I've read here on other threads that there's no difference in measurements before and after break-in. Yet, like you, I heard the differences, even during the break-in period.
On another thread about measurements, I mentioned that I read on a review that despite the different options for selecting filters (on a CDP), they all measured the same on the component. I even cut and pasted the section on it.
Hmmm? Different filter settings resulting in obvious sound differences yet they all measured the same. Could it be that the analyzer could only "measure" the summing of the signal and not the "flavor" of the settings?
I never got a cogent answer for that.
All the best,
Nonoise