Will we ever run out of new music?

Some fun math for your Friday. Vsauce discusses whether or not it’s possible to ever run out of new music, directly challenging the thought that the lack of originality in popular music is due to us hitting some sort of “peak creativity”.

The number of possible combinations of bits that make up a valid mp3 might be significantly less than an admixture of every possible bit combination therein, of course — mp3s have a file structure that must be present to be readable. But for the purposes of this thought experiment, let’s assume a file format like mp3 that already has its header accounted for, and the rest is just a blind read of bits.

Will we ever run out of new music?

10 thoughts on “Will we ever run out of new music?

  1. 1

    Will we ever run out of new music?


    His description of “every possible brief little melody” is not well-defined, because “brief little” isn’t a length. We’ve already done all of these, but not every ordering of all of them, or all combinations of every ordering. That is one thing you’d have to look at, to define what’s possible for one kind of music.

    The giant number of bits in an mp3 is a different issue. Notes in melodies are not “bits.” One note takes lots and lots of bits, because bits are very, very short. mp3s obviously don’t need to be limited to five minutes. That’s all irrelevant.

    And music isn’t simply just different “melodies” which make one “song” different from another. Some music is the juxtaposition of multiple “melodies” at the same time. (Hey, dude, the one telling us all about music: try learning that counterpoint exists first.) Some music doesn’t have a melody at all. Some music doesn’t divide an octave into twelve notes, and some doesn’t divide the octave into the same intervals of twelve notes…. I could go on. But there’s no need to be ethnocentric, or just ignorant in general.

  2. 2

    No, because the vast vast majority of possible sound sequences are unpleasant or just noise. The art is in finding appealing ones amid the universe of nothing special.

    We haven’t even run out of new musical GENRES yet.

  3. 3

    What we think of as music is run by the dopamine signaling structure of our brains, and what we experience as melody is driven by the same circuitry that we use when we match the motion of our hand to the object to which we’re reaching. I learned this in a conference where a new discipline called “cognitive musicology” was being discussed.

    Therefore the mere notion of a universally appealing sound sequence is meaningless, since no standard criteria can be applied. This is not the same thing as saying that any random collection of noise is “music”, any more than I’m saying that any random collection of brain cells is a “mind”. What I do mean is that our brains change and evolve quickly enough that new forms of coherent sound will be continually invented and experienced.

  4. 6

    As consciousness razor @1 pointed out, 5 minutes of mp3 is not a good measure in any context because it’s a compressed format. 5 minutes of mp3 could be 100 bits or 10000000 bits depending on the entropy of the file. In other words, sure, if you compare the bit-lengths of common music (and therefore similarly compressed bits) you might come up with the number of permutations in the video. The problem with this approach is not only the selection bias introduced this way but also that just because there are so-and-so many permutations is not indicative at all of how many of those would even be recognized as music to the human ear and brain.

    The math presented here reminds of the speculation that every possible piece of information can be represented, and thus encoded, as a subsequence of PI, due to its (quasi) random nature. E.g. every single play of Shakespeare’s could be reduced to a start and stop position in the decimals of PI.
    Problem is, the precision required to store or even to find those positions can be practically infinite.

    I guess my point is that while these mathematical shenigans are interesting, they are ultimately not applicable and generally unlikely to grant any insight on the matter at hand, at least when the constraints of the discussion are so ill-defined.

  5. 7

    Perhaps a tangent…

    I think the hard cap on “new music” is legislated by the Copyright Act and subsequent rulings. The Under-Pressure v Ice-Ice-Baby case iirc, determined “melody” the base unit of song copyright as a 4 note phrase, in any key*, of any duration. (Really the intervals are the “magic” bit, so an interval-matching-phrase starting on a different note could be infringing. Could be, because infringement is by default until ruled Fair Use; Fair Use is an affirmative defense.)

    E.g., a song would infringe on Salt n Peppa’s “Push It” if it used just a third of its signature phrase with different timing as its own melody.

    I think this significantly limits the available space of new non infringing music in the US for as long as the Copyright Act continues to have its term extended. Make all the music you like, even if it’s MP3s that are largely just noise. You might get sued and if a judge can’t tell the difference you’re infringing whether bitwise analysis of particular encodings shows variation. (Although, the existing public domain space might be saturated by this definition as well, permitting a vast space to mine without fear.)

  6. 8

    Oh, and you’re probably thinking, “Surely with a combination space this small, collisions must happen all the time!” And you’re probably right. People are plugging away producing songs that are infringing without them ever knowing about it.

    Unlimited copyright kind of makes new music a crime, but everyone is just skating until someone powerful enough notices and screws you.

  7. 9

    Could we run out of possible combinations of music? Maybe, but only if we limit the largest the number of notes in a melody. It’s like computer passwords. If all passwords must be exactly eight characters, then a brute force password cracker has a head start. If password must be at least eight characters but has no limit to it’s length, cracking it will be harder.

    There’s not just the length of the melody to take into account (how many notes or bars), but also time signatures, key signatures, rhythm, beats per minute, etc. And if you add free forms and improvisation, the permutations become so large they might as well be infinite.

    The main reason this is even a discussion is because so many maker of music and muzak (top 40 crap) deliberately limit their forms. More than 90% of all popular music is in 4/4 time, between 60 and 120 BPM, in only a few key signatures based around middle C, and uses nearly all the same instrumentation or computers. How many popular groups do you hear anymore using brass or woodwinds? This isn’t like the 1970s and 1980s when there was a plethora of bands using 6-12 instruments in a song (e.g. Blood Sweat and Tears, Chicago, Sly & Family Stone, Steely Dan, Madness, etc.) and produced an array of sounds.

    Nearly all music today starts from the same place, so it shouldn’t surprise that it all sounds the same. Take a listen to blues music (and early rock and roll), arguably the most limited form of all. Most of it is three chords, twelve bars, the same “walking bass” line, and guitar/bass/drum players. It’s fun to play when you’re a beginner learning how to play music or when you don’t want to think hard, but it’s tedious to anyone with a trained ear.

    Regarding Sheesh’s comment, groups may unintentionally copy what others have done, but I doubt that someone could successfully argue that songs written independently with no knowledge of each other could be deemed infringement. The variety of different popular music in different countries means that could happen. From living in other countries than my own, I’ve heard deliberate copyright infringement go unpunished, local groups copying western pop hits, and I wouldn’t put it past westerners so steal from other countries. There’s a fine line between theft and music appropriation.

    The lack of variety is one of the reasons some groups from the seventies and eighties are still listened to and still play to large audiences now. It’s not just nostalgia, people want to see musicians play. Groups like ELP, Rush, Steely Dan, Madness, Fishbone and many others pushed the boundaries of song content and structure and challenged the listeners and were rewarded with longevity, if not the height of popularity.

  8. 10

    First, I love Steely Dan.

    But Copyright is worse than you think. Music isn’t ruled infringing, it’s infringing by default until it is adjudicated Fair Use (and claiming Fair Use is an admission of infringement, affirmative defense). That’s how the framework is set up; intent to infringe doesn’t matter. Independent “discovery” of a tune doesn’t matter. Publication dates and filing dates matter. The decisions on this really have been quite stupid, and the century long terms amplify the problem.

    The estates or copyright holders of works from the 40s could enlist some musicologists to mine pop music for infringing phrases and be suing people all over the place if the hassle was worth the reward. It’s largely a philosophical concern since we can all agree that since it hasn’t happened yet, it’s not going to even if the existing state of copyright law allows for it. It’s like patent-trolling but for music, I’m sure bloodsuckers have looked into it.

Comments are closed.