(Do you like my silly little featured image? I’ve realized that, for a topic like this, I need something more unique than just a screenshot of the software.)

At long last, the MuseScore 4 update was unveiled to the world this past December. Originally slated for release back in March of 2022, MuseScore 4 finally saw the light of day just before the year ended. Let me share some thoughts on the update before this topic gets away from me.

MuseScore is my notation software of choice, and it has been since I started using it back in 2013. I’ve never seriously considered switching to another software — the big “premium” competitors Sibelius, Finale, and Dorico are all very expensive, and I hated using them in college. I’m vaguely aware of other software like Encore and Forte, but I haven’t spent any time with them.

As far as free software goes, the only other option I knew about circa nine years ago was LilyPond. LilyPond is kind of funky — its UI style is called “what you see is what you mean”, and it works like a markup language, so you create music by entering text which is then spat out into something like a PDF (LilyPond very similar to LaTeX if you ever had to use that in school as I did).

As a high schooler with no coding experience, that was all a bit intimidating (although I’ve since learned that there exist GUI extensions for LilyPond). MuseScore on the other hand is not only free but has long had a reputation for being very easy to learn.

Still, MuseScore has long felt like it was playing catchup to some of the other programs out there, so there were big ambitions for version 4. To start, the app has gotten a bit of a facelift, and the older versions of MuseScore look downright ugly in comparison. A couple of components have been redesigned — setting up new projects is more intuitive, while the infamous “parts” system got a total overhaul (this is what you use to split a conductor’s score into sheet music for individual instruments). 

Some UI redundancies have been reduced, while the two biggest changes are perhaps a new engraving engine to make things look nicer, and a new playback engine that allows for the use of VSTs to make playback sound more realistic and lifelike. You can watch this video to learn more about some of the more substantial upgrades.

Let’s move on to some new features. Writing this stuff in prose is kind of tedious so I’ll just use bullet points.

• Tempo changes (ritardando and accelerando) now affect playback and are highly customizable. 
• After eleven long years, MuseScore now has built–in support for working double and quadruple repeats. Remember what I said about playing catchup?
• Two new notation styles have also been added: Finale Maestro and Finale Broadway. 
• Notation fonts overall also just work better — previously, MuseJazz wouldn’t use proper repeat symbols. You had to drag them from the symbols pallet and line them up over the wrong ones. That’s been fixed and will be a big timesaver for me. 
• You now have the option to automatically number a string of repeats. Unfortunately, there’s no way to change the way the numbers look, and they don’t look too good. So I won’t be making use of this yet (compare it to me manually adding in the numbers with some less aggressive text):
• There’s a new button to add parentheses to noteheads — I no longer have to use keyboard shortcuts to create ghost notes. (although I wish there was a ghost note option built in to make the notes quieter — I don’t bother doing it manually).
• Smaller noteheads are now also easier to add as well, as are cues. 
• Instrument changes now work much better, and seem to work fine even for switching back and forth from a percussion clef. No more kludgy workarounds for cues in my drum scores.

Now for the disappointing news: there are a couple of “advanced” features that haven’t made it into MuseScore 4 yet. One example is over–the–barline tuplets:

This is mainly a drum crops thing, but every amateur drumline composer uses MuseScore, so I reckon this is an important feature. You can still fake it by combining two measures, creating an offbeat tuplet in the right spot, and then manually adding in a barline. It works but will get extremely tedious if you have a big score with a lot of instruments. 

Alternating meter is still a bit limited. The issue is twofold. First, notation. There are basically two ways to do it:

Neither of the options look good in MuseScore:

For the example on the left, the numbers are too close together — it looks like 47/48 instead of 4/4 7/8 (and no, you can’t add spaces to time signatures). Meanwhile, the example on the right should only have one plus sign. As is, it again looks like one giant time signature instead of two connected ones. 

Regardless, we come to the second (bigger) problem: there’s also no effect on note entry and measure duration, so no matter how the time signature looks, you’d still have to make one big measure of 15/8, disguise it as an alternating meter, and add the barlines manually. Or just use 15/8 and hope people can read it. 

Speaking of niche time signatures, there’s no native support for irrational meters like 5/12. 

Now, I’m not terribly surprised, since irrational/tuplet time signatures are a complicated and experimental songwriting technique that most people would be better off not knowing about. Still, it could be a unique selling point.

Per usual, with enough patience, you can fake these. It’s actually gotten easier since you can now add whatever numbers you want to a time signature without changing the actual meter. So just create a measure of 5/8, set it to display as 5/12, and then add an invisible tempo change after doing some math. Or just use a metric modulation like a normal person. 

Speaking of metric modulations (also known as tempo modulations for all you drummers out there who have never read a book), MuseScore has had basic ones for a while, but there is still no support for modulations that use tuplets or tied notes. Again, you can fake them with some custom text, a bit of math, and manual tempo changes. 

You have to get used to manual tempo changes anyway since metric modulations are busted now — the tempo always changes to 120 BPM no matter what. 

There are a couple more specific things that haven’t improved for me yet. Here’s one: you can’t change the thickness of beams. Turning notes into cues with the handy new “cues” button reduces the size of the noteheads, stems, flags… pretty much everything except for the beams, which are the same. It just doesn’t look right to me:

One last nitpick: I still can’t seem to force tuplet brackets to stay horizontal the way you can with beams, and I go crazy trying to figure out if these things are crooked or not. 

Ok, here’s the real final nitpick: when exporting PDFs, MuseScore won’t remember your previous destination. If you want the PDF to go to the desktop, you have to manually set it to desktop every time. 

I think it’s kind of funny that MuseScore still has this problem since head designer Martin Keary criticized Sibelius for doing this. There is one improvement, however: if you’ve used an older version of MuseScore, you’ll remember how spaces in the title of your score were turned into underscores in the file name.

They finally fixed that bullshit. 

Well, that’s enough about features. The big question is how my workflow has held up with the changes.

It’s still the same basic MuseScore, although the new properties box seems to be slowing me down a bit. Notice how there are five buttons and tabs to click on during the search for what I want, in addition to a “show more” menu you can unfurl at the bottom:

As I said last week, I decided to do something drum transcribing to test things out, and something felt… off in MuseScore 4.

Here’s what’s up: in older versions, MuseScore would remember what drum I have selected when I go in and out of “note entry mode”.

My transcription process is usually to do things piecemeal — figure the hands out and then move on to the feet. I go in and out of note entry mode to fix subdivisions before adding the notes. Now, I’m constantly having to reselect the drum I want as part of this process.

It’s… annoying to say the least, and I can’t tell if it’s intentional or just a bug. I suspect the latter since MuseScore 4 still remembers what subdivisions I had selected (now with articulations).

I still feel like there could be smarter ways of putting in notes, but I don’t have the energy to expand upon that since this article is long enough as is. Many of my issues probably come from being a drummer; I can’t just click on the staff where I want notes to go (Finale tries to make this work for drums, and it’s a catastrophe). 

For all the touted improvements to engraving, MuseScore 4 will still do things I can only describe as harebrained. Look at the accents at the end of this line:

Disgusting. 

Check out this repeat symbol, which looks way too small:

And how about single measures that randomly get bumped to the next system? You bet:

Finally, we come to the last topic — the glitches.

I won’t add much here, since the forums and Reddit are full of bug reports as is. Here’re a couple of my favorites. 

• Multiple projects open at once won’t open new tabs, but whole new instances of the software.
• The shortcuts for changing the width of a measure are now broken.
• When you click on a file from Finder, it will only launch MuseScore. You have to open MuseScore first, then find the project from your file browser and click on it there to actually open it.
• Double–clicking an element and trying to position it with the arrow keys does nothing now.
• Autosave does not reliably work. I learned this the hard way last week…

So to conclude, this is a mixed bag. Some of the new features are welcomed improvements, overdue as they may be. It has fallen short for me in some regards, and the obvious regressions in functionality are very disappointing. It’s pretty clear most of the resources went into the playback engine, which a lot of people are excited about. Me… not so much. 

First and foremost, I need MuseScore to be a notation app; I could care less if it sounds good. If I really cared that much, I would just throw the MIDI into Logic. But a lot of people don’t have Logic, so their gratitude is understandable. 

Nevertheless, the playback engine cannot come at the expense of more important features, especially ones that were working fine in the previous version (e.g. autosave). While some glitches from version 3 have been ironed out, there’s a litany of new ones — while it was way over schedule, the general consensus online is that MuseScore 4 was not ready for primetime yet. 

I don’t think this was ready for the world. There were no updates for nine months, with the developers blowing through their original deadline of March 2022 to no fanfare. Some honest and routine updates about the state of the software would be nice, more so than going dark only to deliver a surprise debut that’s clearly not ready. 

Let me be clear: I’m deeply grateful that MuseScore exists, but all in all, I’m not entirely sure if I’ll be using MuseScore 4 full–time. The workflow has basically stagnated with improvements and regressions; the bugs are very irritating; and I still feel like I have to make a lot of manual adjustments for my score to look good (e.g. repositioning flams, which are still too far away from primary notes, and often collide with beams). But hey, at least we got double repeats now, so we’ll see I guess…