Now that pedal down samples can be simulated rather exactly by sympathetic resonance scripts, I'm wondering if convolution is neccesary at all any more.
The PMI website says that two components are necessary for true sympathetic resonance: One is the sympathetic resonance script, and the other is the convolution impulse of the soundboard:
"one is for the body resonance of the piano. A convolution engine excites a resonance model of the ENTIRE piano harp, exciting the string resonances that have a harmonic relationship with the keys struck. Pushing through a "chromatic" piano body impulse, the tone exciting the impulse will determine which of the chromatic resonances will have impact on the sound. If you play an "A" you only hear resonances related to "A" up its overtone series. The frequency components which go "un-excited" will simply disappear underneath the overall sound and be unnoticeable.
I just don't get why you need every note going through the convolution engine on top of having the sympathetic resonance script. Isn't the resonance of the soundboard allready recorded in the dry note samples?