Docker-pulseaudio-micro-record

This is a minimal Docker-Ubuntu configuration that is able to record sounds from a microphone using Pyaudio.

Getting access to microphone from Pyaudio is a requirement in many Speech Recognition Systems and they are usually based on Dockers. It is not uncommon for accessing microphones from those Docker containers to become one of the biggest problems. Reported issues include not having access at all or recorded muted signal.

Requirements

This repo has been tested with Ubuntu 18.04, Python 3.6.9, Pyaudio 0.2.11, Docker version 19.03.11 and Pulseaudio 11.1. However it works with many other configurations.

Instructions

First configure pulseaudio.

In terminal:

sudo apt-get install paprefs

In terminal:

paprefs

In the pop up window (pulseaudio preferences) on "Network Server" tab: Check "Enable network access to local sound devices" and check “don’t require authentification”.

Then restart Ubuntu

Execute test

Clone repository:

git clone https://github.com/ivangtorre/Docker-pulseaudio-micro-record && cd Docker-pulseaudio-micro-record

Build docker:

sudo docker build -t pyaudio .

Execute docker:

sudo docker run -it --rm --network host -e DISPLAY=$DISPLAY -v /tmp/.X11-unix/:/tmp/.X11-unix -v /home:/home --device /dev/bus/usb --device /dev/snd pyaudio

Execute test: The test first will show available input devices and ask the user to input the number of the device to use. Then, it will record sounds from that file during two seconds and save an audio file.

python3 test.py

To check the audio file copy it to outside Docker (replace $user with any other path):

cp pyaudiotest.wav /home/$user/