> Every single one was for a work in the public domain.
While the sheet music is, the performance by other artists - as your own performance - is not. That's the issue the algorithm is having here (not defending).
The automated system would need to "understand" that this is indeed a new performance of a public domain piece of sheet music and not a reproduction of a copyrighted performance by somebody else. Even if you were note for note playing exactly the same (tempo and whatnot) with the same instrument tuned the same way. I think this would be an argument against automated systems. Whether a human could know that my bike-ride was scored by myself and not somebody else is doubtful though.
On the other hand, I do understand that people do not want their individual performances to be used without licence and there may be many such performances.
I said that two actual and copyrightable performances of the same melody are arguably indistinguishable.
Edit: How would an automated system know, that you did not just non-transformatively alter another person's performance, instead of performing yourself?
I have gained the impression that YouTube has two distinct techniques within Content ID, one for claiming melodies, and another for claiming specific recordings.
I’m absolutely confident that this is doing melody matching: these recordings are trivially distinguishable from any professional performance, with completely different instruments and playing styles, and with much lower quality singing.
For example, the most repeated claim was by “AdRev Publishing”, claiming “Crimond (The Lord's My Shepherd) - FirstCom” three times. (That song was also claimed once by “Capitol CMG Publishing and Adorando Brazil” as “The Lord's My Shepherd I'll Not Want”.) The first two times, I was accompanying with a piano-sound keyboard in the traditional four-part harmony—admittedly they sound fairly similar to one another; but in the third, an Indian was playing, using a piano-and-strings sound on a different keyboard, using Indian harmonisation (which is quite different).
I think we even had an unaccompanied song claimed once, matching throwaway0b1’s report.
These are completely different performances from whatever recordings the liars may have provided to the Content ID system. The melody is the only thing they will have in common.
My experience aligns with yours. I had a video of a half-hour coffee-shop-style gig blocked because at one point I covered "Hotel California". This was just me -- one voice, one acoustic guitar played poorly, and was down a whole step from the original key.
I believe that Hotel California is still under protection as a melody, rather than a performance.
Hence covers require either a license or fair use according to law. This differs a lot from the case of performing a 200 year old song whose melody is in the public domain.
Not to say it was wrong what you did, or that Google was in the right for flagging. But Google was legally correct.
(Google suggests controversy around the question of whether hotel California is public domain or not)
> How would an automated system know, that you did not just non-transformatively alter another person's performance, instead of performing yourself?
Yes, exactly. You have really nailed the problem with the current system.
Google chose to build an automated system. That was a choice, not an immutable fact. Google chose not to have a human arbitration process. Another choice. Google chooses not to punish the claimants that abuse the system. Another choice, not an immutable fact.
People aren't saying there must be a perfect automated system. People are saying Google has chosen to employ an automated system that does not meet the actual needs.
The criticism is for Google choosing to exclusively use an automated system, which can never successfully perform this task.
Ideally, the easy cases are automated, and more difficult cases like these are not. Hence, the system would know because of an actual DMCA request, rather than an automated YouTube request.
While the sheet music is, the performance by other artists - as your own performance - is not. That's the issue the algorithm is having here (not defending).
The automated system would need to "understand" that this is indeed a new performance of a public domain piece of sheet music and not a reproduction of a copyrighted performance by somebody else. Even if you were note for note playing exactly the same (tempo and whatnot) with the same instrument tuned the same way. I think this would be an argument against automated systems. Whether a human could know that my bike-ride was scored by myself and not somebody else is doubtful though.
On the other hand, I do understand that people do not want their individual performances to be used without licence and there may be many such performances.