[USER QUESTION] A question about Waveform Truncating

[CalNaTNS]

Hi everyone,

I met a problem with waveform truncating when I wanted to do PE of GW190521 with SEOBNRv5EHM, a newly developed waveform.
Since it is not yet included in gwsignal or LAL, I wrote a function based on bilby.gw.source.gwsignal_binary_black_hole() called SEOBNRv5EHM_fd(frequency_array,mass_1,mass_2,chi_1,chi_2,theta_jn,luminosity_distance, eccentricity,rel_anomaly, phase,**kwds) to call and use SEOBNRv5EHM in bilby, and it worked well.

According to the paper of SEOBNRv5EHM, the reference frequency (The actual f22_start for EOB waveforms) used for GW190521 analyses is 5.5Hz, which is different from the minimum frequency of 10Hz. To do this, I set f22_start=reference_frequencyand use minimum_frequency only for frequency_bounds so that I can generate a waveform starting at f_ref but truncated at f_min.

However, I am unsure how to correctly calculate time_shift. Initially, I used the epoch time of the generated waveform to calculate the time_shift that

dt = 1 / hp.delta_f + (hp.epoch.gpsNanoSeconds * 1e-9 + hp.epoch.gpsSeconds)
time_shift = np.exp(-1j * 2 * np.pi * dt * frequency_array[frequency_bounds])

But I soon realized that this very epoch time (or duration before merge) corresponds to the original waveform starting at 5.5Hz, not the truncated waveform starting at 10.0Hz. How should I handle this problem?

I am also unsure about the actual minimum_frequency used in likelihood evaluation. I once thought of setting a cutoff_frequency directly in GravitationalWaveTransient(), but I found nowhere to do so. Can I set different minimum_frequency values for interferometers and waveform generators (This seems to be possible in pbilby.), and which one is the actual minimum_frequency used for the calculation of likelihood? If allowed, would it be possible to generate a waveform starting from 5.5 Hz but truncate it at 10 Hz when injecting it into the interferometer? I believe this approach could avoid the time shift issue while meeting my requirements.

I sincerely appreciate any guidance on these issues. Thank you very much for your time and help!