Advanced method helpers#

These secondary interfaces expose reusable building blocks for composing or inspecting denoising workflows. Start with the primary estimators on the API reference page when choosing a method.

ASR helpers#

`compute_clean_window_mask`	Compute a retained-sample mask from ASR window statistics.
`fit_rms_distribution`	Fit robust location and scale to RMS statistics.
`process_guided_asr`	Apply a calibrated ASR state with guided reconstruction.
`select_juggler_reference_samples`	Select calibration samples with JugglerASR rules.

DSS building blocks#

Additional iterative helper#

iterative_dss_one

Extract one DSS component by fixed-point iteration.

Segmentation#

`CovarianceSegmenter`	Segment data where windowed covariance changes.
`FixedWindowSegmenter`	Segment data into fixed-length windows.

Component selection#

`auto_select_components`	Select a component count with `iterative_outlier_removal()`.
`auto_select_components_robust`	Combine outlier and knee component counts and return the larger count.
`detect_eigenvalue_knee`	Select components above the largest qualifying score drop.
`iterative_outlier_removal`	Count values removed by iterative mean-plus-sigma thresholding.

DSS convenience variants#

`smooth_dss`	Create an ordinary DSS estimator with `SmoothingBias`.
`narrowband_dss`	Create a DSS estimator with a bandpass bias.
`narrowband_scan`	Scan candidate frequencies with narrowband DSS.
`ssvep_dss`	Create a DSS estimator configured with `CombFilterBias`.

Nonlinear helper functions#

`beta_tanh`	Return the tanh fixed-point coefficient.
`beta_pow3`	Return the cubic fixed-point coefficient `-3.0`.
`beta_gauss`	Return the Gaussian fixed-point coefficient.

iCanClean helper#

null_r2_threshold

Estimate a circular-shift null threshold for squared CCA correlations.

SOUND helper#

compute_sound_ref_best

Compute SOUND with a selected single-channel reference.

SSA helpers#

`ssa_decompose`	Decompose a one-dimensional series into Basic SSA components.
`ssa_w_correlation`	Compute weighted correlations between SSA reconstructions.
`ssa_clean_channel`	Clean one channel by frequency-grouping Basic SSA components.
`local_ssa_clean_channel`	Clean one channel with clustered local SSA reconstruction.