"PulseAudio Is Still Awesome"

[fritsch]

On debian it's libasound2-plugins, that adds a pulse device and maps alsa's default device to it -> done. This is btw. installed by default on every "enduser" distribution. Debian should do the same.

[PiroXiline]

Pulse does't have modes. Likemode for bit perfect audio output from one source & other mode - mixing several streams.
That is why I use alsa devices for bit perfect output.
Web flash use only pulse, only first device. I think HTML5 is the same.
And since Flash grabs pulse even when nothing is playing, pulse grabs audio device. And pulse doesn't hand over device to alsa.
So when I want to play music - I need to kill pulse or close browser just to free audio device that was robbed.

[magika]

Do you have any data to back up that claim? I ask because I might be interested in doing some low latency audio stuff.

Backup what? Latency is fully configurable in JACK by design, see this https://wiki.archlinux.org/index.php/Pro_Audio

Mmh. That's full wrong. We don't need PA to run. If context does not open -> ALSA is used.

Thats not what I experienced so far. Take GuitarPro or Skype: if there is no PA there is no sound. However app that targets ALSA seamlessly integrates into PA setup.

And since Flash grabs pulse even when nothing is playing, pulse grabs audio device. And pulse doesn't hand over device to alsa.
So when I want to play music - I need to kill pulse or close browser just to free audio device that was robbed.

You can run PA alongside ALSA without pulse grabbing device https://wiki.archlinux.org/index.php...ardware_device

[GreatEmerald]

Pulse is immensely useful to me, since I do video recordings, for one, and have a pretty complex sound setup (four sound devices, one of which is DisplayPort audio).

One thing that's still quite annoying is that SimpleScreenRecorder still doesn't have mixing support, so it relies on some PA module trickery to achieve that, and it doesn't always work for some reason (I get sound corruption with noise and echoing). I'm not sure why. Maybe I should play around with the latency setting a bit.

Another issue that happened just today is that for some reason there's no audio playback going through my monitor via DisplayPort, but that's probably due to DisplayPort (its support has been spotty lately, with the monitor refusing to go to sleep, but that was just recently fixed, yet might have broken something else related to sound).

[GraysonPeddie]

I'm just wondering... Is there a way to make Google Chrome use ALSA Loopback? I use JACK as a sound server and an ALSA loopback that connects to JACK is good enough for me, so I would like to eliminate PulseAudio if it's possible. Without PulseAudio, I get no sound.

[magika]

I'm just wondering... Is there a way to make Google Chrome use ALSA Loopback? I use JACK as a sound server and an ALSA loopback that connects to JACK is good enough for me, so I would like to eliminate PulseAudio if it's possible. Without PulseAudio, I get no sound.

Without PA Google Chrome uses default ALSA device out of the box.

[tenente]

Google Chrome works fine without pulseaudio for me, maybe try export ALSA_CARD=Loopback before launching it, it's supposed to be one way to choose which device to use with alsa.

Tbh I don't understand how people find configuring alsa with multiple devices hard, even networking is pretty straightforward. I guess as long as there's no GUI people start to complain. Btw networking never worked for me with pulse, that the primary reason I uninstalled it on my main computer ! But I can understand that for a developper pov the API can be easier to use as fritsch was stating.

[dimko]

PulseAudio is good, exactly because it meets the needs of advanced users. Some people happen to do all this at once:

- play 4 sound sources (Web browser, VoIP, a game, music player) at once with different volumes
- use a 5.1/7.1 USB headset
- use HDMI sound or multiple devices at once

play 4 sound sources - hell you can. Until adobe flash starts pulse audio, than you can do one thing at a time.
use a 5.1/7.1 USB headset - dunno, but normal sound card, PCI-E, 5.1 or more channels(Asus Xonair DX) works perfect, until fucking pulseaudio kicks in.

[TeamBlackFox]

ALSA and PA both are poorly designed:

Basically, ALSA is a broken pile of mistakes from the get-go, because it uses blocking I/O and thus often two inputs to the sound card cannot be negotiated. It also lacks features of OSSv3 and v4. Its config syntax is also cryptic and doesn't always work as intended.

Pulseaudio further clouds the mix by adding another layer of redirection in userspace, which further increases latency, and its main purpose is to cover the holes ALSA leaves in a proper sound API, and to mix all of the different inputs together and feed them into ALSA.

When you combine both layers together you get a system that is unreliable, high latency and needlessly complex. Someone should have instead forked OSS when the license changed, or use its API while using a different system internally, like FreeBSD did - FreeBSD's system is nowhere near perfect, for sure, but it at least works, the worst one has to do on a laptop or desktop is to edit /boot/device.hints and indicate the primary soundcards using a rather simple syntax.

If Linux wants any chance of being a good, reliable desktop, it needs to trash ALSA, PulseAudio, JACK, and maybe even OSS and write a kernel and userspace sound system that is both simple and coherent, yet flexible. I shake my head at the Linux kernel and distros of GNU/Linux for making so many beginner mistakes.

[mdias]

ALSA and PA both are poorly designed:

Basically, ALSA is a broken pile of mistakes from the get-go, because it uses blocking I/O and thus often two inputs to the sound card cannot be negotiated. It also lacks features of OSSv3 and v4. Its config syntax is also cryptic and doesn't always work as intended.

Every piece of hardware you have on your PC can only be controlled by one driver. This is for obvious reasons.

Pulseaudio further clouds the mix by adding another layer of redirection in userspace, which further increases latency, and its main purpose is to cover the holes ALSA leaves in a proper sound API, and to mix all of the different inputs together and feed them into ALSA.

Yep. That's because that's where software dsp (including mixing) belongs. It serves 90%+ of the users' needs. If you have the need for low latency, go with Jack. ALSA may have flaws, yes, but I don't think exclusively locking your sound card is one of them.

If Linux wants any chance of being a good, reliable desktop, it needs to trash ALSA, PulseAudio, JACK, and maybe even OSS and write a kernel and userspace sound system that is both simple and coherent, yet flexible. I shake my head at the Linux kernel and distros of GNU/Linux for making so many beginner mistakes.

What Linux needs is to poop on users that:
- support hardware vendors that don't supply documentation
- can only argue on forums instead of actually opening bug reports and help debug problems
- only see their own special needs as needed to be flawless

Can't you see that PA actually works wonders for most of the people? You may see a large percentage of users complaining, but that is because those who have it working properly don't even bother replying... I for one only had a problem once with my mic's volume that was quickly fixed.
I do believe people complain because they do have problems, but less complaining and more bug reporting would help everyone. Linux gives you tons for free, filling bug reports is the least you can do.