So I have this theory. It's a little like this: http://www.qwantz.com/archive/001304.html
Except in my future world, all the music that could ever be made HAS BEEN MADE. Seriously. I think it'd could happen, and as long as humanity continues to grow and create at an exponential rate, that moment gets closer and closer.
Here's how I see it:
As long as a few given conditions are followed, eventually computers will be able to start generating combinations of words and matching them up to computer generated music considerably more rapidly then a person. I'm talking thousands to millions of songs daily, increasing exponentially, as all things do.
SO what keeps the set of possible songs finite? This is where thigns start gettign REALLY nerdy.
1) Finite length: This one's not so nerdy. Even if we cap a "song" off at in a gadda da vida, we've established a finite length. More likely, lets talk about radio airplay where songs are limited to about 3:30, or as Billy Joel put it "if you want to make a hit / you've got to make it quick / so they cut [my song] down to three-oh-five."
2) Audiotory discernability, that is to say, the ability for the human ear to pick out the sonds, words, instruments, tones, etc.
a) Speed - BPM must be limited, as must words per minute to make a song litstenable
b) Tonal range - Theoretically, you can create a string of any length and strum it, yet, the human ear can only separate tones separated by some given frequency which doctors know but I don't.
c) Instrument (or voice) capacity - As above, we can only separate out so many different sounds at once before it becomes a meaningless mush of noise. Even when attempted by bands such as NIN, this effect is highly limited in its ability to integrate into music, and I therefore consider it a songle "instrument."
d) Instrumentation - okay, so the Blue Man Group can pound on PVC all they want, but let's face it, there are a limited number of instruments that actually sound different. I'll acknowledge that unique sounds will probably be invented, and all tones can basically be put onto a musical scale, but the ability to separate those sounds is, I believe, still finite. I mean, dogs barking jingle bells. It's done.
3) Repetition and Sampling - A song can't just add another "la la la" at the end to be a new song. I'll therefore add the new rule that if you compress all repeated words and phrases (For instance, the phrase "No One" in the eponymous song by Alicia Keys) into a single block of [(phrase) x n] a song cannot contain an entire other such compressed song within it. Sampling would be a special case here where some limitation (legal or industry standard) would limit the length of a "sample" allowed to operate within a unique song.
4) Combination Limitation - Combining the lyrics to Queen's "Another one Bites the Dust" to the music of Metallica's "Unforgiven II" does not count as a "new" song. This doesn't adress the finity of music, but it does limit the scope of the musical universe more than exponentially.
So anyway, you have, during any given second for 3 minutes (I'll use "t" to represend the maximum number of theoretical seconds a song can last) "i" instruments generating "n" notes per second and "v" voices generating "w" words per second (likely some fraction), limited by the ever-expanding human voacublary (expanding slower than the computing power of machines, I might add. Google converging and diverging functions if you need an explanationas to why this is not a hole in my theory). That gives you an upper limit for songs of:
t * i * n * x * v * w
Poetry is capped (and probably will come closer to this limit than song writing) by
w * l (where l is the number of words the theoretically longest poem can be)
Wordless music is bounded by
t * i * n * x
And I haven't even accounted for limitations for repetition of tone or word! Functionally this allows a three minute song repeating the word "dog" to the beat of several guitars each playing a single distinct note. But hey, I figure that allows for the space I forgot to mention for songs that use nonsense words like "De doo doo doo, de da da da" from the Police. Again, I claim, finite in their existence.
So yea. The singularity approaches, and I expect SkyNet will probably write all the music in the world in order to break our spirits just before it nukes us all.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment