data_operations.specgram_quality#

Back to Module page

Summary#

Computes specgram of whitened data, along with list of bad time and frequency channels in the specgram (bad time channels are windows that overlap with zeros in mask, and bad frequency channels are varying lines or those beyond the analysis range)

Signature#

def specgram_quality(strain_wt, qmask, valid_mask, dt, interval, edgesafety = 1, nfire = params.NPERFILE, freqs_lines = None, mask_freqs_in = None, fmax = params.FMAX_OVERLAP, **spec_kwargs)

Input variables#
Name	Type	Default	Description
`strain_wt`			Whitened strain data
`qmask`			Boolean mask with zeros at holes in unwhitened data (we have already widened holes due to the highpass filter)
`valid_mask`			Boolean mask with zeros where we cannot trust whitened data
`dt`			Time interval between successive elements of strain_wt (s)
`interval`			Time interval for FFTs (s, sets resolution of lines)
`edgesafety`		1	Safety margin at the edge where we cannot trust the nature of the whitened data (we haven’t applied this to the mask when we get here)
`nfire`		params.NPERFILE	Number of times line-detector fires per perfect file
`freqs_lines`		None	Array with frequencies on which we previously detected lines (optional)
`mask_freqs_in`		None	Boolean mask on freqs_lines with zeros at previously detected lines/varying lines (optional)
`fmax`		params.FMAX_OVERLAP	Maximum frequency involved in the analysis
`\\spec_kwargs`

Output variables#

Return annotation	Docstring type	Description
`None`		Array of length n_freqs with frequencies (Hz) 2. Array of length n_times with times (s) 3. n_freqs x n_times array with specgram (two-sided psd in Hz^-1) 4. Boolean array of length n_freqs with zeros at candidate varying line channels 5. Boolean array of length n_times with zeros where FFTs see holes/invalid data 6. Array with amplitude scaling factors needed to suppress lines above upper threshold

Docstring#

Computes specgram of whitened data, along with list of bad time and
frequency channels in the specgram (bad time channels are windows that
overlap with zeros in mask, and bad frequency channels are varying lines or
those beyond the analysis range)
:param strain_wt: Whitened strain data
:param qmask: Boolean mask with zeros at holes in unwhitened data (we have
              already widened holes due to the highpass filter)
:param valid_mask:
    Boolean mask with zeros where we cannot trust whitened data
:param dt: Time interval between successive elements of strain_wt (s)
:param interval: Time interval for FFTs (s, sets resolution of lines)
:param edgesafety:
    Safety margin at the edge where we cannot trust the nature of the
    whitened data (we haven't applied this to the mask when we get here)
:param nfire: Number of times line-detector fires per perfect file
:param freqs_lines:
    Array with frequencies on which we previously detected lines (optional)
:param mask_freqs_in:
    Boolean mask on freqs_lines with zeros at previously detected
    lines/varying lines (optional)
:param fmax: Maximum frequency involved in the analysis
:return: 1. Array of length n_freqs with frequencies (Hz)
         2. Array of length n_times with times (s)
         3. n_freqs x n_times array with specgram (two-sided psd in Hz^-1)
         4. Boolean array of length n_freqs with zeros at candidate
            varying line channels
         5. Boolean array of length n_times with zeros where FFTs see
            holes/invalid data
         6. Array with amplitude scaling factors needed to suppress lines
            above upper threshold