(Sorry for resurrecting this thread, I thought it was better than making a new one..)
So I was equally intrigued by the same video, and searched to see if anyone had mentioned it here.. I guess I was hoping Pianoteq might have a secret commandline option to disable the direct sound, or a hidden config, etc., so it would be possible to get a similar effect. (Hey Modartt! I do think this or a "direct note volume" would be a cool setting! I guess it might not be so easy to implement though :-) )
Anyway, since that doesn't seem to be the case, I thought about how this could be achieved by other means.
I hit on the idea of running two instances of PT simultaneously, identically set up except one with the harmonic pedal depressed and the other not, then phase-inverting one of them, so "subtracting" the direct sound. Maybe it wouldn't work exactly but might get somewhere near.
Before proceeding, I tested my phase inverter: I routed a single instance of PT directly out, and also out via the phase inverter (using calf stereo tools on Linux, setting stereo phase to 180, all other settings at default.) As I was hoping, the result was silence, indicating that the phase-inverted signal exactly cancelled the direct signal.
I then went to the next step and started two identically configured PT instances, output one directly and the other via the phase inverter. To my surprise, this did not result in silence. The sound was quieter and thinner, but output was quite audible.
I reasoned at this point that there must be some nondeterministic factor involved in sound production on PT, creating some small random differences between the notes that make phase cancellation ineffective. The only setting I could think of was the humanisation setting in the mallet bounce section, so I turned this off along with all effects. I believe this improved the cancellation somewhat, but not completely.
Regardless I carried on with the experiment, since the non-cancelled sound was fairly quiet, and in any event even with perfect cancellation I might expect "leakage" once I introduced a difference between the two instances by depressing the harmonic pedal since this might well introduce other differences due to details of the modelling.
So with sympathetic resonance turned up to 5 and everything else as described, the "resonance only" effect was mostly achieved. You could hear bits of the direct note, but the sound was predominantly sympathetic resonance, and sounded not entirely unlike the sounds in Paul Barton's video - a bit thinner and less sustained perhaps, but that could be the particular piano model I was using; setting up a piano model for the best "resonance only" sound is an entirely new enterprise, after all.
So.. The point of posting this, anyhow, is partly to enquire if other PT parameters might be worth setting to avoid randomness, partly to ask if anyone from PianoTeq could confirm that there is some randomness (i.e. I didn't just blow my experiment ) and partly just to pass on my idea in case it interests anyone.
Last edited by petez9 (12-05-2018 15:49)