I agree that it still becomes a "garbage in, garbage out" argument. It is especially true if the hi-rez files were not originally recorded in hi-rez, but upsampled or upconverted. That being said, the most incredible demos at CES were original 24/192 recordings (Michael Schnee's at TAD room). Absolutely beautiful recordings, with realistic harmonics that I've never heard other than live.
The double-edged sword that is 24/192 (or 24/176) is that those bandwidths require especially stable and quiet signal paths/power supplies/reclocking mechanisms, etc. Otherwise noise is introduced due to the stresses. From the designer of a leading DAC:
"to extract all of the information in high resolution signals such as 24-bit 176.4 kHz or 192 kHz recordings, the signal source also needs to have extremely low jitter. That is because input signal jitter components form "side-bands" that inject energy into the signal grounds of the DAC. Even though (our DAC) has extremely high input signal isolation the amount of information in a 24-bit 192kHz signal requires microvolt analog accuracy to reproduce and the less jitter "noise" the input signal carries the better.