SoundMeter

SoundMeter is a command-line tool to obtain sound power in real time. It basically turns the audio recording functionality into a sound meter for machines that ship with audio input devices (e.g. microphone). It currently reveals the root-mean-square (RMS) of sound fragments, which is a measure of the power in an audio signal. The actual values also depend on the system settings of sound input.

Features

• A command-line meter that supports triggering upon events
• Monitor API for backend module programming

Dependencies

System packages

On Debian/Ubuntu:

$ sudo apt-get install portaudio19-dev python-dev alsa-utils

On Fedora/RHEL:

$ sudo yum install portaudio-devel python-devel alsa-utils

On OS X:

$ brew install portaudio

PyPI packages

• argparse
• pyaudio
• pydub
• python-daemon

Installation

You can install the package with pip using the following command to allow externally hosted packages:

$ pip install soundmeter --allow-all-external --allow-unverified pyaudio

Or, you can download a source distribution and install with these commands:

$ python setup.py install

Usage

The simplest usage is to run "soundmeter" from command-line without any options:

$ soundmeter

Collect RMS values for 10 seconds so that you will know the sound level in the current environment:

$ soundmeter --collect --seconds 10
Collecting RMS values...
   154  Timeout
Collected result:
    min:        152
    max:        211
    avg:        156

You can set trigger and action for soundmeter.

Stop the soundmeter if RMS is greater than 211 consecutively for 3 times:

$ soundmeter --trigger +211 3 --action stop

Execute trigger.sh if RMS is greater than 211:

$ soundmeter --trigger +211 --action exec --exec trigger.sh

Execute trigger.sh and stop soundmeter if RMS is less than 152 consecutively for 3 times:

$ soundmeter --trigger -152 3 --action exec-stop --exec trigger.sh

Run the soundmeter with trigger and action in the background:

$ soundmeter --trigger +211 3 --action exec --exec trigger.sh --daemonize

Run the soundmeter for 2 minutes and log to meter.log:

$ soundmeter --seconds 120 --log meter.log

Command-line Options

The "soundmeter" command accepts the following options:

-p PROFILE, --profile PROFILE
	config profile (section name)
-c, --collect	collect RMS values to determine thresholds
-s SECS, --seconds SECS
	time in seconds to run the meter (default forever)
-a ACTION_TYPE, --action ACTION_TYPE
	triggered action (stop, exec-stop and exec)
-t THRESHOLD, --trigger THRESHOLD
	trigger condition (threshold RMS and an optional number of consecutive triggering times, which defaults 1)
-e FILE, --execute FILE
	shell script to execute upon trigger (defaults to ~/.soundmeter/trigger.sh), can be configured to pass the "last triggering" RMS value as argument by setting rms_as_trigger_arg to True in ~/.soundmeter/config
-d, --daemonize
	run the meter in the background
--log LOGFILE	log the meter (defaults to ~/.soundmeter/log)
-v, --verbose	verbose mode
--segment SECONDS
	audio segment length recorded in seconds (defaults to 0.5); when specified, it overrides audio_segment_length in ~/.soundmeter/config

Config

Some "dependency-required" parameters can be configured at ~/.soundmeter/config. The default configuration is:

[soundmeter]
frames_per_buffer = 2048
format = 8
channels = 2
rate = 44100
audio_segment_length = 0.5
rms_as_trigger_arg = False

You can have multiple sections in the config file and specify the one to use with the --profile command-line options. The default profile name is soundmeter. For example:

[soundmeter]
frames_per_buffer = 2048
format = 8
channels = 2

[test]
frames_per_buffer = 1024
format = 8
channels = 1

To use the test profile:

$ soundmeter --profile test ...

There is also an input_device_index parameter, which specifies the index of input device to use. If unspecified, it uses the default input device.