I read the paper, but I don't really see where it proves your point? From the summary:
"Although all PA capacitors exhibit high capacitance and mΩ level low ESR values regardless of the capacitor structures, other electrical performance of PA capacitors depends highly on the capacitor structures. The PA capacitors made with loose wound structure show high dielectric loss, high DC leakage current, low capacitance roll-off frequency, and low surge breakdown voltages. The capacitors with “laminated” structure exhibit the most stable dielectric response versus frequency and temperature. All PA capacitors show the same failure mechanism in surge breakdown voltage testing regardless of the capacitor structures. The much tighter distribution in surge breakdown voltages of PA capacitors, when compared to that of a MnO2-based tantalum capacitor, suggests that voltage de-rating may not be necessary for PA capacitors."
There doesn't seem to be any concrete conclusions, other than some capacitor structures breakdown less than others, and the some PA capacitors may not need voltage de-rating?
Interesting read, regardless. Thanks for sharing! For the ZX-507, I feel like the high frequencies have chilled a bit as I've listened to it. Being somewhat sensitive to treble, I don't think that I've grown used to the sound, but it's possible. It sounded pretty good when I first got it, and it sounds even better now after a couple hundred hours of use on the balanced side. The SE side still sounds a bit brighter than the balanced side to me, but I rarely use that port, and it could all be in my mind.