There's nothing wrong with that unless the distortion rises with frequency, which is very common in solid state amps (in which case I agree). So I suspect it does in this case too since every MBL system I've heard sounded bright. I know its not the speakers since I've heard them with amps I know aren't bright.
If you look at the measurements linked above they show I'm right. Distortion begins to rise at about 1KHz (very common turnover point BTW). This cause higher ordered harmonics (such as the 7th harmonic of 1KHz) to be unmasked (IOW, audible). The ear converts distortion to tonality, so this adds a bit of brightness and harshness to the presentation. Apparently Atkinson, like so many others, didn't understand the implication of this measurement.
This is such a common problem that people have come to assume that high feedback is the problem, when really its distortion rising with frequency. That happens because the feedback and the gain of the amp (together called 'loop gain') are not both supported by the Gain Bandwidth Product the amp makes. So in a secondary sort of way, if the feedback were reduced it might be possible to find a value where distortion no longer rises, which makes for a much more listenable amplifier.
If an amplifier is bright and/or harsh, that cannot be construed as neutral. Its a coloration. The fact that this can be caused by very low amounts of higher ordered harmonics seems lost on many designers (and audiophiles for that matter). There is a weighting system for assigning various harmonics the amount of audibility they have. Oddly, this bit was understood back in the 1930s (see the Radiotron Designer's Guide 3rd edition, chapter 5)) and somehow seems lost today.
That being said, i find the treble of these amps very extended but also somehow forgiving at the same time, a rare quality.
