> I hear it bandied around in these articles that music
> notation is awful. Can someone explain to me why?
I've played a lot with alternatives, and disagree. But I can explain why people might object to it, but then defend the settlement.
There's so much harmonic music in our lives that most people take harmony for granted. Separately, someone who was interested in only a subset of western music might decide that western notation was unnecessarily complex. e.g. piano can be represented with a role, single-voice instruments can be represented with solfa, guitar with chord sheets.
The Bobby McFerrin video shows us that human music starts with rhythm and a single-voice pentatonic melody.
Western music is many layers more complex than this.
* Uses a standard template of notes (diatonic scale, aka set of black and white notes on the keyboard). Hence, instead of five tones, we have 12.
* Polyphony/harmony. Multiple voices singing against one another.
* Due to the way the tuning of this scale has been standardised, you can play a piece in the key in C, or in F#, and it is recognisably similar. This is a non-trivial achievement, and is post-Bach.
* The "circle of fifths" and the standardised system of tuning that has been in place since Bach allows us to modulate (change key) during the course of a piece. This is a theoretical accomplishment that builds on the standardised tuning.
The optimisations here are so competently selected and executed that "people" take all of this for granted. But Harmony is a product of engineering.
Western sheet music has evolved to give us mastery over harmony. Tracking harmony requires us to track at least two dimensions: (1) time; (2) chords that sit within it at any point. Then, the system needs to scale to multiple voices, and support complex metadata such as vocal words, note ornaments and musical dynamics.
Western notatation is a strong settlement for capturing harmonic music, and it's optimised for the player's readability. (Consider the rule that says you are not allowed to have a crotchet rest across the middle of a 4/4 bar; you'd have to have two semicrotchets and a tie)
Where I think music could be improved is if it were easier to compose. An analogy: consider if you wrote papers using PostScript. Can it be done? Yes. Is the output effective? Yes. But what you really want is a composition-oriented markup that then renders to the form.
I'm currently playing with an approach to composition that feels like literate programming. You build up melody, voices and chords using ascii representations like solfa and figured bass, and a DSL to glue the elements together.
Having done this, I'll want to render it. Render to Clojure/Overtone so it can be played by a computer; render to sheet music so an orchestra could perform it.
My opinion is that there is only glaring flaw in the western musical system: we count notes from one rather than zero. The thing we call "a third" should have been called "a second". But it's a superficial flaw, an awkwardness that makes thinking about intervals more difficult than they should be but is otherwise harmless.
> why do you think that the chromatic scale is correct?
It's not that it's correct so much as that it allows you to do certain things that aren't otherwise possible. The key thing is modulation. As you step away from the current chromatic-scale settlement, you lose the ability to modulate between keys.
Consider this alternative-world settlement. All composers draft all music by pulling levers. The levers drive a pipe organ. It has a strict tonal scale (the seven white notes on the piano, no black notes). The tones are the same as they are currently: B to C is a half tone, C to D is a whole tone. As they play into this pipe organ, the notes are punched into a piano roll, and it can be played back, or a choir can read it.
In this world, you could represent the early catholic church music, although you'd have to play some of it at a different key than it was originally played. (https://en.wikipedia.org/wiki/Neume)
You could get nice harmony going. You could perform the Tallis canon both as a piece with chords (https://www.youtube.com/watch?v=_PqV-iL44wk) or as a true canon (https://www.youtube.com/watch?v=pMNMKxOyGq8).
In this sparse world, you'd write stuff in alternative musical modes to get variety. As they did. You could arrange the Third Mode Melody. (https://www.youtube.com/watch?v=T8oKEx1-J1w) But you wouldn't get the consoling resolution on the final note.
Now let's consider an evolution. Say you allowed the levers to swing left and right a bit so that you could get blues notes by leaning on levers to move them to side or the other. The world we're in here is kind of equivalent to the renaissance. You could perform a fair bit of Palestrina and Victoria on this instrument.
But the Monteverdi Vespers and everything after it is impossible. Bach, Beethoven, Brahms, Jazz, Beatles, Coldplay and everything in between rely on extensive use of chromatic scale, particularly for modulation.
Bach lived after the main transition, but is interesting because the tuning settlement hadn't been worked out yet. He experimented with the frequency differences between the notes. Some say that (some of) his pieces play better if you tune your keyboard for "mean temperament" rather than the now-standard "even temperament".
The rise of electronic music gives us room to go back and play with these fundamentals a lot. The chromatic scale can be challenged: you could design a microtonal scale that followed the same rules but had more subtlety, without needing a craftsman capable of producing a piano keyboard with extra black keys, or someone to play it. You can represent complex rhythms without relying on a human that can perform them. And you can change the temperament of a piece while it's playing.