Just for argument... At 24 bit sampling depth, there's not going to be much in the way of stair casing in digital recording either. It's there but on an infinitesimal level. That's 16 million possible steps from 0dB down to the floor. At some point a sufficiently fine grained digital system becomes effectively equivalent to an analog system in any practical terms. Since the 'measuing device' here is the human ear, that's far, far beyond the human ear's ability to distinguish one sample level from another.
If you have a -120dB floor, that's 0.0000075dB per step. Most folks can probably barely hear a quarter dB change reliably, so that leaves more than a bit of leeway. Even if you aren't using the full drange (peaking at -6dB, say) that still leaves at a tiny fraction of a dB per sample stepping.
Obviously the sample rate plays into that as well, but you are free to do 24/96K, which would just be so far beyond any human ability to distinguish that it's not worth worrying about. Particularly when, whichever way you record, you are going to put it out to 16/44.1 digital format anyway.
perhaps..but to me tape just sounds better.