sound_weighting_filters 0.9.3

IIR coefficients for A,B,C and ITU_R_468 weighting filters

Two modules which provide the zpk coefficients for the A,B,C and ITU_R_468 weighting filters. These can then be used by signal.lfilter to filter audio signals.

In contrast to other implementations which get the high frequency end completely wrong (because of the bilinear transform) here I have used the matched z-transform which aims to match 1:1 the analogue and digital frequency response.

The precision towards the Nyquist frequency can be increased by using a higher sampling rate but given that audio has such low energy over 10kHz there is probably not much point to it.

On the other hand lower sampling rates are possible for example 8kHz if only low frequency noise has been measured.

Installation

The preferred way to install is with pip / pip3:

pip install sound_weighting_filters

Usage

Import the module:

import ABC_weighting
import ITU_R_468_weighting

This has a single function:

get_zpk(curve='A', fs=False)

• curve defines the weighting filter and can be ‘A’, ‘B’ or ‘C’.

• fs sets the sampling rate of the digital system. Or False for an analogue filter.

and returns zeros, poles, gain of the filter.

For example:

z,p,k = ABC_weighting.get_zpk(fs = 48000)

The same applies to the ITU_R_468 module, for example:

z,p,k = ITU_R_468_weighting.get_zpk(fs = 48000)

Demo plots

Run:

python demo_ABC_weighting.py
python demo_ITU_R_468_weighting.py

Unit Tests

Both modules run unit tests if run as a main program. If the tests are run directly then they also plot the results:

python3 test_ABC_weighting.py
python3 test_ITU_R_468_weighting.py

Credits

• [e-mail], https://github.com/endolith

• Bernd Porr, https://github.com/berndporr/