I just bought a Steinway which sounds like a banjo.

This company. it seems, was doing something like that:  https://www.stereophile.com/floorloudspeakers/99dun4/index.html

X djones 51 
Yes, exactly this is what i'm saying. You got it. 
And you are right, my tiny 1000 euro genelec monitors are able to reproduce sound in a far more natural way as my silly 10k shiny speakers, non matter with which amplification, source, space i let them play. That do not makes any sense. 

It was a hifi shop shop in Stockholm, not a private seller. And i bough it via internet. 

X Eric, squires "pitch, timbre, dynamics and spatiality" is not my jargon, but thanks for the honor.

Thanks tablejockey, I will call this approach "brute force empiricism".

P.S.
I tried but the reseller refused to refund me the purchase. I tried for many months starting from one week after i got the speakers. 

X @ replicnt6, what a shame for making such a mistake! Ahhhh .... Can I try the subjectivity game? For me they are adjectives! Joke. Shame.
Let’s get back to the sound ... Obviously there is no system that can perfectly reproduce the sound of a piano. The same happens with color reproduction in print. Perfect color reproduction is not possible. Despite this, in this sector they do everything possible to get as close as possible to the ideal result. Measure everything possible and agree on certain standards. Converting RGB to CMYK is already a huge problem, because the CMYK space doesn’t have its own light, so it relies on an auxiliary light. More or less the problem of space when playing an instrument on a speaker that has its own space. A similar problem is faced by those who produce a Fazioli who have to grasp a lot of objective aspects, and then, right at the end, there is a small space for the "character" which is part of the differences between a great Yamaha, Steinway, Fazioli and so on. But in this area it seems they don’t want to define any stadard. It would be easy to agree even on multiple stadards. Imagine defining 128 recording standards, using different microphones, media and recording approaches in order to have the masters of something like a "history of audio recording" in controlled environments. You can do a lot of things with something like this. It is also possible to use this material to train neuronal networks in order to improve the digitization of old recordings. Now imagine you have a Steinway Spirio (it can mechanically record itself almost perfectly so you have the same live performance at your disposal as many times as you want), 128 recordings of the same Stainway and the most neutral amplification possible, and a new pair of Speakers in one space that matches the size of the recording space. Now you can start comparing the original with the master recordings played on the new speakers and learn a lot. If you want to agree on standards in an industry, you can raise millions so you can really do something like this with multiple music instruments. In other sectors they always do. In the meantime I have a pair of $ 10,000 speakers that sound much, much, much worse, from every point of view, than my father’s 1986 Bose 505, worth $ 300 driven by a cheap denon. It does not make any sense. P.S. Yes i sent them even back to the manufacturer and they come back almost identical. 

I have been lucky to listen to them a while ago at the home of a musician in Milan, probably the same model you own (they where big). The best sounding speaker i ever heard.

x Brianlucey: I understood my mistake. Maybe I haven’t explained my idea well, or, because other people have understood it, you simply have not understood it. You’re posing the problem as if I wanted to synthesize the sound of a piano ... I don’t want to synthesize the sound of a piano. I would like a speaker to be able to reproduce all the dynamic range, all the timbre and all the spatiality of real instruments. To do this one must have the possibility to compare the original with the reproduction and if there is any discrepancy, precisely identify the origin of each problem. He must also have the possibility to compare all reproduction techniques and all reproduction media, trying to identify the exact peculiarities of each one and its repercussions in the reproduction phase. Today we have many recordings available but they are all made in circumstances so different from each other that we cannot become wiser because the combinations are too many. It would therefore be necessary to remake masters in controlled situations, so that each variation can be related to a repercussion and mold the construction of the loudspeaker around that information.
It’s completely useless that some of you tell me that they are satisfied with their system! You simply are the champions of the silly "brute force empiricism" approach.
If i buy a source, an amp and speakers, i turn them on (ohhh yes! One year of burn in, i know!), play a known track and no instrument sounds like it should then the silly "brute force empiricism" approach is the only way i can go...i can pass the rest of my life thinking about hifi gear and go for the endless the try-and error highway, hoping to find the three components that song wrong enough to copensate each one. All that in my space in my home...completely useless to listen to a speaker at the resellers shop because wen you bring the speaker at your home it sound completely different. It’s your space!, It’s your amp! the cables!
Imagine i go to buy a sax, i play it in the store, i like it, i buy it, i play it at home and i’m disgusted. I go back to the shop to complain and they tell me that it’s probably because i’m playing the sax in the wrong space with the wrong t-shirt.
It seems that we cannot understand each other. It’s fine.
I like to use my head to formulate theories; you’d rather use it to bang it on the "wrong hi-fi component" wall. Brute force empiricism, or how to share a problem and never its solution. Only a logically structured solution can be a shared solution.