I thought there has been tonz of documented testing w/the R9 and it was determined to only use about 15bits of actual signal w/the rest being ass ?
which, always floored me because I think the r9 sounds *awesome* to my ear.
I missed hearing about R-09's 15 bit resolution limitation that might be if someone 'summed' all the 20-20,000+ cycle BANDWIDTH noise to gain a dismal looking equivalent to bits number, AND then
assume that's the end of the story.
This seems more a bad rumor than good science.
Problem with many such
contrived specifications is the popular notion that it means something to perceived audio quality
.
I see similar thinking applied to microphone/mic array schemes.
Let's assume a different more
visual perspective.
As the spectrum graph in last post indicates, the noise floor
at frequency is way below 16bit (96dB?) resolution
in the most audible higher frequencies range by at least 20 dB (-116dB) which is what (20-22+bit???) equivalent bits resolution?
My take is we hear more like what's
visually seen in the graph by listening inside the the full acoustic spectrum to perceive stuff over full hearing bandwidth range of frequencies. If this is true, then there's plenty of (bit resolution depth) headroom
above the noise floor that's not being wasted, not by one bit, as proven we hear stuff happening quite clearly that's deeply buried in noise!
In other words, think of a flute playing a single C note at ~440 cycles (plus whatever harmonics). We can see on the graph (worse case) noise at that frequency is -114dB, and is even lower at higher (harmonic) frequencies.
This should explain
why the R-09 has way more audible 'noise-free'
benefit at full 24bit recording mode than
'rumor-tech' would have us believe.