Translation(s): English - español - Français - Русский
PulseAudio is a network-capable sound server program. A sound server is a background process accepting sound input from one or more sources (processes, capture devices, etc.), that is able to mix and redirect those sources to one or more sinks (sound cards, remote network PulseAudio servers, or other processes).
PulseAudio is installed by default in most Debian desktop environments. Some Desktop Environments like GNOME have switched to the more modern PipeWire in recent releases.
PulseAudio works on top of ALSA.
Contents
- Installing PulseAudio
- Basic Configuration
-
Solving Problems
- Restarting the Pulseaudio Daemon
- Interrupting play in Amarok when running Skype
- Sound level is low or suddenly becomes too loud
- Missing playback devices or audio capture
- Front Panel Jacks not working
- Stuttering and audio interruptions
- Excessive CPU usage and distortion
- Various problems with Skype and Wine
- Disabling daemon autospawn
- Echo test: hearing the microphone
- Stop running the PulseAudio daemon for a while
- Pops when starting and stopping playback or starting audio applications
- Advanced
Installing PulseAudio
As of Debian 12 (Bookworm, in 2023), you rarely need to install pulseaudio by hand as it is a dependency of the most popular desktops
Desktop Environment
Automatically includes pulseaudio?
Cinnamon
Yes
Gnome
No
KDE
Yes
LXDE
No
Mate
Yes
XFCE
No
If you need to install it manually, apt-get install pulseaudio should "just work" in every release since Debian 6 (Squeeze). PulseAudio becomes the default when the package is installed. Everything will use it.
This can be checked by with pavucontrol (in the pavucontrol package). If the sound is registering in the vu vumetre of that program, then it is being routed through the pulse system.
Basic Configuration
Mixers
pavucontrol, an optional package for PulseAudio, allows you to make volume adjustments (entire system or per application). It also allows you to change "mixer" settings like headphone output, mic level, input device, and stereo versus surround sound profiles.
However, you may not need to install pavucontrol as most desktop environments have their own control panels for configuring PulseAudio. Often such programs will not mention PulseAudio and will simply be called something generic like "Sound Settings".
pamixer package is a TUI mixer for PulseAudio.
Surround sound system
Many people have a multi-channel sound cards, but use the speakers for the two channels. PulseAudio has no default settings for surround sound support. To enable all channels, edit the file /etc/pulse/daemon.conf: uncomment default-sample-channels (ie remove the semicolon at the beginning of the line) and set it to 6 if you System 5.1 or 8, if your system is 7.1.
# Default default-sample-channels = 2 # To 5.1 default-sample-channels = 6 # To 7.1 default-sample-channels = 8
After making the changes, restart Pulseaudio.
Solving Problems
Restarting the Pulseaudio Daemon
Since Debian 9, PulseAudio is managed by the per-user instance of systemd. It uses socket activation, which means PulseAudio will start automatically when an application tries to use it.
To reread the config files ~/.config/pulse/daemon.conf and /etc/pulse/daemon.conf, one can restart pulse:
$ systemctl --user restart pulseaudio.service
Or on systems that do not use systemd (including Debian 8 and earlier):
$ pulseaudio --kill $ pulseaudio --start
Interrupting play in Amarok when running Skype
Comment out or remove the line in the /etc/pulse/default.pa
load-module module-cork-music-on-phone
Sound level is low or suddenly becomes too loud
Add a line to /etc/pulse/daemon.conf:
flat-volumes = no
Missing playback devices or audio capture
If PulseAudio does not correctly detect your input / output devices ("sources" and "sinks" in PulseAudio parlance), you can try deleting the configuration files and restarting PulseAudio. This is probably unnecessary overkill, but might help some people.
$ rm -r ~/.config/pulse /tmp/pulse-* $ systemctl --user restart pulseaudio.service
PulseAudio, in its default configuration, will likely want exclusive access to the hardware. It will therefore skip devices already in use by other applications, eg. web browsers and volume control applets. lsof /dev/snd/* will help identify processes you may need to kill off before PulseAudio will use your audio device.
Front Panel Jacks not working
As of wheezy, for some reason, PulseAudio does not see the toggle feature of some cards (i.e. CMI8788 [Oxygen HD Audio]) exposed by ALSA and playing with the PulseAudio interface (pavucontrol) won't bring happiness. The trick, (for now?), is to bring up the alsamixer (or alsamixergui) and the audio output can be switched from the read of the card to the front-panel as well as the mic input.
Stuttering and audio interruptions
If a low-power machine stutters (audio breaks up), you can try adding the following to /etc/pulse/daemon.conf:
high-priority = no nice-level = -1 realtime-scheduling = yes realtime-priority = 5 flat-volumes = no resample-method = speex-float-1 default-sample-rate = 48000
Excessive CPU usage and distortion
Add a line to /etc/pulse/default.pa
load-module module-udev-detect tsched = 0
Various problems with Skype and Wine
Add or uncomment the line in /etc/pulse/daemon.conf
default-fragments = 25 default-fragment-size-msec = 25
Disabling daemon autospawn
A PulseAudio client program will automatically run the PulseAudio daemon if there isn't one running already. If you want to be able to kill the daemon without it respawning, make these changes:-
Copy the client configuration file to ~/.config/pulse
$ cp /etc/pulse/client.conf ~/.config/pulse/
Use your favorite editor to change the line which says ; autospawn = yes to autospawn = no. Be sure to uncomment the line by removing the leading semicolon.
Alternately, to do both the copy and the edit in one step, just cut and paste this into the command line:
sed 's/; autospawn = yes/autospawn = no/' < /etc/pulse/client.conf > ~/.config/pulse/client.conf
Echo test: hearing the microphone
If you are unsure about your microphone setup, you can hear the input from the microphone in real-time by enabling the loopback module (source):
pactl load-module module-loopback
The module will show up in the Recording tab of the pavucontrol program, where the source and volume can be configured. While latency should be low, it should be sufficient to get a feeling of the sound quality as you will hear yourself speak in the microphone. To make the change permanent, add the following line to your ~/.config/pulse/default.pa:
load-module module-loopback
Watch out for feedback! Be ready to lower all volumes in case the microphone picks up the output from the loudspeakers. Naturally, it is better to run such a test with headphones.
On top of this above technique, many conferencing services have built-in echo tests. For example, ?Skype has an echo test. ?Mumble can also be configured to echo audio back (from local or remote) through the Configure -> Settings -> Audio Output -> Loopback test.
Stop running the PulseAudio daemon for a while
If you suspect trouble with a running PulseAudio daemon in your systemd user session, you can switch it off:
systemctl --user mask pulseaudio.socket systemctl --user stop pulseaudio.service
and then to switch it on:
systemctl --user unmask pulseaudio.socket systemctl --user start pulseaudio.service
Pops when starting and stopping playback or starting audio applications
This can be caused by PulseAudio suspending sinks after a period of inactivity. To disable this behaviour, comment out or remove the following line in /etc/pulse/default.pa and /etc/pulse/system.pa files:
load-module module-suspend-on-idle
A better solution is to create the file ~/.config/pulse/default.pa and populate same thusly:
.include /etc/pulse/default.pa .nofail unload-module module-suspend-on-idle .fail
Reference: ArchWiki/PulseAudio/Troubleshooting/Pops when starting and stopping playback
Advanced
Dynamically enable/disable
As mentioned above, all sound will automatically be routed through PulseAudio when the pulseaudio package is installed. These instructions describe how to disable it with the pulseaudio package still installed. Individual users can then reenable it themselves as needed.
The "just works" magic is achieved through configuration files placed in /usr/share/alsa/alsa.conf.d/. Without these files in place, the regular alsa defaults will be used. Therefore, to achieve the default ?alsa behavior with the pulseaudio package installed, divert these files
mkdir /usr/share/alsa/alsa.conf.pulse/ dpkg-divert --divert /usr/share/alsa/alsa.conf.pulse/pulse.conf --rename /usr/share/alsa/alsa.conf.d/pulse.conf dpkg-divert --divert /usr/share/alsa/alsa.conf.pulse/99-pulseaudio-default.conf.example --rename /usr/share/alsa/alsa.conf.d/99-pulseaudio-default.conf.example dpkg-divert --divert /usr/share/alsa/alsa.conf.pulse/50-pulseaudio.conf --rename /usr/share/alsa/alsa.conf.d/50-pulseaudio.conf
Now if a user wishes to use pulseaudio, they can create an ~/.asoundrc file that looks something like
@hooks [ { func load files [ { @func concat strings [ { @func datadir } "/alsa.conf.pulse/" ] } ] errors false } ]
If a user wishes to switch between pulse and non-pulse on a quasi-regular basis, put the above into the ~/.asoundrc.pulse instead and symlink it to ~/.asoundrc when pulse is desired
ln -s ~/.asoundrc.pulse ~/.asoundrc
and remove it when not
rm ~/.asoundrc
be sure also when disabling pulse to kill the server so that other things can directly access the soundcard again
killall pulseaudio
Temporarily suspend and run an application without PulseAudio
You can use the pasuspender utility, if you only need to disable PulseAudio temporarily, to run an application and have it access your audio devices directly.
Run: pasuspender -- yourapplication [yourapplicationoptions]
- Configure your application to access your audio devices directly (e.g. select your sound card ALSA address in an audio player)
While you run pasuspender, other applications won't be able to use PulseAudio. When you quit the application, the default behaviour of PulseAudio will be restored automatically.
See man pasuspender for more information.
Connecting PulseAudio and JACK
It's possible to use PulseAudio alongside the JACK server, to provide JACK access to applications that only support PulseAudio.
Install the pulseaudio-module-jack package; this will install the modules module-jack-sink, module-jack-source and module-jackdbus-detect; the modules will create a "PulseAudio JACK" sink and source that will show up in pavucontrol and in the qjackctl "Connections" window.
- If you want to enable this feature temporarily, load the modules manually:
pacmd load-module module-jack-sink pacmd load-module module-jack-source
If you want to enable this feature automatically, make sure you're running jackdbus whenever you run jackd, e.g. if you start jackd using qjackctl, check "Enable JACK D-Bus interface" in the Setup -> Misc tab. Thus, JACK will tell PulseAudio to load the modules at startup. Warning: you might need to close and re-open qjackctl in order for this change to take effect.
Here are some features of PulseAudio
- High quality software mixing of multiple audio streams with support for more than one entrance (source) and exit (sink).
- Can be used to combine multiple sound cards into one (with frequency rate).
Large set of supported client libraries. Applications that use ESD, ALSA, oss, libao and GStreamer, are supported without the need for any changes. Modules for PulseAudio are available for xmms and mplayer.
- Low latency and accurately measured delay time for recording and playback. Ability to fully synchronize multiple playback streams.
- Network transparency: the application can play or record audio on a computer other than the one on which they run.
- Extensible architecture with modules for jackd, multicast-rtp, lirc and avahi, among others.
External links