sage.data.waveform.taper

Filename : taper.py Description : Short description of the file

Created on 2026-02-08 00:45:34

__author__ = Narenraju Nagarajan __copyright__ = Copyright 2026, ProjectName __license__ = MIT Licence __version__ = 0.0.1 __maintainer__ = Narenraju Nagarajan __affiliation__ = N/A __email__ = N/A __status__ = [‘inProgress’, ‘Archived’, ‘inUsage’, ‘Debugging’]

GitHub Repository: NULL

Documentation: NULL

Functions

fd_low_freq_taper(f, f_min, df, width_bins)

Smooth low-frequency roll-on taper in the frequency domain.

fd_high_freq_taper(f, f_cut, df, width_bins)

Smooth high-frequency roll-off taper in the frequency domain.

fd_taper(f, f_min, f_cut, df[, low_width, high_width])

Combined low- and high-frequency band-pass taper.

Module Contents

fd_low_freq_taper(f, f_min, df, width_bins)[source]

Smooth low-frequency roll-on taper in the frequency domain.

Returns 0 below f_min, smoothly rises to 1 over width_bins frequency bins, and stays 1 above that range.

Parameters:

  • f (torch.Tensor) – Frequency array (Hz).

  • f_min (float) – Frequency at which the taper starts rising.

  • df (float) – Frequency bin spacing (Hz).

  • width_bins (int) – Number of bins over which the taper rises from 0 to 1.

Returns:

Multiplicative taper weights, same shape as f.

Return type:

torch.Tensor

fd_high_freq_taper(f, f_cut, df, width_bins)[source]

Smooth high-frequency roll-off taper in the frequency domain.

Returns 1 below f_cut - width_bins*df, smoothly falls to 0 over width_bins bins, and stays 0 above f_cut.

Parameters:

  • f (torch.Tensor) – Frequency array (Hz).

  • f_cut (float) – Frequency at which the taper completes its roll-off.

  • df (float) – Frequency bin spacing (Hz).

  • width_bins (int) – Number of bins over which the taper falls from 1 to 0.

Returns:

Multiplicative taper weights, same shape as f.

Return type:

torch.Tensor

fd_taper(f, f_min, f_cut, df, low_width=64, high_width=64)[source]

Combined low- and high-frequency band-pass taper.

Multiplies a low-frequency roll-on and a high-frequency roll-off to produce a smooth band-pass window that is 0 outside [f_min, f_cut] and 1 in the interior of that band.

Parameters:

  • f (torch.Tensor) – Frequency array (Hz), shape (B, F) or (F,).

  • f_min (float) – Lower frequency edge (Hz).

  • f_cut (float) – Upper frequency cutoff (Hz).

  • df (float) – Frequency bin spacing (Hz).

  • low_width (int) – Taper width at the low-frequency edge (default 64 bins).

  • high_width (int) – Taper width at the high-frequency edge (default 64 bins).

Returns:

Multiplicative taper weights, same shape as f.

Return type:

torch.Tensor