Fozgometer

... the manual states that channel separation is what's being measured.

"Channel separation" is just a derivative statistic. The only way to quantify it is to measure how much information intended to be in channel A is bleeding into channel B, and vice-versa; i.e., crosstalk.

If you measure crosstalk in each direction, average the results and subtract from a notional maximum signal level, the resulting statistic is called "channel separation". The less crosstalk, the greater the channel separation. They're exactly and inversely proportional.

A statement that one is measuring channel separation but not crosstalk would be nonsense. They're two sides of the same coin.

In a couple of years of adjusting azimuth on multiple cartridgess using the Wally Analog Shop tool, I don't remember ever seeing much change in output (ie, channel balance), but differences in crosstalk (ie, channel separation) were immense.

About 10 years ago I dropped the instrumentation after realizing that I can adjust by ear just about as accurately and far more quickly. As with the instruments, I've never heard a difference in channel balance when tweaking azimuth within reasonable limits (ie, within a degree or so from vertical), but changes in crosstalk from even the tiniest possible adjustment are readily audible.

YMMV, of course...