What progress has Linux audio made in 2025?

[uOpt]

Regardless of whether you think that Linux audio is currently good or bad overall, what progress do you see looking back to 2025?

[uOpt]

As you know, I am generally in favor of Linux, but I see very little in 2025 that advances the cause.

Fender Studio still has that crippled Linux beta and no new ports came out. Harrison probably made a few steps back in their Linux support, not that they answer questions about it. Not aware of major new open source software except from Surge Team, which really carry us right now.

[audiojunkie]

Regardless of whether you think that Linux audio is currently good or bad overall, what progress do you see looking back to 2025?

There has been a ton of progress! I've mentioned several times through various threads many, many things that have improved. Do a search with the KVR search option, and look for Linux as a search word, or search through my posts. You'll find many things listed.

In the meantime, off hand, there is no need for custom realtime kernels anymore--the realtime kernel patches have finally been mainstreamed. Pipewire has replaced JACK and PulseAudio and setting up and configuring your audio is much easier and faster. Prior to some time around 2017 there was no Reaper, no Bitwig, no Tracktion Waveform, no Fender Studio Pro, no native Linux support in the VSTSDK, no CLAP support, much fewer apps, no yabridge, no LinVST3, no audio specific Linux kernel boot configuration parameters, less documentation........ there's much more, but I can't remember off hand. So much has changed that it is a completely different environment than it was just 10 years ago. There is no comparison that even closely matches how much things have changed for the better. Back then, you had to use JACK to connect multiple standalone applications (not plugins) together to make music. Now, these days, Linux stands on equal footing with Windows and MacOS in pretty much everything except the number of applications. Vendors have even been adding class compliant modes to their audio interfaces specifically for Linux users (we don't need special audio drivers). There is so much more hardware available because of this, because class compliance is an accepted standard. Things are much, much better.

[Tiles]

That’s a very generous interpretation of the current state of Linux audio.

There has definitely been progress, but you’re as usual overstating what that progress actually translates to in practice.

Upstreamed realtime patches and PipeWire are real improvements, but they don’t remove the need for tuning or give you the same level of stability and predictability as other platforms. PipeWire improves usability, but it’s still catching up to long-established pro workflows.

The DAW and plugin situation is where your argument really breaks down. Having REAPER or Bitwig Studio is a step forward, but the absence of major tools like Ableton Live, Pro Tools, or Logic Pro defines the platform. The same applies to plugins: native support exists, but the commercial ecosystem is still largely missing. Tools like yabridge help, but a compatibility layer is not equivalent to native support.

Hardware is not “equal footing” either. Class-compliant USB helps with basic I/O, but that’s just a lowest common denominator. On Windows and macOS you get dedicated drivers with better latency, stable performance, DSP integration, and full control software. On Linux, you often lose vendor-specific features entirely or rely on partial reverse-engineered support.

Look at real workflows. On Windows, you install your DAW, add plugins from vendors like Native Instruments or iZotope, install the official driver, and you’re in a fully supported environment. On Linux, you’re working with a limited native stack and bridging the rest, with no vendor support when something breaks.

So yes, compared to 10 years ago, Linux audio is in a much better place. But “equal footing” ignores the core issue: ecosystem, vendor support, hardware integration, and real-world workflows still lag significantly behind.

So when your question is if it is now the time to switch to Linux to make music, better think twice.

[audiojunkie]

As you know, I am generally in favor of Linux, but I see very little in 2025 that advances the cause.

Fender Studio still has that crippled Linux beta and no new ports came out. Harrison probably made a few steps back in their Linux support, not that they answer questions about it. Not aware of major new open source software except from Surge Team, which really carry us right now.

You are looking in the wrong places for the software and news. There has been leaps and bounds more software in the last year. There is stuff coming out almost daily. Bookmark and then check some of these sources daily:

* https://www.kvraudio.com/plugins/linux/newest

* https://linuxdaw.org/

* https://linuxmusic.rocks/

* https://linuxmusicians.com/index.php?si ... 9f8a60d61a

* https://www.reddit.com/r/linuxaudio/

* If you are on Mastodon in the Fediverse, you can follow #LinuxAudio also. New software is posted almost daily there too. Copy this link and check daily:

https://mstdn.social/tags/LinuxAudio

* viewtopic.php?p=9230836#p9230836

* While largely about Ardour, lots of new software and news is posted on the Ardour forum as well:

https://discourse.ardour.org/

There are lots more, but this should be enough to get you started and convince you that there is a ton of software coming out for Linux.

[Tiles]

That’s not really addressing the point, you’re just listing sources of activity.

No one is denying that there is software being released. The question is whether that translates into a comparable ecosystem.

Most of what you’ll find on those sites is small projects, niche tools, or early-stage plugins. That’s fine, but it’s not the same as broad support from established vendors with maintained, production-grade tools.

The gap isn’t “is there new software coming out”, the gap is who is not there. Companies like Native Instruments, iZotope, or Waves Audio are still largely absent with native, supported Linux offerings. The same goes for flagship DAWs like Ableton Live or Pro Tools.

An active hobbyist and indie scene is not the same thing as industry adoption.

If anything, your list highlights the difference: on Linux you need to track multiple community sources daily to discover what might work. On Windows or macOS, the mainstream ecosystem is directly supported by the vendors themselves.

So again, yes, there is more activity than before. But activity is not parity, and it doesn’t put Linux on equal footing in terms of ecosystem, support, or real-world production use.

[ampetrosillo]

Tiles is overly negative (and AI driven). Linux is not a viable platform for prosumer use, unless your needs align with what's on offer. Interestingly, it is much better for traditional studio workflows, where you have a fixed installation with a lot of outboard gear, and the DAW is used as a fancy tape machine + audio router + automation. Latencies on Linux are routinely extremely low, not that it matters much for the aforementioned workflows. Is there some clunkiness? Yes. But this kind of use really only requires the computer to be an appliance.

[Tiles]

This is not about whether Linux audio has improved. It has. The question is ecosystem parity, not progress in isolation.

Written with the help of ChatGPT, at least from what i remember.

[uOpt]

In the meantime, off hand, there is no need for custom realtime kernels anymore--the realtime kernel patches have finally been mainstreamed.

Technically that was a 2024 change (Linux 6.12), but I'll let it stand.

Is any distribution actually using PREEMPT_RT for audio desktop? Last time I checked e.g. the NVidia drivers refused to work with PREEMPT_RT. It was a little rocky.

[uOpt]

This is not about whether Linux audio has improved. It has. The question is ecosystem parity, not progress in isolation.

Well, the question I started the thread with was not about parity, in fact I wanted to avoid repeating the same discussion going on in two other places here right now.

I am specifically curious about progress in 2025, or let's say the last 18 months.

And I think there are good audio interfaces for class-compliant Linux duties (among them my Harrison 32Ci), but I don't think anything was new in 2025.

[audiojunkie]

In the meantime, off hand, there is no need for custom realtime kernels anymore--the realtime kernel patches have finally been mainstreamed.

Technically that was a 2024 change (Linux 6.12), but I'll let it stand.

Is any distribution actually using PREEMPT_RT for audio desktop? Last time I checked e.g. the NVidia drivers refused to work with PREEMPT_RT. It was a little rocky.

I'm sure there probably are. I would especially imagine the Debian/Ubuntu family probably still have default realtime or low latency kernels in their repositories. There is also the Liquorix kernel, as well as a couple of others (although they may be just configured for low latency audio rather than specifically using preempt_rt).

I believe the Arch family still has a realtime kernel, and if I'm not mistaken, they also use the Xen kernel by default, so that should already have improvements over a user that isn't using any of the realtime kernel boot parameters in a generic kernel.

Unfortunately, I don't know much about the OpenSuse distro, as far as what kernels they use.

I may be wrong now, but Fedora hasn't been using realtime kernels as far as I am aware. I've always used (and preferred) generic kernels with realtime kernel boot options.

As far as distributions using realtime kernels, I believe the turn-key purpose designed distros do, such as AV Linux, and LibraZik. Ubuntu Studio uses a low latency kernel similar to what would be found if you used a generic kernel and the realtime kernel boot paramaters--or at least that was the direction they were trying to head to not have to support maintaining a kernel.

[uOpt]

Yes, low latency. But that's not PREEMPT_RT.

Has anybody even given it a comparitive spin with audio?

[audiojunkie]

OK, because you wanted to know, I did more digging specifically for you. Here are a couple of other things that are new for Linux support:

* The Joyned MU16 (The first Class Compliant, Milan-Compliant USB Audio Interface) was relased in November 2025. It is a world first: a class-compliant USB-to-Milan/AVB interface. It enables Linux systems to interface directly with professional audio Milan networks.

It doesn't require drivers for macOS, Windows or Linux. Networked pro audio was previously unreachable from Linux laptops.

* The MOTU 16A (2025 redesign) is fully class compliant and was released in April of 2025. It is a high-channel-count fully class-compliant pro studio interface that no longer requires vendor drivers for basic operation.

* MOTU 848 (First audio interface designed specifically to operate as a class compliant device over USB4, not just as a USB2 fallback) That means that it's not going to be limited to USB2 speeds, but the full speed of USB4.

* Historically, Antelope locked users into proprietary drivers. However, Antelope changed its model with the Antelope Audio Zenith 2 device released in November of 2025. It supports macOS/Windows/iOS and Linux, unofficially through Class Compliance.

* While Yamaha has historically shipped excellent class-compliant hardware, the Yamaha UR-X explicitly not only gave the device a Class compliant fallback mode, but also made the DSP optional (meaning it doesn't have to be added or used), which is exactly how Linux needs it.

But that's not all. There have been new Midi controllers too:

* The Akai MPK Mini IV (USB-C) model was released in 2025. It works with ALSA Midi right out of the box.

* The Novation Launchkey Mini MK4 / Launchkey MNK4 series removed driver dependency entirely. There is no configuration utility required at all when switched to the correct mode. You have access to deep DAW features via standard MIDI messages too.

* The AMC3+ Touch (100mm fader MIDI controller) was also released (June 2025). This has no software dependencies at all, and is USB class compliant and has on device configuration!

Probably the biggest thing out of all of this though, is that between 2025-2026, we've seen that USB-C and USB4 interfaces are beginning to DEFAULT to class-compliant firmware. This is a much bigger thing than just stating some hardware that was released, because it forms a foundation for future releases. Vendors are increasingly shipping "driver-optional" rather than "driver-required" hardware.

[audiojunkie]

Yes, low latency. But that's not PREEMPT_RT.

Has anybody even given it a comparitive spin with audio?

Let me be more specific. I meant "Low Latency Audio", not Low Latency in general.

Every single person who is using Ubuntu Studio (and most probably most other Ubuntu derivatives) are using the Low Latency configured kernel.

Many people are getting sub 5 ms latencies (ROUND TRIP!!) Not everyone, of course, but many.

[uOpt]

Yes, low latency. But that's not PREEMPT_RT.

Has anybody even given it a comparitive spin with audio?

Let me be more specific. I meant "Low Latency Audio", not Low Latency in general.

Every single person who is using Ubuntu Studio (and most probably most other Ubuntu derivatives) are using the Low Latency configured kernel.

Many people are getting sub 5 ms latencies (ROUND TRIP!!) Not everyone, of course, but many.

Of course. But that is nothing new that came with kernel 6.12.

A lack of specific testing also makes it impossible to figure whether the normal preemption models really have a problem that can be solved with PREEMPT_RT.