This is a Python-script to batch-en- and decode audio samples to generate graphs of audio metric scores.
This is a graph generated from ODG-scores as determined with GstPEAQ as audio quality metric.
Note that objective audio metrics augment (but do not replace) actual listening tests. Automated metrics can provide approximate guidance, though.
- An audio quality measurement tool, such as GstPEAQ. The script
peaqer.py
derives its name from the PEAQ audio quality metric. - Some audio en- and decoders, such as Opus tools or LAME etc.
- FFmpeg
python3
, withmatplotlib
installed. On Debian and Ubuntu, this can be obtained with thepython3-matplotlib
package, but installing viapip
should be plenty fine as well.
The script peaqer.py
does not have any command line arguments - all settings regarding the encoding-experiments are set in settings.json
.
Proposed workflow:
- Assemble a directory with
.wav
files. For audio quality testing, usually rather short files with 5 to 20 seconds of audio content are employed. - Copy
settings.json
into that directory and customize it to your needs. - Start
peaqer.py
with the current workdir being the aforementioned directory with.wav
files andsettings.json
. - Be patient while
peaqer.py
en- and decodes the samples files with all encoders and bitrates and invokes the configured audio quality metric tools. This can take a while.