pytimbre.spectral.swipe.swipe

pytimbre.spectral.swipe.swipe(samples, sample_rate: float = 22050, hop_size: float = 256, minimum_frequency: float = 55.0, maximum_frequency: float = 1760.0, pitch_resolution: float = 0.010416666666666666, erb_resolution: float = 0.1, strength_threshold: float = 0)

Implementation of a sawtooth waveform inspired pitch estimator (SWIPE). This version of the algorithm follows the original implementation, see swipe_slim for a more efficient alternative.

[1]

Arturo Camacho and John G. Harris, “A sawtooth waveform inspired pitch estimator for speech and music.” The Journal of the Acoustical Society of America, vol. 124, no. 3, pp. 1638–1652, Sep. 2008

Parameters:

• samples (ndarray) – Audio signal

• sample_rate (int) – Sampling rate

• hop_size (int) – Hop size

• minimum_frequency (float or int) – Minimal frequency

• maximum_frequency (float or int) – Maximal frequency

• pitch_resolution (float) – resolution of the pitch candidate bins in octaves (default value = 1/96 -> 96 bins per octave)

• erb_resolution (float) – resolution of the ERB bands (default value = 0.1)

• strength_threshold (float) – confidence threshold [0, 1] for the pitch detection (default value = 0)

Returns:

• f0 (ndarray) – Estimated F0-trajectory

• t (ndarray) – Time axis

• strength (ndarray) – Confidence/Pitch Strength

• Remarks/Development

• 2023-04-01 - FSM - updated the argument names, and put the parameter helpers into the list