Back a couple of pages, Todd R got it right. We interpret the spec as strictly as we can, because it's a liability not to.
If we let a signal with a badly formatted header come into the A/D, it could turn out to have been a Dolby AC3 signal, which would render as full scale noise. In a product where the user could be wearing headphones with the volume cranked up, we can't take the risk.
Even 48kHz /44.1kHz would produce sample slips which render as nasty clicks. We don't want the audio to contain clicks that you might miss at first - that gives both of us a bad rep.
With professional gear such as the HD-P2, we can relax the limits a little because we expect pro users to always test first.
If mytek or someone can lend us one of their ADCs, we can investigate relaxing at most the "AES header, but unbalanced signal" restriction.
Tom (TASCAM).