I confirm the technical aspects mentioned by Boesendorfer
The ASIO driver allows one to avoid passing through a considerable number of software virtualization layers (many more on Windows than on Linux, I don't know if it exists on Apple) and to access audio playback peripherals more efficiently.
This allows for the reduction of what Simone calls "Lag", i.e., the time elapsed between the moment the key is touched and the moment the generated sound effect is heard.
If this time exceeds very tight thresholds (milliseconds!), the performer perceives the phenomenon and can no longer play well (assuming they normally do!).
Furthermore, since the performer can press many keys in a short amount of time (this is called an increase in polyphony), it is necessary that the computer (which today processes each key consecutively) processes the effect of all pressed keys without falling behind the performer; otherwise, the playback breaks down (keys are pressed that are not played, which is even worse than Lag!).
To this end, the ASIO driver has a buffer (a customizable memory of finite size) in which one part of the driver accumulates the playback information (which we will hereafter call events) and from which another part of the driver retrieves the events to perform the playback. The larger the buffer, the less risk there is of losing events, but the larger it is, the more the Lag can increase, with the risk of what I explained before.
To what has been said now, we add what was said in previous posts, regarding the fact that the sound is "synthesized" on the spot, without resorting to any "sample"; therefore, available power must be ample and/or one must have a minimum of technical knowledge to "adjust" the computer parameters so as not to have negative effects.
It is clear that sample libraries today have advantages over programs that, like PianoTeq, work with physical models, but it is true that - with time and the increasing performance of computers (there are empirical laws that have proven true so far, supporting this statement) - the trend will be in favor of the latter, for a very simple reason: sample libraries can only play sampled sounds and/or modify them slightly.
I will give an example of something that sample libraries cannot do:
We all know how different the sound sensation is if we play two notes - one after another - or play them together, for example, from a timbral point of view: it is observed that the timbre changes as a result of the non-linear interaction of the timbres of the two component notes.
If a model is sophisticated enough to be non-linear, then it succeeds in reproducing such behavior. A technical reason is that the model reasons on frequencies, i.e., on the elements that compose the single emission of a note with a timbre.
A sample library, instead, is based on the recording of the sound (including the timbre) corresponding to the touch of each key at a certain velocity. At this point, being unable to sample all combinations of simultaneous note associations, it would have enormous difficulties (in practice it would be impossible) to reproduce the effect.
The other aspect of PianoTeq that attracts me is that, when appropriately customized, it manages to reproduce an enormous number of instruments, with different timbres and tunings (just the other day I was about to post some examples, from Rag Honky-tonk to Baroque on spinet played with PianoTeq, then I got bored - it was late at night - and I gave up for a well-deserved nap!).
I believe that, if one also wants to have fun, this is an advantage; I consider it not only musical but also and especially economic.
It is therefore obvious that one must have a moderate and gradual approach, experimenting and evaluating products also based on one's own needs and possibilities (musical and economic).
Perhaps that is why, when I was young, I never bought a Ramirez Concerto guitar (> 9000 Euro today!), but I settled for a Takamine (its sound is beautiful!) and a Cuenca flamenco (wonderful). Today I would have the money, but a Ramirez would still be wasted (but is it, Segovia ?!).
But that is just my opinion, everyone can do what they want.
P.S. @ Boesendorfer
The lady we spoke about some time ago was seen, in an evident state of advanced intoxication, around those parts. Was it Teroldego from your cellar or one of your performances with PianoTeq?