I wasn’t intending to poke that bear here, but I easily hear the difference between 16/44 and 24/192. This is why I asked my question. I hear it easily with the GDAC and with a Node2i, plus can hear it with some effort with my AVR. So, I’ve been trying to figure out why. The best technical explanation I’ve found is the paper linked below, actually from dCS back in 1997, measuring the time domain degradation caused when producing the recording.
My belief is that BOTH the 16/44 recording format is degraded relative to 24/192 AND the playback DAC introduces more degradation to 16/44 than 24/192. Better DACs should do less damage. My posted question was seeking feedback from those who can hear the difference and have listened to very good DACs. If the very best DACs make both formats sound the same then I need a better DAC, but if the best DACs still have a significant difference between formats then most of the damage is on the recording side.
I’m very happy with the GDAC, so I’ll choose to believe that Stereophile review.
Here’s the paper I mentioned and my summary.
“A Suggested Explanation For (Some Of) The Audible Differences Between High Sample Rate And Conventional Sample Rate Audio Material”
It is commonly recognized that human hearing is sensitive in the frequency domain up to 18 kHz, rounded up to 20 kHz. It is rarely mentioned that human hearing in the time domain is sensitive to as little as 5 to 10 microseconds. The frequency domain is important for understanding sound and the time domain is important for locating sound. The Nyquist theorem requires sampling at least twice the bandwidth of the signal to avoid aliasing distortion. Some musical instruments have frequency content up to 100 kHz, but filtering would be required even if it were only noise. For CD quality 44.1 kHz sample rate, the recording must be filtered to less than 22.05 kHz to avoid aliasing. Most professional music recording use ADCs at a higher sample rate, which is maintained through mixing, then digitally filtered to reduce the sample rate for the final product. Digital filtering smears audible energy across multiple sample periods. For 44.1 kHz the audible energy is spread across more than 1000 microseconds, 96 kHz spreads it across 100 microseconds, and 192 kHz spreads it across 50 microseconds.