Komp is a beautiful and ambitious new scoring app for iPad

Reviews

The iPad Pro is my favorite computer I’ve ever owned. Nothing else is really even close. Along with the Apple Pencil, it has allowed me to remove paper from my workflows in ways I had never truly imagined. It has even allowed me to get work done in more flexible, intuitive, and even delightful ways.

However, one sticking point for me and many other musicians has been a lack of notation and scoring software for iOS. It seems like such a natural fit for the immediacy and accuracy of writing with the Pencil on iPad. I know I’m not the only one who has looked with admiration and a hint of jealousy at Windows-only StaffPad.

That’s why I was keen to try Komp, a new app from Semitone, which launched as v. 1.0 several days ago on Friday, April 28. This blog previewed Komp in January at the NAMM 2017 show, where the app was being demonstrated in a pre-release state.

With the app now available to the public, it’s time to see where it lines up in the world of iOS scoring apps. Komp isn’t the most powerful in the category, but it might be the prettiest.

Komp’s main editing interface, as it looks on my 12.9″ iPad Pro

The basics

On first launch, Komp greets the user by giving a quick video demo. Unlike many similar videos, this one is short and succinct, allowing you to get right into using the software. Like many other writing apps, Komp also asks if you’re using an Apple Pencil so that it can optimize the speed and clarity of writing. There are a handful of example scores to look at, play back, and edit. Some are complex and some are simple. Creating your own score is as easy as filling in a few broad details in a New Score wizard that will feel familiar to most users of any desktop scoring application.

After completing the setup, you’ll find a blank page with very few tools in front of you. The tool palettes are clean, minimal, and unfussy, allowing you to focus on the score itself. You can start writing here like you might expect.

This is where Komp next distinguishes itself from other apps in this category. Unlike NotateMe, MusicJot, and even StaffPad, Komp will start interpreting handwriting almost instantly, rather than waiting for you to complete a measure.

Komp interprets each new symbol as you write, rather than waiting for the complete measure

This can be both good and bad. If you write slowly, Komp may process your shapes before you’re done. Or if you tend to draw in a different order than Komp expects, it might get a little confused. The writing itself is very fast and accurate. As users of Apple Pencil or Microsoft’s Surface Pen know, not every application handles this gracefully. Komp is as smooth as my favorite handwriting apps on iPad.

In my testing, I found the notation interpreter to do a pretty good job most of the time; but, there were enough misses that I found myself getting frustrated at times. There are also a few symbols that the handwriting recognition can’t handle, such as a sixteenth rest or a triplet. These can be entered — more on that in a moment — but not using the handwriting interpreter.

It’s understandable that some symbols are hard to figure out, since the first step in drawing a sixteenth rest is to first draw an eighth rest. But the handwriting feels so good when it works, it’s a shame to have to pull up short when it doesn’t, and it’s easy to forget which symbols will or won’t work. Sixteenth notes: yes; sixteenth rests: no. Slurs: yes; hairpins: no. Rhythm dots: yes; staccato dots: no. In fact, Komp’s speedy handwriting conversion often caused me to inadvertently add rhythm dots all over my scores just from tapping around to make selections or scroll. A bit annoying, but these are easy to delete by tapping to select (or circling with the Pencil) and tapping the trash icon.

The rendered notation is another thing that distinguishes Komp from its competitors on iOS: it looks really good. Komp uses the Bravura music font from Daniel Spreadbury at Steinberg. This is a positive sign for the future of both SMuFL’s future as a music font standard and the future development of Komp’s engraving options. In addition to the standards-based approach to presentation, Komp relies on MusicXML for its scores. More on that later.

Training the robot

Komp’s creators understand that everyone has their own unique way of drawing a quarter rest. And rather than forcing everyone to learn a standard set of notation-like glyphs, as in the old Palm Pilot Graffiti system, Komp invites users to train the app on however they write basic notes, rests, and accidentals. By writing sixteen quarter rests, you can teach Komp your own process. I found that after doing this with a handful of common symbols, the quality of the recognition improved noticeably. It was still not perfect, but much better than the untrained interpreter.

Training steps
Komp asks you to give training examples based on how you write notes and rests, then shows how well the training worked

Beyond handwriting

A much larger set of symbols can be entered and edited using the tools along the right side of the page.

First is the very stylish and clever radial menu. This circular, floating palette gives access to all the basic note and rest values, accidentals, and more. It can be easily moved around the screen and hidden entirely. Sitting on top of the score with a slight drop shadow, it has a distinct, tactile feel that I really appreciate. I almost wish the other tools, such as articulations, dynamics, and ornaments, were included in the radial menu instead of in their own options along the right panel.

Komp's radial menu
Komp’s radial menu can float around the screen

One really intriguing side-effect of Komp’s more immediate handwriting interpreter is that it’s difficult for Komp to apply a lot of control to your writing. You can write a stem in the wrong direction, and most of the time it won’t be corrected. You can even write measures with the wrong amount of rhythm. Even the demo video shown on first launch has incorrect rhythms.

Again, this is both good and bad. Good for free-wheeling composers like me who are looking for an app that supports open-ended sketching, where meters can be fluid. But it could be bad for conforming to notational norms. I can add extra notes in the violin part, but leave them out of the viola part in the same measure. This makes playback and export potentially an adventure.

Komp wrong rhythm screenshot
Komp seems unperturbed by the extra notes in the violin part

An important thing to understand about writing rhythms and adding other symbols in Komp is that the app is, to put it simply, dumb. By that, I mean that it doesn’t seem to know much about what music should look like. If you’re an experienced and careful musician, this is a boon to creative freedom. You can write even weird rhythms like the one above, trusting your own expertise. However, if you’re less careful, you might find yourself writing nonsensical rhythms and stacking two or three accent marks on the same note by accident.

Komp extra articulations
Komp doesn’t seem to know that it’s weird to have duplicate articulations on the same note

Another side-effect of Komp’s open-ended proclivities is the reflowing of rhythms. As you write fast or complex rhythms, you might find yourself running out of space in the measure, or simply crowding your notes uncomfortably close. Komp has a manual reflow button that is always visible at the top to set rhythm spacing. This is simplified a bit by the fact that there are no system breaks. Komp is always in a one-system scroll view.

The user interface is strikingly modern, as clean as any you’ll find in the App Store, including Apple’s own apps. This really distinguishes it from some of the garish and juvenile designs of some of the other iOS notation apps. It might seem shallow of me to say this, but that’s very important to me. I wouldn’t want to spend a lot of time working in an application that I find ugly. Komp has a user interface and navigation that strikes a good balance between power and flexibility. The typography, colors, and general design language of Komp are a genuine pleasure to work with.

Export

As I mentioned earlier, Komp’s developers opted to adhere to the MusicXML standard for score files. As MusicXML continues to evolve, it’s great to see new applications built around it from the beginning. This helps to avoid platform lock-in, and may appeal to users who aren’t yet prepared to make a long-term commitment to a new app publisher like Semitone.

I tested the MusicXML export by sending out my strange string orchestra score to Dorico and Sibelius. It was a deliberately weird document with a wrong rhythm in the violin part and a bass part written in treble clef.

Komp to Sibelius import
This is what the above score looks like imported using the Sibelius defaults. Sibelius cuts out the extra notes that don’t fit in the measure
Komp to Dorico import
This is what the same score looks like imported using the Dorico defaults. Dorico’s more liberal approach to meter keeps the extra notes and even seems to be using competing meters with the violin 1’s switch to 6/8

As you can see, the results of both the strange rhythm and the dynamics were mixed; but, because each difference does represent something from the original Komp project, I think the differences probably say more about the MusicXML importers than the exporter.

At the moment, there is not a way to take a Komp project to PDF, and that’s okay with me. One thing that the app doesn’t deal with at all is line breaks and layout. For my use cases, I think the team made the correct decision to focus more on the handwriting and music notation than on the document layout features.

Limitations

In its current state, Komp represents a remarkable achievement. Semitone has created a tool that many musicians will be happy to use. This is still a 1.0 release that has its feature limitations and even some bugs. Here is a non-comprehensive list of some things that jumped out to me most:

  • Scrolling through larger scores can be slow and jittery.
  • There are occasional stuck notes if you flip to another app during playback.
  • No iPad split view support, so no ability to look at a PDF score or song text in another app while writing in Komp.
  • No way to rename an instrument, even to add a number like “Violin 1”, “Violin 2”.
  • Copying doesn’t seem to know about clefs. Note the measure above that I copied from the bass clef cello part to the alto clef viola part. Komp keeps lines and spaces, not notes.
  • All instruments are shown at sounding pitch with no way of switching to transposed pitch.
  • No way to write arbitrary annotations on the score. This is one that I think is a perfect companion for an app that interprets handwriting. I’d love to be able to mark up sketches.

Many of these become less important when I think about how I would actually use a tool like Komp. I imagine myself using Komp as a mobile notepad that allows me to import my work to a desktop scoring app when things start to get serious. That means most of these issues could go away pretty quickly after that transition.

Semitone is aware of many of the larger missing features and bugs, and they seem very responsive to users. Shortly before Komp’s launch, they published a feature roadmap which lists some upcoming additions and the order in which developers expect to approach each of them. They’ve even given an indication of how long they expect each of these additions will take, along with a place to submit your own suggestions. While I find the limitations of Komp 1.0 to be frustrating, and in some cases dealbreaking, I am impressed with the openness and ambition expressed by the roadmap.

Pricing

Like many new apps for mobile and desktop, Komp is sold as a subscription: $5/mo. or $50/yr. Each of these is available with a free trial, and if you’re just interested in seeing how the tools work for you, you can create scores of less than sixteen measures without starting a subscription at all.

I know a lot of users are not comfortable with subscriptions, and I understand why. However, I think it of subscriptions like this as a mutual agreement between users and developers: I’ll continue paying for your software as long as you continue maintaining and improving it. With Adobe (Photoshop) or Avid (Sibelius) subscriptions, there’s the user-hostile threat of file-format lock-in. Semitone avoids that ransom model by not using proprietary file formats. Since everything is in the open MusicXML standard, you’re free to go when you like.

Conclusion

After having spent time using Komp for several weeks, starting during its beta period, I can’t say that this is an app that will replace my desktop scoring applications in any foreseeable timeframe. However, I can certainly see its utility as a mobile option when I’m not at my Mac. I’m impressed with the writing tools, and I could learn to live with their imperfections while also trusting that those imperfections would be ironed out over time.

I’ve spent some time with NotateMe and MusicJot. Both are nice and can even do several things that Komp cannot. However, NotateMe and MusicJot feel like tech demos to me in their design and the quality of the resulting score. Komp is the first mobile scoring app that feels like a modern app of the same caliber as the other apps I use on my iPad. It’s not the app for me or for now, but I can imagine it being useful to many people — and I will certainly be keeping tabs on its continuing development.

David MacDonald is a composer and educator based in Orlando, Florida. He teaches music composition at the University of Central Florida.


Comments

  1. Waldbaer

    Interesting review, thanks. I think the main deal-breaker for me is the subscription-only concept, even if the file format isn’t proprietary as you point out. I still think, that the old way of paying for an app and then for noticeable upgrades is better for the users: It leaves more freedom to them and in many cases it gets out cheaper. And by the way: Being bound to a subscription, I as a user have to pay just to use the software – no matter if it still gets developed or not. That is the main thing that feels strange to me… they at least should leave us the choice.

    For the moment I’ll keep with my Surface/Staffpad and desktop scoring apps nevertheless, it’s interesting to see the Apple platform developing, too, though (I come from and still prefer macOS a lot over windows). Using Staffpad I often wish it would convert directly. If a bar is full (let’s say a longer line of eights, not to say adding articulation and stuff), chances are quite big that it won’t convert everything correctly at once. And sometimes it even reads notes wrong (e.g. an f instead of a g) and I don’t notice because I think it converted everything successfully, so I assume it to be correct…

    1. David MacDonald

      I totally agree about the nasty surprises that can come from the “batch” conversions in other apps. I’ve had the same issue you describe. Komp still gets notes wrong; but, I can quickly grab the note and drag it up or down an instant later.

      Regarding subscriptions, I have come around to embrace them, as I wrote above. Mobile platforms in particular are changing quickly, and that means app developers have to continue development after your initial purchase. I think it’s fair to support them during this time. The software sales model we have is a vestige of boxed software sales, and I don’t think it makes sense to apply it universally anymore, especially for such complex applications as Komp. It’s particularly unsustainable for small, indie developers. And even with big publishers, I’m happy to pay for what I use. I love my Adobe CC subscription, and I wouldn’t go back to the box-copy if it was available. I get continuing value.

  2. George Grella

    I will be trying this out, but I’m also wary of the subscription model.

    It seems that you’re not familiar with Notion for the iPad, which also supports Apple Pencil, and for desktop Notion users it syncs via iCloud. If you remember the iPad commercial with Esa-Pekka Salonen, he was using Notion. From your review it appears Notion does some things better and some things worse, which is pretty standard, but it does export to XML and is cheaper – $14.98 for basics and handwriting recognition, though you can buy more instrument packs.

    1. David MacDonald

      Thanks for the comment George. I didn’t mention Notion in the review simply because I don’t have as much firsthand experience with it. I know others use it. Part of me doubts that it is truly EPK’s daily driver, but I know it’s got a lot of useful features, not least of which is its connection to the sibling desktop app. Regarding its handwriting recognition part: this actually doesn’t differ from NotateMe or MusicJot. All three of them use the same off-the-shelf handwriting technology licensed from MyScript. It’s my understanding (which may be wrong!) that Komp has its own tech for this, which is why it can have the realtime interpreter and the handwriting trainer.

      To me, this is a pretty fundamental difference that makes Komp a better long-term investment. I agree, though, there are lots of things that Notion and the other apps do that Komp does not yet have. I think the smarter handwriting along with the MXML export is the one that suits my needs better, as I don’t think any of the iPad apps are even 10% of the way to replacing Sibelius/Dorico in my work.

  3. John Blackburn

    I tried Komp with great hopes, but found it more of a technology preview, albeit a polished one, than an app I’d want to invest in. Accurate data entry with minimal cleanup is crucial for a good workflow, but Komp was misinterpreting the simplest of gestures with such frequency that I was forced to correct its errors or puzzle over why my input wasn’t being understood in the first place, forced to concentrate on the software instead of my music—an intolerable state for any music notation program.

    There is much good about Komp, but it doesn’t yet deliver on the promise of touch-inputted music and judging by this version won’t do so for a long, long while.

    1. David MacDonald

      That’s interesting John. Would you mind sharing what hardware you were using? I had pretty good results with my Apple Pencil, especially after training it on my writing. The biggest misfires I saw were Komp putting a note one line or space away, which is pretty quick to fix (just tap and drag).

      If it’s not working for you, that’s totally understandable. I wouldn’t encourage anyone to invest in software or hardware for what it _might_ do later after an update.

      1. John Blackburn

        iPad Pro 12.9″ with Apple Pencil.
        • Simpler input like a quarter-note on G above middle C worked fine, but trying to enter even a quarter-note on middle C with a ledger line worked less well, the ledger line being alternately misinterpreted as an eighth-note flag, a tenuto marking, a second note resulting in a D+E chord, or not interpreted at all.
        • Drawing a quarter-note than an intended flag to convert it into an eighth-note worked sometimes, but sometimes the flag resulted in a second note chorded with the first one.
        • Often I would write several noteheads in a row then attempt to stem-and-beam them into a running eighth-note figure, only to see my noteheads disappear entirely after failing to be parsed properly. This was particularly vexing, since at minimum my input should have been preserved for any needed alterations.
        • Enter four notes, then try to add a sharp (#) to the third note. Half the time your gesture will fail, the other half the # will appear as expected but without the notes being spaced as needed, leaving the # overlapping the previous notehead.

        The graphic presentation of the music is also problematic:
        • flags extend clear up to the noteheads
        • stem direction is sometimes wrong

        If I were notating Mary Mary Little Lamb, I could certainly get it done with Komp, albeit with, I’ll bet, considerably more futzing than desired, but certainly more complicated stuff would exhaust my patience.

        To be clear, I don’t necessarily fault the developers, who appear to have done admirable work, but rather the medium itself, a touch interface offering great challenges as well as great promise. The promise gets great press but that press often gives the concomitant challenges only cursory mention and frames those challenges as something that perhaps version 2 might resolve, when the reality is that version 20 might well still struggle to resolve them. In other words, music notation is a real bear, a dense thicket of precise bits of ink whose placement is crucial to their meaning, so details matter—*everything* matters.

        There’s enough complexity in any reasonably complex piece of music notation to confound any touch parser, making a graceful recovery absolutely essential and ultimately calling into the question the viability of a touch approach at all. I was hoping Komp would have better solutions and perhaps some day it will, but that day seems a long way off.

        1. David MacDonald

          Right. Like I wrote, it’s not going to replace a desktop app anytime soon, and probably never for anything I write.

          There are certainly many flaws in Komp. I only pointed out the most egregious ones for my own uses; but, there are certainly many more. The stem direction thing is really puzzling. In earlier betas of Komp, it seemed to take whichever stem direction I drew, even if it was an upstem on the top line of the staff. The release version does more correction in that regard, but it’s still a long way from reliable. And yes, ledger lines are very annoying to deal with. I didn’t experience as many issues with accidentals as you describe though.

          I disagree that writing interfaces are doomed to fail for the foreseeable future. I can imagine that this kind of writing could be much more consistent in the near term.

          In the meanwhile, I think it would be great to see some more touch interfaces that aren’t writing based. While mouse entry of notes on desktop notation apps is terrible and tedious. All pros switch quickly to keyboard or MIDI input. I think some kind of tap-to-enter scoring system could actually be pretty fast, perhaps even faster than keyboard shortcuts. Imagine having one hand on a menu to select rhythms and accidentals and another holding a stylus above the staff to enter notes. I think the immediacy of that could compete with the speed of quick keyboard entry.

          I think your conclusion is probably the trickiest part: notation is just hard. The Steinberg team spent three years with some of the best and most experienced developers in the field and delivered a product that still has some users left behind. We all wait. :-)

          1. John Blackburn

            I agree that non-writing based input methods continue to offer promise. Keyboards are themselves complex two-handed devices, so composers can handle great complexity in input. In fact, a physical auxiliary device might well make a huge difference for Komp, freeing it from some of the need to divine the user’s intent.

            I disagree with you on the Dorico team, though. They’ve done a fantastic job with what they’ve implemented thus far and have been very clean on what remains to be implemented. I’ve read previous few complaints about the quality of what they’ve done; virtually all of the complaints seem to be about what the product doesn’t yet offer, which at best is a marketing fault, not a technical one.

            Komp too lacks many things, none of which I chose to comment on though they’re certainly important to anyone deciding to subscribe. Rather, it’s the features that Komp includes but which work inadequately that make it unsuitable for real work.

            Dorico, I’d say, under-promises and over-delivers. The rigor of their considered implementation has thus far impressed that what the product does, it does well, and the product’s solid foundation makes it clear that missing features appearing over time will be equally strong.

            I’ll continue to put it through its paces, but already Komp, in contrast, over-promises right from the start by failing to deliver on its promise that writing your score into the program will prove effective, where “effective” is the key word, requiring enough benefit to counterbalance its inevitable limitations.

            Komp’s competition as with all mobile music notation software is not really Dorico or Sibelius, against which they’ll likely compare poorly, but rather the simple mobility of pencil and paper itself. I hope they keep fleshing out the product, but they must first improve the core input reliability, or it’s a non-starter.

          2. David MacDonald

            I didn’t say anything negative about Dorico. I use and really enjoy it. I just said that they left _some_ users behind who have specific requirements. We’re in complete agreement on the matter.

  4. Jen

    Subscription plan? No thanks…I like to use Notion, MusicJot, Touch Notation on my iPad, and StaffPad on my SP4, as a way to ´disconnect´ from the desktop slavery, but I really despise the subscription model. I received the Komp email inviting me to a free month trial, and as soon I see that it was 50 bucks/year I closed Safari. No and one million times no. I can afford it, but I hate it, I feel used. Forever no.

    Goodbye Komp, welcome every other developer.

  5. Max Tofone

    Does anyone who has tried Komp know if it is possible to insert guitar fingering (e.g. 1,2,3,etc) and the circled numbers with the straight line across to describe the strings?

    Many thanks in advance…

    Max

    1. David MacDonald

      I don’t think so. Especially not if you’re looking for tab. You can submit feature requests on their site, though.

      1. Max Tofone

        Hi David,

        Thanks a lot for your kind and prompt reply!

        I am not interested in Tablature because I tend to notate guitar music in traditional notation but need to be able to put into the score fingering and string number…

        Best, Max

  6. Maestro KÜHL

    Symphony Pro is my favourite. Workflow excellent :)

Leave a Comment

Your email address will not be published. Required fields are marked *