Switched to Linux... how does this work?

I don't have the answer to these questions, particularly not to the "count as one" vs "count as two" license. But I can tell that Modartt is always very helpful, so don't worry about it. I suggest you can start by installing the demo version (which you can then make it "full" by activating the license) and start trying that. Maybe that activation would even automatically work by putting in the activation code (copyable from the top-right of your user area) the without "eating" another seat, I don't know.

FWIW, I've found that the Linux version works like a charm. If you find that's the case for you too, you'll figure it out later the mechanics of what you need to do in your user area.

Hi, Bellyman! I am glad you are happy with Pianoteq 9 on your new Linux-based system. I run Linux full time for over two years, and I have to say it is great. I run OpenMandriva Rock, and Pianoteq works great here.

Have you checked the README_LINUX.txt file? There are instructions there on how to edit a system file, to help with crackling and latency on the system. It solves all issues for me with crackling, and I can put latency really low too if I want. I usually play with 20ms delay, though. It just feels better to me, to be honest.

Cheers! Welcome to Linux.

I am so happy both you and your wife are liberated from the gates of heck. I am really happy that Pianoteq runs natively here on Linux, and it is a big reason as to why I got it for myself, and am using it now as my only piano sound plugin. It works perfectly for me.

You should have a smooth sailing time with Zorin. I prefer a different looking working environment, and for a few other reasons I am using what I am using. It is working great.

Awesome, Beco!

I would be curious to learn more about the other voices in your orchestra, how you put that all together. I have done some jazz trio / quartet kinds of things but haven't done any of that with Pianoteq as of yet. I also have been a user of Reaper (DAW) which needs to be loaded on Linux as well (used it on Windows) and have used that just for some simple stuff (piano demos, looking for gigs). But it would be fun to do a bit more complex things as well.

I'd be interested if you'd care to share. 

Very nice! Thank you for sharing!

I didn't quite understand your earlier post and had thought you created the orchestra electronically and played along with that with Pianoteq. You had an actual orchestra!   

Sound came through quite nicely. I have never been happy with the sound from my phone even though I use the video from it. I've never tried mixing / blending the MIDI rendering from Pianoteq with the audio from the phone's recording, though. Interesting idea.

Do you know which piano sound you used on Pianoteq? Do you have a favorite? I play a way different genre, typically jazz standards from the 1930's to 1950's or so, often the more melodic show tunes made popular in old movies. For those, my go-to has been the Bosendorfer for the past year but am really impressed with a few of the other piano offerings in Pianoteq 9. Pianoteq has come a long way and I like it better all the time. Have had numerous compliments on it when playing it in gigs.

Again, thanks for sharing!

Hi there,

i ran PianoTeq8 since several month on debian12 flawlessly. Today i installed PTQ9, I didn't do a lot of testing yet, but i have a issue with fullscreen-mode. PTQ9 crashes when i try to switch to full screen. (bug report is send to modartt already)

Maybe some of you linux guys can check if full screen works for you. (?)

Regards
Bazzo

I guess, there are not a lot of people caring for fullscreen mode. I mean, i can live without it, but in my usecase, the machine is exclusively for PTQ and it used to boot directly into PTQ8 in fullscreen mode.

I mean the "Fullscreen" Menu-->Window-->Toggle fullscreen mode / (Ctrl+Shift+F).

Regards
Bazzo

Thanks for testing. I tested F11, it does the same as (Ctrl+Shift+F) on PTQ8, and it also crashes PTQ9.

Hi all,

i upgraded (re-installed) my setup to debian 13 an ran PTQ9 in cage/wayland. (Minimal Window Manger)
The same crash occurs. But i noticed something strange, I can toggle Fullscreen while the app is running as long as the UI is not in "Compact Mode". I can toggle "Compact Mode" as long as its not in "Fullscreen Mode". But both together will trigger the crash. Nonetheless, when i start with --fullscreen option, it doesn't start at all and outputs no errors.

Regards
Bazzo

Hi bazzo,

I am not any expert, but when I have some problem like that, I usually run the app from terminal to check what it says. After that you will have some answers about the problem.

Hope this will help you.

Best regards.

First, congratulations on your success!
Second, welcome to the Linux world!
I've been using Ubuntu Studio for five years and I don't plan on ever going back to the slavery of Windows!

I've been running Pianoteq on Linux for I guess over 10 years now. My preferred Linux "flavor" (distribution) for many years is Linux Mint, based on Ubuntu (which is itself based on Debian). Ubuntu Studio is very nice, but it's based on the KDE graphical environment and user workspace, and I'm not very fond of KDE (it's just a personal preference; Linux Mint has its own graphical environment called Cinnamon, and also has variants for several other graphical environments (including XFCE, my favorite (so far) low-demand, non-CPU-hungry graphical environment).

Pianoteq is exceptionally well-designed and compatible with Linux. The main issues are not with Pianoteq itself, but in tuning Linux for audio use. Linux used to use ALSA and Pulse as the audio-management backend, but most Linux distributions (including Zorin) have now switched to Pipewire, which is considered superior. So learning a little about how to set up the system and Pipewire is helpful. There are many useful resources on the Internet, and the user forum at LinuxMusicians.com can be very helpful.
LinuxMusicians Forum: https://linuxmusicians.com/index.php

The main fine tuning for Linux is to install a "realtime" or low-latency kernel (the heart of the Linux OS) that does (or can be configured to) prioritize audio and other tasks that need very fast response times from the kernel. It also helps to eliminate CPU frequency-scaling or to set it to operate a maximum by choosing the "performance" CPU governor.

For conveniently setting the CPU governor, I recommend the app cpupower-gui:
cpupower-gui: https://github.com/vagnum08/cpupower-gui

The cpupower-gui app is usually included in the list of packages for the package manager your "flavor" of Linux uses (Zorin I believe uses the Debian based package managers (apt, apt-get, synaptic) that are used by many Linux distributions such as Ubuntu and Linux Mint (my personal favorite).

--- --- --- --- ---

Note: Ignore the following and refer instead to Vepece's post below--

The linux-lowlatency meta-package (a package that includes and installs several related packages/apps) is a popular option to replace the standard linux kernel with a realtime/low-latency version. The Liquorix linux kernel (originally produced by gamers for gamers, but also widely used for audio work in Linux) also works great.
Liquorix kernel: https://liquorix.net/

Be aware than anytime a new linux kernel is installed, it's necessary to update grub (GRUB is the bootloader, the app that begins at bootup and allows you to choose which linux kernel you wish to use/launch, if not the default). Most of the time, when installing a linux kernel from a package manager, the GRUB bootloader is automatically updated by the installer, but if not (or if you didn't look at the output of the installer to double check, and want to make sure) you could update it manually, using the "sudo" command that temporarily elevates your linux username to "superuser" or administrator) using the commands suggested by Google AI below:

After installation, update GRUB to ensure the new kernel is recognized:
Code

sudo update-grub

Hey,

because I read good intended advise like "Have you checked the README_LINUX.txt file?" and similar thoughts like "Have wondered whether I might have liked Ubuntu Studio a little better for what music stuff I do on the computer." and advise like "1. Installing the Low-Latency Kernel:" I like to share my concern and advise:

The README.txt is obsolete! So are 99% of suggestions in the forum.
This comes from people repeating again and again what they pick up along the way - which is prone to be outdated knowledge. Ive actually participated back in 2023 by myself, trying LowLatency Kernels, 1000hz Rates and different settings back than when pipewire was still in its roots.

But you can skip everything youve read before and take the following as your new "base camp" for further exploration:

So why is the README/suggestions here outdated?
They do not take into regard that Pipewire is de facto the new standard - exp. with Distro Flavors like Ubuntu Studio.
Therefor, the "@audio" usergroup is of no concern any longer - since Pipewire 0.3, since Ubuntu 22 LTS!, since more than 3 years ...

The new usergroup we need to care about is "pipewire" and we need to build around that.

First of all, make sure you run pipewire.
Check if "/etc/security/limits.d/25-pw-rlimits.conf" exists.
If so, you will see it cointains the rtprio, memlock and nice values.

Nice is obsolete: nice (POSIX-Nice) is for RT-Threads (SCHED_FIFO/RR) irrelevant.
But we want to use newer scheduler, which is why the entire config_hz=1000 (aka lowlatency kernel, Ubuntu Studio, etc.) is obsolete too:
We use newer Kernel flags to dynamically make our kernel highres/lowlatency when needed - actually we make them tickless which is beyond 1000hz... - so a regular Linux Kernel >6.11?/6.14 does not need further tinkering.

To check if we got what we are looking for:

$ grep PREEMPT /boot/config-$(uname -r)

look for "CONFIG_PREEMPT_DYNAMIC=y" -> your kernel is dynamic.

So: Your Pipewire makes Jack/audio Group obsolete, and your Kernel the 1000hz. All it takes is to check if your user is in the pipewire Group:

$ groups $USER

if your in pipewire group, your set. if not, add yourself.

Now that we know we are in pipewire group, check your RTprio levels!

$ ps -eLo pid,cls,rtprio,ni,cmd | grep pipewire

If you see FF or RR as designation, this means he tries to acquire the SCHED_FIFO Scheduler and seeks RealTimePriority.
If you see a value of >20, (80 to 90 to 95) you know the pipewire group RTprio are working!

But if you see TS, you know he tries to connect via RTkit -> User not in Pipewire usergroup.
If your value is <=20, even when its in FF or RR, you know your accessing pipewire by rtkit, which is way slower in priority and means greater latency. -> Rtprio via Pipewire Usergroup not working.

Summary:
If you run a stock Ubuntu 24, NixOS 24, or any other modern Distro with pipewire, all it needs is to add your user to the pipewire group - no Ubuntu Studio, and not tinkering with creating audio groups or kernel flags/schedulers/rates is required.

The rest of the README regarding using a stable cpu frequency and performance scheduler are legit, and can be achieved by the suggested commands, or by GUI with packages like CPUPower-GUI which offer convenient profiles. https://github.com/vagnum08/cpupower-gui

Another suggestion the README does not suggest is to use taskset as explained here: https://forum.modartt.com/viewtopic.php...3#p974513)

Another suggestion the README does not contain is that pipewire-jack bridge is no default with many modern Distros, so installing that package, using the JACK Server within Pianoteq, setting up your global "Jack Server" Parameters in Pipewire, optimizing Sample Rates and Buffer Size, to harmonize/optimize the Quantum , and therefor "syncing the timing" in the µs range to fit each other optimal is not explained so far (Im not even sure if the forum contains it?).

Cheers

Hey Stephan,

you are trying to explain the basics of Linux there, mh?
While we are at it: "in a terminal window" -> in a Terminal Emulator There is no Terminal any longer.

Ive actually never stated why we want PW-Jack and whats about the quantum and rates:
Long story short is that when using the PW-Jack Interface, you have more features than with PW-ALSA.
In addition, I use it as kind of "Group" for Apps in Music Production: With using the PW-Jack Server everything connected via PW-JACK will run the same rates and quantum, which can be set globally.

So for the sake of completion to my prior advise:

1. Determ best Audio Output (aka Soundsink):
$ wpctl status           -> list all devices. Check your Soundsinks. They list some ID.
$ pw-cli enum-params <ID> EnumFormat    -> see supported formats/rates. The best is true F32LE (32Bit Float -> native Signal stays the same as long as possible) if you want to output to a digital device (like HDMI, DisplayPort, SPDIF, USB Audio). Some actual analog output can only use S32 (String Integer), so you want that. Many consumer grade devices, like my Dockingstation, or the CPUs own I/O Pins only allow for S24_32 which is the best you can do. Everything less than this (S24, F16, etc.) will have less and less "quality" and requires computation

$ pw-cli enum-params <ID> Format        -> check the actual used Format when a Soundsink is Playing sound.

-> Do this to find at best some F32 or S32 Output device and use it as your output (can be set by various methods, I like Helvum as Patchbay for that) to spare all unnecessary computation.

2. Setup PW-Jack Parameter:
https://docs.pipewire.org/page_man_pipe...hatgpt.com

Create directory for custom config files
$ mkdir -p ~/.config/pipewire/jack.conf.d
$ nano ~/.config/pipewire/jack.conf.d/YOUR-CUSTOM-NAME.conf

This is my actual file content:

jack.properties = {
  rt.prio = 88
  node.lock-quantum = true
  node.force-quantum = 128
  node.latency = 128/48000
  jack.show-monitor = true
  jack.merge-monitor = true
  jack.show-midi = true
}

Save the file - done.
(The name of the file doesnt matter at all, every file in the folder will be loaded and you can create multiple ones).

-> Restart the System (systemctl restart services, logout/login, ... they should work, but safest bet is to restart).

3. Use PW-JACK
use "pw-jack" command to start your Apps like PianoTeq. I use .desktop files anyway for my start menu, therefor it was easy to just put "pw-jack" infront the exec path.
To save you some trouble, this is my actual /home/USER/.local/share/applications/PWJACK_PTQv843.desktop file:

[Desktop Entry]
Encoding=UTF-8
Version=1.0
Type=Application
Terminal=false
Name=PWJACK Pianoteq 8 STAGE
GenericName=Pianoteq 8 STAGE
Comment=Selfmade Desktop Shortcut for Pianoteq 8
Exec=pw-jack "/home/USER/Music/MX/Programs/PianoTeq/PianoTeq_8_Stage_843/x86-64bit/Pianoteq 8 STAGE" %F
Path=/home/USER/Music/MX/Programs/PianoTeq/PianoTeq_8_Stage_843/x86-64bit/
Icon=/home/USER/Music/MX/Programs/PianoTeq/Extras/modartt.png
StartupNotify=true
StartupWMClass=Pianoteq STAGE

4. Enjoy your Harmonized Production!

5. Double Check
use
$ pw-top
to check for the Quantum (Buffer Size) and Rate (Sample Rate) and Format (F32, etc) of your applications. PW-JACK Applications wont show any format -> they run F32LE globally.

Check in PTQ Options for Buffer Size and Sample Rate -> if working properly, it must show the values from ~/.config/pipewire/jack.conf.d/YOUR-CUSTOM-NAME.conf

Why is this solution my preferred?
My non-Music-Production Apps like Firefox, Spotify, VLC Player ... do not get resampled and use native Pipewire (or PW-Pulse or PW-ALSA) -> no performance loss, no quality loss, no glitching.
PTQ, Ardour, etc. use PW-JACK and run at the Rates/Quantum I desire.

Which Rate and Quantum you pick is another question and is smth. I would love to find other "geeks" explaining why their calculation is the best.

PTQ uses Quantum 64, therefor we would like to use 64.

48kHz Rate on the other hand would divide Quantum of 48 to a perfect 1ms block size (48/48000Hz=0,001Hz=0,001s=1ms).

USB2 SuperSpeed Poll Rates of 1kHz (48/1=48) would match this quite well too, newer USB3 8kHz Hardware (its quite some time ago that I researched the MIDI-USB specs, were they even allowing for 8kHz?) would match nicely too (48/8=6) so we have nice whole numbers and no integers which take more computation to solve neatly.

That is actually the reason why we dont want 44100Hz, or the Quantum 64.

That is why we have to find gcd/gcm and look for other combinations that are elegant: The easiest solution would be to increase the Rate to smth like 384kHz and than enjoy that many natural numbers would fit as factor/divisor to receive some neat Quantum value, but this idea is limited by actual computational power.

The 128/48000 on my Laptop offers great Performance - but the latency is quite weird with 128/48000=0,002666s=2.66ms and therefor mismatches with its blocks other I/O blocks like the ones by the USB-MIDI controller.

But thats maybe just me and my desire for elegant numbers - as Im not aware of too many people probing with lsusb -v -d what Intervals/Poll Rates their MIDI Controller actually use, in an attempt to optimize for that.

Cheers