My guess is the artistic director decided on substituting with a unipivot because it looked less boring than a gimballed one ;^)
Is it possible though, Raul, that even gimballed arms aren’t entirely "torsionally stable"? The movements may be small but real due to a necessary amount of bearing "play". I remember, decades ago, Syrinx used offset bearings to try to improve this and resonance behaviour.
In fact I used to own a Syrinx PU2... ;^)
Not surprisingly the LF performance really was something to write home about :)