What's new

You're not allowed to use Content ID if you've used sample libraries in your track, or are you?

Chris S

New Member
Hello everyone!

This is something I've recently come across and I'm not sure if I understand it correctly. I've only found very few information about this topic, so hopefully someone knows what's up with that and can clarify...

Apparently, when you want to distribute your track you're not allowed to opt in for Content ID (on Youtube for example) when you've used sample libraries. At first this seemed weird to me, because I'm 100% sure that there are many tracks on YT and in its Content ID system that make use of sample libraries. After all we do have the right to use them commercially, right? Yes, but as I understand it, this doesn't have anything to do with the Content ID system.

Taken from RouteNote:

Will Content ID be blocked on my release if I use samples? Wed, 9 Feb, 2022 at 12:02 PM

If the samples used in your release are available for anyone to obtain, it can’t be sent to Content ID stores.

If you use any non-exclusive samples in your release, unfortunately, we won’t be able to put it on services with content ID systems running. These stores are YouTube, Facebook, Instagram and Tik Tok.

Non-exclusive samples are samples that anyone could obtain for their own track, whether purchased or available for free.

Content ID systems compare audio in their database against anything that’s already been uploaded. If you use a non-exclusive sample, even though you do have the rights to use it, the system will incorrectly claim ownership of other content using the same sample.

In other words, if the sample you’re using is already part of existing tracks online or could be obtained by anyone to use in their release, inaccurate matches and claims would be generated if a track using the same sample gets sent to Content ID stores.
That's how all distributors see it, or at least very similar.

Content ID systems compare audio in their database against anything that’s already been uploaded. If you use a non-exclusive sample, even though you do have the rights to use it, the system will incorrectly claim ownership of other content using the same sample.
This totally makes sense to me, because no one who uses a sample library from Spitfire, OT etc. is the exclusive owner of those sounds. If I use, let's say, a Trombone fx, EQ it and process it a little bit further, you might argue that the finished sound is unique and not recognisable. And then played in context with the whole orchestra (as you aren't allowed to play samples in isolation as stated in most EULA's anyway) it would most certainly not get detected by the Content ID system. But the fact, that you can process the samples, even beyond recognition, doesn't make you the exclusive owner of it, or does it? You're still the owner of your track as a whole though, sure, and you can still make money off of it (on YT too), but you are not allowed to use the Content ID system then.

Taken from DistroKid:

What Songs are Eligible for YouTube Content ID?

If your release contains any audio (beats, loops, samples, video game sounds, other peoples' music...) that you didn't create yourself, it is unfortunately ineligible for YouTube Content ID.


YouTube is really serious about this stuff. All of the following must be true for a release to be eligible:


  • You created all of the sounds
  • Your release does NOT contain beats, loops, sound effects, or other audio downloaded from sample libraries or other public sources. This includes sounds that are available for free from GarageBand, Ableton, Logic, Fruity Loops, etc.
  • You will not allow other artists to remix your song(s), or use any of the sounds from your song(s) in their own songs. That's because if their remix gets uploaded to YouTube, Content ID may claim it as yours. Or if the remixer opts into Content ID, it may think your version is the same song, and claim it as theirs. It's all very complicated.
  • Your release does NOT contain public domain recordings or songs.
  • Your release does NOT contain sounds from video games, television, movies, etc.
  • Your release does NOT contain any audio from other peoples' YouTube videos.
  • You are the only person who will add this single to YouTube Content ID. Nobody else (band members, producer, etc.) will also try to add it to YouTube Content ID, because that causes a mess.
  • You have NOT used another company, such as TuneCore or CD Baby, to submit this single to YouTube Content ID. That will cause a conflict, and a mess.
But what is the exact definition of "sample" here? Is it a vocal loop from Splice? Is it a sample taken from another song (obviously prohibited without permission)? Is it a Timpani roll? Is it a Cello legato line? Is it a simple kick drum? Is it all of them or none?

So, if I understand it correctly, we cannot use non-exclusive sample libraries if we want to opt in for Content ID. But then how does it come that so many tracks are in the Content ID system that definitely make use of sample libraries? Do the composers not know or are they trying their luck? Or perhaps I'm missing or misinterpreting some crucial information? I hope someone can shed a light upon this topic.

Thanks in advance!
 
Last edited:
I think you're confusing "samples" with "sample libraries". If it were the latter, there would be virtually no music on YouTube. Samples are, for example, an actual edit used from another recording. Rap and Hop Hop music are notorious for this.
Yes, this is correct. They are talking about samples of music written by someone else, or pre-existing loops, or beats. Not sample libraries.
 
I don't know the legal theory behind Content ID, or what Content ID might mean by "sample," but I do know that there have been cases where Content ID has picked up on the use of sample libraries and extended claim over other tracks that use the same sample library, so it would make sense that music using sample libraries might be ineligible to use the system. I would assume "sample" in the case of Content ID does indeed extend to "sample library" because sample libraries also interfere with the system. Still, folks are obviously submitting tracks using sample libraries to Content ID, so make of it what you will. My own view is that Content ID is very poorly designed and implemented for what it is trying to do, but it's the world we live in.
 
I don't know the legal theory behind Content ID, or what Content ID might mean by "sample," but I do know that there have been cases where Content ID has picked up on the use of sample libraries and extended claim over other tracks that use the same sample library, so it would make sense that music using sample libraries might be ineligible to use the system. I would assume "sample" in the case of Content ID does indeed extend to "sample library" because sample libraries also interfere with the system. Still, folks are obviously submitting tracks using sample libraries to Content ID, so make of it what you will. My own view is that Content ID is very poorly designed and implemented for what it is trying to do, but it's the world we live in.
Can you give an example? I've only ever known this happen when people are using the same drum loop, or evolving synth preset.
 
As far as I'm concerned, Content ID is a scourge and I firmly believe you should not, ever, be allowed to file a track for Content ID that isn't 100% made up of original sounds. It's an automated, AI-based piece of gear that isn't made for careful differentiation.

The moment a track using Metropolis Ark's high strings doing a certain progression is filed, that's it for any other tracks using those same strings with a similar progression.

I'm pretty sure EULAs and such will start mentioning you're not supposed to content ID tracks using a library's content before long. It shouldn't have been allowed in the first place.
 
I think you're confusing "samples" with "sample libraries". If it were the latter, there would be virtually no music on YouTube. Samples are, for example, an actual edit used from another recording. Rap and Hop Hop music are notorious for this.
That was my first thought too, but I'm not so sure it's that easy... I find the information I've found to be pretty vague.

For example DistroKid speaks explicitly of sample libraries here:
Your release does NOT contain beats, loops, sound effects, or other audio downloaded from sample libraries or other public sources. This includes sounds that are available for free from GarageBand, Ableton, Logic, Fruity Loops, etc.
TuneCore says this:
You can submit specific tracks to YouTube Content ID for revenue collection that:

  1. Use only your own material for which you have exclusive rights.
  2. Use third-party material for which you have exclusive rights.
Do I have the exclusive rights for my track as a whole? Yes I guess, but do I have the exclusive rights for all the different sounds/material that I've used? Well, anybody can buy the exact same sample libraries that I use, so...

jbuhler:

I don't know the legal theory behind Content ID, or what Content ID might mean by "sample," but I do know that there have been cases where Content ID has picked up on the use of sample libraries and extended claim over other tracks that use the same sample library, so it would make sense that music using sample libraries might be ineligible to use the system. I would assume "sample" in the case of Content ID does indeed extend to "sample library" because sample libraries also interfere with the system. Still, folks are obviously submitting tracks using sample libraries to Content ID, so make of it what you will. My own view is that Content ID is very poorly designed and implemented for what it is trying to do, but it's the world we live in.
Yeah, that's exactly what I'm thinking too. The problem lies within the Content ID system itself, that's for sure. There seem to be many different opinions on what's allowed and what isn't. I've seen some people come up and say they never had any problems with using sample libraries and getting their track approved, but imho that doesn't prove that it's allowed but rather shows the flaws of the system.

But again, I might be missing some information here...
 
For example DistroKid speaks explicitly of sample libraries here:
Your release does NOT contain beats, loops, sound effects, or other audio downloaded from sample libraries or other public sources. This includes sounds that are available for free from GarageBand, Ableton, Logic, Fruity Loops, etc.
Ugh, this is caused by the fact that different facets of the industry use the term 'Sample Library' to mean different things. I'm quite certain the term here is meant to refer to actual 'library sites' such as Splice or Freesound.org, where you can download (and in case of the former, license) individual samples for use in your tracks.

I find 'other audio' to be pretty blanket though, and Sample Libraries does, in fact, happen to include Sample-based Virtual Instruments so yeah.

Death to Content ID I say, deaaaaaaaaaaath!!!!
 
How would YouTube possibly know if you used, say, Spitfire Symphonic Strings layered with OPUS?

I still think they’re referring to actual “samples”, even if it’s a phrase from a sample library.

Either way, Content ID Is a steaming pile of cow manure. I don’t even touch YouTube with a 10 foot pole anymore, especially since a few of the publishers I work with are strict about not posting any of my tracks on YouTube whatsoever.
 
Actually how silly would it be to get claimed if you use a sample library or any other sample in non exposed context...
What would than even be the reason for buying samples and sample library?

My thoughts in short:
1) You literally pay for royality free samples to use them in your track (of course in a creative way by processing and mangling the sound)
2) Your track certainly consists of more than 1 percussion loop you just dragged and dropped without any processing?
3)Submit it to Content ID

EDIT: Of course you can get claimed if something is to exposed... not an expert here I'm certainly not a lawyer xD:emoji_joy:
 
Last edited:
I'm pretty sure it refers to samples from records and re: commercial sample libraries, only the phrase driven ones.
 
Actually how silly would it be to get claimed if you use a sample library or any other sample in non exposed context...
What would than even be the reason for buying samples and sample library?

My thoughts in short:
1) You literally pay for royality free samples to use them in your track (of course in a creative way by processing and mangling the sound)
2) Your track certainly consists of more than 1 percussion loop you just dragged and dropped without any processing?
3)Submit it to Content ID

EDIT: Of course you can get claimed if something is to exposed... not an expert here I'm certainly not a lawyer xD:emoji_joy:
It doesn't even have to be that exposed. It just needs to be something that the algorithm identifies as salient. Google "Content ID false positive." Likely the longer the sample (whether from a library or a loop) and more exposed it is the more likely the Content ID algorithm will seize on it. The folks working on Content ID know that processing samples in various ways can evade the algorithm, so they are working on overcoming that. The more they do that, the more they will pick up sample libraries as false positives too. As far as I can tell, Google doesn't care that they create an increasing number of false positives; Google at this point can afford to piss that class of user off. All Google cares about is that the big media firms are satisfied that their music and sounds are being detected at a high rate, so they remain more or less content with the system.

I mean an obvious solution is that Content ID sets up a way for commercial libraries (whether loops or instruments) to submit their samples for automatic clearance. It's actually a bit weird that this isn't already being done, and I've heard that some media productions require new recordings (no publicly available commercial samples—libraries or otherwise—in final production) precisely because of potential issues from Content ID. (That's anecdotal from one music editor I know working in the industry, and I don't know that he wasn't being a bit hyperbolic in the conversation, which occurred in the context of chit chat at a bar. This was also two years ago, and the technology has changed somewhat in the meantime.)
 
What @jbuhler says ))

In stead of thinking about the context of the sample ie from a library or a record, just think of all sound on YT as subject to the CID robot scanners that scan and scan and scan, and when they find something they think they’ve seen before, they match it to the origin - if the match is fair, they apply the policy given to that underlying original asset.
If that sound recording asset is set to Monetize, its supposed replica can be used but the content is subject to ads and the revenue is shared. If its set to track, nothing is monetized but you get data. If it’s set to block, any content using the sample gets blocked.

The original asset is controlled by whom ever first uploaded that to CID, mind you. So.. if you’re the first one to upload a track to CID, anything that sounds like that or part of it, is subject to being claimed by the CID scanners and your policy is applied.

The people can protest if they like, and that’s for another thread.

So… its about those robots.. anything they might see as a copy, they react to by stamping it as a match. This is why people are discouraged from uploading drones. They sound too alike.

So no hard and fast rules. Just scanners and then what YT feel they have to advice people to do, in order to maximize profits on the platform without pissing of the wrong people as stated earlier in the thread.

So just think like the CID scanners.

Best, MB
 
I've been using it for 10 years and never had a problem or false claim unless someone used parts of my master recordings and uploaded them claiming it as their own cue. I guarantee you that Adrev, and all the other content ID companies know without question that most of their users use sample libraries and its no problem at all unless they start getting lots of false claims.

Content ID is the one tiny bit of leverage that we have against these massive tech platforms that make billions using artist's content and its absolutely essential for defending copyrights in the digital age. It is mostly automated but companies like Adrev have teams that verify and do manual claims all day long.
 
Do I have the exclusive rights for my track as a whole? Yes I guess, but do I have the exclusive rights for all the different sounds/material that I've used? Well, anybody can buy the exact same sample libraries that I use, so...
The wording "non-exclusive" is being used in a very broad way to quickly cover every possible scenario that could trip content ID: audio clips, beats, loops, rhythms, construction kits, MIDI packs, etc. Virtual instrument sample libraries are also non-exclusive, but they are not going to trip Content ID any more than real instruments. If you create a song from scratch using virtual instruments and the digital distributor asks if your song uses samples, select "no".
 
Content ID recognises exposed sound used in other productions and flags the new upload. That's an issue with loops (e.g Apple loops ...) and maybe eventually SFX but if your primary sonic source is a melody you've created you should be fine.
 
As Babylon Waves said so succinctly this only applies to exposed samples of commercially available pre-written material, e.g. as loop libraries, phrases from phrase libraries, etc.

Content ID will not be able to identify the sample library you used, that's silly. If that were the case you wouldn't be able to find video of more than one mockup of a popular film score, 1 mockup from each discrete product. Even then, that wouldn't account for MIDI expression, mixing choices, reverb preference and settings, etc. Not to mention this would mean it could fingerprint any instrument - A vintage Roland analog saw wave in any key, a Prophet Oscillator, Serum, Pigments, etc...

The bottom line is don't put out music making use of commercially available pre-written material. That's the kind of stuff it detects and will flag. You can bury one or two elements in a full arrangement and it won't be able to pick it out because it's an algorithm, and algorithms aren't smart enough to be able to cognitively identify a sound - i.e. 'mentally isolate' a sound and recognize it as something you've hear before in one of your sample libraries. But as already pointed out, the more exposed it is the more likely it will be recognizable by the algorithm.

Ethically: Content ID can go fuck itself. :P
 
I'm sorry, but could this system actually identify a sample library patch? I mean, I guess, MAYBE, if you're using it bare, with no changes. But who does that? And let's face it, one string library sounds pretty much like every other string library with differences so subtle that, when layered into a mix, they're almost impossible to differentiate.

Whatever the case, I guess I'm not enough of a YouTuber to know what ContentID is. But every time I upload music—all of which uses sample libraries—it always passes the copyright check.
 
Top Bottom