Splitting audio ports in Linux for simultaneous playback

Introduction

On windows some laptops have the option to enable the playback of different audio streams on different audio outputs, for example you could listen to a video call in the browser through your wired headphones, while playing music from Spotify on the internal speakers.

Such option can have a different name depending on the sound card or laptop's vendor, for my laptop, an HP Omen 15-dc100xxx with a ALC295 card it is called multistreaming and can be enabled from the OMEN Audio Control program.

Here's some screenshot showcasing that feature on Windows:

And this is how you route programs' audio to an individual output:

On Linux I don't have such an option, so can only listen one output at a time

That is:

• You have to unplug your wired speakers/headphones to be able to play anything from the internal speakers,

• And if you connect a wired speakers/headphones:

  • The internal speaker is muted, hidden from the gui and marked as unavailable

  

  • Audio streams are switched to your wired output without giving you the desired output option like windows does

  

The final result

If you follow until part 2

• Allow selecting and playing audio from the internal speakers while a wired output is connected

If you follow till the end

• Have a separate audio sink for each output
• Play different audio streams on each card port, simultaneously (e.g. Spotify on speakers and Firefox on Headphones)

Note: The program used in above screenshots is pavucontrol

Requirements

• PipeWire
• alsa-card-profiles
• alsa-utils
• alsa-tools
• pavucontrol

1 Preparations, analyze how the card appears on Linux

1.1 Gather some information about the card

Run pactl list cards and save the output somewhere, bellow is a stripped version of mine to keep the important things

Card #46
    Name: alsa_card.pci-0000_00_1f.3
    Driver: alsa
    Properties:
            api.alsa.path = "hw:0"
            device.product.id = "0xa348"
            device.vendor.id = "0x8086"
    ...
    Profiles:
            output:analog-stereo+input:analog-stereo: Analog Stereo Duplex (sinks: 1, sources: 1, priority: 6565, available: yes)
            output:analog-stereo: Analog Stereo Output (sinks: 1, sources: 0, priority: 6500, available: yes)
            input:analog-stereo: Analog Stereo Input (sinks: 0, sources: 1, priority: 65, available: yes)
            pro-audio: Pro Audio (sinks: 4, sources: 1, priority: 1, available: yes)
    Active Profile: output:analog-stereo+input:analog-stereo
    Ports:
            analog-output-speaker: Speakers (type: Speaker, priority: 10000, latency offset: 0 usec, availability group: Legacy 3, availability unknown)
                    Properties:
                            port.type = "speaker"
                            port.availability-group = "Legacy 3"
                            device.icon_name = "audio-speakers"
                            card.profile.port = "2"
                    Part of profile(s): output:analog-stereo, output:analog-stereo+input:analog-stereo
            analog-output-headphones: Headphones (type: Headphones, priority: 9900, latency offset: 0 usec, availability group: Legacy 4, availability unknown)
                    Properties:
                            port.type = "headphones"
                            port.availability-group = "Legacy 4"
                            device.icon_name = "audio-headphones"
                            card.profile.port = "3"
                    Part of profile(s): output:analog-stereo, output:analog-stereo+input:analog-stereo

Card details:

Name: alsa_card.pci-0000_00_1f.3
Ports: analog-output-speaker analog-output-headphones
Profiles: output:analog-stereo input:analog-stereo output:analog-stereo+input:analog-stereo
device.product.id = "0xa348"
device.vendor.id = "0x8086"

As we can see, this laptop has three audio profiles, one for all audio outputs and other for all inputs, with an extra one that has both inputs and outputs (yours may vary)

1.2 Inspect the audio sinks available for our sound card

Run pactl list sinks and save the output, bellow is a stripped version of mine to keep the important things

Sink #47
    State: RUNNING
    Name: alsa_output.pci-0000_00_1f.3.analog-stereo
    Description: Built-in Audio Analog Stereo
    Driver: PipeWire
    ...
    Ports:
            analog-output-speaker: Speakers (type: Speaker, priority: 10000, availability group: Legacy 3, not available)
            analog-output-headphones: Headphones (type: Headphones, priority: 9900, availability group: Legacy 4, available)
    Active Port: analog-output-headphones
    Formats:
            pcm

Sink details:

Name: alsa_output.pci-0000_00_1f.3.analog-stereo
Ports: analog-output-speaker analog-output-headphones
Active Port: analog-output-headphones

As we can see, there is only one audio sink that can play audio on the two outputs individually (yours may vary)

2 disable Headphone jack detection for speakers

Note: alsa-card-profile files are provided by alsa-card-profiles package

1. First copy the contents of folder /usr/share/alsa-card-profile/mixer/paths/ to /etc/alsa-card-profile/mixer/paths/

   mkdir -p /etc/alsa-card-profile/mixer/
   sudo cp -r /usr/share/alsa-card-profile/mixer/paths/ /etc/alsa-card-profile/mixer/

2. Next we need to modify the mixer path in /etc/alsa-card-profile/mixer/paths/ that matches the speakers port, in my case is analog-output-speaker.

   It is crucial to pick the correct one, to verify you did, change description-key value to something else (e.g. by removing the last letter) and restart pipewire (systemctl restart --user pipewire pipewire-pulse pipewire.socket wireplumber) the mixer path file name will show in pavucontrol or pactl list sinks instead of the actual port name:

   Ports:
       analog-output-speaker: analog-output-speaker (type: Unknown, priority: 10000, availability group: Legacy 4, available)

3. Delete all the other files in /etc/alsa-card-profile/mixer/paths/ leaving only your mixer path and the common file e.g:

   sudo find /etc/alsa-card-profile/mixer/paths/ -type f ! -name 'analog-output-speaker.conf' ! -name 'analog-output.conf.common' -exec rm -f {} +

4. Change the description-key back to the default, then:

   • Set state.plugged = unknown inside the Jack section that best matches the wired port name of your card, in my case is analog-output-headphones so I use [Jack Headphone] one:

     [Jack Headphone]
     state.plugged = unknown
     state.unplugged = unknown

   • Comment (by adding semi-colon on start of line) the Element section that matches the port [Jack ...] in my case is [Element Headphone]:

     ; [Element Headphone]
     ; switch = off
     ; volume = off

5. Save the changes and restart the audio server by running:

   systemctl restart --user pipewire pipewire-pulse pipewire.socket wireplumber

   If everything went well you should have the speaker option available without having to unplug your wired device and

   • Should be able to play audio on them (individually)
   • Audio streams should still change automatically from speakers to your wired output and vice versa when plugging/unplugging a wired device

   

Congratulations! We're closer to our final goal, but you may stop here if this was your desired behavior 🙂

3 Splitting for simultaneous playback with alsa firmware patch (for different applications on each port)

3.1 Analyze the card sub devices

First we need to identify and check if our card has more than one sub device and test if sound comes out of them

Run

aplay -l

You'll get something like the following

**** List of PLAYBACK Hardware Devices ****
card 0: PCH [HDA Intel PCH], device 0: ALC295 Analog [ALC295 Analog]
  Subdevices: 1/1
  Subdevice #0: subdevice #0
...

At this point, my card card 0 (ALC295), has only one sub device, you may have more, take note of this information as we'll need it later

We can confirm the playback (output) and capture (input) streams the card currently has with cat /proc/asound/pcm

00-00: ALC295 Analog : ALC295 Analog : playback 1 : capture 1
...

3.2 Making the alsa firmware patch file

Option 1 Using hdajackretask

1. Run hdajackretask
2. In the Select a codec drop-down select your card
3. In the Options section check Parser hints
4. In the Hints list set indep_hp and `` to yes with double click on them.
5. Press Install boot override
6. Open the file /lib/firmware/hda-jack-retask.fw and add vmaster=no below indep_hp=yes
7. Reboot to apply the changes

Option 2 Manually (non-immutable distributions)

Run

cat /proc/asound/card*/codec#* | grep -E 'Codec|Vendor Id|Subsystem Id|Address'

You'll get the codecs in the following format, pick the one that matches your card (mine is the Realtek ALC295):

Codec: Realtek ALC295
Address: 0
Vendor Id: 0x10ec0295
Subsystem Id: 0x103c8575
...

With the above we can start creating our patch file:

1. Create the file (don't copy as is, modify according to explanation bellow):

   /lib/firmware/hda-jack-retask.fw

   [codec]
   0x10ec0295 0x103c8575 0
   
   [hints]
   indep_hp=yes
   vmaster=no

   Below [codec] line we should put the values of Vendor Id, the Subsystem Id and Address of the card, separated by spaces.

   Below [hints] we need to add what is called hint strings

   • indep_hp=yes this will make for our jack output to be detected as an independent PCM stream with its own controls (meaning it will split away from internal speakers as a separate audio sink for us to play specific app stream on it).

   • vmaster=no will disable the virtual Master control so we can control volume on each port individually.

2. Create the following file:

   /etc/modprobe.d/alsa-base.conf

   options snd-hda-intel patch=hda-jack-retask.fw

3. Reboot to apply the changes

Option 3 Manually for immutable distributions using script and udev rule

For immutable distros /lib/firmware/ is not writable. As a workaround you can use an udev rule that sets the hints on boot. This method may cause some noises during boot and is not warranted to be as reliable as the firmware one, if that's the case, suggestions to improve it are welcome

1. Create the file /etc/udev/rules.d/91-pipewire-alsa-port-split.rules (don't copy as is, modify according to explanation bellow)

   SUBSYSTEM!="sound", GOTO="pipewire_end"
   ACTION!="change", GOTO="pipewire_end"
   KERNEL!="card*", GOTO="pipewire_end"
   
   SUBSYSTEMS=="pci", ATTRS{vendor}=="0x8086", ATTRS{device}=="0xa348", \
   RUN+="/usr/local/bin/alsa-split-ports.sh"
   
   LABEL="pipewire_end"

2. Create the script /usr/local/bin/alsa-split-ports-hints.sh and set CODEC VENDOR_ID and SUBSYSTEN_ID with the ones from your card from cat /proc/asound/card*/codec#* | grep -E 'Codec|Vendor Id|Subsystem Id|Address', note how the CODEC variable doesn't have the vendor name (Realtek) because whe are matching against /sys/class/sound/hw*/chip_name

   #!/usr/bin/env bash
   CODEC="ALC295"
   VENDOR_ID="0x10ec0295"
   SUBSYSTEN_ID="0x103c8575"
   HINTS="indep_hp = yes
   vmaster = no
   "
   
   get_codec_hwdep() {
       local codec=$1
       local vendor_id=$2
       local subsystem_id=$3
       local addr=""
       [[ -z "$codec" || -z "$vendor_id" || -z "$subsystem_id" ]] && { echo "ERROR: Not enough arguments given"; return; }
       for file in /sys/class/sound/hw*; do
           if [[ -n "$addr" ]]; then
           echo "$addr"
           return
           fi
           if grep -q "$codec" "$file/chip_name" && grep -q "$vendor_id" "$file/vendor_id" && grep -q "$subsystem_id" "$file/subsystem_id"; then
               addr=$file
           fi
       done
       if [[ -z "$addr" ]]; then
       echo "ERROR: Could not get address for c:$codec v:$vendor_id s:$subsystem_id"
       return
       fi
   }
   # get_codec_hwdep "$CODEC" "$VENDOR_ID" "$SUBSYSTEN_ID"
   hwdep="$(get_codec_hwdep "$CODEC" "$VENDOR_ID" "$SUBSYSTEN_ID")"
   
   if [[ "$hwdep" == *"ERROR"* || -z "$hwdep" ]]; then
   exit 1
   fi
   
   while IFS=$'\n' read -r line; do
   if [[ -z "$line" ]]; then
       continue
   fi
   echo "$line > ${hwdep}/hints"
   echo "$line" > "${hwdep}"/hints
   done <<< "$HINTS"
   
   echo "echo 1 > ${hwdep}/reconfig"
   echo 1 > "${hwdep}"/reconfig
   
   # wait some time to intialize before restoring
   sleep 5
   alsactl restore

3. Reboot to apply the changes

3.3 Verify that the patch works

Run again

aplay -l

You should get something like the following

**** List of PLAYBACK Hardware Devices ****
card 0: PCH [HDA Intel PCH], device 0: ALC295 Analog [ALC295 Analog]
  Subdevices: 0/1
  Subdevice #0: subdevice #0
card 0: PCH [HDA Intel PCH], device 2: ALC295 Alt Analog [ALC295 Alt Analog]
  Subdevices: 1/1
  Subdevice #0: subdevice #0

Now the card ALC295 (card 0), has an extra sub device (2), interesting...

Run cat /proc/asound/pcm, if there is a new sub device and has a playback sub-stream (playback 1) like below you can continue

00-00: ALC295 Analog : ALC295 Analog : playback 1 : capture 1
00-02: ALC295 Alt Analog : ALC295 Alt Analog : playback 1
...

3.3.1 Identify what output each device corresponds to

1. Plug-in your wired audio device then run alsamixer -c0 (replace 0 with your card #number if needed)

   1. Un-mute any muted device (the ones with MM below the volume slider) pressing m
   2. Increase their volume to around 30 if the're on 0
   3. Enable Independent HP if is not enabled
   4. Press Esc to exit.

   

2. Finally, save it by running

   sudo alsactl store

3. Stop any running pipewire services (may need to stop it multiple times if it gets restarted):

   systemctl --user stop pipewire.service pipewire.socket pipewire-pulse.service pipewire-pulse.socket wireplumber.service

4. Run speaker-test -Dhw:0,0 -c2 (replace 0,0 with the card #number and device #numbers from your card from aplay -l)

   For the device 0 (speaker-test -Dhw:0,0 -c2) the sound comes out from the speakers, so device 0 is the speakers

5. Repeat for all the devices for your card and note which device corresponds to which physical output

So my card's outputs are the following (yours may vary):

• Device 0,0 (card 0, device 0) handles Speakers (and also microphones, as we saw in cat /proc/asound/pcm)
• Device 0,2 (card 0, device 2) handles Headphones

4 Making a custom alsa profile for our card

4.1 Now is time to do a recap of all information we have captured

Card details:

Name: alsa_card.pci-0000_00_1f.3
Ports: analog-output-speaker analog-output-headphones
Profiles: output:analog-stereo input:analog-stereo output:analog-stereo+input:analog-stereo

Sink details:

Name: alsa_output.pci-0000_00_1f.3.analog-stereo
Ports: analog-output-speaker analog-output-headphones
Active Port: analog-output-headphones

Card sub devices and playback (outputs) / capture (inputs) location:

**** List of PLAYBACK Hardware Devices ****
card 0: PCH [HDA Intel PCH], device 0: ALC295 Analog [ALC295 Analog]
  Subdevices: 0/1
  Subdevice #0: subdevice #0
card 0: PCH [HDA Intel PCH], device 2: ALC295 Alt Analog [ALC295 Alt Analog]
  Subdevices: 1/1
  Subdevice #0: subdevice #0

00-00: ALC295 Analog : ALC295 Analog : playback 1 : capture 1
00-02: ALC295 Alt Analog : ALC295 Alt Analog : playback 1
...

So in my case I have:

• Two ports: analog-output-speaker analog-output-headphones
• Device locations for the card 0 correspond to:
  • Device 0,0 (card 0, device 0) Speakers (also handles microphones)
  • Device 0,2 (card 0, device 2) Headphones

4.2 Create the profile

1. Rename your mixer path file from 2 disable Headphone jack detection for speakers like below:

   sudo mv /etc/alsa-card-profile/mixer/paths/analog-output-speaker.conf /etc/alsa-card-profile/mixer/paths/analog-output-speaker-split.conf

2. Now create the file /etc/alsa-card-profile/mixer/profile-sets/split-ports-profile.conf

   sudo mkdir /etc/alsa-card-profile/mixer/profile-sets/
   sudo touch /etc/alsa-card-profile/mixer/profile-sets/split-ports-profile.conf

3. Open split-ports-profile.conf with your preferred editor, paste the following and adapt it to your system according to the comments:

   ; This will let alsa generate automatic profiles (e.g internal speaker + microphone)
   [General]
   auto-profiles = yes
   
   ; device-strings describes the ALSA device string(s) that PulseAudio uses to open the device, where "%f" specifies the card number (should always be present in the string).
   
   ; This is the mapping for the internal speaker
   ; If needed, change the 0 in "hw:%f,0" to your sub device location
   ; You can change the description for this and other mappings if you want
   ; in paths output put the name of the previously created custom mixer path
   [Mapping analog-stereo-speaker]
   description = Speakers
   device-strings = hw:%f,0
   paths-output = analog-output-speaker-split
   channel-map = left,right
   direction = output
   
   ; This is the mapping for the jack output (headphones)
   ; If needed, change the 2 in "hw:%f,2" to your sub device location
   ; in paths output put the name of the from card details
   [Mapping analog-stereo-headphones]
   description = Headphones
   device-strings = hw:%f,2
   paths-output = analog-output-headphones
   channel-map = left,right
   direction = output
   
   ; This is the mapping that will handle internal and external microphones, as you could see in `cat /proc/asound/pcm`, the card also had a capture port
   ; in the 0,0 sub device location so let's add it here too (change the 0 in "hw:%f,0" to your sub device location that has the capture port)
   ; All the paths-input names here came from the default.conf profile set and you may have to adapt it if your input port name is not included
   
   [Mapping analog-stereo-input]
   description = Microphone
   device-strings = hw:%f,0
   channel-map = left,right
   paths-input = analog-input-front-mic analog-input-rear-mic analog-input-internal-mic analog-input-dock-mic analog-input analog-input-mic analog-input-linein analog-input-aux analog-input-video analog-input-tvtuner analog-input-fm analog-input-mic-line analog-input-headphone-mic analog-input-headset-mic
   direction = input
   
   ; Broken in parts the profile name means to join
   ; The name of the Mapping containing the analog-output-headphones (output:analog-stereo-headphones)
   ; The name of the Mapping containing the analog-output-speaker (output:analog-stereo-speaker)
   ; The name of the Mapping containing the analog-stereo-input (input:analog-stereo-input)
   
   ; in output-mappings put the name of the output mappings
   ; input-mappings put the name of the input mappings
   
   ; NOTE: Not to be confused width the paths-output/paths-input inside the mapping, we're not using those directly
   
   ; This is the profile that will have the internal speakers + jack output + all microphones
   ; in paths output put the name of the from card details
   [Profile output:analog-stereo-headphones+output:analog-stereo-speaker+input:analog-stereo-input]
   description = Analog Stereo Duplex
   output-mappings = analog-stereo-headphones analog-stereo-speaker
   input-mappings = analog-stereo-input
   priority = 80
   
   ; This profile will have the internal speakers + jack output, but not microphones
   [Profile output:analog-stereo-headphones+output:analog-stereo-speaker]
   description = Analog Stereo Outputs Only
   output-mappings = analog-stereo-headphones analog-stereo-speaker
   priority = 70

4.3 Link the profile to the card

1. Create the file /etc/udev/rules.d/91-pipewire-alsa-port-split.rules (don't copy as is, modify according to explanation bellow). Also, if using the udev rule to apply hints from Using script and udev rule for immutable distros you can use the commented rule instead.

   SUBSYSTEM!="sound", GOTO="pipewire_end"
   ACTION!="change", GOTO="pipewire_end"
   KERNEL!="card*", GOTO="pipewire_end"
   
   SUBSYSTEMS=="pci", ATTRS{vendor}=="0x8086", ATTRS{device}=="0xa348", \
   ENV{ACP_PROFILE_SET}="/etc/alsa-card-profile/mixer/profile-sets/split-ports-profile.conf"
   
   # Use this instead for immutable distributions
   # SUBSYSTEMS=="pci", ATTRS{vendor}=="0x8086", ATTRS{device}=="0xa348", \
   # ENV{ACP_PROFILE_SET}="/etc/alsa-card-profile/mixer/profile-sets/split-ports-profile.conf" \
   # RUN+="/usr/local/bin/alsa-split-ports.sh"
   
   LABEL="pipewire_end"

2. Replace the vendor and device id with device.vendor.id and device.product.id respectively, that you got in 1.1 Gather some information about the card

3. Reboot to apply the changes

5 Verify the split worked

Open pavucontrol and in the Configuration tab select the Analog Stereo Duplex profile for your card

Run

pactl list sinks | grep -E 'Name|Desc|State|Port|device.profile-set'

I everything went well you should have a separate audio sink for each output:

State: IDLE
Name: alsa_output.pci-0000_00_1f.3.analog-stereo-headphones
Description: Built-in Audio Headphones
        device.profile-set = "/etc/alsa-card-profile/mixer/profile-sets/split-ports-profile.conf"
Ports:
Active Port: analog-output-headphones
State: SUSPENDED
Name: alsa_output.pci-0000_00_1f.3.analog-stereo-speaker.2
Description: Built-in Audio Speakers
        device.profile-set = "/etc/alsa-card-profile/mixer/profile-sets/split-ports-profile.conf"
Ports:
Active Port: analog-output-speaker-split

Now you should be able to play different applications on each sink, you can do that with pavucontrol or KDE's Audio Volume widget. Congratulations! 🎉

Stuck or didn't work? Found an mistake or something isn't clear enough?

If you faced any problems or are stuck please open a new issue including the information as described here

Support me

This guide is the result of days of effort hunting for the right information and lots of reading. If it was useful to you consider making a small donation

Credits & Resources

• Docs
  • Linux Kernel Sound Subsystem Documentation
  • More Notes on HD-Audio Driver
  • Proc Files of ALSA Drivers
  • ALSA Documentation
• How to output unique audio to multiple ports/profiles of a PulseAudio card?