@willemj
Don't agree about electronics, although I applaud your skepticism and insistence on rational thought and experiment.
First, I think that you will agree that more processing tends to sound worse. So the first thing I did after buying my Quad ESL's was to build amplifiers with limited rails, which could not drive the speakers into protection. I was therefore free to remove protection circuits. That made a difference, and I'm sure that you can see why. I concluded that electronics can make a difference.
Second, I have demonstrated (single blind, A/B) to a skeptical engineer, that ONE resistor can make a clearly audible difference. I used a discrete, single active resistor, gain control (not a ladder) to control preamp volume, and thus to compare five different types of resistor. The engineer's response was, "Now tell me what you really did." He could not believe that the differences he was hearing were all due to different types of metal film resistor, in one location (gain control).
My wife, who sings in a choir, agreed (single blind, A/B).
Differences were subtle between 4 of the 5, but the worst sound was actively unpleasant. One resistor, mind. Nude Vishay was best, no-name Oriental was second. That led me to standardize on nude Vishay's - every low power resistor in my amps is now Vishay, and also in my new phono/preamp.
For the record, the system was built around one pair of 2805's and one pair of 2905's angled at one radian, all running the Vanderveen transformers with circuit protection disabled (since changed to all 2905's). Amplifiers were DIY solid state mono blocks (since improved). Also for the record, I hold advanced degrees in math and science, and am thoroughly acquainted with the techniques of experimental psychology.