Audiobus: Use your music apps together.
What is Audiobus? — Audiobus is an award-winning music app for iPhone and iPad which lets you use your other music apps together. Chain effects on your favourite synth, run the output of apps or Audio Units into an app like GarageBand or Loopy, or select a different audio interface output for each app. Route MIDI between apps — drive a synth from a MIDI sequencer, or add an arpeggiator to your MIDI keyboard — or sync with your external MIDI gear. And control your entire setup from a MIDI controller.
Download on the App StoreAudiobus is the app that makes the rest of your setup better.
Rambling on iOS Software Development Strategies (aka “How Come Bram Can Do So Much?”)
It’s been quite a week for iOS musicking, no?
In the past several days, we got the apeMatrix release, which can be a gamechanger for several people. Laurent Colson released a new AU MIDI plugin version of a powerful app. KV331 Audio released one of the most eagerly awaited softsynths on our platform (with some AUv3 support). AudioKit Pro and Matthew Fecher got some coverage for the imminent release of Synth One. Oliver Greschke also got coverage for iLep. Bram Bos announced a new extension to the Rozeta line and wrote a very insightful piece about the state of iOS music with a focus on interoperable yet independent modules (as opposed to monolithic DAWs). AUv3 is key to most of these things, particularly in connection with AU MIDI. (As you probably know, the beta versions of AB3 support AU MIDI.)
Some of the things which happened this week have been controversial. And it’s not the most active week in the year in terms of new releases. But, all in all, it’s a good week to think about devs’ strategies.
Part of my thinking is about SynthMaster One. Expressed ambivalence about it. It can sound very good but it doesn’t feel ready for primetime. One refrain is that iterating in the open can be a very effective strategy. But there’s also something to be said about not “biting more than you can chew”.
Keep thinking about the way our good friend Bram Bos develops stuff. It’s not only remarkable that he can create new apps on a regular basis as a side hustle but, at least in my case, his apps have made for a consistently excellent User Experience. They’re rock solid, they’re quite elegant, their features make sense, they deliver useful function, and they play very well with other apps. Those apps which include presets have just enough high quality ones to be inspiring without being overwhelming. Typically, each of these apps does “one thing well”.
And that’s profound.
Part of my problems with SM1 is that it does too much. It’s billed as a wavetable synth and that’s certainly a differentiator of KV331 softsynths, at this point in time (since 2.9 and updates to SM1 desktop). But it also does subtractive. Or uses audio input. It has several internal effects that you can route and modulate in countless ways. It has its own arpeggiator/sequencer. Some people really like its internal keyboard (and, just a few days ago, Bülent was wondering how much of it should be in the AUv3). It’s one of relatively few synths which support MPE (Roger Linn should definitely add it to his useful list). In some ways, it’s The Homer of softsynths. Maybe it’s ahead of its time? Time will tell. “The devil’s in the details.” It’s all about implementation.
To be blunt, SM1 is amazing on paper. My experience with it has been frustrating.
As a non-coder working with engineers for my dayjob, had a kind of epiphany while dabbling in Cycling ‘74 Max: encapsulation is the key to the way some people think (including my boss). As most epiphanies, it’s relatively simple when you think about it. But it’s a huge shift in thinking. Computational thinking is “all about” breaking down complex problems into simple ones.
Works extremely well with music, for obvious reasons. From PureData, Nord Modular, MOD Duo, Reaktor, and Softube Modular to Rack (by VCV) and the whole hardware modular world, there’s a whole lot to be said about modularizing as much as you can (but not more). Want a synth which uses a sequencer synced with Ableton Link to modulate the spread of wavetable index across multiple voices? Write a general-purpose sequencer or use one which already exists, ensure that your wavetable synth accepts general input, you’re almost there. As per elitist scholar Vilfredo Pareto’s “law of the vital few” (aka “80/20 rule”), the last part is far from trivial. Still, it’s vital to encapsulate each problem.
Bram’s case is fascinating. For instance, Troublemaker includes its own “bassline sequencer”, which is in itself top notch, with its killer “mutation” feature and such. He also included a version in Ripplemaker (with few mutations and without slides). Before Rozeta came out, was using Troublemaker to control other apps, as it’s so elegant. Bram was able to encapsulate a version of this sequencer to include it in Rozeta (with just a few missing features like note order mutation). Had Bram released Rozeta Bassline before Troublemaker, my guess is that he wouldn’t have included a sequencer in Troublemaker or Ripplemaker.
Perhaps famously, Bram has pioneered AU MIDI and documented the process of writing these plugins. Ripplemaker accepts input from StepPolyArp Unit without a hitch, in apeMatrix, AUM, etc. You can use Envolver’s AU MIDI to control the aforementioned wavetable index spread in SM1. Not only does that make the whole workflow extremely powerful when you have several of these apps, but it also makes it possible to focus every app on doing “one thing well”. (The matter of marketing modular apps is a whole can of worms. But that too can be encapsulated while keeping in mind the whole landscape.)
Note that it isn’t always an easy call to modularize one set of features or another. Should apps “outsource” its effects? Should every app contain its own keyboard? There are pros and cons to either approach. Thing is, we have lots of options for effects and keyboards, on iOS.
In fact, a cool option for the latter is Midiflow Keyboard for AB3. It includes most of the features you’d expect from a general-purpose keyboard, including scales, velocity, “unisono hold”, etc. (Wish it worked as MPE, but that’s another issue.) SM1 has three keyboards instead of the two in Midiflow Keyboard for iPad (only one of iPhone). But chances are that MF Keyboard would be an acceptable solution for most SM1 users working in AB3.
Dörr hasn’t touched most of his Midiflow apps since releasing them. But, like Bram’s apps, they’re rock solid (in my experience). Haven’t used them much in the past few months, because part of my focus has shifted to AUM. One would guess that Johannes is working on AU MIDI versions of his MIDI filters. That could make a significant difference in the landscape.
Those apps “just work”. They’re not buggy, they don’t require too many brain cycles to setup, they perform well, they make a lot of sense as a product lineup.
If “do one thing well” is most explicitly associated with UNIX, “it just works” is something people have perceived about some Apple products, a little while back. Thing is, though, there are growing noises about Apple deviating from this pattern, in both software and hardware. Apple-focused developer and podcaster Marco Arment has been a vocal critic of Apple in this respect. Even if it’s just a matter of perception, this alleged decline in quality has a huge implication for Apple as a corporation, and for us by extension.
None of this is easy. And we may disagree strongly as to which strategy produces the results we want (in part because we want different results). Thankfully, the iOS musicking landscape allows for all sorts of approaches.
Comments
I’m not so sure I agree that SM1 does, or attempts to do too much. Comparatively speaking it’s on par with the likes of Massive and Serum. And while there are still many bugs to work out, I think its value proposition overall is an advancement for this platform. Curious, if it all worked flawlessly in its current form would you feel it’s still attempting to do too much?
Hear, hear!
One thing that I really wanted from SM1 was to be seduced by it’s basic sound (i.e. oscillators and filters) and I was/am a bit disappointed with that. That’s my issue, and I now understand that it’s strengths lie in all of the other things that it can do. And who knows, it might actually fit better in a mix than other synths whose basic sounds I have a greater affinity for. I’m keeping it, but I’m not “in love” with it. Compare that to Troublemaker, which is a far more limited app in scope and sound, but which seduces me quite easily (and not because I have a nostalgia for “that sound”).
I had the same question, but inferred (perhaps mistakenly) that the OP was saying that it did not excel in any area.
Bram has been doing this for a little while now, quite a few of us may well have used his earlier noisemakers back in the day!
I'm afraid synthmaster one is another must have and I have almost every wave table synth on iOS. There are some very clever things, would recommend watching the tutorial videos that were recently uploaded by the developer. Those with older devices will be pleasantly surprised by the low cpu cost!
Fair enough. We have different standards and, most likely, very different usage patterns.
So we’re agreed, then.
Also fair. Not perceiving that, but that makes sense in a certain way.
Actually, yes.
The Massive example is quite relevant. Got it, as part of Komplete Select during the NI sales. Tried a few patches, twiddled a few knobs, watched a few tutorials… Not doing it, for me. It’s both too deep and not flexible enough for my use cases. Finding it unwieldy.
To me, it feels like it’s cramming too much in the same interface and proprietary format. Not surprising that there are so many presets released for synths like Serum and Massive. But those are part of a specific philosophy for synth design which is quite far from modularity.
Again, to each their own. It’s cool that there’s room on iOS for monolithic synths which still play relatively well with others (thanks to AUv3). But it’s a relief that there’s also room for simpler apps that you can chain with one another.
Funnily enough, was somewhat unconvinced by Troublemaker, before buying it. It was at a time when “that sound” really didn’t do much, for me. But its simplicity and elegance started to seduce me. It’s not “impressive”, but that’s also where you can shift the complexity to some other things. In a way, it’s a bit like a great Eurorack module that you can use in novel ways.
Wish there were iOS analogues to, say, Braids.
Hmm, I’m not sure I follow where this comes from, but I’m curious. Enlighten me. For your use cases what would’ve made SM1 shift more to your liking? If we are talking about a bug-free version of SM1, of which it is currently not. Too much, not enough of .....? I’m genuinely curious.....
Sorry my point wasn’t clear. Part of it is because discussing these things feels like entering a minefield.
My perception of the whole KV331 Audio lineup is that these softsynths do come close to excelling in some areas, especially those associated with wavetable handling (including index spread, which is pretty much hidden in SM1) and modulation. So, sure, there’s a bit of a missed opportunity for SM1 to become a leader in wavetable synthesis on any platform. (Similarly, SM2 could become massively popular in any category if it focused on its strengths.)
That’s Bülent’s game to lose. It’s not really my concern.
My qualms are more about the “deal” made with users. You keep changing your roadmap? Won’t get upset as it’s become a classic move in some parts of the tech world. You ask me to “bear with you” while some bugs are ironed out? Fine, if they ever do get fixed. You tell me the text documentation will be ready later as videos should be more useful? Fair enough; we just don’t perceive things the same way. You ask me to buy into your approach to preset packs? Well, ok, but that might not be where there’s most value in using synths (might as well release soundpacks, à la Ampify). All of these things can be understood. They’re part of a development philosophy which would work well if we were contributing to a shared project. This philosophy just doesn’t correspond to my needs. Or to the way the products are marketed.
Okay, it’s all making more sense now. Thanks for the clarity.
The key thing about SM1 (as opposed to SM2) is that it’s a “one screen” interface. In this sense, it’s simplified. There was a thread about simple wavetable synths on KVRaudio and SM1 was probably my best bet. A further move in that direction would have done a lot to shift my thinking.
For instance, the synth could default to a simpler patch. Ideally, it could contain a bank specifically made for MPE controllers (including software ones on iOS). It could make it really easy to set things up with certain controllers (including MPE ones but also virtual ones like Rozeta modules). It could also include the kind of well-crafted documentation which enables deep learning. It might also make interface elements easier to find by locating them in more logical spots (especially in the AUv3). It could leverage UIkit. It could outsource more functions to other plugins (do we really need a sequencer in the AUv3?). It could streamline the crossplatform sharing of patches. It could allow certain elements to be hidden.
Not that these were my expectations based on Bülent’s public comments or on experiences with SM1 and SM2 on macOS (and frustration with SM Player). But SM1 could have surprised me.
One use case which delights me (say, with my MODEP enhanced Raspberry Pi): you plug a controller in the device, you start playing the last patch you made with all the right settings, you navigate through other patches with a single button, you never have to look at a screen.
Or, you plug Rozeta Bassline into a simple softsynth, you plug the output into a waveshaper and a delay, you use a pedal to control a filter… No need to look at the softsynth’s interface much. You can spend hours playing with all the subtleties of the “instrument” you created in a couple of minutes.
You are hardcore dude!
So I guess the question is whether we'd prefer simpler things that work better right out of the box, or more complex apps that might have to go through some growing pains to get up to par. I don't know the technicalities behind coding Bram's apps. They're fantastic, but maybe it's just more difficult to make a solid SM1? I'm happy we have both. I like SM1, and will be patient as the bugs are worked out.
I love a lot of what Bram Bos is doing, but, you can’t compare Synthmaster One with anything that Mr Bos has been released on iOS...
And, the small team of KV331 also is busy with more than iOS. They may have a lot of work in the pipeline for Synthmaster 3.0 for desktop, Synthmaster One for desktop, Synthmaster 2.0 for iOS (just begun) etc etc...
For me, Synthmaster One for my iPad Pro is reel dope - happy as ever!
Ripplemaker and Troublemaker not getting their sequencers baked into the AUv3 devastated me
I was sooo counting on that. At least SM1 is offering me this 