PulseAudio Creator Lennart Poettering Flames Back

[aleph]

http://www.cio.com.au/article/320807...fp=&fpid=&pf=1

I am not too concerned about most of the criticism and flames that erupt from time to time on various channels. All the big Linux distributions have adopted PulseAudio and it is an integral part of both the Palm Pre and the Nokia N900 devices, as well as Intel's Moblin. That basically means that PulseAudio has been adopted by about everyone who could adopt it. There is not really anyone who doesn't do PulseAudio anymore.

<... ...>

More specifically some applications were still using audio APIs [OSS] that are almost impossible to virtualize. And also PulseAudio makes use of a lot of driver functionality that was previously unused and hence little tested.

In fact, for quite a few parts of the lower level ALSA APIs PulseAudio is the first user of all. And of course, it cannot be a surprise that we expose bugs that were previously unknown in the drivers this way.

It's not my intention to shift the blame around though. PA and the other layers of our stack should not be viewed as independent parts. If PA uses a new or previously unused feature of the drivers then we need to fix the drivers at the same time.

If we make PA expect more correct behaviour from the apps, or that applications stop making particular assumptions about the audio stack, we need to fix the applications at the same time.

In short: PA is not bad. It exposes bugs in lower-level layers, and gets flamed by the end-users.

[adrianx]

He is also less than impressed with Ubuntu...

<rant>
So in the past Ubuntu packaged PA in a way that, let's say, was not exactly optimal. I thought they'd gotten around fixing things since then. Turns out they didn't. Seems in their upcoming release they again did some genius thing to make PA on Ubuntu perform worse than it could. The Ubuntu kernel contains all kind of closed-source and other crap to no limits, but backporting a tiny patch that is blessed and merged upstream and in Fedora for ages, that they won't do. Gah.
And it doesn't stop there. This patch is an outright insult. This is disappointing.
Madness. Not good, Ubuntu, really not good! And I'll get all the complaints for this f**up again. Thanks!
/me is disappointed. Ubuntu, you really can do better than this.
</rant>

http://0pointer.de/blog/projects/pa-in-ubuntu.html

[scottro]

Exposes bugs....or breaks something that used to work?

(Both, actually, but one can see why end users would flame something that does that.)

[JN4OldSchool]

I can honestly say I see his point. Like it or not PulseAudio is here to stay, there is no getting around it. Linux becomes more complicated by the day. While I still believe PulseAudio was always a solution in search of a problem I have not had any problems using it myself. As long as it works then for me out of sight is out of mind. The bigger picture is many of us would like to go back to how things were ten years ago, though with a natural evolution of course. I just feel that way too many projects too a wrong turn and are headed in the wrong direction. It never used to be about mass desktop appeal. But...It is what it is and until something better comes along I am happy we still have choices like Arch where I can weed out all this crap.

[scottro]

Judging from the comments at slashdot, many don't agree with him.

As I saw somewhere, it's been termed the Vista defense--it's the drivers/system/applications, it isn't us. Another more or less stated, though with more expletives, what has often been said. People don't care about the reason, when something as basic as sound, that works in all other O/S's doesn't work, something is really wrong.


The approach to the users is probably what brought the most rancor. If it had been advertised as something that will be really good, need beta testers, everyone would be fighting to use it. As it is, it was foisted upon the users.

[AdamW]

scottro: you're quoting selectively. Instructive quote:

"It's not my intention to shift the blame around though. PA and the other layers of our stack should not be viewed as independent parts. If PA uses a new or previously unused feature of the drivers then we need to fix the drivers at the same time.

If we make PA expect more correct behaviour from the apps, or that applications stop making particular assumptions about the audio stack, we need to fix the applications at the same time."

Lennart tends to have a defensive mindset, which is not surprising given the amount of crap that gets heaped on him for the horrible crime of trying to make audio on Linux not suck so much. What he takes exception to is the fact that some people tend to assume bugs encountered when using PulseAudio are due to PulseAudio being crappily written, which is in general not the case. He's not talking about assigning 'fault' or 'blame', just about making an accurate assessment of where the bugs lie, understanding that Lennart is not in fact writing PulseAudio in order to make your life worse, and making sure things get fixed in the right place.

Red Hat pays Lennart to work on PA because we think it's part of the best bet for having a decent audio story in Linux in future. We also pay Jaroslav Kysela to work on ALSA - he's one of a whole *two* paid ALSA developers in the world (the other is Takashi Iwai at Novell) - because we recognize the kernel layer is just as important a part of the story. Lennart and Jaroslav work together to fix problems in the kernel exposed by PA. It's not a blame-assignment exercise.

[scottro]

I was paraphrasing selectively.

I stand by my last sentence. The biggest trouble with pulse audio was the way it was presented.

As I've also said, it gets better with each iteration, but the big thing, and what seemed to be the major complaint on slashdot and elsewhere is that something used to Just Work and now, for many, it doesn't. (Although lately, as I've mentioned here and there, lately, it has been for me.)

Since it does work for me these days, and I'm solipsistic by nature, I only pay casual attention to it these days. I think the next issue will be that it's gotten harder to remove, but that's only a casual observation.

As I've also said, you personally deserve our thanks too. Despite your modesty, your work at not only handling the bugs, but making great efforts to solve users' problems with it, one these forums and elsewhere, are probably one of the main reasons it's become something that *I* (solipsism again) no longer pay much attention to it.

[Gödel]

Yes, Red Hat pays him to bring PA from experimental status to production worthy. In the mean time, why the hell shouldn't ubuntu appy a temporary ugly patch to get bluetooth working, or disable flat_volumes (I must have suggested it half a dozen times on this forum alone) to cure annoyances with the volume control.

Problem with Poettering and the rest on the fedora project is that they are too restricted by the pure FOSS mentality to aknowledge that third party apps like flash or vlc or ("crappy") oss based apps exist and many people want them to run smoothly. Not everyone is an audio-geekhead obsessed with the cpu cycle count efficiencies achieved by outstanding technical application in Poettering's code.

It's a curious thing to introduce extra complexity in the audio at a low level whilst dumbing down the user interface to a Vista-inspired user-is-dumb level so that ordinary users are just perplexed when the volume levels on certain apps just won't go up, or digital laptop mics just won't turn on etc.

Admittedly PA gets a lot of underserved flak, incompetent implementations such as in the popular media app VLC has caused a lot of misery (simply switching the audio from the PA driver to SDL fixes), but the truth is that linux audio worked well enough for 90% of people before PA, so despite the technical accomplishments underneath, many people just don't get the big deal.

Post a couple of videos Poettering, and with the same hardware, demonstrate the awesome benefits that PA brings. Because if all it adds is obscure stuff like faciliating audio in virtualisation solutions and usb/bluetooth headphones it ain't really that great a leap forward, is it?

And be aware that exposing new driver features which don't work properly isn't that thrilling to the end-user.

For the record, I think PA works well now in Fedora 11, although it still can an annoyance when trying to run audio production apps like fluidsynth, or crappy oss based emulators and the like. But I also think people running fedora should understand the experimental nature of it, and have less cause for whinging than users of other distros.

[RahulSundaram]

Hi,

The problem with the Ubuntu patches in this case the lack of communication with upstream developers and patches that themselves cause more breakage than they fix. This causes a bad reputation for PulseAudio. Also, Fedora included the libflash library by default which was a workaround to the fact the Flash was using ALSA in a incorrect way that caused problems with PulseAudio. Flash was later fixed and Fedora dropped the workaround

In general however, Fedora does prefer to improve the free software ecosystem keeping in mind the longer term benefits over short term hacks. This has consistently yielded good results. Remember how much NetworkManager was said to be the cause of problems even though it was crappy wireless drivers that were actually problematic? Now pretty much every major distribution includes it by default. Same thing goes for PulseAudio and in the near future, Nouveau etc. Fedora has a long history of demonstrating that it benefits everyone if we stick to our principles. In terms of technical contributions, Fedora certainly has done more for the Linux desktop than any other distribution out there.

[scottro]

Of course, since I've been saying how it's working for me, on a fresh install today (on a work machine) no sound till I removed pulseaudio. However, this is really, on many test installs, becoming the exception rather than the rule.

However--on all other installs, no matter what strange things I did, it worked quite well. This particular workstation has some odd card in it too, which might have been a factor. <shrug>.

[Dies]

Man, I must be one of the lucky ones, I don't remember ever having a problem with Pulse. I like it.

The thing I don't like is the dumbing down of volume controls in the latest Gnome. They used to give you a nice interface to alsamixer, now they're trying to *guess* what everyone would like to do with their outputs, which is just plain silly. Oh well, as long as I can run alsamixer from a terminal I'm cool with it.

[AdamW]

dies: there's no such thing as a nice interface to alsamixer. alsamixer is indistinguishable from the portal to hell.

the g-v-c in f11 was quite bare, because it was quite early in the re-implementation process. That's why we put gst-mixer in f11 as a fallback. f12's g-v-c is rather better and has most of the functions people need implemented, as you say alsamixer is always there as a fallback, but hopefully most people won't need it.

[JN4OldSchool]

Well, even though I am hard headed I have come to accept the inevitable; Linux is growing at a much faster pace than it has ever had before. Maybe it is finally really about ready to hit the mainstream. At any rate, there is no point fighting all this new stuff. Most anything is rough to start out with but soon settles in. There is no point in fighting it, it is here to stay.

[Evil_Bert]

Maybe it is finally really about ready to hit the mainstream.

Maybe. IMHO, it's not quite there yet, but it is getting darn close. The F12 beta is looking really good so far.

[Demz]

Pulse is improving an will only get better. i think its just about ready for prime time , same with the nouvoua driver, will only improve over time