Hey Ramses,

i know this and tried that, even with low latency kernels it is impossible to get low enough while recording 40plus tracks at the same time.
Unfortunately we do not need low latency, we need no latency ;-)
And a.f.a.i.k that can only be provided by Totalmix.

Arno

Hey Ramses,

you are wrong, no latency is possible with totalmix because its core routines are implemented into the rme hardware.

i can use Reaper on Linux mint to record, i just need to use a safe buffer size. Why should i use a real time kernel if
it does not help me to avoid latency at all ?  Totalmix is doing exactly what i want, just like a hardware desk.
I just do not want to argue with my clients, i want them to know there is no latency, giving  them the peace of mind they need to perform as good as possible. 

Arno

To your other topic (which is not a linux topic).

It has been discussed several times in the forum that there will be no VST support in TotalMix FX, e.g.
https://forum.rme-audio.de/viewtopic.php?id=28535
https://forum.rme-audio.de/viewtopic.php?id=36598

Also see RMEs answer:
https://forum.rme-audio.de/viewtopic.ph … 13#p139913

In addition to what has been said in those threads, implementing it wouldn't be straightforward at all.

1. There is no actual audio passing through TotalMix FX.
TotalMix acts as an ASIO client and controls routing and certain settings on the recording interfaces — nothing more.
Trying to route all audio flows through TotalMix FX would be completely counterproductive in my opinion.
You’d end up with the same latency problems as when routing through the DAW — so you might as well just do it in the DAW directly.

2. Loading VST plugins directly onto the recording interfaces would also have a major impact on the system design.
Even if you were to go that route, you’d run into the same limitations as UAD and would need considerable onboard DSP power.
And at that point, you can forget about using standard third-party plugins — they would all have to be ported to the specific DSP architecture. How is that supposed to work?

If you don’t want to handle it in the DAW, just use a tool like LiveProfessor and host standard VSTs there.

Thanks for the clarification. Yes, my statements were not sharp enough.

I have thought a little further on that topic.

The question here is also what the expectation is.

Scenario 1:

Existing VST2/VST3 plugins could not be loaded into the FPGA or DSP on the recording interface, as they are not binary compatible.
So, if it is to run on the hardware, then VSTs would need to be developed and compiled for the particular DSP hardware.
The next thing is that it would possibly involve a hell of a lot of work to implement this with the flexibility of a DAW, where you can choose to use the VST as an insert or a send effect.
Another point is that it is rather unlikely that it would be usable on any existing RME device, as you would need additional DSPs for that. Maybe even more powerful FPGA.
So, all these efforts would not give any benefit for customers for their existing hardware and any potential new hardware would be a lot more expensive due to additional components needed and the development of new/additional VST that can be loaded into the additional DSPs.

Scenario 2:

If you want to use existing VST2/3 plugins, then they have to be loaded by TotalMix FX on the PC.
Then the audio flow would need to be routed through the PC.
You would also need many additional channels for the implementation of insert and send effect and the communication between recording interface and PC where those VST would be executed.
Look at the Babyface Pro, inserts (EQ) are implemented in the FPGA, but delay could only be implemented as a send effect which is shared between all of the channels and needs an additional (hidden) channel between recording interface and PC.
For multiple VSTs as inserts and sends you would need a lot of additional channels.
I think this would definitely weaken the stability and concept of the current mature TM FX design.
I would also introduce additional latency that can be quite high depending on the buffer size that you choose.

I started working in AI and needed to use Ubuntu, and I am baffled to find out that RME doesn't support Linux...

Apparently since punkrockdude open this thread in 2018... Wow!?

I read the whole threat, and realized the problem. But also learned that some "other" companies managed to support Linux
with their audio drivers... So, it is still a little ludicrous...

Is there any alternative solution to this absurd situation, like a third-party ASIO driver or something like that that can fix this problem?

I won't be recording anything with it, only playback... I have a UFX...

I'm aware of class-compliant protocol; however, unless I have no choice, I would want to avoid switching back and forth...

cheers

Well, I do not know in OP´s case, but AI could do many things in respect to sound. AI is, beside others, used for signal processing, signal/speach analysis and signal/speech synthesis. And as an aid to setup different plugins....

I am not on Linux, but from what I know, you may use Class Compliant mode of RME interfaces on Linux. Many of RME interfaces do support CC, but not all.

Maybe, worth a check for users owning units listed below and using firewire connection on x86_64/amd64 linux machines (the dependancies for arm64  linux machines need still some polish)...

Check out https://github.com/alsa-project/snd-fir … l-services

Models currently supported:
- Fireface 800
- Fireface 400
- Fireface UCX
- Fireface 802

Any help is welcome for:
- Fireface UFX

More details: https://github.com/alsa-project/snd-fir … s/fireface

I would also love official RME drivers for Linux. Especially since in 2025, more and more people are making the switch to Linux 100%. Yeah, we have tools like WinApps and VMs, but let's face it, it's not as great as having a native solution.
RME makes rock-solid sound cards; I would be sad to switch over to another company just because the drivers aren't available, especially since options with 32 to 64 I/Os are quite rare, which the Digiface USB permits us to have.
Did the position change yet? Are we soon going to be able to just ditch our OS to finally be able to be 100% on Linux?

Just doing the same as above here, would love to see official support as it's the main thing stopping me leaving Windows behind at the moment. It's great that people are making their own solutions but I'd like a bit more of a stable base than that.

I was amazed to find that my Babyface Pro FS and UCX II had better latency than Windows despite being in class compliant mode with zero tweaks done in the Linux OS. I imagine it'd rival MacOS with dedicated drivers.. Proper access to Totalmix is the bigger thing though.