Maestro: An app made of pure nostalgia and rage
A new notation app has appeared on the scene! Maestro for Windows and macOS launched on September 29th and has been generating buzz in the notation space. My curiosity was piqued, so I downloaded it and spent a couple of weeks playing around with it. This article is part review, and part speculation about its future potential. Plus, it doubles as my “complaining about Finale” screed. You’re welcome.
People frequently ask me if I’m a composer, and my answer is always the same: “Not anymore - I just like playing around with the software.” Often they think I’m joking, but I’m really not! I genuinely find a lot of joy and fascination in the process of working with notation.
It's a shame that so many composers and strangers consider their software a frustrating obstacle they must overcome to achieve their artistic goals. They’re very complex programs with a lot of moving parts and dialog boxes and interwoven features that can take years to master, so… I get it. The more time spent fiddling with settings, the less time you have for your craft. There are old-school composers who still swear by the pencil, but they are essentially committing themselves to extra effort as a deliberate artistic decision. For better or for worse, computers have vastly outcompeted our own handwriting ability, therefore functional knowledge of one has become a job requirement. Put another way, in a world with readily available word processors, there’s no reason to handwrite the final draft of your novel. You might as well embrace it!
Western music notation is a highly complex language, encompassing thousands of musical genres, and allowing for deliberate vagueness as well as extreme specificity. As such, there are so many different ways you could potentially build an interface for a notation app. The way a developer chooses to structure their user interface, what they display prominently and what they hide away behind menus and dialog boxes, tells you a lot about their perspective on music creation. You barely notice your word processor while you're using it, but your notation app will actually shape how you engage with music.
GATINEAU, QUÉBEC: Working in Finale v27.
The Newcomer
As someone who likes playing around with software, I was immediately fascinated by a surprise newcomer to the notation space that’s making a splash: Maestro. Created by Italian developer Arduino Martini, the beta launched on September 29th essentially out of the blue. It is the latest app to position itself as the spiritual successor to Finale, and it quickly overtook an obscure Russian-developed abandonware notation app from 2016, also called Maestro, in the Google search results. The name “Maestro” was chosen as a nod to the iconic default Finale font.
The website FAQ states it is created by “group of musicians who are also developers,” but it appears to be primarily Martini’s baby, with at least one other person who contributes to development and manages social media presence. Interestingly, the app is completely free, “and it will stay free” according to the website. Its creation was made possible by leveraging a few recent developments in the notation space: the Standard Music Font Layout, or SMuFL standard, introduced by Steinberg in 2013; the Finale fonts, which were made publicly available under the SIL Open Font License in 2021; the goodwill of collaborators and the community who believe in the project’s potential; and agentic AI. I don’t have any concrete proof that AI was used in its creation, but surely it’s the only way that a project with a small team and no budget could create such a high-quality website and effectively translate it into nine languages. I can’t speculate about whether AI was used to write the code since it’s not my domain of expertise, but Maestro’s interface doesn’t bear any obvious hallmarks of “vibe-coded” apps.
[Maestro 0.9.120]: Maestro has a clean look, and features the return of Simple Entry (shown) and Speedy Entry. It also adds a feature that Finale would have massively benefited from: an Inspector, docked on the right side of the screen, for fine-tuning score elements.
The price tag of $0 is great news for people looking to get into the notation space, but it begs the question of its longevity. Passion projects have a tendency to fizzle out after a while without a financial incentive to keep them going. Unlike MuseScore, Maestro is not open-source, so it risks being abandoned when the initial hype dies down and its primary contributor gets too busy. “We’re [...] considering making Maestro open source,” said Martini in a public Facebook comment from October 9th, “so the community can contribute to its development and help it grow.” That could be a promising trajectory for the app, but it remains to be seen if or when they might pull that lever. In the meantime they are accepting voluntary contributions through PayPal through an understated link hidden in their website FAQ.
I’ll talk about my experience using it, but there’s really not much point in me doing a full deep-dive into the features (or lack thereof) given that it is still being updated on a semi-daily basis. Since the launch date one month ago, the team has already powered through 120 releases. It’s brand new software that’s still in the beta phase, so it deserves some grace. Any feature that isn’t included today could be added tomorrow, and indeed, dozens of my minor gripes were fixed during the process of writing this article.
Instead, I want to focus more on the “why” and “how” of the matter. The Maestro team is very deliberately and aggressively positioning it as the successor to Finale. They invite the comparison at every opportunity: right next to the “Manual” and “Download” buttons on their website is a “From Finale” link, linking to a page exploring all the features inspired by Finale. Upon first use, Maestro presented me with a brief tutorial which prefaced several of its instructions with “as in Finale,” and the release notes are also peppered with the phrase, suggesting that early user feedback is mostly focused on the experience being as Finale-like as possible.
"Cool, another notation app"
I must admit, my first reaction was one of skepticism. Would you expect any different from the guy who wrote an article called “It’s Time to Let Finale Die”? Finale had plenty of time in the sun, but its age eventually caught up with it. With so many impressive technological advancements in competing software, particularly Dorico and MuseScore, the experience of using Finale started to feel clunky and outdated by comparison. I saw the writing on the wall in 2021, when I began to switch away to Dorico as my primary notation app.
I can appreciate that learning an entirely new notation software is an enormous undertaking, something that many busy composers and performing musicians simply don’t have time or interest for, so I have a great deal of understanding for composers who have stuck with Finale out of necessity. However, there is still a sizable amount of people who swear that Finale is still the most effective notation software in existence. These people fall into one of two categories: users who have spent decades appending a vast collection of plug-ins and third-party add-ons to the app in order to solve its myriad issues, and users who simply have no frame of reference. The reality is, an expert Finale user simply cannot keep up with an expert Sibelius or Dorico user without modifying the software beyond recognition.
Does this seem like an app that should be imitated in 2026?
The timing of Maestro’s release also strikes me as less-than-ideal. The perfect release time would have been two years ago, in the wake of the official Finale sunset announcement. By this point, a large part of the user base has already migrated to different platforms and had ample time to learn how to use them. The true ride-or-die Finale holdouts will be difficult to dislodge, with some going so far as to buy a dedicated air-gapped computer specifically for running it as long into the future as possible.
It’s also certainly a tough sell for the tranche of professionals who have stuck with Finale because they’re too busy. Maestro is just too early in its development and is missing too many features to appeal to someone who is working deadline to deadline.
As far as I can tell, that leaves a pretty narrow addressable market, composed of primarily amateur and semi-professional Finale users who are interested in updating but are dissatisfied with the available alternatives. And, I suppose, people who have a cache of Finale files but somehow missed out on the opportunity to upgrade to v27. I suspect we’ll see a plateau in user numbers pretty quickly.
Groundbreaking? Or sleight of hand?
Maestro’s marquee feature is that it can natively open Finale files. This includes the newer .musx format introduced in Finale 2014 and the older .mus format from prior versions, to at least some degree. Finale files are not straightforward plain english lines of code; they’re highly complex tangles of data meant to talk to a software that was built from layers of code spanning three decades. Writing a program from scratch that can pick apart and interpret that data effectively is an impressive technical achievement of its own, and it’s something that many people in the notation software space (myself included) thought was impossible until a few months ago.
Unfortunately, there’s a big gap between what’s “technically impressive” and what’s actually useful. There exists already a widely implemented and popular interchange format called MusicXML, specifically designed for moving music notation from one app to another. The standard is supported by over 260 applications, each of which structure their data differently, so some data is necessarily lost during the conversion process. MusicXML 4.0 was introduced in 2021 and remains a very robust standard that captures things like chord symbols, lyrics, tablature in most situations, and even percussion notation and some linked parts layout data in others.
It is trivially easy to convert a Finale file to MusicXML through Finale itself, or if you don’t have access to a copy of Finale, via Robert Patterson’s Denigma Online tool. So, in order for Maestro’s headline feature to be useful, it must be able to interpret Finale’s .musx files to a higher degree of accuracy than a MusicXML conversion.
Let’s put Maestro to the test. I have a piano work written in Finale 2014, engraved at an amateur level so that it’s reasonably representative of an average Finale project. For comparison’s sake, I have opened in Maestro 0.9.114 both directly from the .musx Finale file labelled sample “A,” as well as a copy imported via MusicXML conversion, which is labelled sample “B.”
There is functionally very little difference between samples A & B. Both were able to successfully capture forced beaming choices, forced stem directions, voicings and cross-staff notes, while neither was able to render the unusual alternating time signature nor capture the original horizontal spacing at all.
The most notable difference is that for sample A, Maestro was able to swap out the old Finale Maestro font for its newer SMuFL iteration, but was for some reason not able to make the same substitution for B. Most of the visual differences can be attributed to the B being shifted to the SMuFL default font, Bravura.
Taking it a step further, here is a sample from a more complex woodwind quintet piece, converted from Finale 27 to Maestro 0.9.114:
In this case, the MusicXML conversion arguably fares better. In sample B, Maestro correctly established the “keyless” or “atonal” key signature, but in sample A it was mistaken for C major, resulting in incorrect key signatures in the clarinet and horn parts. Sample B also correctly captured the line objects from Finale as trills (though the specific accidental was lost), while sample A interpreted them as a line of gibberish. In both of my trials, courtesy accidentals only appeared in sample B and were absent from A. On the other hand, Maestro was able to extract some layout decisions from the Finale file for sample A, including fonts selected in Finale’s Category Designer, locked staff systems, headers, and staff labels (or lack thereof, in this case).
The bottom line: In practice, being able to directly open Finale files feels less like a game-changing new development and more like a neat parlor trick. The conversion is still so rudimentary that it does not appreciably outperform a MusicXML conversion. It is likely to continually improve as the software is developed, but I don’t believe it will ever be able to interpret them flawlessly. It affords the user some convenience, but it is a convenience on par with receiving a bagel slicer while you already own a bread knife.
The Finale Trap
Even if Maestro could interpret Finale files flawlessly, it could be a double-edged sword. It would require the program to recreate most of Finale’s functionality and UI from the ground up, including many of the obsolete and half-baked features. By building a copycat, Maestro risks falling into the same design traps that Finale did. At least the developers of Finale had the excuse of working within the constraints of older tech and different prevailing UI design philosophies.
Take for instance the huge array of lines with different purposes and behaviours that are all lumped together as “Smart Shapes.” The Smart Shapes feature was cutting-edge when it was introduced in the 90s, but it has barely seen a meaningful update since. Broadway copyist Steven M. Alper said he always wished “that some of the Smart Shapes were smarter,” which is quite an understatement. To me, this was the Achilles’ heel of Finale: The Smart Shapes algorithm was too unsophisticated.
To show you what I mean, I want to look at how Finale handled hairpins, a process which was actually a series of cascading issues hard-wired into its ancient codebase. Elements of the score like notes, clefs, time signatures, dynamics, hairpins, and lyrics were all handled through different unrelated processes, which means Finale was unique among modern notation apps in that you couldn’t just select some measures and filter for what you needed. To actually select a hairpin for adjustment, you had to first go to the Smart Shapes tool to enable handles on hairpins, then click and drag a box around them to select them. You had to be creative to avoid selecting handles for unrelated Smart Shapes including slurs, trill lines, glissandi, and pedal lines.
[Finale v27]: Good luck untangling this.
This had an interesting knock-on effect: because single-clicking and dragging was already taken, the process to actually create a hairpin required you to double-click, hold down the second click, and drag, a very unusual UI interaction and major issue for discoverability (I recommend this hilarious segment of Tantacrul’s Finale video for more details on that point). Furthermore, the hairpin drew its X and Y values from the arbitrary point you double-clicked on the page. These values were entirely junk information, as evidenced by the above screenshot which is laden with collisions. Hairpins should have been automatically positioned relative to the bottom staff line, the notes below it, and the surrounding dynamics. This information was clearly available, because it’s positioning data that Finale’s Expressions drew from, and I’ll never understand why this issue was never addressed. This alone resulted in my two biggest time-sinks in Finale: having to find and fix collisions, and adjusting and aligning dynamics, two things that every other notation app gets right by default. I even wrote an article for the Finale blog back in 2021 detailing the lengths I had to go to tackle this issue.
I already see the echoes of these issues in Maestro. For starters, deleting elements is surprisingly tricky; some items will go without a fight, simply by pressing “Delete” on your keyboard. But several others will only be removed if you hunt and peck through the Inspector for a “Delete” button. There were even a few elements I wasn’t able to delete at all, such as glissandi, the initial tempo marking at the beginning of the document, and… any time signature or special barline. To delete a time signature, you have to actually overwrite it with a prior time signature.
Maestro appears to have copied the Finale Smart Shapes functionality beat-for-beat, which is such a blunder that I’ve started calling it “Dumb Shapes.” They mercifully abandoned the “double-click, hold the second click, then drag” interaction, but the Lines tool still falls short in many of the same ways. Currently you need to hold Shift and carefully select a note at the beginning point and end point of the line you want to draw, then click on the slur or cresc./dim. buttons to draw it. It’s just too much clicking, plain and simple, and they have only just implemented a way to draw a hairpin on a single note. Slurs are not really functional yet, drawn by default too flat in a ridiculous aspect ratio-squeezed manner, and the handles don’t offer the granular control needed for manually editing them effectively.
This score was posted as a sample on Maestro’s website, created in a very early beta version. The bare bones are there, but there are issues including the slur arcs, ledger lines, horizontal spacing and consistency of fonts.
Collision detection is still highly questionable; in some cases, adding or removing elements like dynamics, double barlines, and Dumb Shapes didn’t seem to impact the layout at all, causing some of the nastiest collisions I’ve ever seen.
[Maestro 0.9.120]: A 4-voice chorale that was originally created in Finale v25 and imported into Maestro; adding a crescendo (highlighted) resulted in the hairpin bisecting the staff below.
Maestro’s layout functionality is clearly an afterthought, taking second priority after the development of note input functionality. Still, to its credit, global layout tools actually have some advantages over Finale. Granular layout adjustments are still not possible, but it has a kind-of-serviceable “Automatic Layout” button. You can also click “Return to Maestro’s Layout,” which does the same thing but also resets the staff size to Maestro’s quite large but acceptable default of 7.3mm.
[Maestro 0.9.120]: The same passage as the previous example, but with “Return to Maestro’s Layout” selected. It benefits significantly from the additional horizontal space, but kerning is still uneven, particularly noticeable in the third measure.
The Automatic Layout functionality actually overcorrects in a way that reveals how unsophisticated it is. In the following example, it has produced an enormous space between the staves. You can easily deduce how it makes its calculation: it spaces them far enough that the lowest object on the upper staff and the highest object on the upper staff couldn’t possibly collide.
[Maestro 0.9.120]: I’ve added dashed lines to better highlight how Maestro’s crude vertical spacing algorithm works.
I won’t harp on layout further since it seems like the developers are playing catch-up. It didn’t even have icons in the first version I tried (0.9.105).
In other news, they decided to reboot Finale’s tool-based UI. They’ve added keyboard shortcuts to access them, which seems like an upgrade, until you realize it’s actually a missed opportunity: there are 16 tools in the left panel, but only 9 shortcuts to access them. The nine shortcuts include useful tools like selection, note input, and articulations, but also less-useful ones like key signatures (which you might use only once or twice in a project) and the measure tool (which I believe doesn’t actually do anything yet). Dynamics, lines, and tuplets don’t get a shortcut. (I don’t think the tuplets tool works yet either.)
To end this segment on a positive note, the last thing to mention is how they’ve faithfully recreated Simple Entry and Speedy Entry. I really enjoyed using Maestro’s take on these without Finale’s screen tearing visual glitches. They still need a little finesse (please keep the input caret active after I press Ctrl+Z!), and Finale’s “text editor” approach to beat positioning is dated, but I still think Finale’s note input was one of its primary strengths. If they’re going to emulate Finale, this is a perfectly fine thing to keep.
The future of Maestro
I feel like I have done a disservice to the developers by showcasing primarily Maestro's shortcomings and glossing over its strengths. There is genuine potential here and Martini is clearly a very intelligent and driven person. The fact he was able to recreate so many beloved features from Finale and crack the .musx enigma is extremely impressive. The app looks great, it’s lightweight, and it’s mostly pretty fun to use.
It's important to remember that I am approaching Maestro from the perspective of a notation software power user. I’ve spent a lot of time learning the ins and outs of Finale, Sibelius, and Dorico, and I've learned that they're at their best when you approach them on their terms rather than imposing your expectations on them. Maestro is a well designed app, but I think the entire premise is misguided. It’s great that they’re building a new notation app, but they should develop a vision for what they want their notation software to be instead of trying to resurrect an old software without addressing its flaws. Maestro feels like an app made of pure nostalgia and rage, destined to fizzle out when the developers realize the enormity of the task they've set for themselves.
I’d like to see Maestro go open-source at some point in the future, since I think it’s the clearest path to improvement assuming it remains free to use indefinitely. The app has a long, long way to go, and even just a few code-savvy people who believe in its mission could make a big difference. In any case, I encourage you to try it out and see for yourself, especially if you’re a “Finale Refugee.”