Doesn't matter what is present. What matters is what is audible. If that single guitar created low level IMD, then adding a Piano to it could mask those components.I disagree: it is a question of intermodulation distortion which can come through a variety of mechanisms. Literally the more audio sources you have in your recording, the greater the proportion of objectionable intermodulation distortion present. And no: "even order" or "low order" harmonic distortion won't help here.
And you would know to detect doppler distortion in auditioning one loudspeaker to another? I don't think so. I have written on this topic before on another forum. I will just quote one piece and leave it at that: http://www.aes.org/e-lib/browse.cfm?elib=15291An example of intermodulation distortion is doppler distortion which cannot be eliminated even with the most perfect zero distortion driver. You really don't want your mid-range riding on top of the bass cone displacements. Splitting the signal into several 'ways' will help to reduce this and other forms of IMD.
http://www.stereophile.com/reference/1104red/#4UvsqZx3uXiiXORl.97
Read more at http://www.stereophile.com/content/...rtion-loudspeakers-page-3#rYBuAWexhYefkTjz.99
"Using the equation given in Beers and Belar [2] , the
distortion of a 1kHz sine tone played simultaneously
from the same drive unit as the 60Hz, 90dB sine tone
calculated above would be 1.30 × 1000 Hz × 0.012 m =
15%. This is above the 2% threshold of audibility for a
pure tone that is stated in their paper, and also above
that anecdotally stated by Klipsch [3].
However, two sine tones is a very critical signal for
Doppler distortion, and one which is rarely encountered
in normal listening circumstances. The audibility of
distortion is highly dependent of the kind, as well as the
amount, of distortion introduced. Informal listening tests
revealed that the effect of Doppler distortion was
unnoticeable on simulated systems with six-inch drive
units, when the low-frequency roll-off of these systems
was considered. This was true even when its in?uence
was arti?cially doubled. We therefore left it out of the
?nal run-time simulation [of speakers with headphones]. "
My advice remains: focus on timbre of loudspeaker, not these secondary products. The former you can easily hear and take to the bank. The latter, not so much.