Nothing that complicated. Just play the music as is, and measure a section with RTA, for example with Forever averaging, and capture both current and peak at the end of the section. Do that for each speaker. I'm just an amateur, guessing that what you're seeing is some directional sensitivity difference between the mics, and suggesting a possible way to evaluate that.
But before you do that, you've already done just a straight frequency sweep measurement of each individual speaker from inside the car, with both mics, with the mic remaining in the same vertical fixed position? And the sweeps match for each speaker?