Real-Time-Audio-Filtering-using-Python

Platform for Audio Filtering (Digital Filters) in Real-Time using Convolution Theorem and Fast Fourier Transform.

Features

• Users to configure the specification of the filter using impulse response of the system h[n], H(z) Transfer fucntion either by H(z) equation or by giving zeros/poles of H(z), LCCDE coefficients, and cut-off frequency.

• It also have in-built Ideal filters like Low Pass Filter, High Pass Filter, Band Pass Filter, and Band Stop filter.

• Users can also Save the Filtered Audio, and Plot the Frequency Response of the ideal as well as custom filters.

Implementation

• Use Pyaudio to get audio in real time.
• Matpoltlib for visualization.
• Tkinter For UI.

In time domain, filtering is convolution of input x[n] and impulse response of h[n].

y[n] = Σ x[k]*h[n-k]

where is y[n] is filtered audio.

Convolution in time domain is same as product in frequency domain. In frequency domain,
Y(e^jw) = X(e^jw)H(e^jw)

To Convert back into time domain, we have to take inverse dft. Fast and efficient way to take dft is ifft.

y[n] = IFFT(Y(e^jw))

Controls

Demo

Audio Waveform

We have use a low pass filter to filter out high frequency audio.

• When low frequency audio was passed through the filter.Audio was allowed to pass.

• When High Frequecny audio was passed through the filter. Audio was blocked and was not allowed to pass.

• When Ideal Low Pass filter was applied. On decreasing the low cut off frequency the higher frequency audio was blocked.