Real-time Spectrogram



In [1]:

    
import IPython.display as ipd
ipd.YouTubeVideo('ydu12NfZd90', width=600, height=360)









    Out[1]:

Here is how you can create a real-time spectrogram in your terminal using PyAudio.

To see the example in action, run the script in this repo, realtime_spectrogram.py:

python3 realtime_spectrogram.py

The basic idea is simple. For every new audio buffer,

Take an FFT, x_fft, of the audio buffer.
Compute a melspectrum from the x_fft.
Print a string, s, where s[i] is '*' wherever melspectrum[i] is above a threshold.

From here, you can manipulate this basic example to do more sophisticated real-time processing, e.g. involving machine learning models.