Use of time-domain approximant in Bilby ?

[bilby-bot]

In GitLab by @git.ligo:ecm on Nov 2, 2021, 17:25

Dear Bilby developers,

I understand that the WaveformGenerator class includes the possibility of a time-domain approximant:

https://git.ligo.org/lscsoft/bilby/-/blob/master/bilby/gw/waveform_generator.py#L35

however GravitationalWaveTransient likelihood object seems to use frequency-domain approximant only:

https://git.ligo.org/lscsoft/bilby/-/blob/master/bilby/gw/likelihood.py#L42

An object which computes the frequency-domain strain of the signal, given some set of parameters

confirmed by the fact that gw.source.lal_binary_black_hole() is a wrapper around lalsim_SimInspiralFD

https://git.ligo.org/lscsoft/bilby/-/blob/master/bilby/gw/source.py#L365

Is this true? Or is Bilby able to produce a frequency-domain strain from time-domain data?

Best, Eric

PS: we're trying to plug-in our own time-domain waveform approximant.

[bilby-bot]

In GitLab by @git.ligo:colm.talbot on Nov 13, 2021, 14:01

Hi @git.ligo:ecm, the standard way to use TD approximants in Bilby is through SimInspiralFD which can call any waveform model implemented in lalsimulation with standardized conditioning/FFT. If you're model is implemented in lalsim you can just use that interface.

Otherwise, you have two options:

• implement your own data conditioning and FFT inside the function passed as the frequency_domain_source_model to the WaveformGenerator.
• pass a time_domain_source_model to the WaveformGenerator instead. If you then call waveform_generator.frequency_domain_strain internally your passed time_domain_source_model will be called and the returned values will be FFT'd. Note that this doesn't include any tapering so that will have to be implemented by you.

[bilby-bot]

In GitLab by @git.ligo:colm.talbot on Nov 29, 2021, 17:44

@git.ligo:ecm did the previous comment help? I don't think there's any issue we can fix here, this is explained in the documentation.

[bilby-bot]

In GitLab by @git.ligo:ecm on Dec 1, 2021, 14:34

Hi @git.ligo:colm.talbot

thank you for the feedback! Yes, this helps. We're going the 2nd option and may come back with other questions.

Cheers, Eric