3 ms·

"Music copyrights are a minefield, even for old music, because it is often performed and rearranged, both of which would restart the copyright clock." Most of

by mikece 6y ago

"Music copyrights are a minefield, even for old music, because it is often performed and rearranged, both of which would restart the copyright clock." Most of what the choir was flagged on was Gregorian Chant, the settings of which trace their origins to Old Testament worship in the Temple of Jerusalem, possibly as far back as King David's tabernacle singers. The texts are all ancient (200 to 1600 years old) and the "new" settings were composed by William Byrd (1540 - 1623) and Palestrina (1525 - 1594). We're not talking about something that was recently reset... though that would explain why Morten Lauridsen's "O Magnum Mysterium" is almost never heard in Divine Worship (that setting isn't even 30 years old yet).

kube-system 6y ago
There are undoubtably a ton of variations of those chants over the past 120 years and probably millions of recorded performances of those chants. Any of which would be subject to a new copyright on the date of that performance or rearrangement, regardless of the original date of the composition. A work being widely performed is a common situation that complicates attempts to solve the problem automatically. An infinite number of monkeys randomly singing a song will eventually sing one that sounds exactly like yours. How does an algorithm determine the difference in intent between my randomly singing choir of monkeys and your church's performance? It can't, as it compares audio -- it doesn't reason like humans do. Content ID doesn't know who your performers are, who recorded it, which sheet music your performers were reading from, where they made mistakes, and/or what licenses you have. (and FWIW, any given human watching the video wouldn't know much of this either) Two audibly identical works can be non-infringing, and two audibly differentiable works can be infringing. ContentID is an approximation -- it does not and never will have enough information to accurately judge everything.
- kelnos 6y ago
  That argument doesn't hold water when my Pixel can hear a piece of classical music and identify the particular performer/album, not just the title of the music itself. If my phone can do that on its own (Google claims they aren't sending any audio to the cloud), then Content ID can surely do just as well, if not better, and not flag someone's church choir's performance as some recorded, copyrighted work.
  - kube-system 6y ago
    You are missing a big part of the problem space as it pertains to copyright law. That application is fingerprinting a portion of the audio and looking it up for a match in an existing database of copyrighted works for a match. This is the same thing that Content ID does, and it's exactly why it has false positives. Knowing whether portions of audio in a recording match other existing recordings is not enough to determine whether or not something breaks copyright law. You can demonstrate this with Shazam (/whatever the Google one is called). Live performances of any decent quality will often match recording fingerprints in their DB and give you a result of a particular artist/album. The law is simply more complicated than `if(audioMatches) then {illegal}` Sometimes a match on audio is not an illegal copyright infringement. Sometimes audio that doesn't match is illegal copyright infringement. Let's say you play a song from the year 1200 on the violin. You have the copyright to that performance. Now, let's say I play the same song, and I play it relatively well. I have the copyright to my performance. Likely, the audio fingerprints match, and Shazam/Google/Content ID will all say that my recording matches yours -- even though my performance was entirely legal, and we both own copyrights to our respective performances. The information Google is missing is: 1. it doesn't know that the sheet music we were both reading was authored in 1200. It doesn't have this information, it just has the audio. 2. it doesn't know that a different human physically performed the work. It just knows that the audio is the same.
    - cwkoss 6y ago
      I wonder how long those portions of audio are. With portion length and bitrate there is some theoretical number of music samples that could be 'brute forced' to cover the entire content ID space. I wonder if their snippets are long enough to make this an impossible number, or if this theoretical attack on Content ID could be put into practice.
    - mrkramer 6y ago
      >The law is simply more complicated than `if(audioMatches) then {illegal}` >Sometimes a match on audio is not an illegal copyright infringement. >Sometimes audio that doesn't match is illegal copyright infringement. >Let's say you play a song from the year 1200 on the violin. You have the copyright to that performance. Now, let's say I play the same song, and I play it relatively well. I have the copyright to my performance. Likely, the audio fingerprints match, and Shazam/Google/Content ID will all say that my recording matches yours -- even though my performance was entirely legal, and we both own copyrights to our respective performances. >The information Google is missing is that it doesn't know that the sheet music we were both reading was authored in 1200. I'll be honest I don't know anything about copyright but wouldn't solution be for example you as a YouTube content creator can claim the right to that particular song and the other guy as well and know Google has a record in their database that says: John owns right to 1200 violin song (song id) Jack owns right to 1200 violin song (song id) And then IF John's (song id) matches Jack's (song id) don't send copyright strike. Google needs to build YouTube knowledge graph so it has better awareness on information.

by mikece 6y ago

kube-system 6y ago

There are undoubtably a ton of variations of those chants over the past 120 years and probably millions of recorded performances of those chants. Any of which would be subject to a new copyright on the date of that performance or rearrangement, regardless of the original date of the composition. A work being widely performed is a common situation that complicates attempts to solve the problem automatically. An infinite number of monkeys randomly singing a song will eventually sing one that sounds exactly like yours. How does an algorithm determine the difference in intent between my randomly singing choir of monkeys and your church's performance? It can't, as it compares audio -- it doesn't reason like humans do. Content ID doesn't know who your performers are, who recorded it, which sheet music your performers were reading from, where they made mistakes, and/or what licenses you have. (and FWIW, any given human watching the video wouldn't know much of this either) Two audibly identical works can be non-infringing, and two audibly differentiable works can be infringing. ContentID is an approximation -- it does not and never will have enough information to accurately judge everything.

kelnos 6y ago
That argument doesn't hold water when my Pixel can hear a piece of classical music and identify the particular performer/album, not just the title of the music itself. If my phone can do that on its own (Google claims they aren't sending any audio to the cloud), then Content ID can surely do just as well, if not better, and not flag someone's church choir's performance as some recorded, copyrighted work.
- kube-system 6y ago
  You are missing a big part of the problem space as it pertains to copyright law. That application is fingerprinting a portion of the audio and looking it up for a match in an existing database of copyrighted works for a match. This is the same thing that Content ID does, and it's exactly why it has false positives. Knowing whether portions of audio in a recording match other existing recordings is not enough to determine whether or not something breaks copyright law. You can demonstrate this with Shazam (/whatever the Google one is called). Live performances of any decent quality will often match recording fingerprints in their DB and give you a result of a particular artist/album. The law is simply more complicated than `if(audioMatches) then {illegal}` Sometimes a match on audio is not an illegal copyright infringement. Sometimes audio that doesn't match is illegal copyright infringement. Let's say you play a song from the year 1200 on the violin. You have the copyright to that performance. Now, let's say I play the same song, and I play it relatively well. I have the copyright to my performance. Likely, the audio fingerprints match, and Shazam/Google/Content ID will all say that my recording matches yours -- even though my performance was entirely legal, and we both own copyrights to our respective performances. The information Google is missing is: 1. it doesn't know that the sheet music we were both reading was authored in 1200. It doesn't have this information, it just has the audio. 2. it doesn't know that a different human physically performed the work. It just knows that the audio is the same.
  - cwkoss 6y ago
    I wonder how long those portions of audio are. With portion length and bitrate there is some theoretical number of music samples that could be 'brute forced' to cover the entire content ID space. I wonder if their snippets are long enough to make this an impossible number, or if this theoretical attack on Content ID could be put into practice.
  - mrkramer 6y ago
    >The law is simply more complicated than `if(audioMatches) then {illegal}` >Sometimes a match on audio is not an illegal copyright infringement. >Sometimes audio that doesn't match is illegal copyright infringement. >Let's say you play a song from the year 1200 on the violin. You have the copyright to that performance. Now, let's say I play the same song, and I play it relatively well. I have the copyright to my performance. Likely, the audio fingerprints match, and Shazam/Google/Content ID will all say that my recording matches yours -- even though my performance was entirely legal, and we both own copyrights to our respective performances. >The information Google is missing is that it doesn't know that the sheet music we were both reading was authored in 1200. I'll be honest I don't know anything about copyright but wouldn't solution be for example you as a YouTube content creator can claim the right to that particular song and the other guy as well and know Google has a record in their database that says: John owns right to 1200 violin song (song id) Jack owns right to 1200 violin song (song id) And then IF John's (song id) matches Jack's (song id) don't send copyright strike. Google needs to build YouTube knowledge graph so it has better awareness on information.