Soundmeter (Noise Thresholds)

AUTHOR: Cynthia

Outcome

Initial Goal: Sound threshold trigger

Realized Outcome: Yes (portaudio and respeaker bug — inconsistent warning but we're using it to sense not playback or recording)

Hardware Requirements: respeaker USB array for Reachy, but can be any mic input configured for appropriate collection of audio data

Install requirements

Install packages with root privileges, apt-get install

alsa-utils

portaudio19-dev
sudo pip install soundmeter (background process)

Then...

Save below 'Reference Files', config (optional, if needed), trigger.sh bash script, and espeak python script that can be run for random messages
Command line to run to trigger script with threshold (threshold needs to be tuned for your own environment a loud clap or shutting of doors may be 1.5 to 2.0x higher than normal noise levels)

soundmeter --trigger +2000 --action exec --exec ~/.soundmeter/trigger.sh

Also, USB webcam — fswebcam package to take test images (in trigger.sh)

Note: Reference Files on rPi4 / Circuit Launch Setup

Trigger Bash Script — executed upon sound threshold met (trigger.sh)
Profile Variable (—profile [soundmeter] or —profile [test]) pulls parameters from config file
Sample espeak Python script — can be run via trigger.sh file

espeak is currently commented out so trigger is acting only as a test image script

dependency: fswebcam (installed to operate usb webcam for images only)

(resolution explicitly for '352x288' instead of default ⇒ testimage is timestamped for debug)

Compatibility between sampling of recording and playback rates between respeaker and portaudio — it's not set the same for respeaker which generates a non-persistent warning
Take time to test and figure out context of sound in your environment - it's different in each area and echos may be an issue as well as mic differences, and hardware differences
May want to just consider integrate message sending only to let background process run to leave it running (w/o interrupt)

TBD

TBD