I dont actually agree that this is PP's best theremin performance - quite difficult to say which is, they are all so good - but I dont think its this one!
What perhaps makes this "sound like the best" is the vocal intonation from the EH-TM.. I do not think many (any?) other thereminist will manage to match or better what Peter has done in this performance for a LONG time (if ever)... I do, however, think that Peter will better this performance.
IMO, it is probably Peter's singing expierience which gives him the edge - his use of vibrato in particular.. He just knows exactly how to intonate everything in the way a singer would - pitch - vibrato - dynamics .. nothing is amiss or wrong, everything is completely believable and convincing and 'right'.
IMO, the only area where improvement in the vocal synthesis is possible would be if real-time control was given to the performer - so that the voice was not stuck on one vowel ('ah' or whatever) - But however the TM does it, it largely manages to overcome this limitation - the more I listen to the performance, the more convinced I become that some "AI" is involved - this is not a static vowel filter, it is dynamic- I think influenced by frequency, amplitude and harmonic content of the input signal.
I think that Peter has done many performances at least equal or perhaps better than this one - because I have heard him make theremins sing like a human without using the TM.. Ok, the vocal formants were not there in the tone, but everything else was perfect.. Perfect enough that I could hear a 'human' in the theremin tone.
I hope the TM doesn't make good thereminists lazy.
Fred.