I've seen that a few times.
You can always put an inline resistor in the "+" lead of the channel with more output. To reduce the gain by about 2dB, just buy a few 1/4 watt resistors (make sure they're non inductive). If it's really about a 2dB error across the spectrum, adding an inline resistor between 8.2 and 15 ohm should do the trick. If your driver is exactly 50 ohms a 12 ohm resistor should do as a first test, and adjust and replace as needed. I've found the Fostex drivers can vary, I've seen 48 and 57 ohms, so a few trials may be needed (especially if your 2dBis a swag).
A more accurate way (assuming you have a multimeter) is to get a variable resistor and trim it until it sounds balanced, then measure the resistor and replace it with a quality non inductive network of resistors giving the same value...
Hey there Speakers. Have you ever thought of creating a plug with two slots to insert resistors into? It would be a cheap enough item to create and being able to play with impedence would most likely prove irresistable to the community.
Yes. We have a target sound/response, the majority of units are right there first pass, but some take quite a few passes to make the curve, and a some never do. Channel mismatch isn't too common once you rule out the easily cleaned oxidation on the cables...
Yeah. That's much harder to address than a phone that's more or less efficient than usual but otherwise has a similar response.
Maybe sometime... :-) Irresistorable is hard to resist.