I have always worked to get rid of all the power supply noise so that the high level of power supply noise rejection that was required was not so great. In addition, the tolerance of components needed to filter out power supply noise are far less critical than those needed for good amplifier PSRR design. Probably a combination of the two methods would be the best choice. Of course, optimizing for the minimum cost was never the highest priority for our group, nor for our customers. Somebody else can always make something cheaper and not quite as good.
Hi, sorry for my english.
About the power supply rejection, in this old Luxman:
can be see the use of differents supplies for each stage of the power amp in a cheap and good way, the effect in the sound quality are very good.
Replay available now: A handful of emerging network technologies are competing to be the preferred wide-area connection for the Internet of Things. All claim lower costs and power use than cellular but none have wide deployment yet. Listen in as proponents of leading contenders make their case to be the metro or national IoT network of the future. Rick Merritt, EE Times Silicon Valley Bureau Chief, moderators this discussion. Join in and ask his guests questions.