meetplay-py

Info

A simple python3 program that listens to the computer (or microphone) audio, and plays a sound file if it finds a specific word. Originally written for unmuting on Google Meet, playing a sound, then muting the microphone back again, but it can be used for any other purpose.

Works only on Linux.

Used libraries:

• speech_recognition
• playsound
• pynput
• pyaudio

Installation

Debian/Ubuntu requires these packages in addition to pip modules:

sudo apt install portaudio19-dev python3-gst-1.0

Use pip to get the modules:

pip3 install speechrecognition
pip3 install playsound
pip3 install pynput
pip3 install pyaudio

If using Google Cloud Speech API:

pip3 install google-cloud-speech
pip3 install oauth2client
pip3 install google-api-python-client

If using CMU Sphinx:

pip3 install pocketsphinx

For other engines, you are on your own.

Usage:

First of all, check your pulseaudio device names - microphone and output monitor:

pacmd list-sources | grep -e 'index:' -e device.string -e 'name:'

Then, copy the names over to strings audioDevice and microphoneDevice in meetplay.py. For example:

audioDevice = "alsa_output.device.monitor"
microphoneDevice = "alsa_input.device"

If using Google Cloud Speech API, create a project in Google Cloud Console, enable the Cloud Speech-To-Text API, then create a service account and create a key. Download the JSON file and copy its contents to GOOGLE_CLOUD_SPEECH_CREDENTIALS:

GOOGLE_CLOUD_SPEECH_CREDENTIALS = r"""YOUR JSON FILE CONTENTS HERE"""

Be aware that with the Google API only 60 minutes of listening are free of charge.

To use CMU Sphinx instead of Google's API, replace

text = r.recognize_google_cloud(audio, credentials_json=GOOGLE_CLOUD_SPEECH_CREDENTIALS, language='en-GB')

with

text = r.recognize_sphinx(audio)

To set the words the program will listen for, edit the word list in meetplay.py:

# word list to listen for
wordList = ["example", "test", "words"]

For further information, read the comments in meetplay.py.