The rise of Linux desktop is inevitable — it’s time music software developers got on board

[Tiles]

Hopefully not nvidea graphics. Those aren't well-supported on Linux.

AMD neither -.-

[Zeisner]

Why three?

To rule out issues regarding the lack of driver support. And to see if I can replicate the most annoying issues (Slowdowns, lags, freezes, crashes). They can.

Hopefully not nvidea graphics. Those aren't well-supported on Linux.

NVIDIA, AMD and Intel. The two better laptops feature two graphics chips/cards each. Plenty of options to test. Never had much issues with NVIDIA but this wouldn't make me claim that no other Linux user can have any issues with those. Just because something works on my systems doesn't mean it is guaranteed to work everywhere else.

[j_e_g]

pipewire is a step forward.

Noooooooooo. In fact, it couldn't have come at a worse time. All of the major linux native apps had added direct ALSA support. And Steinberg had just made VST for linux, JACK had now been made irrelevant. People were finally able to dodge that configuration nightmare, and instead setup their music software exactly like they do on Win and MacOS. Configure audio I/O and effects (plugins) directly from their music apps. The learning curve for migrating musicians would be minimal. And music devs would have only one API (ALSA MMAP) to target/support.

It was getting easy for migrating musicians and devs. Then Poopwire comes around, and we're back to a separate "Connection Manager" and extra config files, and extra points of failure in the audio chain. It's a colossal setback.

Furthermore, even though we can't kill off Jack or PulseAudio, they also can't be separately maintained any more, because they're now part of PoopWire's code base. (I specifically asked this Wim guy how JACK and PulseAudio maintanence was going to proceed, and he didn't know. Obviously, he had no plans to do it after his initial cannibalization of the codebase,) That was stupid since he didn't need to cannabilize JACK at all. Poopwire could have worked in conjunction with JACK just the way PulseAudio did. All he needed to do was implement PA's "card request" DBUS message. Simple and safe, Noooooo. Instead he adds hundreds of lines of code (and points of failure/complexity) to PoopWire? We're never going to advance linux audio with these kinds of programmers.

No, PoopWire is the wrong approach, and at the wrong time.

[lunardigs]

Here you go. Listen to people like Linus go through the same kind of thing that you will experience.

Good video, and a great analogy here: You post a relevant video from about a day ago, while Tiles posts one from 4 years ago. Oh wait, but it's worse than that; the video he posts is from DebConf14 (2014) ...

[audiojunkie]

pipewire is a step forward.

Noooooooooo. In fact, it couldn't have come at a worse time. All of the major linux native apps had added direct ALSA support. And Steinberg had just made VST for linux, JACK had now been made irrelevant. People were finally able to dodge that configuration nightmare, and instead setup their music software exactly like they do on Win and MacOS. Configure audio I/O and effects (plugins) directly from their music apps. The learning curve for migrating musicians would be minimal. And music devs would have only one API (ALSA MMAP) to target/support.

It was getting easy for migrating musicians and devs. Then Poopwire comes around, and we're back to a separate "Connection Manager" and extra config files, and extra points of failure in the audio chain. It's a colossal setback.

Furthermore, even though we can't kill off Jack or PulseAudio, they also can't be separately maintained any more, because they're now part of PoopWire's code base. (I specifically asked this Wim guy how JACK and PulseAudio maintanence was going to proceed, and he didn't know. Obviously, he had no plans to do it after his initial cannibalization of the codebase,) That was stupid since he didn't need to cannabilize JACK at all. Poopwire could have worked in conjunction with JACK just the way PulseAudio did. All he needed to do was implement PA's "card request" DBUS message. Simple and safe, Noooooo. Instead he adds hundreds of lines of code (and points of failure/complexity) to PoopWire? We're never going to advance linux audio with these kinds of programmers.

No, PoopWire is the wrong approach, and at the wrong time.

Everyone has their own opinion on the subject. There's nothing wrong with having an opinion. However, please note that not everyone agrees with your assessment.




"I’m told I lie constantly when I say Linux is great — which is wild, because I’m apparently “misleading” you with things like: it boots, it runs my audio workflow, it stays stable, it gets updates, and it keeps existing despite being declared dead every year since forever. Also yes, it’s “too hard,” “too niche,” “no devs will ever care,” and “Linus complained once,” so clearly I must be fabricating my entire operating system out of pure delusion."
-----The Delusional Linux Advocate

[Largos]

pipewire is a step forward.

Noooooooooo. In fact, it couldn't have come at a worse time. All of the major linux native apps had added direct ALSA support. And Steinberg had just made VST for linux, JACK had now been made irrelevant. People were finally able to dodge that configuration nightmare, and instead setup their music software exactly like they do on Win and MacOS. Configure audio I/O and effects (plugins) directly from their music apps. The learning curve for migrating musicians would be minimal. And music devs would have only one API (ALSA MMAP) to target/support.

It was getting easy for migrating musicians and devs. Then Poopwire comes around, and we're back to a separate "Connection Manager" and extra config files, and extra points of failure in the audio chain. It's a colossal setback.

I can't think of anything that would put a "migrating musician" off more than the entire audio being taken over by one app but you keep ringing that alsa bell.

Also, pipewire allows configuration from apps. It's not a fault of pipewire that devs are not doing that. Of course, you'd know that if you actually used pipewire rather than just typing on forums about it

Furthermore, even though we can't kill off Jack or PulseAudio, they also can't be separately maintained any more, because they're now part of PoopWire's code base.

Neither of those has seen releases for years. Pipewire on the other hand, fixed ALSA's bad setup of firewire audio so FFADO is no longer needed. So, clearly willing to support out of date things.

[Zeisner]

We're never going to advance linux audio with these kinds of programmers.

Correct. Linux audio is not going to happen, just like Linux image/video processing or Linux office and so on. The Linux desktop has been a fairy tale/joke for three decades now and it will remain a joke for the next three decades.

[j_e_g]

I can't think of anything that would put a "migrating musician" off more than the entire audio being taken over by one app

Yeah, damn that Cubase, Logic, Sonar, Pro Tools, etc. Who the hell needs them? Apparently not linux users.

FFADO is no longer needed.

Neither is firewire. Buy an audio interface made in this century, you cheap git. Then the rest of us won't be saddled with your obsolete bloat.

[lunardigs]

We're never going to advance linux audio with these kinds of programmers.

Correct. Linux audio is not going to happen, just like Linux image/video processing or Linux office and so on. The Linux desktop has been a fairy tale/joke for three decades now and it will remain a joke for the next three decades.

Oh look, another non-user in here telling us what Linux is.
Amazing

[camsr]

FFADO is no longer needed.

Neither is firewire. Buy an audio interface made in this century, you cheap git. Then the rest of us won't be saddled with your obsolete bloat.

Firewire was the "best" then. It was replaced by Thunderbolt which is current.
I have no idea if any newer thunderbolt interfaces have full support in linux.

[Zeisner]

Oh look, another non-user in here telling us what Linux is.
Amazing

Not much to use if all the distros I tested slow down after 30-60 minutes until I have to wait for several minutes just for the start menu to open, even on a modern machine with a strong CPU and plenty of RAM. Some distros like Mint can be woken up again after standby/hibernation, others like CachyOS need to be restarted. The cache system is broken, just like the scheduler. Maybe compiling your own kernel is the solution (Something everyone highly recommended) but I don't care anymore. I rather go back to Atari TOS than trying to turn a Linux desktop distro into something useful. I don't want to tinker endlessly ("Best car ever!"), I want a tool that works if I need it.

[Zeisner]

Yeah, damn that Cubase, Logic, Sonar, Pro Tools, etc. Who the hell needs them? Apparently not linux users.

Of course not. Linux desktop fanboys don't actually work with their distros. They don't make music or mix or render 3D graphics and so on. They only tinker with the OS just like the crazy car guy with his unfinished "best car ever" that will never drive a single meter. Actually finishing it would defeat the purpose.

[stanlea]

Really guys you should take some time for making good music instead of wasting hearth ressources.

[GaryG]

Not much to use if all the distros I tested slow down after 30-60 minutes until I have to wait for several minutes just for the start menu to open

Well that's anecdotal evidence at best, i've never experienced that, been using linux on and off for a few years, exclusively for the last few months for hours at a time. many different distros tested...

Obviously if that's your experience then you'd be daft to continue of course but please try and appreciate that's not a common issue before writing linux off for everyone.

[j_e_g]

Firewire was the "best" then.

It was. But it's now obsolete. I don't think ALSA should be bothered with supporting it. That's just extra code to debug/maintain for the sake of a tiny (and diminishing) number of users. I was opposed to ALSA incorporating FFADO code. It's time to toss the firewire interface in the bin along with the floppy and SCSI drives, and toss away (deprecate) FFADO.

Today, one can buy a $200 to $300 interface with better specs and support. It's worth it.

If someone is still using a firewire interface, then he obviously has an ancient system. He should stick to an ancient version of linux, and not update it. If he's happy with the hardware, then he should be happy with the software already written/tested for that ancient hardware.

I have no idea if any newer thunderbolt interfaces have full support in linux.

It's kind of a moot point. I haven't seen an AMD or Intel currently sold motherboard with thunderbolt support (maybe through usb C, but rarely so).

I would recommend a linux user not invest in thunderbolt devices. And for FIREWIRE, it's a "Oh hell no".