Why using Linux for music production?

[A.O.S.]

Hello,

yesterday, while working on my music project and searching for good VSTs, I suddenly had the thought: "Why do I use Linux for music production?"

The background is, that in most cases, when I am asked what I am using for music production, people are starting to make fun about it and not taking it seriously. Personally, I don't care what other people are thinking about my preferred tools. I don't have urges to be a missionary, too. But sometimes it would be nice to have some hard facts available in a discussion.

My intention with this post is, to focus on "technical" reasons for using Linux. I do not want to warm up discussions like that the big corporations are spying on us, are evil or whatever.

Here comes a list of pros and cons, feel free to add more pros and cons:

Pros:

• It is maybe the most cost-efficient solution to do music production
• There are a lot FOSS DAWs available
• There are a lot of available FOSS VSTs for Linux.
• There are very powerful and excellent FOSS DAWs available
• There are very powerful and excellent FOSS VSTs available
• The sound subsystem is stable and reliable
• The sound subsystem is quite flexible and shows good performance in most cases
• In most cases the general system performance is very good, even with older hardware
• Most hardware will work out of the box or with minimal tweaking
• Most natively available Software is mature and reliable
• FOSS software in particular often has a clear licensing situation versus custom licenses which can be tricky sometimes (like defunct companies with deactivated licensing servers, suddenly changing, removing or adding terms and conditions with drastic consequences etc.)
• High security of investment (see point about licensing and hardware which can be used much longer compared to other OS)

Cons:

• Most fee or commercial VSTs are not natively available for Linux
• If a free or commercial VST is natively available for Linux, its often "half-official" with no support
• Most FOSS VSTs are mediocre and not on par with free or commercial ones
• It can get quite complicated to optimize or fix the sound subsystem if there are issues
• Most hardware isn't provided with Linux drivers or official support
• User interfaces, especially of FOSS Software, can be quite cumbersome

[Impostor]

We can combine audioservers in more ways than on Windows or Mac.

Alsa
Alsa+Jack
Alsa+Pulseaudio
Alsa+Pulseaudio+Jack
Alsa+Pipewire
Alsa+Pipewire+Jack
Alsa+Pipewire+Jack+Pulseaudio (does this work?)

One can consider this either a pro or a con.

[erlkönig]

An important point: a lot of upcomming audio engineers are doing their final projects on linux/foss, so you can rely on industry standards on implementation - you are able to have a look on it, it's foss. From an engineering perspective, that's an unbeatable advantage.

[khz]

Freedom.

[talby]

By priority:

• No forced interruption by OS updates destroying my long term session. It is absolutely irresponsible that other OS after time restarts without my consent and destroys data and even disconnects from the control of external hardware! Ignorantly only allowing to postpone updates for a maximum of 1 month instead of allowing to disable them completely is inexcusable. Good that equipment running in a medical intensive care unit does not depend on such OS. System changes take place when I schedule them, and only when I schedule them.
• No forced hardware upgrades, and maybe losing access to older, still perfectly fine hardware accessories only working with older interface types, and still having access to security updates maybe only provided for newer OS versions.
• Desktop GUI freely configurable, greatly enhancing my workflow possibilities and subsequently productivity.
• No distracting notifications popping up, ALL can be configured completely.
• Why should I for music production change away from my everyday computing environment?

It is not about costs. I pay for good software, honoring the work and time of the developers also by supporting them financially. It is all about establishing my personal workflow in software and hardware and not willing this being ignorantly stolen away from me. This is by the way also the reason why I with preference purchase software which I can activate offline, if undecided between competitive products. If in need to re-install my expensive software, then I do not want to find my license invalid because some activation server is meanwhile offline.

[Linuxmusician01]

Manageability.

In Win/Apple I wouldn't know how to troubleshoot or it's gonna be pretty hard. I can't remember precisely how it was in Windows (switched almost 20 y. ago) but I vaguely remember something about the Registry. Impossible sh*te to fathom. And every setting is tenfold in there somewhere in 100 different locations. What a fail.

Also Win's error messages w/ those hexadecimal numbers: what buffoon thought of that? Nobody in the universe can make heads or tails of 'm. Okay, given the popularity of Windows you can find a solution within seconds on the internet for such an error code but it's far from ideal.

When I switched to Linux I finally understood error messages because they are text written by humans, not hex numbers! Admittedly it's not easy to understand 'm at first but after a while you will. And googling Linux errors is the same as Win errors: sometimes the "solution" is wrong and makes things worse, sometimes it works and you learn what was actually wrong.

Oh, and I hate downloading executables to fix things (I'm looking at you Windows). No idea what they actually do.

So why Linux for music? No Registry, no hex code error messages. Unfortunately Linux is ever changing it's way - and has alternative ways - of "doing things". Examples: X.org vs. Wayland, Gnome ver. 3 breaking good 'ol ver 2 support, VST (yes: there's Linux VST and Windows VST!) vs. LV2, PipeWire forcing me to re-learn stuff, your fav. app of yesteryear not in the repo's anymore, etc., etc. When I think of switching back (to not have to re-learn stuff over and over again) then I think of the Registry and hex code errors. Brrrrrrr.

Oh well. computers... Can't live with 'm can't live without 'm.

[A.O.S.]

One can consider this either a pro or a con.

I think its more a pro. Although I think that pipewire with jack plugin is the way to go in these days, its nice to have the possibility just to use Alsa for everything.

An important point: a lot of upcomming audio engineers are doing their final projects on linux/foss, so you can rely on industry standards on implementation - you are able to have a look on it, it's foss. From an engineering perspective, that's an unbeatable advantage.

I am not so sure about this point. I am under the impression that everyone, once reaching the "professional stage", is switching to Windows or Mac. It seems that it is seen more professional when someone uses Windows or Mac for something. Which is at least absurd.

Good that equipment running in a medical intensive care unit does not depend on such OS. System changes take place when I schedule them, and only when I schedule them.

I worked in this sector. You would be surprised how often Windows is used, or devices are dependent on Windows machines. The only exceptions are found with vital devices, but even there it seems that some manufacturers are keen to make Windows work in such environments, which is kinda absurd.

No distracting notifications popping up, ALL can be configured completely.

As someone, who can be easily distracted, I absolutely agree.

This is by the way also the reason why I with preference purchase software which I can activate offline, if undecided between competitive products. If in need to re-install my expensive software, then I do not want to find my license invalid because some activation server is meanwhile offline.

Sadly, the trend is going to online activations etc. I hope that FOSS and music hardware will keep the companies in the audio sector away from subscription models etc.

Manageability.

Very valid point. I am a Linux user since the late 90s, but I always had to work with Windows, because of my job. To 95% the fastest way to solve a Windows related problem is just to re-install Windows. Trying to solve a problem, even with the Microsoft knowledge database, is almost every time futile.

[Psychotronic]

A easy path to stable and upgradable half assed perma computing.

[bluebell]

The answer to "Why using Linux for music production?" is the same as the answer to "Why Linux?"

If you want to control your computer and your data instead of being controlled then Linux is for you.

[runiq]

Because I'm using Linux anyways and I can't be arsed to get back into Windows just for a hobby.

[asbak]

Freedom
Ability to use old audio hardware on recent Linux distros (drivers typically not available in walled garden OS's)
Less spying and datamining
Less coercion than on walled garden OS's
Less annoying subsystems offering to "assist" and "help" you ( the real aims are to control and program and datamine you )

[tavasti]

Because I'm using Linux anyways and I can't be arsed to get back into Windows just for a hobby.

This is my case also. I've had been linux user for 20 years when I started with music. Now been learning 10 years many aspects of making music with my linux system.

[Psychotronic]

• Most FOSS VSTs are mediocre and not on par with free or commercial ones

It's morning, I'm having a coffee. Let me write a little rant.

Most FOSS Plugins do feel cheaper, but let's consider first our own human bias. Our brains assign worth to things we paid money for, because we are biased by growing up in a capitalist society, it feels natural to us. Secondly, i presume most FOSS devs for pro audio tools just consider programming something that works right and "ideal". Most of the commercial plugin world tries to sell you the next big hardware emulator plug instead. Software that is deliberately programmed to emulate the wonky characteristics of some hardware unit. The next neve plug / the next ssl plug, the next oh so great hardware synth emulator, because it sounds soooo... Anyways. A lot of the products on that market are just fake, there is even YouTube defluencers that are uncovering the bs that's being sold, very entertaining. (Don't get me wrong, there are some good commercial plugs out there, but providers like u-he also release for linux, so this is not really an issue. If you need a diva, just buy one, i did too)

But what's the actual difference between what you can find in your run in the mill FOSS plug and those. There is two main areas, the UI optics and handling, trying to emulate the hardware in detail, if done right. Okay, there is this comp that has that secret trick where you punch in all the buttons and then it did this and it wasn't intended by the manufacturer, and so on. Do we need to emulate that? Why not build a better UI that exposes the feature if integrated. Or if you dial in this at 6 o'clock then it sounds exactly like that on my hardware. Fun fact, but not on mine dude! Hardware differs from unit to unit, especially hardware made in the middle to late of the last century. Why try to emulate it. Why not build something that can do more, or fixes the problems of handling the unit. Anyways, that's a long discussion. But in the end, if you know what you need your tools to do, if the tool is capable you will get it and most FOSS tools in that regard are absolutely capable.

The second of the two main areas is emulation of wonky or unexpected behavior in sound, stuff that surprises us creatively, that feeds our drive to play. Most of this has to do with emulation of saturation behavior of hardware parts. So people spend extra money on software and computer hardware to emulate the imperfect behavior and resulting sound of hardware. Now, there is for example the airwindows project that also found its way into cardinal(vcv rack vst clone) from what i saw, but you can also just use them directly. Those plugs don't have a implemented UI, so they use the default UI of your plugin system. The library contains lots of little tools and helpers, but most importantly, there is a ton of little plugs in there that handle some sort of saturation or "wonky" emulation and just that. Now you can take your plugin chain to the next level, consider it a circuit and add some little weird behavior everywhere you want. So you're not bound anymore to some overpriced dev trying to sell you the next SSL Console EQ and stuff. This is where the fun really begins.

So are FOSS plugins really cheaper in build quality? No, they are just built without the general BS marketing in mind.

Think about it. What do you really need to make music? Minimize this down to the feature set, not to a set of plugins. Most of this can be done with FOSS. And yes, there are commercial plugs that do something that makes them worth being industry standard. I get that, the question is, are you an engineer that needs these tools or not. Is your daily workload so high that you need tools that speed up your daily output to the max?

If yes, then why the hell are you reading this. Get back to work on your mac. Time is money! Go go go!!!

[A.O.S.]

Most FOSS Plugins do feel cheaper, but let's consider first our own human bias. Our brains assign worth to things we paid money for, because we are biased by growing up in a capitalist society, it feels natural to us.

You are right, but I mean the "audio characteristics" or the effectiveness of a VST.

I can give you some examples of what I mean.

One example would be ZynAddSubFX before/without Zyn-Fusion. Its an extremely powerful and superb sounding synth engine. But, for me at least, I couldn't really work with ZynAddSubFX, except the patches that come with it. The UI is just so unbelievable cumbersome. Here, Zyn-Fusion is a quantum leap towards an efficient and user-friendly UI.

Now, there is for example the airwindows project that also found its way into cardinal(vcv rack vst clone) from what i saw, but you can also just use them directly. Those plugs don't have a implemented UI, so they use the default UI of your plugin system.

It's interesting that you are mentioning the airwindows project. This is an absolutely great project. I like the minimal UI approach, because their VSTs are often extremely efficient to use. It's the opposite of the example with ZynAddSubFX above.

That's the point I mean. If you take aside all the marketing silliness of the commercial VSTs, there is often one point still valid, they are more efficient to use.

It's my approach to keep my workflow efficient. Yes, I am not a sound engineer, its just a hobby for me. But I have very limited time for my hobbies at all, that's why I want to be efficient and not wasting time for achieving certain sounds.

Another point is, that commercial VST tend to be more aggressive with altering the sound, which is for some genres better.

Fun fact, but not on mine dude! Hardware differs from unit to unit, especially hardware made in the middle to late of the last century. Why try to emulate it. Why not build something that can do more, or fixes the problems of handling the unit. Anyways, that's a long discussion. But in the end, if you know what you need your tools to do, if the tool is capable you will get it and most FOSS tools in that regard are absolutely capable.

Absolutely yes, there are impressive FOSS projects out there. For example, K1v and the DSP53600 project. Both are sounding virtually the same as the Hardware. A negative example for commercial Hardware emulation IMHO is the Roland D-50 VST. They should have called it LA-Synth engine or something else, but not D-50. It's missing character and sounds too perfect. Just how LA synthesis is implemented today.

[Psychotronic]

I agree that there are a lot of synths out there that are overblown in terms of feature richness. But that's not just FOSS projects that have that problem sometimes. Look for example at NI Absynth. Most people use it as a preset bitch with integrated random knob. Everything else with it is to complicated. I even think u-he diva is too feature rich. If i want a modular system, i use one.

The design idea of a Synthesizer comes down to feature reduction. I've been in that pit as a developer for a bit myself a decade ago, developing reaktor 5 synths. It really sounds weird saying it like this, but i do it anyways, in gaming terms: feature reduction is like removing the fog of war to an area, you can see your sound target clearer and if the human brain gets bored by the available targets, then it will abuse the device and do really unintended stuff with it and that is where the real creativity lies. Just bombarding a user with options thickens the fog to a state of total blindness. Users don't want to be blinded.

The interface of a Synthesizer needs to guide the user into fast and enjoyable discovery of sweet spots. As inventor of the device you have to define what those sweet spots sound like. Simple as that.

Some FOSS projects definitely give you a full blown Synthesizer Engine with all the soldering spots accessible, instead of a condensed down sweet spotty device. This forces the end user into the role of a developer, lots of end users don't want to be in that role, so the adoption rate for the project is low. But it also doesn't need to be high, this then is a project for people that want to be in the developer role and can use it to develop their own condensed down sweet spots configurations for their end users without getting down too deep into a low level programming pit, it is FOSS code after all and can be used by others. All a matter of perspective and time. A good example with zyn fusion you already gave, but even that one is "too foggy" for my taste. Condense it down to the accessibility level of a nord lead but with your own vision of sound.