3 ms·

All the Music – Every Possible Melody

3 points by harel 3y ago

timoteostewart 3y ago
Short story “Melancholy Elephants” by Spider Robinson about a future in which all the songs are already taken. http://www.spiderrobinson.com/melancholyelephants.html http://www.spiderrobinson.com/melancholyelephants.html
Rochus 3y ago
> A diatonic scale has only seven pitches, and every melody ever created must work within that seven-note constraint. So the universe of possible melodies forms a grid. Well, the diatonic scale is only one of 66 possible scales with seven notes (without rotation symmetric variants); in western music a lot of these scales next to the diatonic one are in regular use (altered, acoustic, hungarian minor, etc.). And you can also add many other scales with more or less than seven notes, which are in regular use in western music. If you also consider non-western music (which is also subject to copyright) a lot more scales become relevant. If you want to check yourself, here is a tool to do so: https://github.com/rochus-keller/MusicTools/tree/master/ScaleAnalyzer https://github.com/rochus-keller/MusicTools/tree/master/Scal... And these are only one-octave scales. There are also scales which change over octaves (e.g. in Jazz), and melodies usually extend over more than one octave. And this is only the pitch domain. If you also consider rhythm and phrasing, the number of useful possibilities is virtually infinite (countably infinite, but still infinite). The problem from my humble point of view is not the (mathematically) objective difference between melodies, but the (subjective) perception of similarity by the judge and his/her experts. It would therefore not be possible to create legal certainty even if it were really possible to publish all the possibilities in advance, as the human arbitrariness of our legal system still exists. But the whole thing is undoubtedly an excellent marketing idea for a lawyer.
- harel 3y ago
  I think that any reduction in potential litigation of unintentional melody "infringement" is blessed. Even if this resolved to only 1 less lawsuit it's earned it's keep. Our culture of litigation is not a to our credit.
  - Rochus 3y ago
    I agree that the present culture of copyright litigation is not a to our credit. But I don't think the presented approach can improve that; although the authors themselves are lawyers, the approach suffers from a considerably more serious flaw than the ones I have mentioned so far; the primary problem is likely that the published melodies have been generated by an algorithm and are therefore not protected by copyright law, and thus neither in order to invoke a pre-existing work in legal proceedings. But the real problem from my humble point of view is the perversion of copyright law in order to appease corporations and publishers so that it can even be applied to individual melodies (i.e. ideas instead of works).
    - harel 3y ago
      The "created by algo" point is unclear to me. Is the removal of copyright only applicable (at this point in time) to content (or music) generated by LLMs and the likes of them, or also to something like All the music where the melodies were coded into existence by a human. In other words (which might not make sense); In the AI "traceroute" between "algo" to "output", there is one "hop" - the human who coded the algorithm.
      - Rochus 3y ago
        In most countries copyright laws only protect work that was created by a human being; e.g. the US Copyright Office will refuse to register a claim if it determines that a human being did not create the work. It doesn't matter if an LLM or a simple program or calculation generated the music; in both cases the creator was no human being. There were cases in the press recently where this rule was confirmed for DNN generated works. Also note that copyright law protects "expression of ideas" (i.e. works), not the ideas itself (nor compilations of ideas), and that the work must possess "at least some minimal degree of creativity." It is also debatable whether a single, short melody can be considered a work or a mere idea. Unfortunately, in the last twenty years, the courts (and in some cases also the law) have begun to soften the concept of a work even further and to allow even trivialities to be considered protected works (a similarly decadent development could be observed years earlier in the patent system). Although this pleases some corporations and publishers, it has not exactly helped to improve legal certainty (nor innovation).

3 points by harel 3y ago

Short story “Melancholy Elephants” by Spider Robinson about a future in which all the songs are already taken. http://www.spiderrobinson.com/melancholyelephants.html http://www.spiderrobinson.com/melancholyelephants.html

Rochus 3y ago

> A diatonic scale has only seven pitches, and every melody ever created must work within that seven-note constraint. So the universe of possible melodies forms a grid. Well, the diatonic scale is only one of 66 possible scales with seven notes (without rotation symmetric variants); in western music a lot of these scales next to the diatonic one are in regular use (altered, acoustic, hungarian minor, etc.). And you can also add many other scales with more or less than seven notes, which are in regular use in western music. If you also consider non-western music (which is also subject to copyright) a lot more scales become relevant. If you want to check yourself, here is a tool to do so: https://github.com/rochus-keller/MusicTools/tree/master/ScaleAnalyzer https://github.com/rochus-keller/MusicTools/tree/master/Scal... And these are only one-octave scales. There are also scales which change over octaves (e.g. in Jazz), and melodies usually extend over more than one octave. And this is only the pitch domain. If you also consider rhythm and phrasing, the number of useful possibilities is virtually infinite (countably infinite, but still infinite). The problem from my humble point of view is not the (mathematically) objective difference between melodies, but the (subjective) perception of similarity by the judge and his/her experts. It would therefore not be possible to create legal certainty even if it were really possible to publish all the possibilities in advance, as the human arbitrariness of our legal system still exists. But the whole thing is undoubtedly an excellent marketing idea for a lawyer.

harel 3y ago
I think that any reduction in potential litigation of unintentional melody "infringement" is blessed. Even if this resolved to only 1 less lawsuit it's earned it's keep. Our culture of litigation is not a to our credit.
- Rochus 3y ago
  I agree that the present culture of copyright litigation is not a to our credit. But I don't think the presented approach can improve that; although the authors themselves are lawyers, the approach suffers from a considerably more serious flaw than the ones I have mentioned so far; the primary problem is likely that the published melodies have been generated by an algorithm and are therefore not protected by copyright law, and thus neither in order to invoke a pre-existing work in legal proceedings. But the real problem from my humble point of view is the perversion of copyright law in order to appease corporations and publishers so that it can even be applied to individual melodies (i.e. ideas instead of works).
  - harel 3y ago
    The "created by algo" point is unclear to me. Is the removal of copyright only applicable (at this point in time) to content (or music) generated by LLMs and the likes of them, or also to something like All the music where the melodies were coded into existence by a human. In other words (which might not make sense); In the AI "traceroute" between "algo" to "output", there is one "hop" - the human who coded the algorithm.
    - Rochus 3y ago
      In most countries copyright laws only protect work that was created by a human being; e.g. the US Copyright Office will refuse to register a claim if it determines that a human being did not create the work. It doesn't matter if an LLM or a simple program or calculation generated the music; in both cases the creator was no human being. There were cases in the press recently where this rule was confirmed for DNN generated works. Also note that copyright law protects "expression of ideas" (i.e. works), not the ideas itself (nor compilations of ideas), and that the work must possess "at least some minimal degree of creativity." It is also debatable whether a single, short melody can be considered a work or a mere idea. Unfortunately, in the last twenty years, the courts (and in some cases also the law) have begun to soften the concept of a work even further and to allow even trivialities to be considered protected works (a similarly decadent development could be observed years earlier in the patent system). Although this pleases some corporations and publishers, it has not exactly helped to improve legal certainty (nor innovation).