@gs5556 - that's not really true. The caps are charging when the rectified voltage from the transformer is higher than their current stored voltage, and discharging otherwise.
With 60hz power and assuming a full bridge rectifier, the caps will get a charging pulse every 120th of a second. Once the caps reach the steady-state voltage after power-up, if the load on the supply is fairly light, the rectified voltage will only be high enough for current to flow into the cap for a small percentage of the time, but you'll still always be charging or discharging the caps since there is always at least a small current draw from the amp.
The amount the voltage changes on the capacitors every cycle will depend on the capacitance and the current draw. Assuming a low bias class A/B amp, this can be significantly different between low-power operation and high power.
With a class A amp, the amp is always drawing a lot of current, so the difference between idle and fairly loud playback will be much less, but the amount of current flowing into and out of the caps will be considerably higher.
Lots of current flowing into and out of the cap puts a lot of stress on it and will shorten it's life, so the power supply caps will last longer if the amps current draw is low.
I suspect that any operation of the caps will keep them working optimally, but while I believe strongly that capacitors that pass audio signals change noticeably during break-in, I'm not sure that this applies to power supply caps. It's quite possible that power supply caps can respond quicker to changes in current requirements after break-in, which would significantly change the sound of an amp.
