300 ft of cable isn't helping in your case. Check the impedance of each of the twisted pairs and you'll see what I'm talking about. Bill Whitlock covers that nicely in his app notes on CMRR. What Engineering field did you practice?
I read it very briefly by skipping some details (It has been so long I ever dealt with Op AMP...) but very interesting fact about balanced circuit.
In shielded twist pair, I seem to understand potential creation of unequal capacitance between two inner conductors in reference to the shield and this will be amplified by Diff Op-Amp... This is very good...
Maybe the problem we are having with keyboard connection is related to ground terminated both end. I will take a look at it.
I guess I need to isolate GND at DI.
Not familiar with MIC circuitry, but my gut feeling tells me shield can not be removed at receive side for passive mic and pre interface...
What about active mic like phantom? Is ground pin in XLR used for 48V return?