It is possible to objectively test even with music. This is done by sampling the output of the device that is being tested, and then subtracting the original signal that has been filtered to match the frequency response (including phase response and delay) of the recorded audio as accurately as possible. The result contains all distortion, noise, jitter, and whatever other artifacts that have been added.
Here is an example of this. Difference extraction can be used to disprove claims that an amplifier performs great with sine waves, but fails badly with music.
I shouldn't be surprised if it can be tested objectively with music. Has it been?
It can also be "tested" subjectively by listening.
What if it works fine in an objective test, but sounds mediocre?
(I was already aware of what you did in that thread . . . )