Linux Audio Plugin Development (LAPD)
Niels Ott
Linux Audio artist interview with Niels Ott
This interview was conducted by Amadeus Paulussen in 2025.
Dear Niels, firstly, I would like to thank you for taking the time to do this interview. 🙌
From your social media activities, I learned that you have been deeply involved with Linux Audio for a very long time and are not afraid to tackle even complex setups around live mixing concerts via Linux.
A while ago I wrote an article for my blog in which I reported on what my everyday life with Linux looks like today and that I am very satisfied with it. In this article, however, I also noted that I have not only had to learn a great deal about Linux to achieve that, but also, admittedly, sometimes consciously avoid situations that I know can (currently) be problematic.
Could you briefly describe the work you use Linux for?
My main occupation is being a solo entrepreneur in event technology with a focus on sound. But I also do full service, sometimes hiring freelancers for things such as building stages or fancy lighting beyond basics. A laptop running Linux accompanies me for almost all of my jobs.
I also do recording. Which today means mostly doing live recordings of shows and mixing these down afterward. Recording, like in producing a record, is not something clients request anymore too often. I believe that the reason for this is that Spotify and other streaming services have completely cut the revenue stream for lesser-known artists. Hence, there is no motivation for booking external services, eventually spending money for giving away your music for free. Be it as it may, all my remaining mixdown happens on a Linux machine.
During the pandemic, I did plenty of video livestreams and also multi-camera video productions. The situation is a bit the same as with music streaming; it's not something artists can make money of unless they are full-time YouTubers. The livestreaming I still do today is mostly for business and local politics. Again, be it as it may, all video editing happens on Linux, and also for the remaining livestreaming, Linux is on board again. Even though, mainly for controlling and monitoring things, in all event tech scenarios, as in making music, I rely on dedicated hardware a lot.
Which mentions another important thing in my life, which is being a musician. I play the electric bass guitar, synthesizers, and the piano. Almost 20 years now in one rock cover band. In varying projects, I do other, more experimental / avantgarde music full of improvisation or even with a relationship to noise. I did a couple of solo records in earlier days. Some I de-published over the years, but all of my solo releases that ever existed were made on Linux.
How do you perceive Linux in the context of music production today, more than 25 years after you started using it? Are you positive about developments surrounding the Linux Desktop and technologies such as PipeWire, Wayland, or even Flatpak?
I never thought Linux would get this great for audio production. At the same time, I also think there is still a long way to go.
When I made the switch to Linux in 1999, the world of personal computing was quite different. USB 1 was fairly new. Affordable disk space and processing power were limited. Wikipedia states that 1999 was actually the year with the first number one hit produced entirely in a Digital Audio Workstation, using ProTools' dedicated audio processing hardware, which was an expansion card you'd put into your computer. 24-bit-capable sound cards were not common for home users. Linux back then supported some sound cards, like some of the Sound Blasters by Creative Labs. You'd carefully check which chipset was on a card, because there would be drivers, like for some Yamaha chips, and if the card had that, it would likely work. Linux back then was providing the Open Sound System (OSS) interface to applications for using sound.
Now, let's do the time travel: 2025, we have USB everywhere; people are often not using bulky grey boxes with loud fans but fancy laptops. Disk space is virtually unlimited, at least for the needs of audio production, and so is processing power. USB defines a common standard for audio (as FireWire did, by the way). So even without regard to Linux, the situation has improved quite a bit.
Now with Linux, all standard components of a computer are usually supported today. If you install Ubuntu Studio (the distribution I use), you are ready to rumble. Unless a big grain of salt hits you in the face, which it did for me, since both major audio interfaces I use are not properly supported by PipeWire yet. Yet, earlier ALSA, then JACK, and now PipeWire are huge steps. Especially PipeWire. In my opinion, it became mainstream a little too early, but then again I see that developers probably needed a large user base in order to get things really going. What PipeWire will eventually be is what users of Windows and macOS must continue dreaming of.
Also on the software side, we have seen so much evolution. Of course Ardour and Mixbus, but also other DAWs stepping in, like Reaper and Bitwig Studio. In the video domain, things are a little sad still; basically, there's only Lightworks and DaVinci Resolve. While the latter is an industry standard, my nerd powers did not yet suffice to get it to run on an Ubuntu-based distribution. Kdenlive has become a mature tool, too, but to me seems to be oriented more towards the amateur video enthusiast.
New packaging mechanisms such as Snap or Flatpack are a thing I see critically. So, will we, like on Windows, distribute a separate Java Runtime Environment along with every Java program? Just as an example. Of course, it makes things a lot easier, since sorting out dependencies for a multitude of distributions in the end may keep commercial developers from supporting Linux. But it re-creates issues from other OSes that Linux never had.
What was your motivation for initially starting with Linux in the first place, if I might ask?
As mentioned, this was 1999; I was still in school. Me and some others chose to channel part of our teenage rebellion into bashing Microsoft and Bill Gates (who back then was still CEO of MS). This was the time of SuSE Linux, which you'd buy as a CD set, with a discount for students. Another important reason to find Linux interesting was that it catered to our nerd interests: you'd have to conquer it by editing weird config files, writing little scripts, doing hacks, and getting into nerdy things. We felt superior because we could use a command line even via Telnet, while we arrogantly viewed our teachers and most parents as clueless mouse pushers. It felt like an empowerment to our little teenage minds in a stupid world of school and grown-ups. And sometimes it still feels like this today – the same way I today still love Another Brick in the Wall by Pink Floyd.
However, seen less romantically, a few things had happened around that time, and these are things still relevant today. Firstly, KDE 1 showed up, and for the first time, Linux had a desktop environment that was not crap. (GNOME entered the scene a little later.) ApplixWare Office was distributed by SuSE, and so doing actual ordinary work was possible. Secondly, with VMware, the first virtual machine that would run Windows inside Linux was available. The standard dual-boot scenario could be abandoned.
Hence I used SuSE Linux as the primary OS and moved Windows 98 into VMware. I had started my first business at that time, and I could do all the office work with Linux.
I wasn't thinking about any recording or media production back then. But already in 2002, things had started to change a little, and I started experimenting with creating my own SoundFont sample libraries, for which Sound Blaster cards had internal processing on board.
Can you briefly outline your current Linux Audio setup for us? What hardware and software (operating system, apps, plugins) are you currently using?
Let me start with a little shocker for the religious Linuxer: my Linux Audio setup relies on native Windows plugins, which I run via LinVST / Wine. At least this is true for mixdown.
I have two setups, basically. I use an elderly Dell workstation with a Xeon CPU, AMD GPU, 32 GB of RAM, and 9 TB of SSD space for mixing audio and editing video. This old beast has an RME Digi9652 PCI card inside. It connects to a Yamaha 01v96i digital mixing console via ADAT Lightpipe, so I have 24 channels in and out at 48 kHz.
All my instruments that I play connect to this console. But for mixdown work, I work entirely in the box, so the console basically acts as a monitor controller in that case. I use Harrison Mixbus 10 Pro, and sometimes Ardour. Lightworks for video editing. The most important plugins to me are Voxengo Elephant, FabFilter ProQ3, Sknote Disto-S, and Voxengo Voxformer from the Windows side, and the old XT series by Harrison from the native side, including their lookahead limiter and multiband compressor. For reverbs I use impulse responses in the LSP IR plugin, mostly from the Bricasti M7, but also ripped from the REV-X reverbs in my 01v96i console.
Are there native replacements for these Windows plugins? For some, definitely. But switching would require me to check them carefully, e.g., to really check the limiter acts properly on transients and doesn't color the sound and prevents inter-sample overshoots. So, keeping these old plugins alive in LinVST is less of an effort for me than switching.
Important apps from the Windows side are Hofa DDP Player to check CD disk images before delivery and A Control by ADAM Audio for adjusting the DSP in my speakers.
Unfortunately, neither the 01v96i (via its USB port) nor the RME Digi work with PipeWire at the moment, so I still have a hybrid setup with JACK and PipeWire.
The other setup is my laptop, an elderly Lenovo thing with a touch screen, 32 GB of RAM, VirtualBox with Windows 10, and 1 TB of SSD space. This is what I use in my event tech job; it often sits in the dust and moisture of open air and suffers. On my laptop, I use only PipeWire.
Both systems currently run Ubuntu Studio 24.04 with a PPA providing a more recent PipeWire.
What are your wildest Linux audio dreams? What do you think needs to happen to make Linux even better suited for music production?
First of all, we should be more precise about how we define music production with a computer. What I do is mostly mix down live recordings. So, my workflow is more or less what used to be the workflow with a multitrack tape machine, a large mixing console, and some compressors and reverbs. Mixbus (Pro) is a superb tool for this job. Its internal processing is very quick and convenient to use. Plus, there are enough suitable plugins, both native and Windows plugins, that can be used.
However, for a great many people, music production actually means constructing their tracks or songs in a DAW. Production blends with the process of composing music. Some will want to work completely in the box with software instruments, outboard equipment merely consisting of one or two microphones for vocals and maybe an acoustic guitar. Compared to the Windows or macOS world, there are not so many options in software synthesizers, and even more important, some signature plugins are missing. Also, people clicking together their songs in a MIDI editor may favor DAWs other than Mixbus.
Linux enthusiasts may not like to read this, but users that have been using Melodyne or Kontakt Sampler in the past are unlikely to switch to an operating system that does not easily support these. Some of these killer plugins, such as FabFilter ProQ, can be run on Linux via LinVST. Others are very hard to get to work, if at all, such as UAD or Waves. We can praise FLOSS alternatives as much as we want; in the end there will be no way around these tools running on Linux if we want to attract more users.
Thus, my wildest dream would be that all major software manufacturers make the step towards releasing their plugins for Linux.
How would you describe the balance between frantic problem solving and relaxed, satisfying work with Linux Audio today?
In general, I enjoy satisfying work with Linux Audio and Linux in event tech.
One might call it left-brained vs. right-brained… In general, I keep apart activities such as computer system maintenance but also office duties such as accounting or managing jobs from activities that need creativity or at least a lot of sensual perception, such as playing instruments, mixing recordings, or even mixing live shows. This is something I'd like to share: Do not switch modes. If you're doing a mixdown, do not install new fancy plugins in the middle of the process, and don't get interrupted by any kind of maintenance. Don't re-solder the defunct cable; just use another one in that moment.
Setting up my current Ubuntu Studio fully and properly took me around three workdays, which were accompanied by quite a bit of frustration. The hard part was learning that neither of my two audio interface options work with PipeWire. Also, some Ubuntu principles make things uneasy. Such as: No, we don't update your PipeWire; please wait or upgrade to 24.10 (with no long-term support). (There's a PPA to work around it.)
Over the last 25 years I have changed. I used to start as a computer nerd who loved getting lost in technical details. Today I am an audio nerd who does not want to get interrupted by any technical crap happening. At least in my life, Linux interrupts me a lot less than other OSes do. But also it's an OS I am very much used to, so I can handle little interruptions without too much thinking.
And what is your experience with Linux in the context of event technology and live shows?
It doesn't work without running Windows applications. Many of which work in Wine, some require original Windows in VirtualBox. Sad story, true story.
Many applications in event technology are actually just user interfaces on a PC. Like Yamaha's R Remote, which is a remote control for their digital stage boxes. Or Allen & Heath's AHM System Manager, which is the only way of operating/managing their AHM DSP system. There's plenty more of that type. Luckily, most of them run in current versions of Wine! One exception to the game is Behringer/Midas (practically equal), who offer native Linux software for mixing consoles such as the X32/M32 and their X-Air/M-Air series. Funnily, their M32 editor crashes on saving to file on my system.
Heavier applications might not work in Wine. I haven't managed to run the Blackmagic Atem Control in Wine yet, which is a tool to remote-control their video switchers. I doubt Harman System Architect for setting up everything Soundcraft or Crown or dbx would work. The same may hold true for Shure Wireless Workbench to manage / measure wireless mic frequencies. But I haven't tried yet.
One of the most important tools, however, is one we all know: VLC. Because people always confuse the audio guy with a DJ. If they don't show up with playbacks without telling you in advance, they will at least ask for music in breaks of the show.
Open Broadcaster Software (OBS) is also another killer app that I use a lot. Never for actual streaming; this is left to dedicated hardware. But for analysis reasons, such as false color monitors, LUFS-I loudness metering of live streams, and sometimes also capturing/recording video from some devices. I always have a Blackmagic UltraStudio HDMI/SDI capture device with me, which comes with a Thunderbolt 3 kernel driver by Blackmagic.
Audio engineers will miss SMAART for analyzing their PA system, but they might find its most important features in Open Sound Meter, which is free even.
I made my own lighting console in QLC+. It spits out Art-Net, which basically is DMX lighting commands wrapped in network messages. In the beginning I used QLC+ on an old Windows 10 laptop because that was the only touchscreen I had; now I use it on Linux. QLC+, however, is quite strange to people working in professional lighting; these folks might prefer MagicQ by ChamSys. And with these two, I have also mentioned all lighting control software for Linux that I know of.
Now what's really missing on Linux is Dante. For those not familiar with the topic, Dante is an Audio-Over-IP solution for event tech and studio work. It can operate at audio latencies down to 0.25 msec. All Dante communication unfortunately happens with dedicated hardware chips, which are sold by Audinate. Windows and Mac users find a virtual sound card available, which is Dante over a standard networking interface. Also, Dante Controller, the software for managing and routing audio in your network, is only available for Windows and Mac, even though it is a Java application internally. Audinate also does have Dante libraries for embedded Linux systems, so they have the code working, but apparently they don't want to make a virtual sound card. There are a couple of PCIe cards with 64 or 128 channels of I/O that have Linux drivers. To my knowledge, there are only two multichannel USB-C Dante interfaces on the market, one by Yamaha, which is USB class compliant, and the RME Digiface Dante, which isn't. I myself use a 2×2 channel Dante AVIO USB to connect my Linux laptop to Dante networks in live shows.
(PipeWire can do AES67, and Dante chips can do AES67, too. I'll spare you the details; it's a risky endeavor in flexible setups – one single unsuited network switch can break the whole thing.)
What I also do quite often is hook up a mixing console via USB to my laptop. I like to feed playbacks digitally into digital mixing consoles and not via messy 3.5 mm holes with lousy converters.
Other popular digital audio communication means include MADI and Waves SoundGrid, which exists for Linux in a dedicated Linux-based Waves server but not as a virtual sound card; the same holds for the popular Waves plugins. MADI is popular in radio and TV broadcasts, and some RME MADI cards are supported by Linux. I never went into the topic of AVB, but there is hope that a spin-off of AVB called Milan will take off – it is designed to be open, standardized, and manufacturer-independent.
Note that this entire section focuses a lot on what does work and what doesn't – if you were on Windows, your expectation would be that everything of this works 100% for sure. And it would. It's all about manufacturers not making their software for Linux. Linux as such, including PipeWire, is easy.
Does the fact that you use Linux ever play a role in your professional work, or does it sometimes even pose challenges?
One quite stupid challenge is that I don't have Microsoft PowerPoint. Because in business events, it's all PowerPoint. Speakers will use PowerPoint for 100% of their talks, and there is no way around this. In this case I rent a Windows laptop with an Office 365 license from a friendly competitor.
For the mixing of recordings or editing of videos, I'm in the lucky situation that my clients are not tech-savvy. The final result is important for them. They don't care about the way there. This year I worked on an album and a DVD sound-wise, originating from 4h30m of live-fashion 36-channel recording. The client sat next to me in front of Mixbus 10 Pro, and we discussed changes to the mix and editing points, but he didn't even notice that this was not Windows or Mac.
Sadly, there are things I cannot do on Linux. If you want to use recent RME audio interfaces, such as the DigiFace Dante, you just can't. Also, my old Fireface 400 is well supported, but ffado-mixer is just a pain to use compared to TotalMix. Then it's time for macOS. Challenges with event tech have been mentioned already.
However, there are also things that actually work better with Linux. Recently in a mixed show, I had a MacBook on stage with a person operating it to fire off playbacks. People would bring their playbacks like 5 minutes in advance to their part. We used a MacBook because we wanted to run QLab, which is the industry standard for firing audio and video and even DMX playbacks. QLab is Mac-only. Turns out, there are memory sticks that macOS dislikes but Linux does like. Turns out, QLab uses macOS libraries to play back audio, and those won't play back files in the Opus codec. Without my Linux laptop and Audacity, this wouldn't have worked.
Also, when hooking up a Windows laptop to a mixing console via class-compliant USB, you'll find it needs special drivers to work properly, which I find annoying. Not an issue usually with PipeWire on Linux, just flick the pro audio switch for that device, and there you go.
How do you assess the audio market in terms of Linux support? Do you see a positive trend in terms of hardware and software vendors increasingly adapting to support Linux?
Apart from witnessing posts on Mastodon, I don't actively follow this market. Usually I'm in need of a solution and then check if there's something I can do with Linux. Talking about trends here would be just guessing from my side.
According to my conversations with quite a few vendors who already support the platform, Linux does not usually mean significant additional support effort for them. The much-stressed Linux fragmentation, therefore, seems to be a bit of a myth. On the contrary, macOS seems to generate the greatest support effort for most of the vendors I spoke with. What do you think are the main reasons why many manufacturers are still hesitant to support Linux?
I'm afraid that this argument isn't as strong as it looks. Let me present a nasty hypothesis: users who are music nerds but not computer nerds will go for a Mac. Because it's all so easy on the Mac, at least this is the common wisdom. But all users run into some problems, and because macOS users are not computer nerds, they will call for support much earlier.
Now, currently Linux users may be deeper into computers and operating systems. In fact, when I had issues with my Mac, not even Apple support could help me – yet my local Linux User Group could, simply because these folks have a very deep understanding of what goes on inside an operating system. So, currently, Linux users will not need as much support, plus they will find more support in their communities.
However, once Linux becomes more mainstream for audio production, it will attract a lot of less tech-savvy users, and the demand for vendor support will go up.
Yet, this is a hypothesis of mine, not a fact.
Vendors often tell me that entering the Linux market would need to be worthwhile for them. Do you see the Linux desktop market continuing to grow, and what advice would you give to vendors (if any) who are still unsure?
I have a stupid idea about this: We'd need a new manufacturer that defines the next industry standard thing that everybody uses and makes this available for Linux, too. This could break the chicken-egg circle of Linux not being used because there are no signature plugins, and no signature plugins being ported because Linux is not being used enough. But this is unlikely to happen.
Perhaps on Windows and macOS, vendors focus on special plugins with unique selling points. On Linux we still have trouble with the default plugins everybody uses. Example: Skimming through LinuxDAW.org I can see quite some fancy compressors, but I cannot spot one single 1176 clone, which is one of the most-wanted compressors – I even have a simulation of it inside of my live mixing console! And I use it live, especially for mixing Rock music. (But this now is somewhat redundant with my call for signature plugins.)
So my not-so-stupid idea would be: If you've got some boring bread-and-butter plugins in your stock already, port them to Linux, starting with the old clones of famous studio hardware.
Do you recommend other musicians use Linux, or do you think most of them are better off with macOS or even Windows at this point?
I have quit being on the mission of talking people into using Linux. Basically, what I do is show that it can be done, often on Mastodon, sometimes in person.
Do you mainly work with musical material from third parties, or do you also make your own music? If yes, can you share some links to tracks of yours with us?
Funnily, other people work on my own music in Samplitude or Cubase on Windows, while I work on other people's music in Mixbus 10 Pro on Linux.
I'm planning to record solo again, then of course on Linux. I have plenty of ideas, but it turns out that I am lacking a process of developing ideas into pieces that fit what I feel should be my standard these days. It's certainly more of an artistic challenge than a technical one.
I'm hoping very much this will change. So currently my own music is, again, mostly live recordings in collaborations with others. And it turns out that in both of my steady music projects, there are other people on board who enjoy working on these recordings in their DAWs.
An old production possibly worth mentioning may be my EP Cyborgs, from 2013. It was the first production for which I did the mixdown entirely in the box.
Another production, a little younger, dwells on my experiences with the German Agentur für Arbeit (German Federal Employment Agency), which is a remix of their hotline music as well as a music video – but please don't tell people about it; I haven't gotten sued for it yet.
Do you personally know many other musicians who work with Linux?
I regularly meet one other event tech who is also a musician, who's using Linux at home but not in event tech. And that's about it. There's one other guy in my region I am aware of doing recording with Linux, but we meet only occasionally. With regard to the number of persons I know personally who are working in event tech, or recording, or both, two is not a lot.
What fascinates and motivates you about your job?
It's always sound, and often it's music.
I'm fine when I get to transport the sound of an artist or a speaker in a pleasant way. Over the PA into the audience. In a recording or video. I always try as best as I can to make them sound great. I constantly grow with this task. If music interesting to me adds to this, I turn from quite fine into happy. Sometimes there are these acts that just blow you away.
Mentioning some I had mixed live each only once: The Little Unsaid or THIER or Suchtpotenzial. And then there are also bands I mix regularly that entertain me and whose music I enjoy, such as HASA or Mandala Movie.
Mandala Movie have just pre-released their Album Sudhaus Sessions on Bandcamp. I recorded their music in a live session setup with almost 40 channels, eventually resulting in 56 channels and buses of processing in the mixdown session happening in Mixbus 10 Pro.
Turns out, as a musician, I'm also very much driven by sound as such. Thus my love for synthesizers. But also when I sit down at the piano, I sometimes might just push the sustain pedal and hit one note and listen to all the strings resonating. I'm very lucky that with Universum ORS, I found two other guys that share this type of fascination.
You told me that you enjoy working with hardware and do so a lot. What technologies do you use to connect the hardware to the computer? Just an audio interface and a DAW, or are there other things involved?
I already talked about my means of audio I/O. So one other thing I use is MIDI. I have a simple M-Audio MIDI-USB interface hooked up to a vintage Kawai MAV-8 MIDI patchbay. Some instruments, such as Waldorf Rocket, are hooked up to my PC via USB, and so is a simple master keyboard. At some point, I made the decision not to buy any synthesizers or keyboards without a 5-pin MIDI connector.
Even more important is distributing MIDI clock to FX units and arpeggiators. Even my 01v96i mixing console can sync its delay FX to MIDI Clock. I use that delay a lot.
Actually, when generating sound, I use mostly outboard gear. Software synthesizers do not appeal to me. Even when using VCV Rack, I'm fascinated by its countless sequencer modules, but not very much by its oscillator or filter modules. So I route out MIDI notes and clock from VCV Rack. It works well as a drum sampler, though. I'm in the lucky situation of owning a small collection of hardware synthesizers, each of them having a very different character. Not saying that using hardware synths is problem-free, especially when they reach the vintage age, you might need to upgrade your electronics repair skills.
What also fascinates me a lot is tweaking natural sounds with crazy FX, such as the Soma Cosmos Drifting Memory Station or even a semi-damaged Uher tape machine. But this now is not related to Linux at all.