Efficient realtime IIR filter in Python

This is an IIR filter class which performs sample by sample realtime processing of data. It's very efficient because it's not using any indexing operations internally. The class instance acts as the memory of the filter so that it remembers its past.

Import

Use the standard python command to import it:

import iir_filter

Calculate the coefficients

Use your favourite scipy IIR design command and export the coefficients as an SOS:

sos = signal.butter(order, [cutoff(s)], '[filter type]', output='sos')

Create an instance

The constructor takes the sos chain as an argument:

f = iir_filter.IIR_filter(sos)

Perform filtering sample by sample

Filtering is sample by sample by processing the samples as they arrive, for example from an ADC:

sample = f.filter(sample)