Its just a headroom issue. Didn't see the other post you are refering to but here's why it may be of interest to someone. This is a quote taken from a Groove Tubes manual: (Flamesuit on

) Maybe some designers on the list would like to comment...
"It seems that some of the new 24-bit digital systems
demand rather high input voltage in order to reach full-scale
modulation. Evidently, in order to make the signal-to-noise
improvements inherent in the new 24 bit systems meaningful,
some converter hardware suppliers have raised the A-to-D input
drive voltage requirements to levels beyond the capabilities of
what most popular op-amp based outboard gear can provide. This
was foreseeable, since, at present and practically speaking, there
will always be about ½ to 2 micro-volts of noise present across a 20 kHz audio bandwidth at the input of the new 24-bit highbandwidth converters, unless exotic and expensive (and impractical) cryogenic methods are used to reduce it. The noise
origins are a combination of the real-world effective signal source
impedance being presented to the converter’s analog inputs, and
nearby spurious digital or analog noise sources, which may be
leaking into the converter inputs due to circuit layout issues, or
proximity to other noisy system elements, and so forth.
When 1 microvolt of input noise is multiplied by about 16.8
million, which is the approximate number of steps in a 24-bit
system, the resulting number is 16.8 volts. This implies a
minimum drive voltage requirement of 16.8 volts RMS (sine wave)
or about 48 volts peak-to peak for meaningful 24-bit full-scale
modulation under ideal circumstances, not counting dither. Many
real world converter systems will generally need more drive than
16 Vrms. Surprise! You’re out of drive headroom when hitting your
24-bit converter with a signal from outboard gear using
monolithic op-amps running off of +/- 15 to 18 volt supply rails!"