Menu
What's new
By:
Filters

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

I will try to explain.

You can use the samples libraries without any problems; they are created for that.

The problem you are referring to is exclusively if you use ( mostly loops ) very indentifiable and very exposed and this could create Content ID problems being stupid algorithms. But the loop must be really exposed and protagonist, and have been used in the same way by another user in another track.

So if I use a loop or sample even tonal, but that is inside the arrangement there is no problem.


Otherwise a good 60% of the sample libraries will not be possible to use.

These things are always very unclear, to discourage the use of samples and Sampling....without rights included in commercial material.

To make a point, you are not allowed to put into your track a sample recorded by you taken from another track, a live music recording, from you tube or any other sources, or any other material that you are sampling without permission.

The sample libraries ( the commercial sample libraries , often not the FREE products ) are a different thing, because when you buy a sample library you are buying a license to use it, also in a commercial track.
 
Last edited:
StefanoM said:
Otherwise a good 60% of the sample libraries will not be possible to use.
Click to expand...
Yes, but Content ID doesn’t care about whether it affects how and whether we can use sample libraries. At the moment Content ID is not generally seizing on these sorts of things so they are generally safe to use but there are vocal phrases in your Ethera libraries, for instance, that because they have a certain extent and are quite striking, content ID might very well seize on if a song used one in a particular way.
 
jbuhler said:
Yes, but Content ID doesn’t care about whether it affects how and whether we can use sample libraries. At the moment Content ID is not generally seizing on these sorts of things so they are generally safe to use but there are vocal phrases in your Ethera libraries, for instance, that because they have a certain extent and are quite striking, content ID might very well seize on if a song used one in a particular way.
Click to expand...
Absolutely not my experience. The only way is to use as mentioned the sample Without anything else.
If we take the example of the vocal phrases they are absolutely not problematic like other sample libraries. if you use them into a context. If used in the correct way. The vocal phrases of Ethera or Vocalise or any other library are mostly vocal phrases--a series of musical intervals. Besides the fact that any Ethera for example has more than 1000 different vocal phrases ( and are not MUSICAL LOOPS with a strong and incisive melody/harmony structure), that you can combine, a rather infinite possibility of combining them, so you understand well that we are talking about a nonproblem.

I've created I think over 500 tracks for different publishers including Disney and Marvel and Fox, and I woked with some YouTube's guys , I know how it works .Some times they are simply errors in the algorithm.

I also have many customers who use these libreries for Netflix, and other , but never had a problem.


I never had a Content ID report using sample libraries. Then it's all about knowing how to use them, but that's a different thing. Otherwise a rhythmic part of an ostinato or a series musical of'intervals would be a ContentID problem for each song.

To have Content ID problems you need 2 tracks to have exactly a VERY IMPORTANT and sufficiently long portion of very exposed and exactly identical material ( or Musical Loops ) but often Just change tempo and/or 1 semitone and the problem already disappears.
 
Last edited:
As already stated, using what we call "sample libraries" to create original music is not an issue. Using pre-recorded musical loops is definitely an issue.

One issue I have encountered, however, is music assigned to me that is *not* mine and generates strikes against other people. I never found out the details but I think it was a CDBaby issue. Caused issues on Spotify as well.

When I first converted my YouTube channel to an "official" artist channel there was all kinds of stuff that appeared on my page that was *not* mine (there's a thread around here somewhere about it). It took a long time for it to go away and I'm still not sure all the links are removed because the algorithm *really* attaches itself to historical data. There's still stuff on Spotify that appears under my name that is not mine. It's a common problem on Spotify unless you're a major artist with a dedicated brand manager with links at Spotify.

One example of how the problem persists: many years ago, a bunch of the music assigned incorrectly to me was labeled by the YouTube algorithm as children's music (it was not). I figured that out because I've tried to use Google Ads a few times to promote my channel and nearly all of my ads get shown on children's YouTube channels because the algorithm thinks that's what my music is. I've contacted Google about that several times over the years and there's no recourse: Google thinks my YouTube channel is mostly for 5-year-olds because the algorithm was trained that way and it won't re-learn.

rgames
 
StefanoM said:
Absolutely not. The only way is to use as mentioned the sample Without anything else. If we take the example of a vocal phrases they are absolutely not problematic as are most sample libraries. If used in the correct way. The vocal phrases ti Ethera or Vocalise or any other library are mostly vocal phrases--a series of musical intervals. Besides the fact that makes me that any Ethera has more than 1000 different vocal phrases, a rather infinite possibility of combining them, so you understand me well that we are talking about a nonproblem. I've created I think over 600 tracks for different publishers including Disney and Marvel and Fox, never had a Content ID report using sample libraries. Then it's all about knowing how to use them, but that's a different dis
Click to expand...
If you feature it as a melodic excerpt and Content ID picks it as something salient, it will trigger in another piece that Content ID decides is salient and sufficiently similar. This is how most loops and phrases get caught as false positives by Content ID and it's why Content ID wants samples of all kinds excluded. Now it may be that in almost no cases will your libraries trigger, but there is no functional difference between the vocal phrases in Ethera and picking up a pack of vocal phrases and other loops that are properly licensed for reuse.

There is, I should add, no proper or improper use of your libraries ("correct way") except as is set forth in the EULA. You are the developer but you can't determine how anyone uses your libraries so long as they follow the EULA.

Since you are writing for Disney, Marvel, Fox, I wouldn't expect you would have any issue with Content ID at all, since the system is designed to protect the IP of those companies and companies like them. Those who use the same samples in your collection that you did for Disney, Marvel, and Fox are far more likely to experience problems, especially if they are using them in a similar way in a similar genre.

In any case, too many are relying on the common sense of how they think Content ID should be working and not looking at what Content ID is doing and that it is becoming increasingly sensitive, identifying both more infringements but also more false positives.

@Polkasound might be right that the Content ID terms are written overly broadly as a CYA move, and that Google doesn't really expect that folks using sample libraries as we know them will exclude their music from Content ID on that account, but we're all going on a leap of faith and being asked to certify that we did not use commercial "samples," even when we did. I expect to see this emerge as a real problem over the next few years.
 
jbuhler said:
If you feature it as a melodic excerpt and Content ID picks it as something salient, it will trigger in another piece that Content ID decides is salient and sufficiently similar. This is how most loops and phrases get caught as false positives by Content ID and it's why Content ID wants samples of all kinds excluded. Now it may be that in almost no cases will your libraries trigger, but there is no functional difference between the vocal phrases in Ethera and picking up a pack of vocal phrases and other loops that are properly licensed for reuse.

There is, I should add, no proper or improper use of your libraries ("correct way") except as is set forth in the EULA. You are the developer but you can't determine how anyone uses your libraries so long as they follow the EULA.

Since you are writing for Disney, Marvel, Fox, I wouldn't expect you would have any issue with Content ID at all, since the system is designed to protect the IP of those companies and companies like them. Those who use the same samples in your collection that you did for Disney, Marvel, and Fox are far more likely to experience problems, especially if they are using them in a similar way in a similar genre.

In any case, too many are relying on the common sense of how they think Content ID should be working and not looking at what Content ID is doing and that it is becoming increasingly sensitive, identifying both more infringements but also more false positives.

@Polkasound might be right that the Content ID terms are written overly broadly as a CYA move, and that Google doesn't really expect that folks using sample libraries as we know them will exclude their music from Content ID on that account, but we're all going on a leap of faith and being asked to certify that we did not use commercial "samples," even when we did. I expect to see this emerge as a real problem over the next few years.
Click to expand...
No.
Not really. No IP blocking or anything weird.

Then,
First of all I'm a professional Composer and Sound Designer, and now I'm talking like Composer and Sound Designer not like developer.

And Yes, There is a correct way! And The correct way is an Intelligent way of using a sample library.

Like that People that say that Damage is over used, Well it depends on how you use it.

A Vocal Phrase is a little SHORT cue of improvvisation and vocal musical interval.. not a real MELODY, again it is impossibile what are you saying.

Then if you Take this little short pieace and only this one, exposed and THAT IS your TRACK for 3 minutes, and other guys do the same Thing... is we have a problem. But that is not a professional way of using ANY sample library. And we are talking of Content Id problem, so we are talking of Professional and not Amateurs..

Simply if you read my bad English well, the problem occurs for the reasons I explained.
 
Last edited:
Polkasound said:
The recording of your song is 100% your own.


You don't own the samples themselves which remain contained in the library. But you are the master copyright holder of the recording on which you used the library.


Again, the terms "samples" and "sample library" are defined differently in different aspects of the music industry. I have been commercially releasing original music online for years, and every one of my songs uses virtual instrument sample libraries. I do this under the understanding that virtual instrument sample libraries are not what the digital distribution arm of the music industry refers to as "samples" or "sample libraries".

It's possible I could be wrong, and that to the letter of the law, "sample libraries" DOES include virtual instruments sample libraries. But I'm 100% confident that the spirit of Content ID rules fully allow for, and warrant, the use of virtual instrument sample libraries... just like everyone is expected (and even unofficially encouraged) to drive a few MPH over the posted limit.

If you get flagged by Content ID, it's not going to be because you used JB Violin instead of hiring Joshua Bell. It's going to be because an algorithm deemed elements (melody, rhythm, groove, pattern, etc.) of your song to be too close to someone else's song, or it detected a sound clip that was used in someone else's song which may require licensing. This is why pre-recorded song elements (loops, beats, patterns, etc.) run a much higher risk of tripping Content ID. If you're not using prerecorded loops/beats, you have nothing to worry about. Relax and release your music to the world. :)


Don't worry. In this forum especially, there is a lot of respect for musicians like you who cross their Ts and dot their Is because they don't want to do anything wrong. But one thing you'll learn is that most rules in this industry are not designed to prevent people from traveling 56MPH in a 55MPH zone. They're designed to prevent reckless driving. It's pretty apparent you're a safe driver. :)
Click to expand...
Thank you very much for further clarification!

Polkasound said:
But one thing you'll learn is that most rules in this industry are not designed to prevent people from traveling 56MPH in a 55MPH zone. They're designed to prevent reckless driving.
Click to expand...
I hadn't thought of it that way before. This is something I'm most certainly going to remember!

jbuhler said:
but we're all going on a leap of faith and being asked to certify that we did not use commercial "samples," even when we did. I expect to see this emerge as a real problem over the next few years.
Click to expand...
Yeah, this is likely going to be a problem if nothing changes. Although, you've already provided a potential solution for it:

jbuhler said:
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.
Click to expand...
 
StefanoM said:
No.
Not really. No IP blocking or anything weird.

Then,
First of all I'm a professional Composer and Sound Designer, and now I'm talking like Composer and Sound Designer not like developer.

And Yes, There is a correct way! And The correct way is an Intelligent way of using a sample library.

Like that People that say that Damage is over used, Well it depends on how you use it.

A Vocal Phrase is a little SHORT cue of improvvisation and vocal musical interval.. not a real MELODY, again it is impossibile what are you saying.

Then if you Take this little short pieace and only this one, exposed and THAT IS your TRACK for 3 minutes, and other guys do the same Thing... is we have a problem. But that is not a professional way of using ANY sample library. And we are talking of Content Id problem, so we are talking of Professional and not Amateurs..

Simply if you read my bad English well, the problem occurs for the reasons I explained.
Click to expand...
How long does a sequence of notes (or time-span of pitches) have to be for ContentID to pick it up as a copyrightable melody?

Would backing percussion be enough to prevent it from matching the two melodies? Percussion and chords?
 
AnhrithmonGelasma said:
How long does a sequence of notes (or time-span of pitches) have to be for ContentID to pick it up as a copyrightable melody?

Would backing percussion be enough to prevent it from matching the two melodies? Percussion and chords?
Click to expand...
If we are talking of musical content ( so no improvvisation or vocalization ) with a strong melody line could be also 8 misures.. but must be a loop/phrase with a strong melodic content.

So we are NOT talking of Music Interval, Vocal improvvisations, or Harmony..

For the matching is important also the Tempo and The Pitch.

Same melody but different Pitch or Tempo could mask the similarity and the matching.

Of Course As I said is important also the context. If the loop or phrase is EXPOSED or not.

The percussion or chords don't mask the problem too much, but how the melody loop/phrase is mixed into the mix, if is dominant or not, that is important.

If you use a Musical ( Melody) Loop or a Melody Phrase, with a different chords progression the recognition could if the melody is so important.

But as mentioned it must be exposed
 
jcrosby said:
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
Click to expand...
Huh ... I wouldn't be so sure about that anymore. Given the results that services like Lalal.ai already achieve today, it's only a matter of time before algorithms are able to differentiate between individual musical elements as well as (or even better than) the human ear.

-> https://www.lalal.ai/
 
Just had the same problem. Music distributor, Horus, send me a mail, "this release is not suitable to be sent to Content ID platforms. This is because we have reason to believe this release may contain non-exclusive samples/beats/loops, which cannot be sent to Facebook/Instagram, YouTube Content ID, and SoundCloud Content ID. Please change your services selection to exclude YouTube, Facebook and SoundCloud. (Please note: your release will still be distributed to YouTube Music and SoundCloud).If you have exclusive rights to the samples/beats/loops used in your release, or have created the music from scratch, please email the distribution team on [email protected] to provide proof of this. "

This is for a song containing logic and native instruments loops and one shots. They couldn't possibly be described as exposed.
 
Spiral-M said:
Just had the same problem. Music distributor, Horus, send me a mail, "this release is not suitable to be sent to Content ID platforms. This is because we have reason to believe this release may contain non-exclusive samples/beats/loops, which cannot be sent to Facebook/Instagram, YouTube Content ID, and SoundCloud Content ID. Please change your services selection to exclude YouTube, Facebook and SoundCloud. (Please note: your release will still be distributed to YouTube Music and SoundCloud).If you have exclusive rights to the samples/beats/loops used in your release, or have created the music from scratch, please email the distribution team on [email protected] to provide proof of this. "

This is for a song containing logic and native instruments loops and one shots. They couldn't possibly be described as exposed.
Click to expand...
Do any of the loops have melodies that would fall under copyright protection? Content ID matches melodies....
 
Spiral-M said:
Just had the same problem. Music distributor, Horus, send me a mail, "this release is not suitable to be sent to Content ID platforms. This is because we have reason to believe this release may contain non-exclusive samples/beats/loops, which cannot be sent to Facebook/Instagram, YouTube Content ID, and SoundCloud Content ID. Please change your services selection to exclude YouTube, Facebook and SoundCloud. (Please note: your release will still be distributed to YouTube Music and SoundCloud).If you have exclusive rights to the samples/beats/loops used in your release, or have created the music from scratch, please email the distribution team on [email protected] to provide proof of this. "

This is for a song containing logic and native instruments loops and one shots. They couldn't possibly be described as exposed.
Click to expand...
It is quite possible the loops/sound FX, One-Shots were registered with CID by the vendor, not sure why but it may be an attempt to stop bad actors creating loops from their libraries and then claiming the content for revenue on various streaming platforms or AI harvesting. It's the only reason I could think of but CID has become fairly sophisticated as of late.

You can use virtual instruments in your libraries but pre-created/recorded melodies are certainly a no-go
 
Last edited:
Spiral-M said:
Just had the same problem. Music distributor, Horus, send me a mail, "this release is not suitable to be sent to Content ID platforms. This is because we have reason to believe this release may contain non-exclusive samples/beats/loops, which cannot be sent to Facebook/Instagram, YouTube Content ID, and SoundCloud Content ID. Please change your services selection to exclude YouTube, Facebook and SoundCloud. (Please note: your release will still be distributed to YouTube Music and SoundCloud).If you have exclusive rights to the samples/beats/loops used in your release, or have created the music from scratch, please email the distribution team on [email protected] to provide proof of this. "

This is for a song containing logic and native instruments loops and one shots. They couldn't possibly be described as exposed.
Click to expand...
There are reports that Content ID can now in some cases identify samples around a second or shorter in length, even when pitch shifted, time stretched, and with other processing.

I keep saying that this is headed to be a major problem for those who use commercial sample libraries for any music that is associated with anything that will be subject to Content ID, because although the use of sample libraries (and copyright free samples, loops and such) is completely legal, it affects the ability to monetize any product that is flagged by Content ID. This will all need legal sorting out (we are not talking about copyright here but the ability to monetize under the terms of Content ID, and those rules are different: you can have a perfect legal right to use a sample but still not be able to use it because the rules of Content ID make it impossible). Maybe this legal sorting out will happen, but there will be a cost to it, and it's not clear who is going to pay to sort it all out.

Right now the onus is all being placed on composers, and of course that is not sustainable for low margin work like library production music, where there isn't currently enough money available to pay to have an orchestra record, certainly not for the bulk of it.
 
Nevermind sample libraries. Consider real instruments. Here's a perverse thought experiment...

I'm going to record a piece of music that contains just my Taylor guitar, strumming 1 bar of every common chord and register it with Content ID.

Now anyone who records strummed chords with that model guitar is going to get their music flagged by Content ID as infringing on my work.

Then I'll do the same playing the ubiquitous Travis picking pattern. Then playing chords and common accompaniment patterns on my Yamaha piano. And notes on my Rickenbacker bass. Then I'll move to my Martin guitar. And so on and so on...

Pretty soon, I'll control the vast majority of acoustic-based music recorded with common instruments that's uploaded to YouTube, etc. thanks to fingerprinting technology that identifies the sounds of my instruments playing ubiquitous musical parts.


This scenario is far-fetched and childishly simplistic, but given how we are assuming Content ID operates, it's not out of the question and has the potential to break the entire online music biz. So I think the reality is more nuanced and less foreboding than we are speculating.
 
Just to avoid you wasting to much time on this.. YT’s CID registers everything that is released on the platform, and thusly everything that comes after, with similar waveform patterns, rhythm or melody in all forms or which ever identifier the YT black magic can utilize current day etc, will be subject to interventions, claims etc. So far it’s working towards very similar things, and yes @NekujaK sometimes it will match unrelated recordings that sound similar, which is why drones or temple bowls are best avoided. Too many variables in guitar takes though.

So two things… needless to say: make your own stuff whereever you can, and if you are using stuff that other people use too, then just shoot your distributor an email documenting your license to use the stuff.

If they don’t respond, change distributor.

And @jbuhler, if it really does become a major obstacle down the road, we could consider making a joined vi-composers distribution (maybe a branch of one of the majors - with enough onboard they’d do that) and make sure there’s an easy path to clearing these things fast. There are solutions to this thing.

Best, M.
 
Byrial said:
YT’s CID registers everything that is released on the platform
Click to expand...
I don't believe that's true. The rights holder is responsible for registering their work with Content ID, if they choose to - it's not automatic because YT has no way of knowing who the rights holder is for every piece of music that's uploaded, therefore can't properly register it automatically.
 
Byrial said:
And @jbuhler, if it really does become a major obstacle down the road, we could consider making a joined vi-composers distribution (maybe a branch of one of the majors - with enough onboard they’d do that) and make sure there’s an easy path to clearing these things fast. There are solutions to this thing.
Click to expand...
As I understand it's the shows that are demanding the clearance of Content ID. I don't see how vi-composers distribution gets around that problem, since the issue is the flagging by Content ID. Also any work around will have costs, and it's not clear who pays the costs.
 
NekujaK said:
I don't believe that's true. The rights holder is responsible for registering their work with Content ID, if they choose to - it's not automatic because YT has no way of knowing who the rights holder is for every piece of music that's uploaded, therefore can't properly register it automatically.
Click to expand...
Sorry for the confusion. However, YT does scan everything across the platform. But that to the side, of course you’re right: the only way to know who owns what is to tie the trawling and the CID together. But that’s besides my point.
 
You must log in or register to reply here.
Top Bottom