UbuntuThis forum is for the discussion of Ubuntu Linux.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
pulse is now mature and works flawlessly most of the time
( unlike back in the fedora 6 days when it was NEW)
now that said if you are using KDE a glitch might cause hair pulling out
but that is the desktop issue
yes you can remove it
but you will need to replace many things also and likely will need to manually configure the os and programs to use jack or just alsa
Now I can understand if you have problems with Pulse Audio but it would be best to explain what those problems are. Just removing Pulse Audio and replacing it with something else may be more trouble than it is worth.
Personally, I rarely have issues with Pulse Audio and I am running Fedora 25 KDE Spin. Of the rare issues, I have had (loss of sound) a quick log out and log back in again usually fixes the problem. Obviously, I always check my Audio and Video system settings when something like this happens.
I think the last time I had an audio problem (Fedora 24 KDE spin and months ago) was when I inadvertently had more than three audio streams but after logging out then in again I could not duplicate the issue.
I run without Pulse on openSUSE 42 and have done so with earlier versions. I have fairly high end audio equipment and using straight alsa gives me fully independent control of each of the surround channels and the shape of the sound. I have never been able to get that with Pulse. Most of my music is 24 bit flac at 192KHz. My movies mostly use 5.1 or 7.1 surround. My other computers with a simple two speaker setup are fine with Pulse, but if I want to enjoy the sound of a favorite song or good movie, Pulse is out.
Removing and blacklisting Pulse is fairly easy with Yast in openSUSE. If you are using apt, should I assume that you run Ubuntu? I am not really knowledgeable in that area.
yes you can remove it
but you will need to replace many things also and likely will need to manually configure the os and programs to use jack or just alsa
This begs the question about programs that do require jack and alsa anyway.
eg. If one is into something like RoseGarden, FluidSynth, Hydrogen, etc, and maybe wanting mixer features such as one gets with gnome-alsamixer.
Do they all live happily with PulseAudio?
One can have, it seems, audio, or even for the more demanding preferences of some folk, "better" audio, without PulseAudio present at all!
Maybe I misunderstand, but my Mint distro comes with PulseAudio pre-installed. That does not prevent me from installing other audio stuff (like jack and alsa). Audio in Linux, (at least for me), ranks as darn tricky to understand and get right.
The similar problem does occur with Suse Leap 42.2.
After uninstalling pulseaudio, the sound does not work any more.
In every explanation I have read, PulseAudio is a "middleware" layer engine that goes between the Linux kernel ALSA hardware drivers that actually allow sound hardware to make a noise, and the user level libraries and application layers, meaning to have things like Mplayer or VLC, or whatever you have fun with (TuxGuitar, Audacity, etc.)
When you uninstall PulseAudio, it may be there is still left in operation various scripts and configuration files that your distro set up as PulseAudio to be the "default", and expecting it to be present.
ALSA,"Advanced Linux Sound Architecture" must always be there. Check you have not inadvertently removed it, or a needed library, as a dependency along with PulseAudio. So far as I know, ALSA, and all the drivers those dedicated guys attach to the kernel, is not something PulseAudio can easily replace, and is not supposed to. We can use almost any sound hardware without even thinking about "downloading a correct version driver" as like in Windows.
Check the setup of whatever applications you are using to make the sound. They may still be set to PulseAudio, instead of feeding to alsamixer. Some may have it set in "preferences". Others may need to have a configuration file edit.
Without PulseAudio, install at least alsamixer, or gnome-alsamixer, maybe alsamixergui, and experiment to switch through some channels that make a noise. You are likely to hear nothing until you enable a channel, and turn up it's volume control, or un-mute a mic.
I have removed PulseAudio before, and then I made music sound by playing MP3 in Audacity, or VLC while listening in headphones. Then I used alsa mixer to make it happen.
I dare say you may not just have it so easy, depending on what the distro configuration scripts were doing, but there is enough how-to information out there to manage this without intimately understanding every audio feature one can have in a Linux system.
Then, to start to feel better, this explanation from back in 2007 is still helpful.. https://www.linux.com/news/why-you-s...start-doing-it
That feels like PulseAudio propaganda polemic. There are folk who are not convinced!
For those who need to be giving over the entire computing resource to a high-end application, with ultra low-latency access, (try real-time DSP phase-shift keying demodulation signal processing), then one does not much care one cannot have the pop MP3 playing at the same time. Some computers still find it hard to cycle the display, keyboard polling and all else at 1kHz for Jack server. If you use Jack at all, you likely know a lot about audio already, and can be teaching me stuff! If you have a drum synth, and are into multi-channel 192kHz high digital audio recording/mixing studio, then maybe PulseAudio has it's downsides.
I did not find it too much problem to get sound going without PulseAudio. Keep in mind some volume control default apps might need to be dumped, or re-configured. PulseAudio, being a layer, inevitably has some overhead. I am sure there are experts here who can tell you about that.
> That feels like PulseAudio propaganda polemic. There are folk who are not convinced!
Similar to systemd, except that systemd took over the whole linux stack whereas pulseaudio
only took over audio stuff.
@shevegen - Indeed, and thank you. That is a perceptive description. I do not have the knowledge, and I get confused when I read much on the internet that seems contradictory about the merits of some "new" layer.
The whole business of audio is mangled with the motivations of many interests that would seek to implement "their" variation, and for Linux, it is the tricky line between building on previous work, and fundamentally altering the scheme.
For me, the "audio" channels were not being used for music, but for frequency down-converted demodulated phase shift keying data, with the baseband maxed out. There had to be enough sheer speed to leave room for "other" services, like servicing keyboard interrupts. Infrequent things like keystrokes or mouse pulses can be serviced at the higher priority, well in time to return to the channel data.
I have not enough life nor patience to learn and debate the merits of things like systemd, other than to know it is likely to have a coding overhead. For my needs, I settled for dedicated outside DSP hardware delivering to a Gig-E network port, and I left PulseAudio, and all it's ways to making sound for music or video playback.Setting up audio, especially if you want to play with high end recording, mixing and sound manipulation, is actually quite complicated stuff!
It's mainly the desktop environment... Debian with XFCE doesn't require PulseAudio. I haven't actually tried removing it from Xubuntu but I'd say that's your best bet if you want to go ALSA only.
This begs the question about programs that do require jack and alsa anyway.
eg. If one is into something like RoseGarden, FluidSynth, Hydrogen, etc, and maybe wanting mixer features such as one gets with gnome-alsamixer.
Do they all live happily with PulseAudio?
One can have, it seems, audio, or even for the more demanding preferences of some folk, "better" audio, without PulseAudio present at all!
Maybe I misunderstand, but my Mint distro comes with PulseAudio pre-installed. That does not prevent me from installing other audio stuff (like jack and alsa). Audio in Linux, (at least for me), ranks as darn tricky to understand and get right.
I totally agree that Linux ranks high in the trickery to understand and get working properly. Especially Bluetooth.
Maybe I misunderstand, but my Mint distro comes with PulseAudio pre-installed. That does not prevent me from installing other audio stuff (like jack and alsa). Audio in Linux, (at least for me), ranks as darn tricky to understand and get right.
Very common misunderstanding. PulseAudio is a parasite layer between ALSA and sound applications. If you remove ALSA then PulseAudio will not work. Although you can use OSS4 instead of ALSA.
Emerson posted a clear and accurate description of what PulseAudio is - basically it is unecessary middle-ware, installed between the well-designed ALSA (Advanced Linux Sound Architecture) code, which is what lets the Linux operating system interact with the audio hardware, and the user. ALSA and other drivers (example, video and TV-card drivers) are installed in the Linux kernel (a wise and sensible place to put them, since they have to be small, fast and reliable), and they interact with the devices that one installs in the machine. You can typically see and verify operation of devices, by looking in the /dev directory. But if you are using a modern Linux (eg. CentOS or Redhat variants, and I assume Ubuntu and others), there now is no longer a /dev/dsp device. That device, (dsp = digital sound processor, IIRC), was easily accessable in earlier Linux implementations. You could copy a random-noise file to it, and it would output "PHSSSSSSH..." out of your computer speakers, and in this way, you knew your sound-card was correctly configured.
But with PulseAudio, there no longer is a proper ability to access /dev/dsp directly. If you run MPlayer and/or Mencoder to access/process video and sound files, you find this out real quick. Nothing works right, until all your code is re-written to invoke and manage the complexity of PulseAudio. In my opinion, it is an ugly, and toxic piece of manufactured complexity, designed to make it much more difficult to manipulate video and audio files. If you think for a minute, you might come up with a few reasons why certain people would be *very* interested in making that happen. I doubt that PulseAudio was an evil plot by the American copyright police to damage Linux and frustrate Linux users. That would be silly (and we all know that the modern forces of fascism are busy doing other things, once the DMCA became US law, and net-neutrality was nuked... But the original poster's question is a very good one. Can we remove PulseAudio?
I run CentOS 6.6 and CentOS 7.4, and a few older Fedora boxes. PulseAudio is miserable, annoying, useless middleware. Yes, if I had a bunch of high-school kids logging into a bunch of virtual machines on a multi-server cluster or something, the security concerns might justify it being there. Of course, you don't want to give low-level access to devices to just anyone on the machine. But on single-user workstations, PussAudio just gets in the way, and stops my video stuff from working right. Right now, on a CentOS box, I want to use Mencoder to capture a video-stream - with SOUND and Video TOGETHER (what a novel idea, eh?). I have rebuilt all the code from souce, made sure all the software has all the needed codecs (codec just means: COde and DECode - ie. take an analog signal and turn it into some kind of digital format). I have all the software rebuilt, and am 100% certain I have all the needed functionality - but the machine in question (running CentOS 6.6), which can display the video feed with sound and vision (thru a bit of hackery), simply cannot be tailored to capture the video feed WITH SOUND. It works fine WITHOUT audio - and I can see and confirm the audio-track is being laid down in the .AVI and/or .MPG or other 4CC type file. I have tried both Mencoder and ffmpeg. I have experimented with ALL the codecs and libraries (-mp3lame -lavc -pcm, tried -raw, and some others) and fiddled the incoming video feed in about a hundred different ways, and - oh, did i mention I have recompiled *EVERYTHING* from source, just to check and verify things? It all works. I can play and display almost any video format. MPlayer works fine, and most of the codecs are compiled-in, and all runs happy on my GNOME desktop. But the hideous, awful and annoying changes introduced (ah, I am thinking: PulseAudio) in the newer versions of Linux MAKE IT IMPOSSIBLE FOR ME TO DEBUG THIS BECAUSE THERE IS **NO** /dev/dsp DEVICE. Now this, to me, is what we call: A RETROGRADE CHANGE. It is a crappy, nasty, functionality-wrecking "enhancement" ("hey! It's a feature, not a bug!") that damages the whole idea of what Linux was and is supposed to be.
Look, I think Linux is one of the best things ever built since North American Aviation built the Valkyrie back in 1964. I started running Slackware on an IBM/PC back in the last century. Linux has been great, excellent engineering. Well, until systemd was introduced... (but I am a guy with a DecSystem 20/20 emulator running on one of my Linux boxes, so maybe I am not objective..)
Anyway, the whole point of this epistle, is to ask for someone to - please - try to answer the OP's question. Is there an reasonably easy way to completely remove PulseAudio from say, a CentOS box? I gave up on Fedora, as it has become a test-bed for bad ideas now. Redhat has a smart business model, and CentOS is not bad (once you figure out wtf "upstream" actually means..). But I would really like to have a simpler Linux - with the gunk and bloat like PulseAudio stripped out, and lower-level device access put back in. The morons of the world can use Apple and empty their wallets onto silcon valley desktops if they want to. I won't complain. I would just like to get a little work done, and get back to having clear, immediate access to my own computer and it's attached devices.
The ALSA stuff alone was well thought out, and works (worked) quite well, and was reliable. Any low-level device access is going to be complex and sensitive - that is the whole point. The PulseAudio program buys me nothing, but it costs me the ability to control my stuff and make my computers do what I want them to. Emerson absolutely nails it perfectly, in his January post:
systemd - no
pulseaudio - no
I have taken more words to say the same thing. I would plead: "Stop wrecking Linux!". I have a CentOS 7.4 box, which I used to build TensorFlow from source. What an undertaking! The bloat is grotesque, the complexity absurd, and the gain is small - so small, I couldn't even find it. The Gnome desktop has been dumbed-down to the point where it is a joke. TensorFlow required about 100 special-purpose packages - Bazel, cmake, a bloat-boat of Python stuff, and for what? So I could find out that everything runs as 32-bit floats anyway, since it is designed to run on the GPU! - and which give different answers each time! - and this great 64-bit environment has no tools or facilities to let me fix the SELinux awful/horrible restrictions, which make it impossible for simple, needed stuff to run (like, on Fedora-9, there is a nice GUI-based tool to let you configure the SELinux-booleans and such, so to make SELinux less of a toxic rathole of restriction - so you can at least leave the damn thing running and maybe get some security gain). All these system-configuration features are GONE from the desktop gui - just VANISHED in CentOS 7.4. The desktop is for an intelligent monkey or something - it is not a legitimate scientific workstation anymore. <sigh>
Anyway- it would be very helpful if someone could write up a note on how to pry PulseAudio out a modern (or semi-modern) Linux implemenation. I bought a MacBook and installed all the Xcode bloat on it to build TensorFlow there also - just to see if it was any better. It wasn't, sadly. I'm done with Microsoft and Apple. I have a truckload of work I want to do on either PyTorch or TensorFlow - but I got so frustrated, I'm now thinking I might build my own stuff from some older code I have which works actually quite well.
But it would be *really* nice if I could get the bloat and restrictions (like PulseAudio, in this case) out of the way.
If someone knows where this has been documented, pls post a link.
If I can engineer my own solution, I will post it here. The problem is that modern distros seem to have integrated this nasty thing tightly. Pulseaudio reminds me of the "face-hugger" in Alien. Yes, you can still breathe. But it is a tad more difficult, don't you notice?
- Rus.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.