Question: Is it possible to run pyaudio on Ubuntu 22.04 under WSL2 with Windows 11?

[EdSeymore]

Would like to access the microphone and speakers from python3 application. I tried various approaches without success. I also was unable to get gradio to function with python3 on ubuntu. Also, the speakers do not honor my request to default to HDMI speakers when a USB micropohne with an earphone port is present without the earphone attached.

[bulutthecat]

I am having the same issue, is anybody planning to fix this? I use WSL to test applications before I deploy them to edge compute platforms.

[butzhang]

I have the same issue

[butzhang]

Update: I later bought a M3 pro as a workaround

[ZFYL]

sigma move
btw i found a way here: https://www.linuxuprising.com/2021/03/how-to-get-sound-pulseaudio-to-work-on.html
but still a lacking solution as of 2024

[EdSeymore]

Please see #9896 for additional information.

I have it running successfully with many ALSA lib messages being displayed.

[butzhang]

I am at the same place as you and have successfully installed pyaudio. and getting a stream of error from ALSA lib

[bulutthecat]

Please see #9896 for additional information.

I have it running successfully with many ALSA lib messages being displayed.

There is an issue with ALSA lib, so far I have found a couple alternatives, but I would have to rewrite my program to support them. #9896 is only one of the issues regarding microphone usage in WSL. unless anybody knows of any good alternative I think we will have to wait.

[EdSeymore]

We should report this issue as a error with the WSL2 implementation. Having both Windows and Ubuntu running on the same computer seems like a very good arrangement going forward. Microsoft should put forth the effort to insure success of this implementation. This should not be left as an exercie for the end user.

[bulutthecat]

We should report this issue as a error with the WSL2 implementation. Having both Windows and Ubuntu running on the same computer seems like a very good arrangement going forward. Microsoft should put forth the effort to insure success of this implementation. This should not be left as an exercie for the end user.

I agree, I have had similar issues before with passing on USB devices to WSL2. It would be a better idea to approach this more from a VM standpoint (proper pass-through for USB devices and disk mounting, along with GUI application support) then an in house OS implementation. At least then we could work with Xwindow and GUI applications without needing to start a server (and install a desktop environment). I am going to switch to a Raspberry PI for testing until these issues get resolved. I would love to see Microsoft pushing their Hyper-V virtualisation and making a better interface for WSL2 through their existing Hyper-V platform (which should be able to support all of this anyways).

[Jeronymous]

At the time being, has anyone succeeded to have pyaudio working in WSL2?

[ZFYL]

not yet
but it's getting frustrating
maybe we can have some kind of an audio service that feeds the audio stream from windows to a wsl2 audio client, that is a microphone

[gprieto]

Hey guys, some progress on my side!

I found out that if I run pyAudio with sudo, it does manage to list available devices.

The following works if I run it with sudo python list_devices.py

import speech_recognition as sr

for index, name in enumerate(sr.Microphone.list_microphone_names()):
    print("Microphone with name \"{1}\" found for `Microphone(device_index={0})`".format(index, name))

I have tried adding myuser to the audio, pulse and pulse-access groups but nothing has changed.

I also found out that the pulse audio server from WSLg is owned by user pulse and while myuser is able to record and play through it, sudo doesn't.

#this works
arecord test.wav
aplay test.wav


#this doesn't
sudo arecord test.wav
sudo aplay test.wav

Maybe @EdSeymore this shines some light for others to come up with the missing pieces. But I feel it should be working...

On a side note, I recompiled the linux kernel enabling snd-dummy and snd-aloop modules for ALSA. Again, issue with permissions as sudo is able to see those virtual devices while myuser does not:

arecord -l
arecord: device_list:274: no soundcards found...

sudo arecord -l
**** List of CAPTURE Hardware Devices ****
card 0: Dummy [Dummy], device 0: Dummy PCM [Dummy PCM]
  Subdevices: 8/8
  Subdevice #0: subdevice #0
  Subdevice #1: subdevice #1
  Subdevice #2: subdevice #2
  Subdevice #3: subdevice #3
  Subdevice #4: subdevice #4
  Subdevice #5: subdevice #5
  Subdevice #6: subdevice #6
  Subdevice #7: subdevice #7
card 1: Loopback [Loopback], device 0: Loopback PCM [Loopback PCM]
  Subdevices: 8/8
  Subdevice #0: subdevice #0
  Subdevice #1: subdevice #1
  Subdevice #2: subdevice #2
  Subdevice #3: subdevice #3
  Subdevice #4: subdevice #4
  Subdevice #5: subdevice #5
  Subdevice #6: subdevice #6
  Subdevice #7: subdevice #7
card 1: Loopback [Loopback], device 1: Loopback PCM [Loopback PCM]
  Subdevices: 8/8
  Subdevice #0: subdevice #0
  Subdevice #1: subdevice #1
  Subdevice #2: subdevice #2
  Subdevice #3: subdevice #3
  Subdevice #4: subdevice #4
  Subdevice #5: subdevice #5
  Subdevice #6: subdevice #6
  Subdevice #7: subdevice #7

[r0kk]

If it will be of help to anyone, following worked for me after a lot of debugging in Ubuntu with sounddevice python library:
• Windows: Make sure that under Privacy & security, Microphone, permissions are set ON for Let desktop apps access your microphone
• WSL2 Ubuntu: sudo apt install libasound2-plugins

wls and os version: WSL version: 2.3.26.0 Kernel version: 5.15.167.4-1 WSLg version: 1.0.65 MSRDC version: 1.2.5620 Direct3D version: 1.611.1-81528511 DXCore version: 10.0.26100.1-240331-1435.ge-release Windows version: 10.0.22631.4460