So, my notation software of choice, MuseScore, got a pretty big update a couple of months ago with version 4.5 (although v4.5.2 is currently the most up-to-date version of the app). I figured this would be a good time to… erm, reevaluate the music notation software landscape with Finale’s recent end (which, lest we forget, I had a front row seat to).

I’ve been using MuseScore for a little over ten years now, and everything I’ve shared on the site has been done in MuseScore. I started with MuseScore because it’s free, and I came to grips with it pretty quickly — it is indeed quite easy to learn. When I went to college, I spent time with Sibelius, Dorico, and Finale, putting me in a tough spot: do I make the investment (of both time and money) to jump to something more serious, or do I stay with MuseScore until my sheet music needs justify such expenses?

It was also a little over ten years ago that I first used Finale, when I was a junior in high school, as a computer lab on campus had Finale installed. It was a surreal moment since I finally got my hands on the ultra-high-end notation software, the one that all the pros use. And all I did was fart around with drum grooves for about 30 minutes.

Well, the circumstances of Finale’s fate forced my hand somewhat, and I am sticking with MuseScore for the foreseeable future. Maybe in another universe, I had more of a relationship with Finale.

I have been a bit hard on MuseScore in the last year or two. The developers have some serious ambitions for the software, which is admirable, but the app is far from perfect in somewhat inscrutable ways. 

I must be clear that I am very grateful MuseScore exists, and I don’t know to what extent I would be transcribing or sharing anything related to sheet music had it not been for MuseScore. It also feels unfair to criticize the software — it has not regressed in any significant capacity, and while the developers do make money for being involved in this project, it is and will (hopefully) remain free in a market where things tend to be very expensive.*

The program has come a long way since I first started using it in 2013. Beams in particular were quite weak, as v1.3 could not join 32nd notes and 16th note triplets with an 8th note beam, leaving them quite an eyesore. Other improvements have been steady, but slow; it took over ten years to get working double repeats in MuseScore (with older scores requiring kludgy workarounds prior). This makes it a little tough to be optimistic about the implementation of upcoming features.

To be fair, before MuseScore’s sheet music database jumped to a freemium model in the late 2010s, it’s hard to know what resources were at the developer’s disposal (MuseScore.com is the primary source of income for the app).

While MuseScore is somewhat of a trailblazer, the shortcomings are still plentiful — MuseScore is the only scorewriter out of the major four three that does not allow for copying and pasting of time signatures. Not all of us are notating prog metal, but it does become irritating for me personally when I do triplet-based 4–2‑1 exercises, which switch between 12/8, 6/8, and 9/8.

There are some more outlandish things that most scorewriters struggle with, such as alternating time signatures (although the aforementioned copy/paste issue would make kludgy workarounds easier) and over-the-barline tuplets (which Dorico does actually support). 

The MuseScore development team does seem to have this in mind, as a blog post from the end of last year mentions the limitations of MuseScore’s measure-based architecture — although, interestingly, over-the-barline tuplets were mentioned as a minor issue that would conveniently happen to get fixed after the current architecture gets blown up. Larger (but unspecified) improvements would also come, but I can only speculate if alternating meters or even copying and pasting time signatures would be part of that.

I guess this at least helps to explain why playback has taken such a priority — you might as well do something while you rebuild the backend, and that is considerate, even if most playback features are not as important to me as actual notation is. 

Up until the 2020s, MuseScore was not taken seriously by professional engravers, who would insist that scores in Finale or Sibelius would look better. But what exactly does that mean? Is there something magical about the premium scorewriters?

Well… no, not really, as all of them will do silly things by default. People say that Finale and the like “look better” because the big selling point of Finale, especially, was that, while very burdensome to learn, it has an outrageous amount of modularity and flexibility. This was famously evidenced by the Conquering Finale web series — the penultimate episode shows off the ridiculous number of parameters for simply styling ties.

Indeed, there are a lot of wacky things that you can do in Finale, things that you would imagine were done in some sort of graphics editor, and not a notation program. MuseScore has quite a way to go before it will come close to having those capabilities.

Yes, most people will never use them, but it does give me some pause. Where do you draw the line between outlandish, experimental sheet music and copy/pasting time signatures? Because right now, MuseScore cannot do either. MuseScore can’t even let you choose whatever font you want for guitar tabs, nor can it use third-party music fonts; another ostensibly overdue enhancement that was supposed to have been added in 4.5, but has been (rather quietly) postponed.

And then there are MuseScore’s bugs, and there are a lot. This is probably what fills me with the most unease, as MuseScore currently is as buggy as ever; if anything, the bugs seem to be growing at an alarming rate.

Software is complicated, and bugs are inevitable, but it gets a little disturbing when basic functionality stops becoming a guarantee. The ⌘B saga really did a number on me — before version 4, ⌘B would make text bold when text was selected, but would otherwise add a measure to the score. From version 4 up until recently, the ubiquitous shortcut for making text bold would not work, instead always adding an extra bar to the project. 

It’s a shame to see such issues undermine MuseScore’s famously gentle learning curve. New Mac users are in for a shock when they see that the shortcut for adding measures seemingly requires a key that doesn’t exist:

It’s supposed to be the insert key… which Mac systems don’t have. Also, the “Insert at end of score…” shortcut is just wrong. It’s still ⌘B.

I can’t help but be reminded of a college professor who told her class and me quite bluntly that Adobe software like Illustrator would get a hotfix every month or two, fixing a flurry of bugs and introducing as many more, and that’s just life. Now, anytime MuseScore gets updated, I find myself on edge as to “what will break this time?” Scrolling through menus is among the latest things to go haywire.

Once more, I don’t want to write invective here, and given the circumstances, it might seem odd to spend so much time complaining about a free piece of music software. Naturally, the end of Finale has left many people feeling exposed to a certain vulnerability; the realization that a good chunk of your music career relies on a niche and proprietary program. I remember the stress of having to deal with my drumstick of choice going out of production last year, and while it eventually came back, I’ve realized how much trouble I would be in if Vic Firth discontinued the AJ5 for good. 

I do typically use a lot of open source software; my website runs on WordPress (that’s .org, not .com), and I use Inkscape and GIMP for most of my image needs (neither one of them is great, but they’re enough for what I do). I have access to Motion and Final Cut, but video isn’t a big part of my life right now.

The only proprietary software I make regular use of is Logic Pro, another old piece of software; strictly speaking, it originates from 1987 as a MIDI sequencer and, funny enough, a scorewriter (that feature is still in Logic, but that’s not why anyone would buy the program). 

The only high-end software Apple has discontinued recently was Aperture, a photo editor that met its demise in 2014. Apple claimed that the features of Aperture would morph into the Photos app, but many say that was a spurious ambition, and that Photos has yet to fully catch up, if it ever will. 

Logic is quite popular, and I don’t know to what extent Aperture was widely used in the face of other software. But I suppose one day Apple might decide it doesn’t want to spend the resources on Logic anymore. One of the prices of proprietary software is that you are isolated from development.

Open source DAWs aren’t a thing — the closest we can get is Audacity, which is great as an audio editor/file converter for quick and dirty stuff, but it has a long way to go before it resembles a modern DAW.

You can at least crack open Logic projects to find the WAV/AIFF files within. Kind of a nightmare if it’s a large recording, but at least you can get to them.

People now have decades of Finale work, thousands of project files that need to be converted to MusicXML for posterity. Todd Bishop (as part of a Finale article that was both hilariously and understandably cynical) did an interesting demonstration with a transcription of his, first as its original Finale file, then as MusicXML files in Dorico and MuseScore.

Clearly, MusicXML is imperfect, and you would need some sort of static reference (such as a PDF) to know for sure how the score is supposed to look. 

Turning PDFs of sheet music into MIDI/MusicXML is a promising but experimental field in the world of music technology, and each score that goes through such a conversion seems to produce different results. Maybe it’s another opportunity for AI to actually contribute something meaningful to the arts…

I’m sure some people have wondered if MakeMusic should have just blown Finale up and rebuilt it, perhaps reevaluating how the UI worked. It’s another calculation, and would have been similar to Apple hitting the reset button on Final Cut Pro back in 2011, which was quite a kerfuffle and turned many people off the program for good. But Apple is a big company with a lot of resources, and the world also has more of a need for video editing software than music notation software. 

As far as what’s been added in MuseScore 4.5, drum entry has been overhauled, and you now use something resembling a MIDI drum pad to enter notes:

This replaced a system that worked similarly to the pallets you would use for adding other things. The old system was a bit clunky, exacerbated by the fact that part of it broke in recent updates, and it would forget what drum you had selected as you went in and out of note entry mode.

Some new note entry modes have also been added in an attempt to make life somewhat easier for those used to Finale. I can’t say which method I prefer myself — entering notes has always been somewhat laborious for me, and I try to copy and paste whenever I can.

It’s taken me some time to learn how to use this new drum pad effectively. It’s not as customizable as I would like, and when moving in and out of note entry mode, the UI either seems to cover up the measure I’m working on or lurch the score around accordingly.

I don’t know if I just haven’t learned the best approach, or if it’s just an enduring problem with notation software. I can’t help but suspect the latter; the Finale promo video from 1988 spends most of its time showing off how you can use MIDI to enter notes efficiently, with engraving coming off as something of an afterthought. Anyone who has used Finale within the last 20 years will feel that the opposite became the case — scores would look great if you had the patience to make them look so. I found this insightful comment on an upload of the promo:

What’s most interesting to me about this 36 years later is the focus on connecting playing and scoring, trying to make it automatic and effortless… which it certainly still is not. Almost everyone has used Finale to manually write and publish scores. And dozens of companies are still trying to do what this video talks about.

Speaking of Finale videos, MuseScore product head and aspiring surreal comedian Martin Keary shared his long-awaited Finale video a few weeks ago:

Keary has become one of the most public faces of MuseScore, and it’s interesting to see some of his older videos about things like Dorico to get a sense of his design philosophy. It’s also… interesting to see how Keary has a bit of a bone to pick with GIMP. GIMP has its issues, but it does let you do things like search for fonts, which MuseScore used to be able to do, but can no longer. I said MuseScore has not regressed significantly, but it has regressed.

I haven’t talked about usability (outside of the omnipresent interference from bugs) because MuseScore is still pretty solid in this regard, but there are nevertheless some clunky elements. There’s an unclear hierarchy with styling the score’s footer and elements within the footer (like page number), while changing time signatures will split up or combine subsequent measures, which, aside from irritatingly removing your styling, may leave you surprised when bars go missing. Adding measures is easy enough, except the return key doesn’t work in the add measures dialogue box. So that’s annoying.

And, of course, if you’re on a Mac and try to open two scores at once, the program just launches itself twice.

Revisiting older scores with the new update also makes it clear that MuseScore does a lot of things that are just bizarre, especially when left to its own devices. The rests for the hi-hat pedal here are just all over the place, requiring manual adjustment:

More often than not, MuseScore will try to cram too many measures into a system, which can create a tower of accents, a frequent sight:

A lone measure on a system is often too small:

Or too big:

Many drum-centric things are missing from the program, like ghost notes and symbols for open, closed, and half-open hi-hats. Adding and styling these elements is tedious and time-consuming; I use chord symbols for these hi-hat marks, and they never seem to look quite right, no matter how I position them. While sticking text can be added, you’re better off using lyrics when you can, which are much easier to work with.

I could go on, but it’s wearisome to do so — I might just have to document myself using the software, which I have been considering anyway, to provide insight into my transcription process. 

Look, here’s the bottom line: MuseScore has a lot of problems, wether it’s bugs, UI problems, catching up to other programs, or simply how it behaves in very strange ways… but it’s still a gift to a planet that probably doesn’t deserve it, and is immune to a lot of problems that it’s competitors have — anyone who has the motivation for work on MuseScore can do so, and take it in any direction they want. 

I guess I just needed to let out a ~3000-word sigh. I am, at least, not alone. Far from it. Not only that, but it has recently come to my attention that there is some sort of schism(?)† within the MuseScore development team. I found a fork of MuseScore 4 called MuseScore 3 Evolution for “users still do not feel ready to migrate to MuseScore 4, often because of missing features [emphasis mine!] which are important for their work, or because they dislike the new UI.”

Hmm… maybe we’ll have to revisit this later.

* ^ I am ignoring MuseGroup for this article, which is a whole other beast to talk about.

† ^ “Schism” might be a totally inaccurate description of the motivation behind this fork.