Yeah, @Rochus is right. Once the action has passed the escapement, its entire point is that you lose all control, else you'd get thudding, reversing the point of the fortepiano's innovation.

That's not to say us pianists have other tricks up our sleeves though. Staccato, slightly delayed playing, damper pedal (which is not just binary but continuous), slower release of key (because you can slowly apply the damper giving it a slight ringing effect), tempo changes, and other techniques get you quite a broad range of feeling. It doesn't however mean that timbre is orthogonal to volume--they are coupled by velocity.

Also, this post doesn't go into methods, so I worry about lack of ABX testing, whether it's a single note or a piece, whether they could see the pianist, etc. Perhaps they addressed that in the paper that they'll publish...

Edit: another thought: why is this even using subjective listeners? You can just measure velocity and run an FFT to see whether they can make a timbre separate from volume and velocity.