I think this is what it boils down to. Please anyone correct me if I am off base:
(1) You need a test LP that has a band where signal of a known frequency is recorded for one channel only. Lets say it's the R channel. Play the band and measure V in the R channel. Now play that same band and measure V in the L channel. That V represents crosstalk of R channel information into the L channel. The difference in V can be converted to db. Formulae are on the internet.
(2) Do the exact same thing, only this time play another band on your test LP that drives only the L channel at the same frequency,. Measure V in the L channel and V in the R channel and again calculate the difference in db.
(3) Adjust azimuth so that you get the largest possible absolute values of db for both sides. It's always a compromise.
(1) You need a test LP that has a band where signal of a known frequency is recorded for one channel only. Lets say it's the R channel. Play the band and measure V in the R channel. Now play that same band and measure V in the L channel. That V represents crosstalk of R channel information into the L channel. The difference in V can be converted to db. Formulae are on the internet.
(2) Do the exact same thing, only this time play another band on your test LP that drives only the L channel at the same frequency,. Measure V in the L channel and V in the R channel and again calculate the difference in db.
(3) Adjust azimuth so that you get the largest possible absolute values of db for both sides. It's always a compromise.