LALSIMULATION: Inconsistency in what f_lower means with Phenom waveforms
There is an inconsistency in lalsimulation about how the f_lower parameter is interpreted. The new waveform file format suggests that the f_lower that we send to lalsimulation means "The waveform will start at the point that the 2,2 mode will reach this frequency" (so that the 4,4 mode would actually start at a higher frequency). However, the Phenom waveforms interprets f_lower as "start all modes from this value".
This is then a problem, as if I ask for a lalsimulation to generate me a time-domain version of Phenom[XPHM] it will underestimate how long the waveform actually is. The duration calculation used is a 2PN 2,2-mode only calculation, and the same calculation is used for all modes.
It would be good to have consistency with this. Could Phenom maybe use different f_lowers for different mode, to match what I think is the convention? ... I think many parameters have been made consistent across approximants in the last few years, would be good to also do the same with f_lower.