PulseAudio/サンプル

提供: ArchWiki
2017年3月14日 (火) 20:44時点におけるKusakata (トーク | 投稿記録)による版 (Kusakata (トーク) による版 8245 を取り消し)
ナビゲーションに移動 検索に移動

デフォルトの入力ソースを設定

利用可能な入力ソースの一覧を表示するには:

$ pacmd list-sources | grep -e device.string -e 'name:'
name: <input>
 device.string = "hw:2"
name: <oss_input.dsp>
 device.string = "/dev/dsp"
name: <alsa_output.pci-0000_04_01.0.analog-stereo.monitor>
name: <combined.monitor>

システム全体のデフォルトとして設定するには、使用する入力ソースを /etc/pulse/default.pa に追加してください:

set-default-source alsa_output.pci-0000_04_01.0.analog-stereo.monitor

一時的に使用するには:

$ pacmd "set-default-source alsa_output.pci-0000_04_01.0.analog-stereo.monitor"
ヒント: デフォルトのソースは、コマンド内で @DEFAULT_SOURCE@ として参照できます。たとえば: $ pactl set-source-mute @DEFAULT_SOURCE@ toggle

デフォルトの出力ソースを設定

index の前に * が付いている新しいソースの名前を確認:

$ pacmd list-sinks | grep -e 'name:' -e 'index'
  * index: 0
	name: <alsa_output.pci-0000_04_01.0.analog-stereo>
    index: 1
	name: <combined>

システム全体のデフォルトとして設定するには、以下を /etc/pulse/default.pa に追加してください:

set-default-sink alsa_output.pci-0000_04_01.0.analog-stereo

設定したら変更を有効にするためにログアウト/ログインまたは PulseAudio を手動で再起動してください。

ノート:
  • デフォルトに設定されている sink には、インデックスの前に * 印が付きます。
  • sink の番号付けは永続的ではないため、default.pa ファイル内ではすべての sink を名前で識別する必要があります。
  • 実行時に素早く識別したい場合 (音量の制御など)、sink 名の代わりに sink index を使用することができます:
    $ pactl set-sink-volume 0 +3%
    $ pactl set-sink-volume 0 -- -3%
    $ pactl set-sink-mute 0 toggle
    
  • 通常の音量 (100%) を不必要に上書きしないようにするには、サウンドを管理するための別のユーティリティを使用する方がよいでしょう。詳細については フォーラムスレッド を参照してください。
ヒント: デフォルトの sink は @DEFAULT_SINK@ で指定することができます。たとえば: $ pactl set-sink-volume @DEFAULT_SINK@ +5%

HDMI とアナログの同時出力

PulseAudio は、複数のソースに同時に出力することができます。例えば、あるアプリケーションでは HDMI を使用するように設定しつつ、別のアプリケーションではアナログを使用するようにすることができます。また、複数のアプリケーションで同時に音声を受け取ることができます。

$ aplay -l
**** ハードウェアデバイス PLAYBACK のリスト ****
カード 0: Intel [HDA Intel], デバイス 0: ALC889A Analog [ALC889A Analog]
  サブデバイス: 0/1
  サブデバイス #0: subdevice #0
カード 0: Intel [HDA Intel], デバイス 1: ALC889A Digital [ALC889A Digital]
  サブデバイス: 1/1
  サブデバイス #0: subdevice #0
カード 0: Intel [HDA Intel], デバイス 3: HDMI 0 [HDMI 0]
  サブデバイス: 0/1
  サブデバイス #0: subdevice #0

もしくは pacmd コマンドを使用すると:

$ pacmd list-sinks  | grep -e 'name:'  -e 'alsa.device ' -e 'alsa.subdevice '
	name: <alsa_output.pci-0000_00_1b.0.analog-stereo>
		alsa.subdevice = "0"
		alsa.device = "0"

このような構成では pavucontrol の Configuration > Internal Audio で選択されているものがデフォルトのデバイスになります。Load pavucontrol > Configuration を選択し、プロファイルとして HDMI を選択してください。

アナログデバイスをセカンダリソースとして設定するには、/etc/pulse/default.pa の最初の方に以下の行を追加します (他のモジュールよりも先にロードさせるため):

### Load analog device
load-module module-alsa-sink device=hw:0,0
load-module module-combine-sink sink_name=combined
set-default-sink combined

PulseAudio を再起動し pavucontrol を実行して「出力デバイス」タブを選択します。3つの設定が表示されます:

  1. Internal Audio Digital Stereo (HDMI)
  2. Internal Audio
  3. Simultaneous output to Internal Audio Digital Stereo (HDMI), Internal Audio

MPlayer, VLC, mpd などの PulseAudio を使用するプログラムを起動し「再生」タブに切り替えます。実行中のプログラムからドロップダウンリストを使って3つのソースのうちの1つを選択することができるはずです。

詳しくは PulseAudio FAQこちらのスレッド を参照してください。

HDMI 出力の設定

As outlined in ftp://download.nvidia.com/XFree86/gpu-hdmi-audio-document/gpu-hdmi-audio.html#_issues_in_pulseaudio unless the HDMI port is the first output, PulseAudio will not be able to have any audio when using certain graphics cards with HDMI audio support. This is because of a bug in PulseAudio where it will only select the first HDMI output on a device. A work around posted further down is to first find which HDMI output is working by using the aplay utility from ALSA.

The original title for this section indicated the problem is specific to nVidia cards. As seen in this forum thread other cards are affected as well. The rest of the section will use an nVidia card as a case-study but the solution should carry over for people using other affected cards.

HDMI 出力の確認

Then find the working output by listing the available cards

# aplay -l
sample output:
 **** List of PLAYBACK Hardware Devices ****
 card 0: NVidia [HDA NVidia], device 0: ALC1200 Analog [ALC1200 Analog]
   Subdevices: 1/1
   Subdevice #0: subdevice #0
 card 0: NVidia [HDA NVidia], device 3: ALC1200 Digital [ALC1200 Digital]
   Subdevices: 1/1
   Subdevice #0: subdevice #0
 card 1: NVidia_1 [HDA NVidia], device 3: HDMI 0 [HDMI 0]
   Subdevices: 1/1
   Subdevice #0: subdevice #0
 card 1: NVidia_1 [HDA NVidia], device 7: HDMI 0 [HDMI 0]
   Subdevices: 0/1
   Subdevice #0: subdevice #0
 card 1: NVidia_1 [HDA NVidia], device 8: HDMI 0 [HDMI 0]
   Subdevices: 1/1
   Subdevice #0: subdevice #0
 card 1: NVidia_1 [HDA NVidia], device 9: HDMI 0 [HDMI 0]
   Subdevices: 1/1
   Subdevice #0: subdevice #0

適切なカードでテスト

Now a list of the detected cards is known, users will need to test for which one is outputting to the TV/monitor

# aplay -D plughw:1,3 /usr/share/sounds/alsa/Front_Right.wav

where 1 is the card and 3 is the device substitute in the values listed from the previous section. If there is no audio, then try substituting a different device (on my card I had to use card 1 device 7)

PulseAudio を手動で設定して Nvidia HDMI を検出

Having identified which HDMI device is working, PulseAudio can be forced to use it via an edit to /etc/pulse/default.pa:

# load-module module-alsa-sink device=hw:1,7

where the 1 is the card and the 7 is the deivce found to work in the previous section

restart pulse audio

 $ pulseaudio -k
 $ pulseaudio --start

open the sound settings manager, make sure that under the hardware tab the graphics cards HDMI audio is set to "Digital Stereo (HDMI) Output" ( My graphics card audio is called "GF100 High Definition Audio Controller"

Then, open the output tab. There should now be two HDMI outputs for the graphics card. Test which one works by selecting one of them, and then using a program to play audio. For example, use VLC to play a movie, and if it does not work, then select the other.

サラウンドサウンドシステム

Many people have a surround sound card, but have speakers for just two channels, so PulseAudio cannot really default to a surround sound setup. To enable all of 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. For a 5.1 setup, or 8 for a 7.1 setup etc.

# Default
default-sample-channels=2
# For 5.1
default-sample-channels=6
# For 7.1
default-sample-channels=8

If your channels are not correclty mapped or the volume controls for the individual channels do not work as expected in pavucontrol, and you have a HDMI and an anlog soundcard, then try to add the following line to /etc/pulse/default.pa

load-module module-combine channels=6 channel_map=front-left,front-right,rear-left,rear-right,front-center,lfe

Note that this example is for a 5.1 setup.

After doing the edit, restart PulseAudio.

フロント/リアの分割

Connect speakers to front analog output and headphones to rear output. It would be useful to split front/rear to separate sinks. Add to /etc/pulse/default.pa:

 load-module module-remap-sink sink_name=speakers remix=no master=alsa_output.pci-0000_05_00.0.analog-surround-40 channels=2 master_channel_map=front-left,front-right channel_map=front-left,front-right
 load-module module-remap-sink sink_name=headphones remix=no master=alsa_output.pci-0000_05_00.0.analog-surround-40 channels=2 master_channel_map=rear-left,rear-right   channel_map=front-left,front-right

Make sure to replace alsa_output.pci-0000_05_00.0.analog-surround-40 with the sound card name shown in 'pacmd list-sinks'.

Now you have 2 additional sinks which can be used separately. You can choose 'sink_name' freely, as long as there is no sink with that name already. The 'remix' parameter controls whether the audio should be down-/upmixed to match the channels in the sink.

ヒント: If pulseaudio fails with master sink not found, comment out the remapping lines, start PulseAudio and verify your card output is set to the one you specified (e.g. analog surround 4.0). Alternatively, try using a sink index instead of a sink name.

7.1 を 5.1+2.0 に分割

Similar to the example above, you can also split a 7.1 configuration into 5.1 surround and stereo output devices. Set your card to 7.1 mode, then add the following lines to /etc/pulse/default.pa:

 load-module module-remap-sink sink_name=Surround remix=no master=alsa_output.pci-0000_00_14.2.analog-surround-71 channels=6 master_channel_map=front-left,front-right,rear-left,rear-right,front-center,lfe channel_map=front-left,front-right,rear-left,rear-right,front-center,lfe
 load-module module-remap-sink sink_name=Stereo remix=yes master=alsa_output.pci-0000_00_14.2.analog-surround-71 channels=2 master_channel_map=front-left,front-right channel_map=front-left,front-right

Make sure to replace alsa_output.pci-0000_00_14.2 with your sound card name, get it by running 'pacmd list-sinks'. This configuration will use the front/rear/center+lfe (green/black/orange) jacks for the 5.1 sink and the side (grey) jack for the stereo sink. It will also downmix any audio to stereo for the stereo sink, but will not touch the 5.1 output.

ヒント: If pulseaudio fails with master sink not found, comment out the remapping lines, start PulseAudio and verify your card output is set to analog surround 7.1. Alternatively, try using a sink index instead of a sink name.
ノート: The remix=yes parameter will only work if you also have enable-remixing = yes in your /etc/pulse/daemon.conf (default).

LFE リミックス

By default, PulseAudio remixes the number of channels to the default-sample-channels; however, it does not do this for the LFE channel. To enable LFE remixing, uncomment the line:

; enable-lfe-remixing = no

and replace no with yes:

enable-lfe-remixing = yes

then restart Pulseaudio.

ネットワークを介して PulseAudio

One of PulseAudio's unique features is its ability to stream audio from clients over TCP to a server running the PulseAudio daemon reliably within a LAN.

To accomplish this, one needs to enable module-native-protocol-tcp.

TCP サポート (networked sound)

To enable the TCP module, add this to (or uncomment, if already there) /etc/pulse/default.pa on both the client and server:

load-module module-native-protocol-tcp

For this to work, it is a requirement that both the client and server share the same cookie. Ensure that the clients and server share the same cookie file found under ~/.config/pulse/cookie. It does not matter whose cookie file you use (the server or a client's), just that the server and client(s) share the same one.

Note: If experiencing trouble connecting, use (on server)

pacmd list-modules

TCP サポートと匿名クライアント

If it is undesirable to copy the cookie file from clients, anonymous clients can access the server by giving these parameters to module-native-protocol-tcp on the server (again in /etc/pulse/default.pa):

 load-module module-native-protocol-tcp "auth-ip-acl=127.0.0.1;192.168.0.0/24 auth-anonymous=1"

Change the LAN IP subnet to match that of the those clients you wish to have access to the server.

Zeroconf (Avahi) パブリッシング

PulseAudio のリモートサーバーを PulseAudio Device Chooser (pasystray) に表示させるには、適当な zeroconf モジュールをロードして、Avahi デーモンを有効化してください。

On both machines, run:

$ systemctl start avahi-daemon
$ systemctl enable avahi-daemon

On the server, add load-module module-zeroconf-publish to /etc/pulse/default.pa, on the client, add load-module module-zeroconf-discover to /etc/pulse/default.pa. Now redirect any stream or complete audio output to the remote PulseAudio server by selecting the appropriate sink.

If you have issues with the remote syncs appearing on the client, try restarting the Avahi daemon on the server to rebroadcast the available interfaces.

ローカルの X クライアントで使用する PulseAudio サーバーを切り替える

To switch between servers on the client from within X, the pax11publish command can be used. For example, to switch from the default server to the server at hostname foo:

$ pax11publish -e -S foo

Or to switch back to the default:

$ pax11publish -e -r

Note that for the switch to become apparent, the programs using Pulse must be restarted.

何もかもが失敗しているような場合

The following is a quick fix and NOT a permanent solution

On the server:

 $ paprefs 

Go to Network Access -> Enable access to local sound devices (Also check both 'Allow discover' and 'Don't require authentication').

On the client:

 $ export PULSE_SERVER=server.ip && mplayer test.mp3

ALSA モニターのソース

To be able to record from a monitor source (a.k.a. "What-U-Hear", "Stereo Mix"), use pactl list to find out the name of the source in PulseAudio (e.g. alsa_output.pci-0000_00_1b.0.analog-stereo.monitor). Then add lines like the following to /etc/asound.conf or ~/.asoundrc:

pcm.pulse_monitor {
  type pulse
  device alsa_output.pci-0000_00_1b.0.analog-stereo.monitor
}

ctl.pulse_monitor {
  type pulse
  device alsa_output.pci-0000_00_1b.0.analog-stereo.monitor
}

Now you can select pulse_monitor as a recording source.

Alternatively, you can use pavucontrol to do this: make sure you have set up the display to "All input devices", then select "Monitor of [your sound card]" as the recording source.

特定の出力のモニター

It is possible to monitor a specific output, for example to stream audio from a music player into a VOIP application. Simply create a null output device:

pactl load-module module-null-sink sink_name=<name>

In Pulseaudio Volume Control (pavucontrol), under the "Playback" tab, change the output of an application to <name>, and in the recording tab change the input of an application to "Monitor of <name>". Audio will now be outputted from one application into the other.

PulseAudio と JACK

最新の方法

This configuration only works with jackdbus (JACK2 compiled with D-Bus support). It also requires the pulseaudio-jack package. Add to /etc/pulse/default.pa:

load-module module-jackdbus-detect

As described on the Jack-DBUS Packaging page:

Server auto-launching is implemented as D-Bus call that auto-activates JACK D-Bus service, in case it is not already started, and starts the JACK server. Correct interaction with PulseAudio is done using a D-Bus based audio card "acquire/release" mechanism. When JACK server starts, it asks this D-Bus service to acquire the audio card and PulseAudio will unconditionally release it. When JACK server stops, it releases the audio card that can be grabbed again by PulseAudio.

module-jackdbus-detect.so dynamically loads and unloads module-jack-sink and module-jack-source when jackdbus is started and stopped.

If PulseAudio sound does not work, check with pavucontrol to see if the relevant programs appear in the playback tab. If not, add the following to ~/.asound.conf or /etc/asound.conf to redirect ALSA to PulseAudio:

pcm.pulse {
    type pulse
}

ctl.pulse {
    type pulse
}

pcm.!default {
    type pulse
}
ctl.!default {
    type pulse
}

If it still does not work, check with pavucontrol in the playback tab and make sure the relevant programs are outputting to PulseAudio JACK Sink instead of your audio card (which JACK has control of, so it will not work).

新しい方法

The basic idea is that killing PulseAudio is a bad idea because it may crash any apps using PulseAudio and disrupt any audio playing.

The flow of how this setup works:

  1. PulseAudio releases the sound card
  2. JACK grabs sound card and starts up
  3. script redirects PulseAudio to JACK
  4. manually send PulseAudio apps to JACK output (pavucontrol may come in helpful for this)
  5. use JACK programs etc
  6. via script, stop redirecting PulseAudio to JACK
  7. stop JACK and release sound card
  8. PulseAudio grabs sound card and reroutes audio to it directly

With QJackCTL, set up these scripts:

pulse-jack-pre-start.sh set it up as the execute script on startup script

#!/bin/bash
pacmd suspend true

pulse-jack-post-start.sh set this one up as execute script after startup

#!/bin/bash
pactl load-module module-jack-sink channels=2
pactl load-module module-jack-source channels=2
pacmd set-default-sink jack_out
pacmd set-default-source jack_in

pulse-jack-pre-stop.sh "execute script on shutdown"

#!/bin/bash
SINKID=$(pactl list | grep -B 1 "Name: module-jack-sink" | grep Module | sed 's/[^0-9]//g')
SOURCEID=$(pactl list | grep -B 1 "Name: module-jack-source" | grep Module | sed 's/[^0-9]//g')
pactl unload-module $SINKID
pactl unload-module $SOURCEID
sleep 5

pulse-jack-post-stop.sh "execute script after shutdown"

#!/bin/bash
pacmd suspend false

古い方法

The JACK-Audio-Connection-Kit is popular for audio work, and is widely supported by Linux audio applications. It fills a similar niche as PulseAudio, but with more of an emphasis on professional audio work. In particular, audio applications such as Ardour and Audacity (recently) work well with Jack.

PulseAudio provides module-jack-source and module-jack-sink which allow PulseAudio to be run as a sound server above the JACK daemon. This allows the usage of per-volume adjustments and the like for the apps which need it, play-back apps for movies and audio, while allowing low-latency and inter-app connectivity for sound-processing apps which connect to JACK. However, this will prevent PulseAudio from directly writing to the sound card buffers, which will increase overall CPU usage.

To just try PA on top of JACK, have PA load the necessary modules on start:

pulseaudio -L module-jack-sink -L module-jack-source

To use PulseAudio with JACK, JACK must be started before PulseAudio, using whichever method one prefers. PulseAudio then needs to be started loading the two relevant modules. Edit /etc/pulse/default.pa, and change the following region:

### Load audio drivers statically (it is probably better to not load
### these drivers manually, but instead use module-hal-detect --
### see below -- for doing this automatically)
#load-module module-alsa-sink
#load-module module-alsa-source device=hw:1,0
#load-module module-oss device="/dev/dsp" sink_name=output source_name=input
#load-module module-oss-mmap device="/dev/dsp" sink_name=output source_name=input
#load-module module-null-sink
#load-module module-pipe-sink

### Automatically load driver modules depending on the hardware available
.ifexists module-udev-detect.so
load-module module-udev-detect
.else
### Alternatively use the static hardware detection module (for systems that
### lack udev support)
load-module module-detect
.endif

to the following:

### Load audio drivers statically (it is probably better to not load
### these drivers manually, but instead use module-hal-detect --
### see below -- for doing this automatically)
#load-module module-alsa-sink
#load-module module-alsa-source device=hw:1,0
#load-module module-oss device="/dev/dsp" sink_name=output source_name=input
#load-module module-oss-mmap device="/dev/dsp" sink_name=output source_name=input
#load-module module-null-sink
#load-module module-pipe-sink
load-module module-jack-source
load-module module-jack-sink

### Automatically load driver modules depending on the hardware available
#.ifexists module-udev-detect.so
#load-module module-udev-detect
#.else
### Alternatively use the static hardware detection module (for systems that
### lack udev support)
#load-module module-detect
#.endif

Basically, this prevents module-udev-detect from loading. module-udev-detect will always try to grab the sound card (JACK has already done that, so this will cause an error). Also, the JACK source and sink must be explicitly loaded.

QjackCtl とスタートアップ/シャットダウンスクリプト

Using the settings listed above, use QjackCtl to execute a script upon startup and shutdown to load/unload PulseAudio. Part of the reason users may wish to do this is that the above changes disable PulseAudio's automatic hardware detection modules. This particular setup is for using PulseAudio in an exclusive fashion with JACK, though the scripts could be modified to unload and load an alternate non-JACK setup, but killing and starting PulseAudio while programs might be using it would become problematic.

ノート: padevchooser in the following example is deprecated. It is replaced by pasystray

The following example could be used and modified as necessary as a startup script that daemonizes PulseAudio and loads the padevchooser program (optional, needs to be built from AUR) called jack_startup:

#!/bin/bash
#Load PulseAudio and PulseAudio Device Chooser

pulseaudio -D
padevchooser&

as well as a shutdown script to kill PulseAudio and the Pulse Audio Device Chooser, as another example called jack_shutdown also in the home directory:

#!/bin/bash
#Kill PulseAudio and PulseAudio Device Chooser

pulseaudio --kill
killall padevchooser

Both scripts need to be made executable:

chmod +x jack_startup jack_shutdown

then with QjackCtl loaded, click on the Setup button and then the Options tab and tick both "Execute Script after Startup:" And "Execute Script on Shutdown:" and put either use the ... button or type the path to the scripts (assuming the scripts are in the home directory) ~/jack_startup and ~/jack_shutdown making sure to save the changes.

PulseAudio と OSS

Add the following to /etc/pulse/default.pa:

load-module module-oss

Then start PulseAudio as usual, making sure that sinks and sources are defined for OSS devices.

chroot から PulseAudio (e.g. 32-bit chroot in 64-bit install)

Since a chroot sets up an alternative root for the running/jailing of applications, PulseAudio must be installed within the chroot itself (pacman -S pulseaudio within the chroot environment).

PulseAudio, if not set up to connect to any specific server (this can be done in /etc/pulse/client.conf, through the PULSE_SERVER environment variable, or through publishing to the local X11 properties using module-x11-publish), will attempt to connect to the local pulse server, failing which it will spawn a new pulse server. Each pulse server has a unique ID based on the machine-id value in /var/lib/dbus. To allow for chrooted apps to access the pulse server, the following directories must be mounted within the chroot:-

/run
/var/lib/dbus
/tmp
~/config/.pulse

/dev/shm should also be mounted for efficiency and good performance. Note that mounting /home would normally also allow sharing of the ~/.pulse folder.

PulseAudio selects the path to the socket via XDG_RUNTIME_DIR, so be sure to drag it along when you chroot as a normal user using sudo (see Sudo#Environment variables).

For specific direction on accomplishing the appropriate mounts, please refer to the wiki on installing a bundled 32-bit system, especially the additional section specific to PulseAudio.

PulseAudio サーバーの自動生成を無効化

Some users may prefer to manually start the PulseAudio server before running certain programs and then stop the PulseAudio server when they are finished. A simple way to accomplish this is to edit /etc/pulse/client.conf and change autospawn = yes to autospawn = no, and set daemon-binary = /bin/true. Make sure the two lines are uncommented as well.

/etc/pulse/client.conf
autospawn = no
daemon-binary = /bin/true 

Now you can manually start the pulseaudio server with

$ pulseaudio --start

and stop it with

$ pulseaudio --kill

You may also have to move or delete a .desktop file in /etc/xdg/autostart if it exists.

ステレオからモノにリマップ

Remap a stereo input-sink to a mono sink by creating a virtual sink. It would be useful if you only have one speaker. Add to /etc/pulse/default.pa:

load-module module-remap-sink master=alsa_output.pci-0000_00_1f.5.analog-stereo sink_name=mono channels=2 channel_map=mono,mono
# Optional: Select new remap as default
set-default-sink mono

(replace alsa_output.pci-0000_00_1f.5.analog-stereo in the sound card name shown from pacmd list-sinks)

Switch player between virtual mono sink and real stereo sink.

左/右チャンネルの交換

Swap/reverse stereo channels by creating a virtual sink. Add to /etc/pulse/default.pa:

load-module module-remap-sink sink_name=reverse-stereo master=alsa_output.pci-0000_00_1f.5.analog-stereo channels=2 master_channel_map=front-right,front-left channel_map=front-left,front-right
set-default-sink reverse-stereo

(replace alsa_output.pci-0000_00_1f.5.analog-stereo in the sound card name shown from pacmd list-sinks)

module-remap-sink documentation

PulseAudio as a minimal unintrusive dumb pipe to ALSA

Some people do not want to run PulseAudio all the time for various reasons. This example will turn the full fledged audio server into an unobstrusive dumb pipe to ALSA devices that automatically starts and stops itself when done, allowing applications that requires PulseAudio to fully function while not touching any ALSA setting nor setting itself as the default ALSA device.

This configuration tells native PA clients to autospawn the daemon when they need it, then the daemon is configured to autoexit as soon as all clients have disconnected. The daemon itself uses a plain simple static configuration that uses your configured pcm.!default ALSA devices and nothing more. No replacement of ALSA's default, no playing with mixer levels, nothing but record/playback. Also make sure pulseaudio-alsa is not installed so standard ALSA clients don't default to pulse. alsamixer functions properly as well as any other ALSA clients. Also make sure common frameworks like Xine, Gstreamer and Phonon are configured to use ALSA: by default if they detect PulseAudio is installed they will try to use it before ALSA.

/etc/pulse/daemon.conf
# Replace these with the proper values
exit-idle-time = 0 # Exit as soon as unneeded
flat-volumes = yes # Prevent messing with the master volume
/etc/pulse/client.conf
# Replace these with the proper values

# Applications that uses PulseAudio *directly* will spawn it,
# use it, and pulse will exit itself when done because of the
# exit-idle-time setting in daemon.conf
autospawn = yes
/etc/pulse/default.pa
# Replace the *entire* content of this file with these few lines and
# read the comments

.fail
    # Set tsched=0 here if you experience glitchy playback. This will
    # revert back to interrupt-based scheduling and should fix it.
    #
    # Replace the device= part if you want pulse to use a specific device
    # such as "dmix" and "dsnoop" so it doesn't lock an hw: device.
    
    # INPUT/RECORD
    load-module module-alsa-source device="default" tsched=1
    
    # OUTPUT/PLAYBACK
    load-module module-alsa-sink device="default" tsched=1 
    
    # Accept clients -- very important
    load-module module-native-protocol-unix

.nofail
.ifexists module-x11-publish.so
    # Publish to X11 so the clients know how to connect to Pulse. Will
    # clear itself on unload.
    load-module module-x11-publish
.endif