I'm not surprised that a critical listener is more attracted to piano on the Harbeths. It is the defining glory of the Harbeth sound - in my opinion.
You know, the piano has a similar characteristic to the human voice. I think that is why it is such a universally loved instrument - it 'talks' to us with a voice and soul just like a human. Reproducing human voice - and hence the piano accurately is
by far the most difficult task for a loudspeaker. That's becuse we all know how a real live human sounds, and many of us know how a live piano sounds too. No degree-level acoustics expertise needed to hear and judge a great piano in action.
I have mentioned before and I state again: the fundamantal element in a loudspeaker system that governs (or limits) the ability of a speaker to 'sing' naturally
is the material from which the bass/midrange driver cone is formed. Please real that last sentence again. Twice. It is so desperately important to grasp that idea. Harbeth's RADIAL™ cone material has, amongst other properties, extreme stiffness. That means it acts more like a piston, and that is exactly what you want for a speaker cone: as much of the cone moving together as one harmonius push/pull element as it tries to trace-out the micro detail of (for example) the piano tone especially as it decays into silence between the notes.
Contrast the pistonic behaviour of RADIAL™ with that of a soft-dome drive unit. First, the soft-dome is, by definition, made from knicker elastic (or similar), which is a soft, floppy, woven cloth. Bad news. Second, to give it the illusion of stiffness (and damping) it is coated with some sort of sticky gunk which lies unevenly on top of the fabric. That means it is impossible to make two domes which behave the same because they will be physically different. Have a close-up look at the glossy, injection moulded Harbeth RADIAL™ cone and compare that with a similar close-up of the fabric and see for yourself.
There is absolutely no justification to use coated underwear fabric as a diaphragm material. It's simply the wrong solution. Our stiff injection moulded cone made from an engineering plastic is by far the best way to ensure that the micro-tones are sent as sound out into the room, not flopping around in the woven doped mesh of a soft diaphragm eventually lost as heat.
Anyone able to take close-up pictures of the Harbeth cone v. a fabric cone? No need to mention brand names; the issue is that fabric bass/mid drivers just do not meet the primary requirements for high fidelity transducers, namely: repeatable, rigid and optimally damped at a microstructural level. The fact that a midrange dome buckles under the slightest finger pressure proves the point.
You wouldn't make car engine pistons from lightweight woven elastic (even if the could withstand the temperature) and nor should you make speaker cones. Structures that move (including engines, planes, bullets and midrange cone/domes) especially structures like domes/cones which are under great acceleration - should only be made from rigid, reliable, strong, homogenous materials.
I've previously covered the sonic signature of cone materials
here. I think I specifically mentioned polypropylene as a sub-optimal acoustic material, but cloth domes fall into a similar category.