Regarding the above linked article (did he see some previous chat here?
), I'm very aware of the points raised but that's not really the real issue because noise can be filtered off and higher sample rate allows more gentle filtering. I understand in broad sense that is something similar in the common DAC design to have oversampling. If you read the DAC data sheets different incoming sample rates can be handled differently with different algorithms and clock, for example a 44.1kHz is 8x oversampled and a 192kHz is 2x oversampled. So the end result could be rather similar.
Charles Hansen did comment on filtering in the Ayre BDP thread and he concedes that 176.4kHz is preferable. The other point I've seen raised elsewhere is DSD is equivalent to about 100kHz in resolution - I can't remember where I read this or if this is correct. There is a graph done by someone showing DSD and 24/96 PCM have very similar waveforms on a test tone. Now you can argue 88 is close to 96 and 80 is close to 88 and you're on a slippery slope.
All in all, it is a safer bet to have higher rather than lower sample rates to start with. I'm not saying that's the only thing that matters. Even to this day AFAIK many AVRs/pre-pros can't process 196kHz. Sending DSD to AVR is a complete mystery because you just don't know what the AVR does to the PCM conversion.