Um, how could these "comps" possibly tell you anything about the usefulness or value of a modification? You have a band making a certain sound; that sound has whatever qualities it has (loudness, balance of highs/mids/lows, distortion, etc.), which a given listener might consider to be a nice set of qualities or perhaps not.
You put microphones in particular places in that room, but they're not absolutely identical microphones in absolutely identical places. You then make recordings on two different recorders and you play them back later to compare them.
I'd say the chances would be quite small that you would ever hear identical sound from the two recordings even if the recorders were absolutely identical. Digital recorders certainly can sound identical if all the variables are controlled carefully enough, but that probably won't be the result that you'll get because of the separate microphones and mike positioning, and analog gain settings which aren't exactly alike.
So what does the result tell you? If one recording has nicer sounding low frequencies than the other one, for example--that could be due to differences in frequency response, absolute amplitude, distortion or several room and miking variables, for example. Even if you feel that you like the results better on (say) the modified recorder, it could be because that band's PA system or the room they were performing in had a peculiar characteristic and the modification has another peculiar characteristic that's working to counter or limit that effect. Or it could be that the two effects reinforce, and you happen to like the mixture.
Either way, the difference that you hear has no predictive value. The next recording you make may have--no, will have--a different set of characteristics and problems. Just because a certain medicine helps you to recover from one illness does not mean that you should take it for every one of 1,000 other illnesses.
There is no one sound-altering treatment that always makes things sound better, or else you could route your signals through it an infinite number of times and each time, the result would keep sounding better and better--an obviously absurd situation. But if a circuit modification actually helps a recorder to perform better (i.e. make its output more nearly equal to its input), then that can be determined on a test bench far more easily and reliably than by ear, where there are so many other factors that can't be controlled.
--best regards