Sherlock Holmes needs clues

Judit,
In my experience, clipping sounds like the ch in challa or chutzpa at both the top and bottom ends as the amp begins to clip. Increasing the volume decreases the intelligent audible spectrum as the distortion starts from the outside edges towards the middle.

The distortion caused by clipping does disappear when the gain is decreased.

Frankly, if the phase angle and impedence is brutal at the frequencies you are mentioning the amp would actually only be putting out a fraction of its "FTC" output. An example would be a 200 wpc SS amp - into 8 ohms - which actually is only putting out 6 watts or so at this nexus. Thus clipping occurs even though the volume is low. (This is why, in part, so many of the 70s SS amps of high power did not drive real speakers as well as a 35 watt Dynaco tube amp.) Am I being clear?

Sherlock Holmes has spotted Dr. Moriarty ... Today I took 7 piano CDs to an audio store and listened to an entirely different system than my own - Magnepan 3.6, Levinson CDP and pre, and Audio Research tube amps.

I must tell you how surprised I was to find that this system had exactly the same problem with piano that I was hearing in my own listening room. This has convinced me that Plato may have hit the nail on the head. Digital playback is where Dr. Moriarty hides. Either there or in my ears.

The problem with brass playback I am having, appears to be unrelated, as it did not occur during today's listening session.

Hmm...lots of midrange drivers (esp 5") have a rising frequency response between 300 and 1k or so, which will thin out a piano,sometimes with a "cuppiness" coloration on human voice, too. For that reason I've preferred 6.5" midranges as a rule for inexpensive speakers. I've designed both 5" and 6.5" two-ways, and have a Steinway B in my room as a reference, and could NEVER get the 5" to sound as neutral as the 6.5. (Yet a GREAT 5" CAN be made to sound neutral, of course, as was accomplished Verity Audio in their Parsifal Encores, for example). Also note that room reflections around 800-1k are VERY position sensitive. Try moving your listening position just a few inches to see if the mids even out. That may tell you to play around with sidewall-damping schemes like sofas/pillows (what I use), etc. I do NOT buy the CDP suggestion as the source. There are other issues ("digititis", etc.) with CDPs, but a rising frequency response in the midrange isn't one of them!
Good luck.

Judit-
Could you please list the CDs that show this phenomenon? I'd like to try it on my system and report back. Perhaps you could describe where, timewise/trackwise, the phenomenon occurs?
Thanks,
IJ