Azimuth 2020

Bit hard to follow the conversation to know if any conclusions were reached:

Magnified, the grooves in a turntable may look straight, but they are still curved with the record which causes the cross-talk from one channel to be higher than the other with perfect azimuth. Stylus shape will impact this hence likely why high end cartridges will exhibit this more. Tuning azimuth based on averaging cross-talk will not be ideal. I do optically with a first surface mirror with a graticule, and if I am ambitious, I pull out the test record and do a high frequency sweep to make sure channels are matched.

With an offset tone-arm, usually the azimuth will change as the cartridge is raised and lowered, but the change will be very small over the likely height differences that would ever be encountered while playing, or what you put under the stylus while setting azimuth optically.  Note, I said "usually". For most tone-arms, the pivot angle is perpendicular to the main arm. For some tone-arms, the pivot angle is closer to perpendicular with the offset portion. For some, it is somewhere in the middle. If the pivot angle is perpendicular to the offset portion of the arm, the azimuth will not change as the cartridge is raised and lowered. VTA of course will always change as the cartridge moves up and down.

Robert, You wrote, "I pull out the test record and do a high frequency sweep to make sure channels are matched."
Matched for what?  Balance or crosstalk?

Matched for output level on a high frequency test sweep. If the azimuth is off, you will get left to right output level variations as the frequency changes at high frequencies.

Yes, but changing azimuth will also alter crosstalk, and azimuth has far less of an effect on channel balance vs crosstalk.  So a little change in balance makes a big change in crosstalk relationships. 

Azimuth has a significant impact on channel balance when you start looking at frequencies out past 10K, and especially past 20Khz, which you can get on the Ortofon test records (and Denon if you can find them).

If you balance the crosstalk, so it is the same on both channels, then your azimuth is likely off. The forces are slightly different for the left and right channels, even with perfect alignment and that contributes to slightly different crosstalk. If you balance the frequency response, then you are getting consistent groove tracking on both channels.