Partially agree with all of you... Sounds/Noises of nature: they are spontaneous and the only difference in capture is represented by the type of equipment used. Speaking of musical instruments, in my opinion, the discussion is very different and one should separate the listener's impressions from the musician's impressions. Let's put ourselves in the listener's point of view. We can state as true that it is truly difficult to recognize performances recorded via digital samples rather than analog recordings, and this has already been demonstrated in another topic where a real challenge to find the analog was conducted. The search could then shift toward searching for sound imperfection or possible imperfect note intonation to uncover the analog. I will tell you, when I can choose an audio product from the '40s-'50s, I prefer it 100 times over a contemporary product, precisely because being perfect does not necessarily mean being better. Moreover, this ARTIFICIAL perfection generates great mental confusion that completely fails to meet the listener's expectations in live recordings, where there can be error and where there can be an imperfect sound rendering, perhaps due to environmental factors such as humidity, especially for woodwinds and strings, or absence of reverb and thus sonic dryness, or perhaps the opposite: too much natural reverb. Non-professional listeners, accustomed to the aforementioned recordings, are therefore led to judge a beautiful performance negatively just because there was a wrong note or because of the aforementioned "audio rendering". The ear is therefore shaped to recognize only that which is perfect, yet if we look closely at Michelangelo's paintings, approaching them shows clearly that it is not a photograph. There you have it, art is made by humans, not by computers.
Speaking from the musician's point of view, personally, I adapt to playing on digital pianos, while still realizing that I have to give up many things. One among many is the vibrato pedal, a large part of voice polyphony, and generally everything that makes the difference in expressivity.
Returning to an audio fidelity type of discussion, it must be said that the sample can be better than the acoustic as long as the estimation is performed on a single note. Already in the case of dyads the discussion changes because, from a purely mathematical point of view, the sum of the waves of two samples (e.g., C - E) is different from the sum of the waves of the same dyad recorded acoustically, because the first sum does not take into account the interaction between partials. I am not at home right now, but tonight I would like to let you hear the sarabande from Bach's second part for violin. You will hear the difference with your own ears.
The recording is by violinist Alessandro Marino who, if he wishes, can intervene and give us his points of view...