Context Navigation

Frequently Asked Questions

Some of the answers need updating! Also, some questions are too long for the outline, and they get truncated.

Uncategorized questions
Compatibility with other software
Troubleshooting
I want to do this specific thing…

Uncategorized questions

How does PulseAudio compare with ESOUND/aRts/NAS?

PulseAudio is sound daemon similar to ESOUND and NAS, but much more powerful. aRts is a realtime-synthesizer-cum-sound-server, i.e. it does much more than PulseAudio. However, I believe that PulseAudio does what it does much better than any other free sound server.

Is PulseAudio a GNOME program?

No, PulseAudio has no dependency on GNOME/GTK/GLIB. All it requires is a UNIX-like operating system and very few dependency libraries. However, the accompanying GUI tools are written with gtkmm, i.e. require both GLIB and GTK.

I want to write a new driver for PulseAudio, are there any docs?

Currently, only the client API is documented with doxygen. Read the source and base your work on a simple module like module-pipe-sink.

The pulseaudio executable is installed SUID root by default. Why this? Isn't this a potential security hole?

PulseAudio activates SCHED_FIFO scheduling if the user passes --high-priority=1. This will only succeed when executed as root, therefore the binary is marked SUID root by default. Yes, this is a potential security hole. However, PulseAudio tries its best to minimize the security threat: immediately after startup PulseAudio drops all capabilities except CAP_SYS_NICE (At least on systems that support it, like Linux; see man 7 capabilities for more information). If the calling user is not a member of the group pulse-rt (which is required to have a GID < 1000), root rights are dropped immediately. This means, you can install pulseaudio SUID root, but only a subset of your users (the members of the group pulse-rt) may make use of realtime scheduling. Keep in mind that these users might load their own binary modules into the PulseAudio daemon which may freeze the machine. The daemon has a minimal protection against CPU hogging (the daemon is killed after hogging more than 70% CPU for 5 seconds), but this may be circumvented easily by evildoers.

Is PulseAudio capable of providing synchronized audio playback over the network for movie players like MPlayer?

Yes! Unless your network is congested in some way (i.e. transfer latencies vary strongly ('jitter')) it works perfectly. Drop me an email for experimental patches for MPlayer.

What environment variables does PulseAudio care about?

The client honors:
- $PULSE_SINK (default sink to connect to)
- $PULSE_SOURCE (default source to connect to)
- $PULSE_SERVER (default server to connect to, like ESPEAKER, see ServerStrings for an explanation of the format)
- $PULSE_BINARY (the binary to start when autospawning a daemon)
- $PULSE_CLIENTCONFIG (path to the client configuration file).
The daemon honors:
- $PULSE_SCRIPT (default CLI script file run after startup)
- $PULSE_CONFIG (default daemon configuration file)
- $PULSE_DLPATH (colon separated list of paths where to look for modules)

How do the PulseAudio libraries decide where to connect to?

The following rule applies: (see ServerStrings for an explanation of the format of server specifications)

If the the application using the library specifies a server to connect to it is used. If the connection fails, the library fails too.
If the environment variable $PULSE_SERVER is defined the library connects to that server. If the connection fails, the library fails too.
If $DISPLAY is set, the library tries to connect to that server and looks for the root window property PULSE_SERVER for the host to connect to. If PULSE_COOKIE is set it is used as authentication cookie.
If the client configuration file (~/.pulse/client.conf or /etc/pulse/client.conf) sets the server address, the library connects to that server. If the connection fails, the library fails too.
The library tries to connect to the default local UNIX socket for PulseAudio servers. If the connection fails, it proceeds with the next item.
The library tries to connect to the default local TCP socket for PulseAudio servers. If the connection fails, it proceeds with the next item.
If $DISPLAY is set, the library tries to connect to the default TCP port of that host. If the connection fails, it proceeds with the next item.
The connection fails.

Why the heck does libpulse link against libX11?

The PulseAudio client libraries look for some X11 root window properties for the credentials of the PulseAudio server to access. You may compile PulseAudio without X11 for disabling this feature.

What is this RTP/SDP/SAP thing all about?

RTP is the Realtime Transfer Protocol. It is a well-known protocol for transferring audio and video data over IP. SDP is the Session Description Protocol and can be used to describe RTP sessions. SAP is the Session Announcement Protocol and can be used to announce RTP sessions that are described with SDP. (Modern SIP based VoIP phones use RTP/SDP for their sessions, too)
All three protocols are defined in IETF RFCs (RFC3550, RFC3551, RFC2327, RFC2327). They can be used in both multicast and unicast fashions. PulseAudio exclusively uses multicast RTP/SDP/SAP containing audio data.
For more information about using these technologies with PulseAudio have a look on the respective module's documentation.

Why did you rename Polypaudio to PulseAudio?

Please read this blog story for an explanation.

Compatibility with other software

What about ESOUND compatibility?

PulseAudio is a drop in replacement for ESOUND. That means: you can load a esound compatibility module which implements an ESOUND compatible protocol which allows you to use most of the classic ESOUND compatible programs (including the command line programs like esdcat).

Can I integrate PulseAudio in my GLIB/GTK/GNOME application?

Yes! PulseAudio comes with a GLIB main loop adapter. You can embed both the client library and the daemon (!) into your GLIB based application.

Can I integrate PulseAudio in my Qt/KDE application?

Yes! PulseAudio uses a main loop abstraction layer that allows you to integrate PulseAudio in any program that supports main loops. Unfortunately there is no adapter for Qt publicly available yet.

What about compatibility with NAS?

Is not available (yet?). It is doable, but noone has implemented it yet.

What about compatibility with aRts?

Is not available. Since aRts is as synthesizer application you'd have to reimplement very much code for PulseAudio. It should be easy to implement limited support for libartsc based applications. Noone has done this yet. It is probably a better idea to run arts on top of PulseAudio (through a PulseAudio driver for aRts, which nobody has written yet). Another solution would be to embed PulseAudio in the aRts process.

Can I get OSS and ALSA applications to work with PulseAudio?

Yes, you can! OSS applications are handled using the padsp utility shipped with PulseAudio:
padsp myapp <arguments to myapp>
ALSA applications are handled using the PulseAudio backend for alsa-lib. Documentation and the backend itself are shipped in the alsa-plugins package as part of the ALSA project.

Is there a Win32/Windows driver that act like a PulseAudio client?

Not exactly, but there use to be a ESoundD driver like that, and PulseAudio can connect to a ESoundD client.
This project is WinESD at WinESD. This software was developped for NT4 and Windows 2000. It has been reported to work under Windows XP connected to a PulseAudio server (One computer under Microsoft Windows XP using WinESD, connected to another computer under Microsoft Windows XP using PulseAudio). This project looks unmaintained and last version was "pre-alpha". It still works, but with restrictions :

Sometimes, the sound is "broken". It look like it's when two processes try to emit sounds in the same time. Perhaps it's different from one computer to another.
There is no easy installers as for July 2007, neither for WinESD nor PulseAudio for Win32/Windows. It's a "Let's play with configuration by hands" installation.
You need to unactivate your sound card if you already have one on the computer running WinESD. This will need a reboot. This is undocumented on the WinESD site.
Switching from one driver to another will need to play with Activate/Unactivate drivers, which will cause reboot.

Troubleshooting

I get this error message: "Connection refused"

It is a network level error, which is caused by libpulse trying to connect to the daemon. The most likely reason for this error is that pulseaudio isn't running.

I get this error message: "Invalid argument"

A likely reason for this is that no devices have been loaded. Stop pulseaudio (run "pulseaudio -k"), and start it again with "pulseaudio -vv". That will print verbose output of the daemon startup. If you can't figure out what might go wrong yourself, send a link to the whole output of the command to the irc channel (see Community#IRC). (Help is not guaranteed. Patience is a virtue.)

I often hear noises when playing back with PulseAudio, what can I do?

There are two possible solutions: run PulseAudio with argument --high-priority=1 and make yourself member of the group pulse-rt, or increase the fragment sizes of the audio drivers. The former will allow PulseAudio to activate SCHED_FIFO high priority scheduling (root rights are dropped immediately after this). Keep in mind that this is a potential security hole!

I have a surround sound card, but PulseAudio uses just the front speakers!

Many people have a surround card, but have speakers for just two channels, so PulseAudio can't really default to a surround setup. To enable all the channels, edit /etc/pulse/daemon.conf: uncomment the default-sample-channels line (i.e. remove the semicolon from the beginning of the line) and set the value to 6 if you have a 5.1 setup, or 8 if you have 7.1 setup etc. After doing the edit, restart pulseaudio.

When sending multicast RTP traffic it is recieved on the entire LAN but not by the sender machine itself!

Pass loop=1 to the sender module!

Hmm, tcpwrap (aka /etc/hosts.allow, /etc/hosts.deny) support doesn't work properly for me!

Have you compile PulseAudio with tcpwrap support? Are you using the right tcpwrap service name? PA uses five different names for its five protocols:
pulseaudio-native
pulseaudio-simple
pulseaudio-cli
pulseaudio-http
esound

Why doesn't the system-wide mode detect my default.pa file? Or, why does the system-wide daemon exit telling me there are no modules?

As of PulseAudio 0.9.11, the system-wide daemon looks for a separate pa file by default: /etc/pulse/system.pa

If you want to continue using default.pa for system-wide mode, just symlink /etc/pulse/default.pa to /etc/pulse/system.pa.

Why do I not see remote sinks in control applications?

Sending a stream to a remote server is done through control apps such as pavucontrol or padevchooser (marked obsolete). The streams to choose from are selected using zeroconf (avahi) publishing and discovering of services. If pavucontrol or padevchooser does not list a sink on a remote server the problem can be that the server is not reachable or the zeroconf mechanism is not working.

1. Make sure your host is reachable, firewall allows connections, cables are plugged in, whatever, ...

2. Check if you can reach the remote server using the PULSE_SERVER environment variable from CLI: run a pulseaudio configured application like this:

$ PULSE_SERVER=<hostname> <command>

If your host does not start playing you need to fix the network protocols supported by your server. Refer to protocol sections in the Modules page, refer to ServerStrings page, refer to FAQ item How do I use PulseAudio over the network?.

Assuming that your host started playing we can probably limit the problem to zeroconf.

3. Check that module-zeroconf-publish is loaded on the remote server using paman:

$ PULSE_SERVER=<hostname> paman

and that module-zeroconf-discover is loaded on the local server, also using paman.

If either are not, load them and see if the sink is discovered.

4. Check, using avahi-browse (in avahi-utils for debian based distros), that services are discovered okay. Use following command on both the local and the remote end to dump all (-a) services found on the network and the terminate (-t) the avahi-browse process after dumping:

$ avahi-browse -ta

If in this step the remote end does not show the PulseAudio services locally then check avahi configuration for things like disabled publishing for example. Consider restarting the avahi process. Make sure firewalls allow the avahi traffic. ...

I want to do this specific thing…

I want to run PulseAudio only when it is needed, how do I do this?

Set autospawn = yes in client.conf. That configuration file may be found either in /etc/pulse/ or in ~/.pulse/.

How do I list all PulseAudio modules installed?

pulseaudio --dump-modules

Add -v for terse usage instructions.

How do I use PulseAudio over the network?

Just set the environment variable $PULSE_SERVER to the host name of the PulseAudio server. Alternatively you can modify ~/.pulse/client.conf or /etc/pulse/client.conf and set default-server= (See ServerStrings for an explanation of the format, see FAQ# 18 for all locations you can specify the server to use) For authentication you need the same auth cookies on all sides. For that copy ~/.pulse-cookie to all clients that shall be allowed to connect. Alternatively the authorization cookies can be stored in the X11 server. The server must have module-native-protocol-tcp loaded, with the argument loopback=0 set. Note that this can use pretty much network bandwidth (around 1.3Mbps for 44100Hz sound). If you get choppy sound, try downsampling it.

I saw that `SIGUSR2` provokes loading of the module `module-cli-protocol-unix`. But how do I make use of that?

A brilliant guy named Lennart Poettering once wrote a nifty tool for that purpose: bidilink. To connect to a running PulseAudio daemon try using the following commands:
killall -USR2 pulseaudio
bidilink unix-client:/tmp/pulse-$USER/cli
BTW: Someone should package this great tool for Debian!
New: There's now a tool pacmd that automates sending SIGUSR2 to the daemon and running a bidilink like tool for you.

How can I use PulseAudio as an RTP based N:N multicast conferencing solution for the LAN?

After loading all the necessary audio drivers for recording and playback, just load the RTP reciever and sender modules with default parameters:
load-module module-rtp-send
load-module module-rtp-recv
As long as the PulseAudio daemon runs, the microphone data will be streamed to the network and the data from other hosts is played back locally. Please note that this may cause quite a lot of traffic. Hence consider passing rate=8000 format=ulaw channels=1 to the sender module to save bandwith while still maintaining good quality for speech transmission.

How can I use PulseAudio to stream music from my main PC to my LAN with multiple PCs with speakers?

On the sender side create an RTP sink:
load-module module-null-sink sink_name=rtp
load-module module-rtp-send source=rtp.monitor
set-default-sink rtp
This will make rtp the default sink, i.e. all applications will write to this virtual RTP device by default.
On the client sides just load the reciever module:
load-module module-rtp-recv
Now you can play your favourite music on the sender side and all clients will output it simultaneously.
BTW: You can have more than one sender machine set up like this. The audio data will be mixed on the client side.

How can I use PulseAudio to share a single LINE-IN/MIC jack on the entire LAN?

On the sender side simply load the RTP sender module:
load-module module-rtp-send
On the reciever sides, create an RTP source:
load-module module-null-sink sink_name=rtp
load-module module-rtp-recv sink=rtp
set-default-source rtp_monitor
Now the audio data will be available from the default source rtp_monitor.

Can I have more than one multicast RTP group?

Yes! Simply use a new multicast group address. Use the destination/sap_address arguments of the RTP modules to select them. Choose your group addresses from the range 225.0.0.x to make sure the audio data never leaves the LAN.

Can I use PulseAudio to playback music on two sound cards simultaneously?

Yes! Use module-combine for that.
load-module module-oss-mmap device="/dev/dsp" sink_name=output0
load-module module-oss-mmap device="/dev/dsp1" sink_name=output1
load-module module-combine sink_name=combined master=output0 slaves=output1
set-sink-default combined
This will combine the two sinks output0 and output1 into a new sink combined. Every sample written to the latter will be forwarded to the former two. PulseAudio will make sure to adjust the sample rate of the slave device in case it deviates from the master device. You can have more than one slave sink attached to the combined sink, and hence combine even three and more sound cards.

Can I use PulseAudio to combine two stereo soundcards into a virtual surround sound card?

Yes! You can use use module-combine for that.
load-module module-oss-mmap device="/dev/dsp" sink_name=output0 channel_map=left,right channels=2
load-module module-oss-mmap device="/dev/dsp1" sink_name=output1 channel_map=rear-left,rear-right channels=2
load-module module-combine sink_name=combined master=output0 slaves=output1 channel_map=left,right,rear-left,rear-right channels=4
This is mostly identical to the previous example. However, this time we manually specify the channel mappings for the sinks to make sure everything is routed correctly.
Please keep in mind that PulseAudio will constantly adjust the sample rate to compensate for the deviating quartzes of the sound devices. This is not perfect, however. Deviations in a range of 1/44100s (or 1/48000s depending on the sampling frequency) can not be compensated. The human ear will decode these deviations as minor movements (less than 1cm) of the positions of the sound sources you hear.

How can use my Windows box to play the sound from my Linux box?

Download the windows binary of pulseaudio from the download page.
In the pulseaudio directory, create a file default.pa whith a content similar to this one (you may have to adjust the auth-ip-acl to reflect your network's settings):
```
load-module module-waveout

load-module module-native-protocol-tcp auth-ip-acl=127.0.0.1;192.168.0.0/16
```
In pulseaudio applet: PA applet -> Default Server -> Other -> Enter the hostname or ip of your windows box.

How can I reverse my left and right speaker channels?

This is the same as "reverse stereo", where the left and right channels are to be swapped.

cat /proc/asound/cards and use the name string for the device you wish to use (the one in square brackets)
Edit /etc/pulse/default.pa and comment out module-hal-detect and module-detect lines.
Search for the commented-out line that starts "#load-module module-alsa-sink", uncomment it and change it to

load-module module-alsa-sink device=hw:[answer from step 0] channel_map=right,left

Restart the pulseaudio deamon by running pulseaudio -k; pulseaudio -D

How do I record stuff?

(Question added 2009-03-14)

Selecting the input device

Most recording programs use ALSA. In those cases you need to set the input device in the recording program to be either "default" (if you have PulseAudio set up as the default ALSA device) or whatever name you or your distribution has set up for the pulse plugin for ALSA. If the recording program offers input devices that have a more descriptive name, including the sound card name or something like that, don't choose those, because in those cases the recording program will access the sound card directly (that may not be such a bad thing, actually, if you don't have any real reason to record through PulseAudio).

If the recording program uses ALSA, and you configure it to use the pulse plugin (as described in the previous paragraph), then the real device you will use depends on which source (source is a PulseAudio term for input devices, physical or virtual) is the default. If you're not happy with the default, you can change the default for example with pavucontrol, but there's another way to choose the device. When the PULSE_SOURCE environment variable is set when starting the recording program, the environment variable will be used to select the source. So, if we assume that you want to use the fictious myrecordingapp program, you would start it in the console like this:

PULSE_SOURCE=<source_name> myrecordingapp

<source_name> is of course the name of the source you want to use, but how to find out what the name is? This command lists all available sources:

LANG=C pactl list | grep -A2 'Source #' | grep 'Name: ' | cut -d" " -f2

The names can be rather cryptic, but hopefully you can figure out which name identifies which source.

Which program to use for recording

I don't do any recording myself, so I'm not an expert on what is available, but here is some information anyway. If you know more, please edit this! The situation for PulseAudio support in GUI programs doesn't look too good. Audacity is quite nice, but getting it to work can currently be a hassle. See PerfectSetup#Audacity for more information. I gave Jokosher 0.9 a quick try recently, but didn't get it to work properly. In my experience, command line programs work much better (you can of course edit the file with any program you want after recording it). The two programs I have tested are arecord (comes with ALSA) and parec (comes with PulseAudio). I'll give an example command for both:

parec:

parec --device=alsa_input.usb_device_46d_990_BF15FF18_if2_alsa_capture_0 --format=s16le --rate=44100 --channels=2 | sox --type raw -s2L --rate 44100 --channels 2 - --type wav output.wav

As you can see, in addition to parec, we use sox. That is done because parec generates only raw data, which is not nice to work with. sox is used to add the WAV header to the file. For information about parec and sox parameters, see parec --help and man sox, respectively.

arecord:

PULSE_SOURCE=alsa_input.usb_device_46d_990_BF15FF18_if2_alsa_capture_0 arecord --format=S16_LE --rate=44100 --channels=2 output.wav

That command assumes that PulseAudio is the default ALSA device. If that's not the case, but instead PulseAudio is assigned a device name "pulse", add -D pulse somewhere to the command (see man arecord for more information about the parameters).

And the arecord command seems to not be working because the PULSE_SOURCE seems to be out of order. But the parec is working.

How do I record other programs' output?

(Question added 2009-03-14)

After reading the answer to the previous question, all you need to know is that every sink automatically has a monitor source, which you can use for recording. To list all monitor source names, run

pactl list | grep -A2 'Source #' | grep 'Name: .*\.monitor$' | cut -d" " -f2

How do I play sounds from my chroot to my host's Pulseaudio daemon?

(Question added 2009-04-29)

First, ensure that pulseaudio is working and playing audible sound within your host. If it is not, look through the rest of this wiki and make sure it works.

Setup of the chroot is outside the scope of this wiki, and should be done in whatever manner is recommended by your distro. Pulseaudio then needs to be installed within the chroot. A default install should be sufficient, no need to run the daemon or change any of the .conf files. Finally, Pulseaudio needs access to certain folders within the host system, the following list should enable all functionality on a default Pulseaudio install.

/var/lib/dbus
~/.pulse
/tmp
/dev/shm

/var/lib/dbus contains a reference to the machine-id which pulse uses to identify the local instance of the daemon. ~/.pulse contains various files created by the daemon, one of them, named <id_number>:runtime, symlinks to a folder in /tmp, hence both these folders should be accessible. /dev/shm is needed for proper running of pulseaudio, to prevent audio glitches.

Most guides to setting up chroots (for example, 32-bit chroots in a 64-bit host) will already have /home and /tmp mounted, at least. Some would include /dev/shm, but very few, if any, would include /var/lib/dbus. You can manually mount the appropriate directories using the following commands, where /path/to/chroot is the path to your chroot.

mount --bind /var/lib/dbus /path/to/chroot/var/lib/dbus
mount --bind /home /path/to/chroot/home
mount --bind /tmp /path/to/chroot/tmp
mount --bind /dev/shm /path/to/chroot/dev/shm

If you use a helper app for your chroot, such as schroot, you will need to change the mount options for that helper app instead. Here is an example of a /etc/schroot/mount-<name_of_chroot> file, including the necessary options for Pulseaudio within the chroot:-

# mount.defaults: static file system information for chroots.
# Note that the mount point will be prefixed by the chroot path
# (CHROOT_PATH)
#
# <file system>	<mount point>	<type>	<options>	<dump>	<pass>
/proc		/proc		proc	defaults	0	0
/proc/bus/usb	/proc/bus/usb	none	rw,bind		0	0
/dev		/dev		none	rw,bind		0	0
/dev/pts	/dev/pts	none	rw,bind		0	0
/dev/shm	/dev/shm	none	rw,bind		0	0
/sys		/sys		none	rw,bind		0	0
/tmp		/tmp		none	rw,bind		0	0
/var/lib/dbus	/var/lib/dbus	none	rw,bind		0	0
/home		/home		none	rw,bind		0	0

Download in other formats:

Plain Text

Context Navigation

TracNav

Home

About PulseAudio

PulseAudio Community

Download PulseAudio

Documentation

Developer Documentation

Frequently Asked Questions

Uncategorized questions

How does PulseAudio compare with ESOUND/aRts/NAS?

Is PulseAudio a GNOME program?

I want to write a new driver for PulseAudio, are there any docs?

The pulseaudio executable is installed SUID root by default. Why this? Isn't this a potential security hole?

Is PulseAudio capable of providing synchronized audio playback over the network for movie players like MPlayer?

What environment variables does PulseAudio care about?

How do the PulseAudio libraries decide where to connect to?

Why the heck does libpulse link against libX11?

What is this RTP/SDP/SAP thing all about?

Why did you rename Polypaudio to PulseAudio?

Compatibility with other software

What about ESOUND compatibility?

Can I integrate PulseAudio in my GLIB/GTK/GNOME application?

Can I integrate PulseAudio in my Qt/KDE application?

What about compatibility with NAS?

What about compatibility with aRts?

Can I get OSS and ALSA applications to work with PulseAudio?

Is there a Win32/Windows driver that act like a PulseAudio client?

Troubleshooting

I get this error message: "Connection refused"

I get this error message: "Invalid argument"

I often hear noises when playing back with PulseAudio, what can I do?

I have a surround sound card, but PulseAudio uses just the front speakers!

When sending multicast RTP traffic it is recieved on the entire LAN but not by the sender machine itself!

Hmm, tcpwrap (aka /etc/hosts.allow, /etc/hosts.deny) support doesn't work properly for me!

Why doesn't the system-wide mode detect my default.pa file? Or, why does the system-wide daemon exit telling me there are no modules?

Why do I not see remote sinks in control applications?

I want to do this specific thing…

I want to run PulseAudio only when it is needed, how do I do this?

How do I list all PulseAudio modules installed?

How do I use PulseAudio over the network?

I saw that SIGUSR2 provokes loading of the module module-cli-protocol-unix. But how do I make use of that?

How can I use PulseAudio as an RTP based N:N multicast conferencing solution for the LAN?

How can I use PulseAudio to stream music from my main PC to my LAN with multiple PCs with speakers?

How can I use PulseAudio to share a single LINE-IN/MIC jack on the entire LAN?

Can I have more than one multicast RTP group?

Can I use PulseAudio to playback music on two sound cards simultaneously?

Can I use PulseAudio to combine two stereo soundcards into a virtual surround sound card?

How can use my Windows box to play the sound from my Linux box?

How can I reverse my left and right speaker channels?

How do I record stuff?

Selecting the input device

Which program to use for recording

How do I record other programs' output?

How do I play sounds from my chroot to my host's Pulseaudio daemon?

Download in other formats:

I saw that `SIGUSR2` provokes loading of the module `module-cli-protocol-unix`. But how do I make use of that?