Having previously operated an LHYaudio switch followed by a 10m clocked Etherregen after a Draytek router I out of a whim added a humble Netgear GS 305E for not even $50 in the chain ahead of the LHYAudio switch. All Cabling is ONTI with a Xangsane braided silver ethernet cable between Etherregen and Server.
The effect was not subtle: major uplift in bass performance in tightness and definition, ‘being there’ spatiality and impulse rendition. I live in Central London and probably have more incoming crud on the cable, RMI and EFI than more remote locations. Given that I now daisy chain 3 switches and a router to achieve optimal results I cannot but wonder whether any of the individual component designers have fully understood the task at hand in providing an optimum ethernet audio signal to the server. So far as I can make out individual component tests become borderline meaningless. But then some would argue that bits are bits: So not!