Realtime FIR filter implementation in Python

This project is a simple FIR filter implementation in python to process signals in a realtime mode, sample by sample. The FIR filter is implemented in a python class that can be used in any of your python projects.

Import the filter

Use this command to import the fir class in your python script.

  import fir

Calculate the coefficients

You can extract yout filter coefficients from the FIR filter design benchmark and design file:

    coeff = signal.firwin(NTAPS,[f1/(fs*2) ,f2/(fs*2)], pass_zero = 'bandpass', fs = fs)

In the design process, implementation requirements must be taken in accordance to which behaviour is required on your application. Here is where your expertise and knowledge about DSP come into play. I can provide some recommended literature to understand the world of digital signal processing:

1. http://www.dspguide.com/
2. Discrete-Time Signal Processing by A. V. Oppenheim and R. W. Schafer.
3. Theory and Application of Digital Signal Processing by Rabiner and Gold.

Instantiate the filter

You can create an instance of the FIR filter by calling it:

    myFIR = fir.FIR_filter(b, NTAPS)

Filtering Flow

In the script, a combination of sine waves can be created from the function provided. In the case of the example provided, a combination of a 1 and 50 Hz sine waves are provided.

For filtering sample by sample:

  y = np.zeros(SAMPLES)
    for i in range(SAMPLES):
       y[i] = myFIR.dofilter(mysignal[i])

And you obtain something like this:

The frequency spectrum of the filtered signal is the following