Great, now consider that assuming a JND of 0.1 dB is extremely generous to the opposing argument, and the actual JND is likely higher, pushing the threshold for speaker distortion even lower in order for DAC/amp distortion to have any audible effect when combined with the former. (The 0.1 dB figure comes from some evidence a 0.1 dB/octave spectral tilt is audible as Floyd Toole says in his book, but spectral tilts are changes our ears are most sensitive to. I'm not aware of any scientific evidence of a JND for distortion as low as 0.1 dB). Also:
Tests of the above described in the
link in this comment put that at around -60 dB at the extreme.
Only because you've been told it is. Past a point as explained above, lower distortion numbers have no audible consequence.
If anyone is claiming distortion at -87 dB is audible, then they can demonstrate this by successfully passing a blind test with music using the excellent
Distort software and Foobar's ABX Comparator component, and posting the log with verifiable checksum. Extraordinary claims require extraordinary evidence.
@Mad_Economist has even done (most of) the work for you, with the last sample on
this page simulating the iFi Zen DAC using Distort, which has a similar distortion profile (bit better) then the X3800H. And be sure to check out all the other samples and simulations of other worse performing DACs/amps on that page and see if you can hear their distortion. It certainly is an eye (or rather ear) opening experience as to what all these figures mean in practice when it comes to audibility.