mne_denoise.dss.SpectrogramDenoiser#

class mne_denoise.dss.SpectrogramDenoiser(threshold_percentile: float = 90.0, nperseg: int = 256, noverlap: int | None = None, mask: ndarray | None = None)[source]#

STFT-mask denoiser for iterative DSS.

Parameters:

threshold_percentilefloat, default=90.0: Percentile of STFT magnitude used as the adaptive threshold; bins above it receive gain one.
npersegint, default=256: STFT segment length.
noverlapint or None, default=None: Number of overlapping samples. None uses nperseg // 2.
maskndarray or None, default=None: Fixed time-frequency mask. When supplied, it replaces adaptive mask calculation.

denoise(source: ndarray) → ndarray[source]#

Compute or apply an STFT mask to a source.

Parameters:

sourcendarray, shape (n_times,) or (n_times, n_epochs): Source time series; columns of 2D input are processed separately.

Returns:

ndarray: Reconstructed source with the input shape.