On an acoustic piano how would noises be influenced by key release velocity?
The velocity of the key release directly influences the velocity with which the damper is lowered onto the string (assuming the sustain pedal is not pressed).
So a quick release lowers the damper quickly, a slow release lowers it slowly.
The sound of the string being damped is different depending on how quick or slow the damper is lowered onto it.
DPs can attempt to model this varying sound if they know the release velocity.
This can only be modeled as an approximation however, because of the differences between acoustics and (most current DPs):
On acoustic pianos, the dampening of the string finishes
at the moment of the full key release.
On a DP however, it starts
a the moment of the full key release.
Imagine a very very slow key release, let's say so slow that between the moment when the damper begins to touch the string and the moment the key is fully released (i.e. the damper sits fully on the string) a whole second passes. During that second, you will hear the acoustic effect of the damper touching (and dampening) the string more and more, until the sound finally ceases at the end
of that second (or even a bit earlier), which is the exact moment when the key is released fully.
On a DP, the release velocity is also measured during this second just before the key release, but he note off message that makes use of this information is only generated at the moment of full key release. So if the DP would try to model the same key release sound of the string being damped more and more over the course of the a whole second, the DP would have to model this sound effect during the second that follows
the full key release - while on the acoustic it would happend during the second before
the full key release.
So using the key release velocity to generate different sounds for the key release is actually not optimal. Well, in real life, key releases are of course usually much faster, so it probably won't be noticeable, but still, there is
a deficiency inherent to this method.
The AvantGrands (and likely also some of Yamaha's acoustic silent systems) are currently the only DPs that I know of who try to address this deficiency by re-purposing the after touch MIDI event (as was discussed elsewhere - and don't confuse this with the normal pressure-sensitivity after touch and also not with the same term used to regulate acoustic pianos).