-captioning-an-audiobook-illegal-major...

Reply All also has an interesting story about how impossible is it to have ANYONE accountable AT ALL for DMCA takedown machinery that harasses people: https://gimletmedia.com/shows/reply-all/76h5ro A person got strikes on his internet connection due to previous owner and there was noone in the insane chain of IP strike machinery that could be appealed to or own up to the mistake.

The current state of copyright is so broken that it's downright evil.

This is all historically bound up. Audiobooks and printed books used to be separate physical things. The publishing rights ecosystem was built assuming this.

Now that they're fungible, its not making much sense any more. But millions of publishing contracts exist, that enforce the distinction. What do we propose to do about it? Without dismantling fundamental contract definitions.

20 years ago I made the observation that the whole intellectual property domain was fundamentally based around all of the individual aspects of intellectual property being cleanly and completely distinguishable from each other, e.g., this is a book, this is a song, this is a movie (which is made of this song and this script), etc. But even then it was becoming clear these categories were getting fuzzier and fuzzier and in many cases outright merging, and the categories weren't going to just go together nicely because of the various contradictory assumptions in them.

The worst then, and I think still today, is computer software, which is covered by both patents and copyright. Copyright has always covered them, in the last 20 years, patents have covered more and covered less at times, but still have always covered some of it.

I'm actually a bit surprised that I've still seen no hint of the legal system directly grappling with this problem. It just keeps getting worse. I've been keeping my eyes out but I haven't seen anything.

And this is another example; nobody wrote the copyright laws for the YouTube case. If you get down to it, you have this intricate interplay of video, audio, automated systems generating some content, automated systems checking them, an extra-judicial system for mediation where I will not say the incentives are 100% misaligned, but they are certainly nowhere near 100% aligned with what they should be, with a lot of mismatches between YouTube, the copyright holders, and the video producers in terms of powers and responsibility (that is, there are many places in the system where there is power without responsibility, and responsibility without power), and we're trying to adjudicate it on 20th century law that was already, at times, kinda cobbled together. (Don't look, but we're already 1/5th of the way into the 21st century.) It's madness.

>I've still seen no hint of the legal system directly grappling with this problem. It just keeps getting worse. I've been keeping my eyes out but I haven't seen anything.

IP is modern capitalism's goose that lays the golden eggs. I suspect that they are somewhat nervous to do the kind of root and branch reforms that would be neccessary to solve the problems you have outlined. Part of the genius of open source licences is that they are built out of some of the aspects of law that capital is most wary of breaking.

> but we know they cannot because of the LAWS that are protecting IP owners

I don't think that's a fair assessment. I played a bunch of less well defined tactical competitive games when I was younger and even though the rules were not perfect we were able to yield outcomes that sufficiently would be satisfactory to everybody and in many cases also representing of the skills of the competitors, i.e. good games.

The problem with "real world games", e.g. like this question of "who gains the most from a creation?" is not that the rules are not perfect, but that there are many players who don't have the slightest intention to play good games. There is good argument for why the creator should gain the most, or his family, or the company that produced the company in a market worthy format. There is even an argument for that the person should take it who through his cunning or strength is simply more able to take the gain. But there is a big difference between competitors trying to win while enjoying the game and those who want to win no matter the cost to everyone, including themselves.

On the gaming table you simply exclude these people, but in the real world you can't.

On Youtube they will probably not take down a video if nobody claims its protection rights. So someone came and claimed these videos.

And that's the problem in my eyes. If everybody is trying to create an enjoyable environment for everybody else nobody would attempt to "generate every possible sequence of notes [for personal gain]" because it wouldn't even be fun even for that player for long.

The problem with awarding only complete works is: bad people exist. Laws need to have enough flexibility so that truly devious people are covered by the law as well.

Example: I copy 99% of a popular song, but the first 3 seconds of my song = me playing a trumpet at full blast. Technically it's not a copy of the complete works, so I can argue in court that it's a sub-section of the original song. No judge/jury would agree with me, but my point is shown in this example. Even if you specify your definition sub-section to be < 50%, you would still be unhappy, because I copied half your song. What sub-section is OK?

The YouTube/AI issue is that loose/vague laws assume a certain degree of "common sense" among the population. When you try to give a loose task to a computer, it must make assumptions (either with AI, or pattern matching) and will invariably screw up. Computers are not human brains so they don't have "common sense"... yet.

And the flip side is that if you wanted to write the 1812 overture you couldnt because that quotes La Marsailles.

I'd argue dotcom domain squatting has a big part in the lack of names available. Eventually, most startups that could use them will go bust and sell the domain 5-10 years down the line anyway, it's outliers like google.com that are always going to be taken.

New namespaces like .app should help with the problem, but at the end of the day, the culture itself has a limited namespace.

I've always gotten fairly lucky with .coms for my projects (managed to snag feedsub.com recently, had minfolio.com at one point, etc..), though the ones I've let slip ended up being squatted upon for thousands of dollars because of their potential (I think) branding value, which is a shame because it could've been put to better use since I couldn't follow through with a project.

BTW "Google" is a kind of good, inventive, unique name, rather unlike to be squatted. I wish more companies came up with unique names that are not common words.

But have you recently visited pets.com, mail.com, delivery.com, or whatever were the generic names commanding millions in 1999? Pets.com in particular just redirects to a different domain, more trademark-able.

Yeah I chose a really poor example admittedly.

> Try naming a business. NOTHING is available. If you want to run a competing business in a saturated market, the naming options are almost non-existent, especially if you want to name your company something that embodies what your business does.

To be fair, this is only really a problem for a new tech-ish company. Most other companies are geographically limited and can operate just fine sharing an operating name with a different company. There are lots of different companies operating as "John the Plumber" (seriously, Google it), "Rose & Crown Pub", or "Pho Noodle House".

The US government has two options as I see them:

1) Reform the copyright system (unlikely due to powerful lobbyists and lawyers)

2) Eventually lose control over the industry to other countries as consumers vote with their feet.

You missed:

3) Bribe and capture legal system so everyone is forced to pay to the big copyright owners and you can't vote with your feet anymore.

I suppose that's the current reality right?

I'm imagining a future where everyone (re)discovers torrenting. Or Russia/China make a Spotify competitor with entirely stolen content and the government officials there turn a blind eye to Universal's rights.

"generate every possible sequence of notes"

do you have any idea how many songs that would be? your computer won't be finished in this lifetime


When constrain the search space based on commonly-used scales in Western music the space collapses dramatically. Given that there have been successful copyright claims made on short phrases I think the idea of exhausting the search space for popular western music becomes a lot more plausible.

the actual computation was left as an exercise for the reader, but the issue is not the length of the scales, it is how they can be combined. take 10 notes and combine them in 16 ways. two small numbers, but how long to count to all 10^16 possiblities? or rephrased: how many days in 10 quadrillion seconds?

> I'm sure YouTube would love to be more lenient...

I downvoted you because this is not true. YouTube has built their own takedown system that goes above and beyond the legal requirements. They could be more lenient to users and follow the law but instead they made it easier for content to get taken down.

The problem is that there's evidence that YouTube was purposefully uploading and allowing copyrighted content, which means they are no longer fully protected by the safe harbour provisions of the law, and risked losing the Viacom lawsuit. ContentID came as a compromise from that.


Please don't break the site guidelines by insinuating astroturfing or shillage without evidence. A user posting an opposing view is not evidence.

This is an important rule. I've written at length about it, if anyone wants explanation: https://hn.algolia.com/?query=by:dang%20astroturf&sort=byDat...


This is a website primarily dedicated to the creation and growth of corporations.

I used to use music from musopen.org that contains recordings that are either public domain or creative commons, typically from the Air Force band, or various small town orchestras or college or even high school orchestras for background music for my little space game videos, but dealing with the copyright claims became too much of a hassle, so now I just rapidly strum some chords on an acoustic guitar and run it through paulstretch and throw some reverb on it to get some ambient space music sound.

Yeah, I do the same thing with my Korg Kaossilator. For ~$100 (I have the original, not the updated one), you get a phrase sequencer that just about anyone can learn enough in about an hour to be able to generate their own ambient music.

I feel like music education for children starting with pianos, guitars, and recorders is backwards. Get them playing music first. The interest in instruments will come.

This happens to me all the time. As an amateur pianist, whenever I sit down and play a piece and upload it, it nearly always gets a copyright flag -- even if it's 150 years old or more!

Content ID is never going to be perfect (so there always needs to be an appeal mechanism).

But I'd always expected that false matches would be due to some kind of hash collision, not different recordings of the same piece.

Because YouTube doesn't know what the composition is, it only compares to other performances (which are under copyright).

And the fact that two different pianists could interpret and perform a work identically enough to generate the same signature Content ID uses is... astonishing to me, from my classical music background. If I'd had to guess, two musicians normally probably couldn't achieve that even if they tried.

So this seems to just be a crazy statistical fluke? Otherwise classical performances would be getting blocked left and right, given the hundreds/thousands of different performances of the same underlying music? Or is something more advanced that audio fingerprinting going on?

Content ID is designed for flexibility as well: it has to catch a lot of transformative that might be applied to try to bypass it. Perhaps you've noticed that actually evasive material has increasingly terrible effects layered in as the years go on... (I remember recently finding a cartoon that looked like it was a zoomed in camcorder recording from a CRT... But emulated in effects layers.) I've also seen things subtly slowed down or sped up, strangely cropped, bordered, then given animated borders once that stopped working.

Between an incessant evasion arms race and a content generation industry all to happy to profit from false claims, I just think to myself, "this is why we can't have nice things."

You can tell it's not a fluke if you look at the copyright claims that are sometimes available in the "Show More" section. Frequently, it will show claims that recordings of the same piece by other performers are part of the video.

Here is one example: https://www.youtube.com/watch?v=RxDCgFFBk20

Two of the movements have correct copyright claims for the performance, but another of the movements has a claim for a recording by a different conductor and orchestra. I see this all the time.

Like another person said here, we don't know how Content ID works or what exactly it's looking for, but it's clearly not an "exact match" of sound waves or something: it has to catch covers of compositions that are still under copyright, for example [1].

[1] https://diymusician.cdbaby.com/youtube/posting-cover-songs-o...

I had a recording of the Revolutionary Etude, filmed on a phone (iirc), played on an out of tune Acrosonic piano (more a piece of furniture than a functioning piano), with no pedal usage, get copyright matched with a Sony published recording. I disputed the claim and that was it.

It's about as far from a professional recording as you can get on YouTube, except for some piano recital videos featuring ugly children inanimately reading their way through the sheet music.

Hash collisions? No, I don't think you understand how these content matching systems work.

well, that does not seem to be a very constructive answer. mind sharing some info/references on this specific one, or any others ? because I though it's a blackbox type of software

I mean how exactly they work isn't public knowledge. But we can be sure they don't just hash anything. It would be too easy to just add small amounts of noise and defeat them.

Here's a reference I googled for you https://www.toptal.com/algorithms/shazam-it-music-processing...

I think you misunderstood the parent comment ( or I misunderstood your reply to that). He talks about about a signature (which is pretty much what it comes down to), not a file hash as in here's a hash of the file contents. That would be meaningless.

As for the reference you provided I know how shazam works, they pretty much made a blog post about it. This is not that. Not by a long shot.

I can't think of a useful copyright matching technique that wouldn't flag different renditions of the same song. That's my point. The failures of these techniques wouldn't be hash collisions but songs that legitimately sound similar.

Rousseau found a way around the block. He uploaded the video again in a way that caused it to be claimed, rather than blocked, by a different company. He'll still have to dispute it, but YouTube is not obliged by law to take down content that only received a claim. The video is no longer blocked.

This is a shame. I love Rousseau's performances of classical pieces. It really reminds me how little I am capable of playing the piano. ;)

On a slightly more serious note, this is what happens when you force a content host to bypass the normal flow of copyright claims.

Youtube seems to have gone to a ridiculous extreme on filtering copyright material. They blocked a non-public video of my daughter's Irish step dance recital, because the live band played three traditional tunes in the same order as found on some commercial recording. It took months to get the "copyright holder" to grant us permission; ridiculous.

YT doesn't, and can't, have enough resources to correct every such situation, so they err on the side of the copyright holders. When they're receiving a billion submissions a day, the innocent ones caught in the net are just a rounding error.

I'm wondering whether Bitchute might take up the slack here.

I have a slightly tangential question: can someone link to articles and writeups that discuss how YouTube and other platforms efficiently disambiguate between a valid/allowable recording and a prohibited recording, when the content between valid/prohibited is very similar or near identical in the most obvious features (e.g. rhythm and key).

Answer: They don't, and you're guilty until painstakingly proven innocent.

I don't disagree with your cynicism about the flawed and slow disciplinary process. But it's still in their operational best interests to improve and iterate on their auto-classifier algorithms.

Another huge problem are the companies making money simply calling copy-right strikes.

Rousseau is actually active on HN, curious to see what he says about it here. What a ridiculous DRM implementation by YT. I understand that with such a large volume of data, it's a difficult challenge. But it should not be difficult for an algorithm to figure out that Schubert or Liszt are in the public domain.

They are but the recording is not: it's copyright belongs to the performer. Unfortunately YouTube can't figure out the subtle differences between performances that identify them.

The headline is slightly inflammatory in its reference to the "200 year old composition." Exactly to parent's point: a contemporary recording would be neither 200 years old not in the public domain.

I'm not sure I agree with inflammatory. It seems that the common attribute between the two materials is that it is based on the same 200 year old composition and not much else. Like if IT takes your keyboard and gives it to someone else because they are the same keyboard. Completely missing subtle differences like the serial number or the fact that on one the space is worn out and on the other the tab key.

Composition is entirely different from recording. Composition is the written score, which in this case is ~200 years old.

I have a feeling that if Google thought they could get away with making it less sensitive, they would have done it.

They surely can't be happy about censoring legitimate content on their platform by mistake.

Where do you get that piano visualizer? Looks helpful for learning and playing.

Have a look at Rick Beato on Youtube


Maybe he's referring to the video breaking down the bullshit Flame vs. Katy Perry lawsuit about Dark Horse,


He also says one of his videos was de-monetized because Dark Horse was bleeding through someone's headphones in one of his videos.

He does fairly deep analysis of recording techniques and music theory on his channel. He claims that even playing a few bars of music from certain artists himself on an acoustic or electric guitar will cause his videos to become de-monetized.

DMCA, ain't it fun?

DMCA as intended would be better here. The person who thinks they found something infringing should have to proactively send a takedown, which should trigger a process that can be appealed. What Google has created is a proactive automated takedown system with minimal recourse.

Applications are open for YC Winter 2020

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