I have a Scarlett Solo and I have found that with the linux-rt kernel, I get no noticeable latency recording audio, whereas there’s a few ms delay on the main linux kernel.
I am able to run other applications and play games just fine on linux-rt. Is there any reason not to just make this my default?
Correct me I’m wrong, but i think in practical sense of real time kernels is that you could have a program take over much of the processing power, limiting your ability to multi task. I believe this raises security and stability concerns for every day use, but is fine for Audio.
Keep in mind you can have multiple kernels installed and just choose which one to boot into. This way you can save the rt kernel for your audio stuff and then switch back into the normal kernel when you’re done.
As an anecdote, I use the Scarlet with Guitarix and dont notice any difference between the RT and normal kernel myself. Both have had little latency for me and the RT Kernel seemed stable enough. It’s possible you may have something running the background that’s holding you up in the normal kernel. Read the Arch Wiki article on performance enhancements for Audio https://wiki.archlinux.org/title/Professional_audio