I was surprised to see Julian's results, considering that I've been getting great sound with no audible distortion from my 32 ohm Philips cans. I suspect it's because they're very efficient (published spec is 100db at 1mw), and I don't listen anywhere near that loud. I stuck my phone mic into the ear cup with a db meter running (no idea how inaccurate this is, but hoping it's ballpark). At my loudest listening levels it measured around 85 db.
If I'm doing the math right, this would be somewhere in the neighborhood of 100µΩ, which according to Julian puts the THD+N below -70db:
View attachment 100531
Still not super impressive, but I know from listening tests that I probably can't hear it.
I compared to the headphone output on my Mac Pro tower. At higher levels the computer output showed less distortion than than the Clarett, but it sounded significantly worse. Thinner and brighter. So for whatever reason the computer's output seems to be eq'ing the sound in an unfavorable way. This was not a blind test nor was it objectively level-matched. But I was getting the same strong impression from low to high output. And what I was listening for was distortion (which I couldn't hear from either source).