The origin of those tales lies, almost certainly, in the fact that attenuators are employed to deal with overload situations (which is, of course, their primary purpose) and the result of removing an overload is a drastic reduction in all manner of distortion products that occur as a direct result of that overload.
Whether you prefer the sonic mess that an overload entails is yours to call, but that's the reasoning behind that particular old wives’ tale.
Thanks - I suspected as much.
I wonder if I am more missing something though? (I suspect I am not). All the tests on here seem to be done with an output voltage of circa 4v. Whilst the DAC is feeding a preamp or headphone amp with its own volume control that may be all well and good. But surely that's not the normal use-case? Won't most people will be using the DAC to drive their power amps or headphones directly - or else why would all these DACs have volume knobs and headphone outputs?
So the 4v test scenario is entirely unrepresentative of normal listening conditions. A DAC may achieve a very good SINAD measurement of say 110db but only at a volume setting which may never be used.
In real life, maximum (damned loud) volume might be a good 10 or 15 dB quieter and normal listening maybe another -15 dB quieter still. So the SINAD at normal listening levels could be perhaps more like 80 dB?
So a DAC such as the RME ADI-2 DAC with its adjustable reference output level, whilst not topping the charts at 4V output, could very well be significantly better than other DACs which need to be turned down by -30dB for normal listening?
What puzzles me is why is none of this mentioned in the reviews? And why do the other DAC manufacturers not incorporate similar output attenuators?
As I say, perhaps I am missing something?