I have to agree with Doug. Moreover, crosstalk, it seems to me as a layman, cannot be discussed without reference to a level of signal. The Signet Cartridge Analyzer guide advises one to use (any of several possible test LPs) with a 1kHz steady state tone on one channel and nothing on the other. Then you set the Analyzer to receive the recorded signal and set that value to "0db" on its meter. You then set the Analyzer to receive the signal on the ideally negative channel, and the meter shows a value on the negative side of its 0db marking. Thus you can say you have negative X db of crosstalk into the channel that is receiving no intentional signal. Then you repeat the process using another band on the test LP that has signal encoded in the opposite channel. Because you have referenced your result to 0db in both cases, you can say that the level of crosstalk in both channels is equal, or not. And you can discuss magnitude in a meaningful way. Also, as Doug inferred, the results will be very different at different test frequencies. It's usually best around 1kHz and worst at the frequency extremes.
So, I don't know what the quoted part of the Foz manual means exactly, except I guess it would be acceptable to not know the reference level, so long as one did not change the level when comparing crosstalk into channel A with crosstalk into channel B. The signal level, whatever it may be, HAS to remain constant it seems to me, when doing that, if you really want to be able to say you have equal amounts of crosstalk in both channels.
TKetcham, The only other thing in your last post that I would mention is the business about "balance". If you are talking about achieving equal amounts of crosstalk in each channel, then I understand, and I do think that's what you meant. But if you are talking about "channel balance", as in equal gain in both channels, adjusting azimuth is not a way to get there.
