A few thoughts:
(1) Balance using each possibility (e.g., receiver test tones, player test tones, Avia test tones), and write down the results. If the differences are not significant (say, less than 1 dB), then you don't really need to choose.
(2) If there is a significant difference, then you can either (a) split the difference for the affected speaker(s) or (b) use the settings most appropriate for the source whose audio quality you value most. I used the latter approach, but then I don't watch much TV and our radio reception is nonexistent.
Good luck!
