plotSpectrogram

maelzel.snd.plotting.plotSpectrogram(samples, sr, fftsize=2048, window='hamming', winsize=0, overlap=4, axes=None, cmap='inferno', interpolation='bilinear', minfreq=40, maxfreq=16000, mindb=-90, axeslabels=False, figsize=(24, 8), method='specgram', yaxis='linear')[source]

Plot the spectrogram of a sound

Parameters:

  • samples (np.ndarray) – a channel of audio data

  • sr (int) – sr of the audio data

  • fftsize – size of the fft, in samples

  • winsize (int) – size of the window, in samples

  • window (str) – string passed to scipy.signal.get_window

  • overlap – number of overlaps. If fftsize=2048, an overlap of 4 will result in a hoplength of 512 samples. Use None to use a sensible value for the number of samples.

  • axes (Axes | None) – axes to plot on. If None, new axes will be created

  • cmap – colormap (see pyplot.colormaps())

  • minfreq – initial min. frequency

  • maxfreq – initial max. frequency. If None, a configurable default will be used

  • interpolation – one of ‘bilinear’

  • mindb – the amplitude threshold

  • figsize – if axes is not given, use this size to determine the figure size. The value should be a tuplet (width, height)

  • yaxis – one of ‘linear’, ‘log’

  • method – the actual routine to plot the data, one of ‘specgram’, ‘specshow’ (based on librosa)

  • axeslabels – draw labels for the axes

Return type:

Axes

Returns:

the matplotlib axes object