Hacker News Reader

Comment by rektide | Hacker News Reader

Some really really good bluetooth improvements in here. Really nice. I'm ongoingly surprised how active PulseAudio releases have been in the past couple years.

It does make me a bit nervous for PipeWire. I want PipeWire to succeed, radically, but with PulseAudio continuing to make really good advances it feels like expectations/needs for PipeWire keep going up. The additional bluetooth capabilities here are an example. There's a bunch of new devices that have special handling by pulseaudio. The new udev based configuration makes pulseaudio even easier to configure. Hard to imagine replacing such a robust sound system!

View on HN

I'm not afraid for Pipewire future; they seem to be moving faster and delivering results sooner.

Even the bluetooth improvements, new for this release: they are already in Pipewire. On the other hand, it took Pulseaudio _two_ years to go through a merge request of a developer, that implemented it.

You can read the entire f-up for yourself: https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge...

Christ, what a clusterfuck. I will give them this though: they tell you what to expect in the CONTRIBUTING file instead of the usual bullshit about how welcome patches are.

Could someone explain to me why Linux needed all these different audio systems?

I remember when ALSA was the successor to OSS. But there were also others I can't name. JACK too, but I think that's still in use for professional sound production (not sure why anyone would use Linux for that). Then came PulseAudio which was The Future 2, which had a bunch of incompatibilities with existing programs that either caused latency or an absence of audio. But it got better over time. What was it, a little over a decade since PulseAudio got fully rolled out?

Now there's yet another project called PipeWire.

Frankly, this is one big reason I ditched Linux. The illusion of "choice" is more like a bunch of people just one-upping each other for the sake of it. I've seen it time and again with things like the FFMpeg vs Libav debacle, the various spinoffs of GIMP that didn't pan out, most desktop environments, and so on. Actually, at least there's sort of a better argument for infinite DEs, but I don't see audio needs being all that different between various users.

Fragmentation is definitely a problem on Linux systems, but I'll take freedom with the resulting fragmentation over someone making the choice for me. Especially when that choice is almost guaranteed to be suboptimal for some subset of users, with no practical alternative.

> Frankly, this is one big reason I ditched Linux.

If you want to not care, dont. Just run Fedora. You'll get a faster, lower latency, lower-resource audio daemon. Improvements happen, and the distros mostly make this all seamless for your user experience. You should work on your attitude foremost if for some reason improvement bothers you.

> more like a bunch of people just one-upping each other for the sake of it.

If they are one-upping each other, wouldn't that be according to some metrics or cause? Wouldn't that be a gain? Yes, we are trying to one up each other, exactly, for the sake of getting better. You should be delighted by this. Instead... ?

Choice is amazing. Freedom is amazing. The shifts & changes are signs of vibrancy & caring.

PipeWire giving us fast, pluggable audio & video transport with zero-copy- using new Linux DMA_BUF capabilities that were created after PulseAudio was began- is critical innovation. PipeWire is a drop in replacement for PulseAudio _and_ JACK.

I for one have always found the progress exciting & compelling, have enjoyed learning of what's happening, hearing how people think things could be better. I have a progressive outlook. To me, if you want a conservative mindset, that's fine: there are distributions which package & build good out of box experiences & that have solid upgrade paths to help you not have to worry. Typically they shield you from the jank, keep you from having to worry about choice or transitions. Nothing is perfect, and some users did & do experience problems with PulseAudio initially, but I believe the "silent majority" was enjoying a much improved working sound experience very quickly & happily, and that the dissent/reactionary-ism is outsized.

OSSv3 was a kernel API supported by a bunch of Unixoids including Linux. It had some limitations, and while OSSv4 addressed them OSSv4 was proprietary, and so on Linux it was replaced by a new thing: ALSA. (OSSv4 was eventually released under the GPL, 5 years later, but by then the ship had sailed.)

ALSA is low-level enough that it's necessary for a layer on top of it. Well, maybe… enough people think so. So that's where Pulse and Jack come in.

Pulse and Jack do the "same" thing, but target different use-cases and so have different design constraints; Pulse serving consumer audio and Jack serving pro audio. The creator of Pulse wrote a good article about this, about why Pulse and Jack aren't competing with eachother[1].

So for a long time that was the choice: Pulse or Jack (and whichever you choose, it's using ALSA on the backend).

The same article goes on to write "One question asks itself though: can we marry the two approaches? Yes, we probably can, MacOS has a unified approach for both uses. … let's not forget that we lack the manpower to even create such an audio chimera." Well, that's what PipeWire is. It's a system that fills both use-cases. Additionally, it implements both the PulseAudio and Jack APIs, and so users don't need to care which API a given application was implemented with, and can just use PipeWire as a drop-in replacement for both.

[1]: http://0pointer.de/blog/projects/when-pa-and-when-not.html

More reading: http://0pointer.de/blog/projects/guide-to-sound-apis.html

Before PulseAudio there was ESD.

https://en.wikipedia.org/wiki/Enlightened_Sound_Daemon

In my experience pipewire is better than pulse ever was.

Reply on news.ycombinator.com