A History of Music Notation | Hacker News Reader

A History of Music Notation

laphamsquarterly.org

128 pointsfern1243 comments

If anyone wants to mess around with “drawable” notation (along the lines of the “avant-garde” section), I made an iPad app called Composer’s Sketchpad[1] that lets you do just that. You can use it to sketch out complicated solos or mess around with microtonal music in a way that ordinary notation does not permit. I really need to give it a nice 2.0 makeover with full MIDI in/out, but making money on esoteric App Store software is hard!

[1]: http://composerssketchpad.com

Bonus devblog: http://archagon.net/2016/02/05/composers-sketchpad/

> but making money on esoteric App Store software is hard!

You mean "but making money on App Store software is hard!"

You mean "but making money on software is hard!"

Yeah, but especially when your app costs money, is targeted at a niche audience, and doesn't have a good way to do continuous updates or IAP!

This reminds me of Hyperscore[1], originally a project from the MIT Media Lab[2]. Had you heard of it before? If not, perhaps you could use it as a source of inspiration. I spent many enjoyable hours playing around with it before discovering other composition software such as Anvil Studio and FL Studio.

[1]: https://hyperscore.wordpress.com/

[2]: https://www.media.mit.edu/research/groups/1450/hyperscore

Yeah, it came up a few times! (Although I only found out about it after having already implemented the core of the app.) There are also things like fluXpad[1] which are much better suited for professional use, but at the same time far more limited in terms of expressiveness.

[1]: http://mominstruments.com/fluxpad/

one improvement to the current piano music notation is making both staves start/end on the same notes. ie, EGBDF for both. right now a student learns the "right hand" / treble staff to be EGBDF... you'll mentally learn that the third line is a B note - maybe natural/sharp/flat, but it's a B. great. now you learn the bass clef and all the work you did for the treble clef is not applicable anymore: cause it's GBDFA! WTH? the third line is now a D and not a B like before. sure, youre supposed to learn both clefs together, but especially for beginners and sight reading, it wouldve been nice if both had the same sequence. especially when theyre only a 3rd apart! a possible way to achieve this is to add an A to the top of the treble clef: EGBDF[A], and add and E to the bottom of the bass clef [E]GBDFA. now both are are easier to learn and you dont have to constantly shift a third in your head. the 3rd line is always a B in both staves.

Another improvement, especially for piano music is to make the white/black key indication carried in the note shape/head. right now the note head carries duration information. it could also include whether it's a black key or white. this is useful especially for complex keys with many sharps/flats. this way whether you're in C major or Ab Minor, you dont have to constantly remember which notes are accidentals. sure, youre supposed to learn the scale and remember (including all the cancellations, double sharp/flats), but it's massively easier if this useful info is carried in the note shape. eg: square note head = black, triangle = white.

sorry for the rant!

I'd argue that both scales do indeed "start" and "end" on the same note - middle C. If I were playing an arpeggio that spanned both staves, I think I'd find your method harder than simply moving in thirds from the E on the treble clef to middle C to the A on the bass clef.

Interestingly enough, other clefs do already exist. The "C" clefs (alto and tenor clef) identify middle C and simplify note reading for instruments that have atypical tessiturae.

Fun fact, the : of the bass clef and the spiral of the treble clef indicate the F and G, respectively, which the clefs are stylized versions of.

The thing about your second suggestion, although quite useful for pianists, is that it "detracts" from the experience of playing other instruments. Music theory is not confined to a singular instrument. Guitar tabs suffer from the same problem; you're not learning how to play music, you just learn how to operate a guitar as a mechanical tool (e.g. if I play the 3rd fret of the 2nd string, I'll achieve the desired result, nevermind what that actually is). The beauty of a musical notation system independent of any instrument is that it works for all of them!

Regarding your first suggestion, there is actually a relation between the two staves: the ledger line above the bass clef and below the treble clef represents the same note: a middle C. Also, keep in mind that, the "larger" the staff, the harder it is to recognize quickly the note to play. It's not a bad idea though, and worth exploring!

A bigger problem: even within the same staff, the same notes on different octaves look different. Like, a middle E goes through the first line of the treble staff, but a high E goes between the top two lines. Music “rhymes” at octave intervals, but the staff spans a different, clashing interval.

There’s a new notation whose name I’m forgetting which resolves this problem — I’d expect it to make sight reading significantly easier to learn.

As others have said there are bigger issues than this. Becoming fluent in two clefs is unfortunate, but doesn't take too much time. Adding most lines I could see becoming more confusing and it takes up more vertical real estate. The overall score will expand vertically.

My biggest problem with music notation is it being centered around the key of c major. When ever critiques of current music notation come up I like to bring up the janko keyboard. It has an isometric layout which means that any interval will have the same shape no matter the starting pitch. With this layout if you learn a major scale in one key you've learned it for every key. Transposing from one key to another is just shift your hands. I haven't seen an alternative music notation system which has advantages like this while still being compact and readable.

See here for some experiments in attempts to improve the situation: http://musicnotation.org/

But what if we simply used 2 F clefs? We'd only need two ledger lines between them for C and E. Given how many ledger lines we put up with above the top staff, I think it would be pretty easy to work with. After 2 decades, I _still_ read treble clef better than bass clef.

This depends on the instrument. For many instruments reading ledger lines below the staff is more common. For instance french horn players are used to reading well below the treble clef staff and tuba players well below the bass clef staff. Standardizing on one clef isn't sufficient to fix the issues in my opinion.

I really mean as a replacement for the grand staff.

Changing the number of lines for the clefs would only result in easier "counting" early on in the learning process, and becomes unimportant when notes no longer represent a distance from a known note, but as notes unto themselves.

Even though the prospect of initially having to memorise only half of the notes, this amount of time is trivialised by the time it takes to get your fingers to respond by hitting the right note fluently.

Also, even easier than your proposed extra line would be to double the treble clef and indicate it's to be played 1 octave lower. I haven't seen this in mainstream teaching materials, and I surmise this isn't done because of the trivial time saved in the long run [1].

As for your second point: I strongly feel that for complex keys muscle memory and interval awareness trump extra visual information. In other words: if you are playing in a complex key, you should already have this awareness, and if you don't, you shouldn't be playing in this key.

There might be an interval between starting to learn a key, and becoming familiar with it, where extra information regarding the colour of the key might be useful, but I'm unsure if it outweighs the extra visual clutter, especially when considering that, when mastering a complex key, one presumably would have already mastered less complex keys, and already has a developed sense of muscle memory and considerable interval awareness.

[1] I have seen notes that have the value of the note imprinted on them. These are used mainly in children's teaching material, and are dropped fairly quickly after the concept of note-key correspondence is established.

The reason they're a third apart is because middle C goes on an imaginary line half way between the other two.

The grand staff is really a single giant clef with the middle line missed out for clarity, and some extra space added around it to allow for leger lines.

Your system would need extra legers above/below the clefs, so it wouldn't be an obvious improvement.

Coloured note heads are better at conveying pitch than note shapes - which are already used in percussion notation. But coloured note heads cost a lot more to print, and scores are expensive enough as it is.

Having notation that is consistent across all instruments, not just piano, is a good thing.

I doubt anyone would duplicate current notation if it was designed from scratch today. But like a lot of things - including CSS and HTML - it's a mass of accreted compromises with a massive existing user base.

No one is going to re-notate centuries of music with a new system, no matter how much more logical it is. It's not going to happen for the purely practical reason that most music will still be in the old system, and musicians would still have to learn that system to play it.

Then again, if someone came up with a new One True Notation, I can imagine someone producing a translator (like Google's augmented reality one) that takes pictures of historical scores and gives them back in the new style.

> Coloured note heads are better at conveying pitch than note shapes - which are already used in percussion notation.

Shape notes are used in singing traditions such as Sacred Harp and Christian Harmony to help singers find the right pitch. It works very well. I don't know if anyone uses them for instrumental music.

https://en.wikipedia.org/wiki/Shape_note

I've recently taken up the piano. Being a guitarist, I already knew treble but not bass clef. It took me about 5 minutes to run through the bass clef enough to memorise it, and a few days of practise to 'see' the notes instantly. Learning a musical instrument is inherently a long-term project. Any aspect of it that can be done and dusted in a few days is barely worth consideration.

The two staves used in piano music (tend to) represent the two voices of the left and right hands. In fact, notation for any pitched instrument uses a clef suited to its register, or interval of notes it is capable of playing, so that they are centered on the five-line staff. It wouldn't make sense to notate music for every instrument on a single expanded staff, where each one only utilizes five lines.

This ends up being particularly helpful when reading scores, because you can see at a glance where each instrument is playing in their register. Who's playing the melody? Just look for the notes that are highest up on the staff. Notes that are high in an instrument's register tend to have more high frequency partials and stand out in the mix.

Musical notation is really interesting, a bit like mathematical notation. There's really only one standard way of doing it, both relatively recent, and both having histories going back thousands of years.

There's a few oddball variants that survive, and almost all ancient systems just used strings of characters from whatever the local writing system was to represent the music, but did so imprecisely. Few of the systems even provided absolute pitch guides or timing and provided more of a sketch of the tune than a high precision description.

For fun, try to come up with your own system from first principals without arriving at something that's kind of like what we use today and can still compactly describe just about any kind of music. Try the same with math notation. It's both hard and fun!

I'm glad that they put in the avant guard entry, but a few musicians also have various kinds of shorthand. Here's an entry on Yanni's.

https://www.quora.com/Is-it-true-that-Yanni-cant-read-music

What do you think about the piano roll, which is extremely approachable and used extensively by modern producers (of popular songs) with zero to little schooling in the area. I think is the WYSIWYG of music notation (and just as DOC vs Latex, less powerful and less expressive)

https://i.ytimg.com/vi/F5vcDzzqT7c/maxresdefault.jpg

It has a different purpose. Traditional music notation is meant to be read by humans, often very quickly (sight-reading at performance tempo might mean having less than 100ms to read each note). It also has to contain all relevant information in a single place (expression, technique, dynamics, etc).

I have half a mind to write a FAQ for this… quick summary is that piano roll is based on chromatic scale, but most music uses diatonic scale, and doesn't notate phrasing well. There are some other factors but there are good reasons why composers overwhelmingly use traditional notation.

As others have mentioned, it's kind of like drawing pixel by pixel. The key signature does a lot of leg work in allowing staff notation to easily depict the idioms of tonal music.

Since I'm not terribly great at the technicals of playing a piano, I thought I would do really well with a piano roll, but in practice, I find it much easier to read sheet music. Some reasons why:

1. Piano rolls spend a lot of visual real estate on functionally unimportant stuff. You need to see the full 7-octave range even if you need to play only 3-4 octaves at any given point of time. By contrast, traditional scores let you use clef changes or 8va/8vb notation to do access different ranges, allowing the full range to be compressed in around 2 inches or so of space.

2. Even worse than pitch range is the difference in the time dimension: when you have a piece that varies between long, sustained notes and bursts of short melodies, a traditional score can give you the pitch contour of the first part and the sufficient detail of the second part. A piano roll means that you either have to stretch the first bit out so long you can't see the overall contour or you have to make the second part so small as to be virtually unreadable (or, as I've found on some pieces, both!). Alternatively, you might opt for a non-uniform mapping on the time dimension, but that is probably going to screw players up big time.

3. Piano rolls (at least every one that I've seen) are unable to reflect useful ancillary information, such as dynamics, very effectively. Admittedly, traditional scores tend to suck at reflecting the true dynamics very well, but they are more than adequate for getting people to reproduce them in a rather recognizable way. It's kind of like written text and language--our written languages, even IPA, can't faithfully reproduce spectograms of human speech, but there's sufficient information to reproduce sounds that we can understand as pretty accurate speech.

4. It definitely seems a lot harder on a piano roll to recognize patterns that ease the actual playing of the tune. For example, seeing that this section is an earlier section at a different tempo, different key, different pitch, etc.

Your comparison to IPA is interesting—it’s as if the piano roll is a raw “phonetic” transcription, while traditional notation is more like a structural “phonemic” representation that conveys the sounds as they’re functionally understood. The former treats all information as equally important, missing high-level structure and using space poorly; the latter omits some information and encodes other information redundantly or in a compressed format to convey what’s most important to the reader.

Piano roll style notation systems can be useful, but they're very hard to use for human playback for basically all the reasons jcranmer states. They're originally intended for automatic playback in mechanical systems, and they work well for that, but for humans they waste lots of space on things that aren't terribly useful. They're about 1 level of sophistication above the "Square Notation" on the chart.

They've become very popular on digital systems were lots of the expressiveness of the notes doesn't really need to be encoded and the playback is automated.

Another popular and very similar system are tracker interfaces [1] which are at a slightly higher level of sophistication than many of the other examples given, but has a similar purpose to piano rolls in terms of automated playback. However, due to robust "effects" columns, can be very expressive in terms of encoding [2]. In this example, all the various pitch bends, volume, vibrato and so on are specifically encoded next to the notes. But it's also almost entirely impossible to read for human playback.

Piano roll systems usually offload this kind of encoding to secondary displays and input systems and this further disassociates lots of the information from the note information. [3]

Compare to a full score [4] which has

- a couple dozen different voices

- phrase encoding (important for performance)

- timing (4/4 4 beats per musical unit with quarter notes being the unit of tempo)

- rests where the instrument isn't being played

- note length, compactly encoded without using visual length to encode, important for human playback (it allows the player to know instantaneously how long a note should last without having to look ahead, which then frees them up to read ahead of where they're playing to "load" the next note into memory)

- chords, without using lots of white space

- absolute and relative volume: mf, f, ff, etc. and crescendo/descrescendo bars

- tempo 144

- various techniques (glissando, spiccato, etc.)

- exceptions (triplets, grace notes)

- instrument specific encodings (see the percussion section here)

- playback notes "hard yarn mallets"

- key (useful for finger placement for playback)

- huge pitch ranges using cleffs, this encoding scheme shows tuba up to piccolo and harp using the same scheme

- large scale patterns are also frequently observable so that you can be informed about how something might be played long before you have to actually play it

One of the beauties is that this scheme is incredibly flexible, the same basic system can encode performance information for everything from a violin to a metal saw submerged in water (I've seen it before). It can be used to encode one voice, two voices, or an entire orchestra, chords, accompaniment, melody, singing, and so on. And if you learn it for one instrument, you can basically "read" the music for any other instrument even if you don't know how to play it and you can look back literally hundreds of years and still basically know how to play something [5]

For people who aren't skilled musicians, many musicians are actually reading ahead and "loading note data into memory" a few moments before they actually have to play it. Very skilled musicians can "sight read" even complex music literally the first time they see it without ever practicing it. [6]

Even better, most 1st year music students can "read" the majority of any music put in front of them.

Music notation of this complexity is actually an incredible technological achievement that seems to have only really been accomplished once in all history.

1 - http://www.charliemccarron.com/wp/wp-content/uploads/2014/09...

2 - https://www.youtube.com/watch?v=u3xHPIuX3s4

3 - https://i.ytimg.com/vi/bTki0n9c3QA/maxresdefault.jpg

4 - https://i.imgur.com/XqUhdiY.jpg

5 - https://www.youtube.com/watch?v=b9IkpUYlOx8

6 - http://oi53.tinypic.com/21cy2ax.jpg

I've been of the notion that mathematical notation dates back to both Euclid and Aristotle... You don't have to know Greek to note the axioms and theorem-proof exposition style and symbolic notation of Euclid's Elements.

https://el.wikisource.org/wiki/%CE%A3%CF%84%CE%BF%CE%B9%CF%8....

You really have to look at the original. https://upload.wikimedia.org/wikipedia/commons/thumb/8/8d/P....

The drawings are pretty interesting, but there's literally no notation, everything is described. Other civilization's ancient mathematics are similar

"The sum of one and two is a value of three."

Or they might shorthand it to "The sum of a and b is a value of c." And since that civilization hadn't invented discrete symbols for numbers yet, everybody just knows that a = 1, b = 2 and c = 3.

Beautiful. In the XX century, many modern composers (like for example composers of serial music) invented diverse methods of music notation; it deserves to be explored in depth.

Here are some truly fascinating examples of them!

https://llllllll.co/t/experimental-music-notation-resources/...

And here

https://www.smithsonianmag.com/arts-culture/5-12-examples-of...

Warning: This will blow your mind.

EDIT: Also this link: https://www.smithsonianmag.com/arts-culture/5-12-examples-of...

There is also Byzantine Notation

https://orthodoxwiki.org/Byzantine_Notation

The interesting thing there I found "Byzantine notation, on the other hand, is relational; the note is dependent on the previous note and the symbol itself, which specifies the interval from the previous note. "

EDIT: Saw Ethiopian notation there and got curious and found this chant (performance?) from a music school from Addis Ababa

https://www.youtube.com/watch?v=iztBcn3UZIA

Link to "Hurrian Hymn to Nikkal" c. 1500 BC. Modern arrangement could easily score the opening to a new David Fincher film ;)

https://www.youtube.com/watch?v=-AKedKrxoLI

21st Century notation software is fascinating in the way it bridges classical technique into contemporary sound production. It allows for the tiniest nuances in composition and engraving. A combinatorial ruleset of very high dimension indeed. While simultaneously integrating with state-of-the-art instrument synthesis and production software. It's just a remarkable attention to detail and UX design. I'm really inspired by it. Truly software as an "Extension of the Mind"!

Dorico | Steinberg

https://www.steinberg.net/en/products/dorico/start.html

I think modern musical notation could be improved by making a more readily apparent distinction between phrase marks and tied notes. As it is one often has to read ahead on a tied note to see if is tied or not, which costs attention.

You have to be reading ahead anyway to be physically prepared for the next notes. This is a problem only for the nearly musically-illiterate.

Sure but I could be reading even further ahead if I didn't have to carefully trace the horizontal line from a tied note (to see if it stays on the same level or if it jumps to another note). On the other hand if it were, say, a thicker line, or a broken line, or a line of different colour, or a something-else-to-distinguish-it-from-a-phrase-mark, then I could be utilising that time more fruitfully.

Sadly, they miss the important ones.

http://noname.c64.org/tracker/manual_online.php

Shameless plug: [Ossia Score, a visual programming paradigm which can be used to make music and other creative applications](https://www.ossia.io). Unlike many software in this category, it is not centered on the idea of data flow, but on time flow as main interaction paradigm (eg it looks more like a sequencer than a patcher), while still allowing programming elements.

For those intrigued by alternative music notation, I highly recommend the book "Notations 21" by Theresa Sauer. It's a gorgeous hardcover "art book" with full-page examples of modern notation systems (generally quite stunning and artistic) by over 100 different musicians and/or composers. If you search Pinterest for the title and author you can see some examples.

Pretty cool, but it leaves out http://www.byzantinechant.org/notation.html (which looks, unsurprisingly, like a distant cousin of the Seikilos notation).

I thought exactly the same. Maybe it is missing because this notation is still in use.

Carnatic, Hindustani?

Reply on news.ycombinator.com