maad.features.spectral_quantile

maad.features.spectral_quantile(s, fs, q=[0.05, 0.25, 0.5, 0.75, 0.95], nperseg=1024, roi=None, as_pandas=False, amp=False, **kwargs)[source]

Compute the q-th quantile of the power spectrum. If a region of interest with time and spectral limits is provided, the q-th quantile is computed on the selection.

Parameters:

s1D array: Input audio signal
fsfloat: Sampling frequency of audio signal
qarray or float, optional: Quantile or sequence of quantiles to compute, which must be between 0 and 1 inclusive. The defaul is [0.05, 0.25, 0.5, 0.75, 0.95].
npersegint, optional: Length of segment to compute the FFT. The default is 1024.
roipandas.Series, optional: Region of interest where peak frequency will be computed. Series must have a valid input format with index: min_t, min_f, max_t, max_f. The default is None.
as_pandas: bool: Return data as a pandas.Series or pandas.DataFrame, when amp is False or True, respectively. Default is False.
amp: bool: Enable quantiles amplitude output. Default is False.
kwargsadditional keyword arguments: If window=’hann’, additional keyword arguments to pass to sound.spectrum.

Returns:

Pandas Series or Numpy array: Quantiles of power spectrum.

Examples

>>> from maad import sound, features
>>> s, fs = sound.load('../data/spinetail.wav')

Compute the q-th quantile of the power spectrum

>>> qs = features.spectral_quantile(s, fs, [0.05, 0.25, 0.5, 0.75, 0.95], as_pandas=True)
>>> print(qs)
0.05    6029.296875
0.25    6416.894531
0.50    6632.226562
0.75    6890.625000
0.95    9216.210938
dtype: float64

Compute the q-th quantile of the power spectrum and its amplitude

>>> qs = features.spectral_quantile(s, fs, [0.05, 0.25, 0.5, 0.75, 0.95], amp=True, as_pandas=True)
>>> print(qs)
             freq       amp
0.05  6029.296875  0.000007
0.25  6416.894531  0.000067
0.50  6632.226562  0.000087
0.75  6890.625000  0.000022
0.95  9216.210938  0.000004