Converter complement:
Babyface specs:
PCM4104 specs (48 kHz):
DR 118dB(A) typ, 112 dB(A) min
THD+N -100 dB
The datasheet indicates THD of about -107 dBr. THD+N seems to be limited by an increase in noise floor at high levels.
No idea whether the part is subject to 0 dBFS hard-limiting, but on a computer that should be easy to avoid anyway.
As typical for RME, they seem to have gotten pretty close to converter spec. I'm assuming the trusty NJM4580 is holding DAC THD back a bit, some improvements in measured performance are likely to be feasible with a higher-performance part. Internal supplies presumably are +/-5 V. The output may well be fixed level with digital attenuation only, so expect line output noise levels of ca. -100 dBu(A) (8 µV(A)).
I'd say there is little wrong with the DAC if you can get such a Babyface at a good price. It's still better than a typical CS4272 job (Scarlett 2i2 gen3 etc.). There is little doubt that a MOTU M2 beats it in measured performance though, and the Audient iD4/14 MkII (even if its output tops out at +12 dBu) would be a hot contender as well.
On the ADC side I would prefer the old AK5385B to a new-fangled AK555x though (dynamic range may be a few dB less but digital filter performance is not even close), so points for the Babyface there. (I'm sure the ADC input stage could use a bit of help, given that the AK5385B is rated at -103 dB THD+N.) Same goes for build.
Now the PCM2505 mic preamp, while excellent in dynamic range (EIN still is -119 dBu at +9 dB gain, where max input level should be +3 dBu), tops out at a so-so EIN of about -123.5 dBu, though there is no indicated of it being A-wtd, so could be 20 kHz flat or approx. -125.8 dBu(A). Great for condenser but not quite the last word for dynamic mics... though when you can pick up a Klark Teknik mic booster for 35 bucks, it's not such a super big deal.