Skip to content

Fix odd-m modes bug in LVCNR

Vijay Varma requested to merge vijay.varma/lalsuite:LVCNR_odd_m_modes_bug into master

When fRef>0, the old separation vector was pointed from BH1 to BH2. This causes the odd-m modes to be off by a minus sign (orbital phase rotation of pi). Now the separation vector points from BH2 to BH1.

Here's an example of the bug, demonstrated by comparing LVCNR to itself with fRef=0 and fRef=0+eps: https://git.ligo.org/waveforms/reviews/NRHybSur3dq8/blob/master/scripts/lvcnr_bug.png

Here's the same figure after the fix: https://git.ligo.org/waveforms/reviews/NRHybSur3dq8/blob/master/scripts/lvcnr_bug_fixed.png

Script to generate these plots: https://git.ligo.org/waveforms/reviews/NRHybSur3dq8/blob/master/scripts/LVCNR_start_freq_bug.py

Merge request reports