Traditional Music Notation is part of the problem

traditional Music notation is part of the problem

To try to make Traditional Music Notation (TMN) easier to learn, I put on my User-eXperience (UX) designer hat, and began throwing away the elements of it which would get a UX pro fired from their job. I was left with nothing. That meant I couldn't just fix one thing and have good UX. I had to throw it all away and start over from scratch. On this page, I catalog over a dozen of those reasons.

In lieu of reading it all, you can watch it in this video which is about 33 minutes long. Because it is served through YouTube, it may show ads or make recommendations to other videos that 1) I don't know about, 2) I can't change, and 3) are based on your video viewing history and preferences.

#1 Arbitrary icons

To represent rhythm alone, a dozen unique non-intuitive icons must be memorized. I mean, think about the symbols: "oval, stem, filled oval, wind-blown flags, flags with no wind, dot, hat, squiggle, rotated/stretched parenthesis, upside-down hat, half an eye, circles connected to tilted-stem by flags" (Well you tell me what they look like to someone who hasn't spent years learning them!) None of these icons has any bearing to the physical world, so we call them "arbitrary".
All that simply to mark when to start and when to stop. If users won't be trained formally, then it's not 'ok' to require them to memorize arbitrary icons.

"leading" is many things, but one thing it must be is to "show someone the way ahead of time", with emphasis on the "ahead of time" part. When a worship leader performs in-time with the music, they are very specifically not ahead of time. There is a part of the brain which can infer when something else will happen at a precise moment in the future... (And it's not the 'logical thinking' part.) So there is actually a solution to this problem.

#2 LIsts to memorize

Communicating pitch also has an icon-memorization problem, such as the gothic "gs", "F", "C", "#" and "b", (Oh, you didn't know the treble clef symbol was the letters "g" and "s" written in a gothic font? hmm... #3 illegible fonts). But the real problem with pitch is primarily in the implicit lists. The half step intervals of the major scale (2,2,1,2,2,2,1) must be memorized and frequently recalled by the names of the notes (another list) (c, d, e, f, g, a, b) (anything seem slightly odd about beginning an alphabetical list with "c"? #4 alterations to traditional conventions), and then the notes are placed in entirely different positions on the staff, depending on the clef.

Some of you are thinking "but that's just the way music is", or "but there's a historical reason for this" or "the eccentricities of music are what make it beautiful", then I would say, "The practical result is that TMN is so hard to learn, that most people don't, and won't.". (And yes, I had to work through the 5 stags of grief to come to terms with that reality.)

#5 Multiple ways to do the same thing

In TMN, many concepts can be achieved in multiple ways each, so viewers have to learn multiple ways of doing the same thing.

A note's duration can be extended by a dot or by a "tie". A viewer must learn both. And by the way, "ties" looks exactly like "slurs". And "extend the duration by half" dots look just like "shorten the duration by half" dots, they're just in ever so slightly different places.

Notes can be marked as accidentals by either flats or sharps. Heck, notes in the key signature can be marked as either sharps or flats.

Triplets can be indicated either by a square bracket with a "3", or by compound time.

One lesson we learned when moving apps to the iPhone from the Mac is: if there's only 1 way to do something, instead of 4, then 75% of the clutter is just gone. So a good musical user-experience would have only 1 way to notate any particular feature.

#6 Missing information

TMN makes up for redundancy by leaving out other information entirely. Most hymnals, for instance, don't have repeats at the ends, even though they have multiple verses.

Unable to change their layout once printed, hymnals can’t communicate the rest measures or modulations inserted by the music ministers. Causing the hymnal to fail at its primary purpose: unambiguously communicating when to sing what.

#7 Anti-proportionality

When there is more of something, we typically see more of it. So when we see more of something, we think there is proportionaly more of it. However, for the durations of notes, the amount of ink spent on the note is inversely proportional to the duration of the note. I.e., the more ink: the shorter the duration; the less ink: the longer the note. Whole notes have half as much ink as half notes, but double the duration. Quarter notes again, approximately double the ink, but half the duration. Adding a flag halves the note duration, but adds more ink.

For people to intuit how much of something there is, as the amount of the thing increases the visual indicator needs to increase, not decrease.

#8 Useless Flexibility

Take a look at time signatures. Sure, they look like a fraction, but fractions were actually invented later. That lower number tells you which kind of note gets the beat. But that's useless. If we want to make notes shorter, we can add a flag to the stem, ad infinitum... or at least until no one can reproduce the rhythm. Or if we want to make notes longer, we can use ties. So what is the use? It doesn't serve one anymore. Some say we need it for compound time, like 6/8. But in 6/8, the eighth note doesn't get the beat, and there aren't 6 beats in the measure. 3/8 get the beat, and there are 2 beats. Neither of those numbers show up in this "time signature". Nowadays we have triplets, and could notate all compositions which use compound time as triplets. but no... they've got to stick to using a meaningless notation, "6/8" to mean 2 x 3x8ths. Once I'd realized just how loose worship leaders are with their rhythms, and I had created a rhythm system which is both infinitely precise and non-discrete, I realized I had no need for a distinction between triplets or compound time. What I needed always was one thing: "the beat". Almost everyone knows what "the beat" is. Once I had made this design choice, the flexibility of the lower number in the time signature was meaningless. But of course, it took me a full year after creating my first prototype to realize that. And I think that level of unawareness of the terrible User eXperience from TMN is fairly typical amongst classically-trained musicians.

#9 cumbersome inflexibility

Frequently, TMN tries to fix one mistake by doubling down in another. Consider pickup beats. While there are many popular songs without pickup beats, a majority of popular songs have pickup beats. TMN doesn't have a way to make sections break (or lines break) anywhere other than measure lines; for instance, repeats always happen at measure boundaries. However, songs with pickup beats always break at places other than measure lines. Measure lines were about the one feature I kept from TMN, but I don't place these kinds of restrictions on them. I knew from my survey results, that 2/3 of the people who would be making the changes to the songs that the worship leader wanted (skipped verses, inserted turns, repeated sections) would need to be made by people who had no musical training. To do this in a song with pickup beats would require careful attention to detail which they wouldn't be trained in how to do. Musicians figure it out with intuition, guided by a first measure without enough beats to match the time signature. So SingAccord would have to be programmed with my intuition – to be able to identify and divide a song into its sections by itself, with no help from the user. Instead, the user can drag & drop sections into a "form", with as many repeated sections in any order they want, and SingAccord will automatically stitch together the pickup beats the right way. Without inventing a dizzying array of arbitrary icons and #10 foreign language #11 abbreviations in to indicate what to do.

#12 ILLEGIBLE Patterns

Researchers were stunned to learn that there are brain cells (neurons) in the retina of the eye, which compress the data that's sent to the brain. One of their functions is to detect repeating patterns. To the brain, they only send the image of a single element in the pattern and the extent of the region occupied by that pattern. Our brains literally never get all the image data. Depending on the person, the limit of the number of repeating elements to trigger this pattern filter is between 3 and 5.

This means the 5 staff lines of the traditional music staff are at the limit of human perception.

Attempts at 'chromatic' musical notation systems which use 6 or more lines can't even be seen by humans.

#13 Ratio of size / live thickness is too large

The human eye has a tiny region in the middle called the "fovea" where visual acuity is best. It's only 5 degrees wide. Near the edge of this region, the sharpness of our vision falls off rapidly, and continues falling off rapidly in the (surrounding) parafoveal region.

The result is that for average vision, people literally cannot read the fourth line of the lyrics and see which line of the musical staff a melody note is on at the same time. Professional singers are taught to look at a point half way in between, but chances are your congregation hasn't heard that.... if they even have hymnals.

A new music notation would not only have to account for legibility of content on the big screen for folks in the back, but also for the fact that not all the information on the screen can be seen in full resolution at the same time.

#14 Unwritten Rules

Consider how accidentals are notated. The musician must instantly memorize a key signature at the beginning of each line, which contains an oddly-sorted list of alterations to notes. While reading the line, the musician must note any additional symbols and memorize them, overriding any line-specific alterations. BUT, when a measure line is encountered, he must forget the temporary alternation, and revert back to the one for the line. A very complex task. Worst of all, these rules don't appear printed on the page, so unless someone tells you, you aren't going to "figure it out".
#15 In fact, even when someone does tell you the rule, it's still hard to actually do it. It's so hard that music for beginners often doesn't use key signatures, but instead explicitly notates every accidental individually.
"Shape notes" owes its entire existence to the removal of the accidentals rules, but as so often happens in TMN, fixing one mistake means doubling down in another, by creating more arbitrary icons to memorize.
Oh yeah, and the order of the accidentals in the key signature? It's sorted by another unwritten rule, the "circle of fifths".

#15 Shows Un-perceivable and Un-actionable info

Only a handful of people have 'perfect pitch'. Most other folks only have 'relative pitch' which allows them to kind-of carry a tune. But if you played them a note, and five minutes later played a note 1 step different, they wouldn't know.

Notating music for instruments in perfect pitch makes sense: an E♭ is played in the same position on a trombone or the same fingering on a trumpet regardless of whether the key signature is A♭ or B♭. But to the human ear, the 4th and the 5th chords sound very different.

A well-designed user experience places emphasis on what most people use, and deemphasizes advanced features. While TMN concentrates on presenting information that almost no one in the congregation can make accurate use of, even if they can read it.

So SingAccord uses a relative-pitch system. Each note is notated against the intervals in the key directly. The key note is notated in a deemphasized way, so that the few people with perfect pitch can still make perfect use of it.

#16 Can't adapt to new artwork

I don't believe in re-directing the congregation's attention from the content of the lyrics to a psychedelic background video on the screen during worship, but I do believe carefully-crafted artistic content can help set the mood for a song, or section of a song. But traditional music notation is stuck in black and white with Baroque styling. And worse yet for projecting on a screen, it isn't designed to have an image background. With only 1 pre-defined color (black), when laid over a background image TMN can't guarantee legibility by providing internal contrast. It can't be themed to feel at home or stand out from the background content.

A new music notation, designed for a projected screen, would have to take into account the artist's use of color & background imagery to set the mood.

#17 The problem with slide shows is... slides

And it doesn't stop with the music notation. I had to figure out how to solve those pesky presentation-software program problems: like when the volunteer is half a second late and the slide doesn't change and suddenly nobody knows what to sing. Going into this, I did not expect that the problem with slide show software was the slides. I used to be the guy running the slides. But it was. So I got rid of slides. So SingAccord never creates these "all or nothing" moments. The next line is just always there, ready to be sung. Be honest, when I first said "slides are the problem", you probably thought I was crazy. But after you see the demo now, maybe do you think continuing to use slides is crazy?

#18 ADvanced Math

Consider the system of decreasing the duration of notes based on the number of flags. The formula is duration = 1 beat * 2^(-#of flags). In other words, the number of flags is the negative exponent to 2 times 1 beat. Negative exponents is pretty advanced math, but musicians are expected to perform it multiple times per second. No wonder you occasionally see those articles in the news about how some musician is doing some difficult complicated technical/ science/math job. Probably only the people who can handle the complicated science / tech / math jobs can handle the music notation.

the courage to change the things I can

At first, I thought I was the only one who admitted we have a problem with TMN. It turns out that many, many people have figured out there were major problems in music notation. In the last 400 years, more than 457 documented attempts have been made to fix TMN. It's just that almost none of them have any experience in User eXperience... Until now. (Although, 4 of those old guys had good ideas.) TMN's founding principles are based in a technology that is obsolete: pen & paper, and on a complete lack of any expertise in visual communication or user experience. Guitar tabs proved that even a half-way solution could be very successful if only it agreed to throw out the mistakes of TMN.

Many, many musicians have spent many, many years learning traditional music notation. Others haven't, such as many lead guitarists who can only read guitar tabs. I have, in the past, even used the phrase 'musically illiterate'. However, when I took a dispassionate & professional look at the user-experience of TMN (as I was taught to while getting my Ph.D. in Aerospace Engineering), I discovered that it is truly very poorly designed. And it couldn't be tweaked. Decades of research of dozens of international researchers have gone in to trying to tweak it to make it better, and their results were zip, zero, nada. They had little to no expertise in User eXperience (UX), and they believed in the same fundamental problems that had lead TMN to be so bad in the first place.

But kids could still play 'Guitar Hero', so there had to be a solution...