Actually I was referring to the 'delay' in getting the global negative feedback around to the input again. If this is a significant source of distortion it seems like an impulse test of some sort would excite it, maybe?