[trobbins]

I'd say both words are valid. The forward gain of the amp will have frequency response dips and bumps and roll-offs, and with global feedback applied the designer can play with the level of dips/bumps/roll-offs to achieve more stable operation (ie. large gain and phase margins) - imho that is circuitry level compensation for an outcome of 'better' stability, but can also be traded off for extension of frequency response (eg. extension of bandwidth before loop gain starts to fall, or extension of 0dB/180deg cross-over frequencies).

Just basic instrumentation is what most people have had for many decades. I guess we have become luckier over the last decade or so in that 2nd hand expensive instrumentation is now cheap and/or relatively cheap soundcards and PC software allow diyers to easily do stability related testing that was once only achievable with quite expensive instrumentation, or access to such instruments. It's not that a gain/phase plot of an amp's frequency response can't be done from 1Hz to 1Mhz at home, it's just that it takes time and effort, and I am always amazed that Williamson did that as a young 20-year-old back in circa 1945.