pymodalib.algorithms.harmonics module

Detecting harmonics.

harmonicfinder(signal: numpy.ndarray, fs: float, scale_min: float, scale_max: float, sigma: float = 1.05, time_resolution: float = 0.1, surrogates: int = 10, parallel: bool = True, crop: bool = True) → Tuple[numpy.ndarray, numpy.ndarray, numpy.ndarray, numpy.ndarray]

Harmonic-finder algorithm.

Parameters:

• signal (ndarray) – [1D array] The signal to analyse.
• fs (float) – The sampling frequency of the signal.
• scale_min (float) – The minimum wavelet timescale.
• scale_max (float) – The maximum wavelet timescale.
• sigma (float) – (Default value = 1.05) Used to define the timescale increment. For example, sigma = 1.05 is a timescale increment of 5 percent.
• time_resolution (float, optional) – (Default value = 0.1) The time resolution.
• surrogates (int) – (Default value = 10) The number of surrogates.
• parallel (bool) – (Default value = True) Whether to parallelize the algorithm, which provides a significant speed boost in many cases.
• crop (bool, optional) – (Default value = True) Whether to crop the results, removing the NaN values around the left and bottom edges.

Returns:

• freq (ndarray) – [1D array] The frequencies.
• res (ndarray) – [2D array] The raw harmonics.
• pos1 (ndarray) – [2D array] The number of surrogates which the raw harmonics are higher than at each point.
• pos2 (ndarray) – [2D array] The raw harmonics relative to the mean and standard deviation of the surrogate distribution.

Notes

Author: Lawrence Sheppard.

[1]	L.W. Sheppard, A. Stefanovska and P.V.E. McClintock, “Detecting the harmonics of oscillations with time-variable frequencies”. {doi:10.1103/PhysRevE.83.016206}