Surprising to read this return to this subject--I've been thinking about it too, and thinking about sympathetic resonance, and listening to real and sampled pianos and PianoTeq with a focus on the cabinet and soundboard response\rumble and the sound of sympathetically vibrating strings. I'm wondering if the location of the sound and the way it spreads might be further explored. I've come up with more questions than answers, and a longish post:
1. About the soundboard and cabinet response: Part of the difficulty is of course that the original model was for a classical, distant perspective, from which the vibrations from the soundboard would seem more grouped together and coming from a single location. The same is true of the cabinet vibrations. In any case, my impression--what I seem to hear and what I assume from common sense--is that although the entire soundboard and cabinet shake, rattle, and roll, the evolving sound of the rumble is at first localized near the string and then spreads out. In other words, taking just the soundboard as a resonator (forgetting the filtering and reflective properties): the vibrations from the string first vibrate the nearest area of the soundboard, but the vibrations are quickly transferred throughout all of the cells of wood until they reach the surrounding bracing, and then the standing waves are established. In Pianoteq, the sound from the soundboard seems a little more permanently localized near the string--the partials are correct and at the right amplitude, but the sound doesn't spread across the entire soundboard. Insofar as there is a problem, it's just with the narrowly defined location of the sound? In other words, could the desired rumble be achieved by fairly simple means, something along the lines of what was done several versions ago in changing the apparent location of the "strings"--by having the sound spread out more in all directions, very fast, and then stay spread as it dies? (I'm not sure if the amplitude of each part of the soundboard changes much as the vibrations bounce off the bracing and return to form the standing waves, and I'm not sure how much we can perceive the isolated locations of each standing wave, but a listener sitting close to the piano might need to hear, without consciously hearing, that the partials are slightly spread across a narrow area of the soundboard.)
2. Similarly, the sound of the Global resonance (the cabinet and harp) is also more stable in the model than it is on a real piano: it seems to come from, and stay at, a given location. I don't want it to pan, exactly, but instead to both stay where it is and pan. Is there a name for that--a widening stereo and vertical spread that responds to velocity, perhaps? This approach might be more efficient, in terms of coding, than a convolution reverb approach. (However, I would still like to have the harp resonance decoupled from the body resonance, with impedance and filter controls for each. The current coupling and lack of controls greatly limit the user's ability to shape the sound.)
3. About the sympathetic resonance: Is it true in real pianos that the closer an undamped, unstruck string is to the struck string, the more it will vibrate in sympathetic response to the vibrating string? Seems as though it would--an undamped D next to middle C would receive much more force from the vibrations of a struck middle C than the F below it. I understand that the amplitude depends largely on the extent to which the unstruck note contains partials of the struck note. However, the simple force of the hammer on the string, which causes the transient noise, does seem to set off vibrations on neighboring notes on real pianos--even on damped strings, for pianos that have a less than perfect damper. (The ones I often love.) I'm not sure I hear the effect of this relative closeness in the model though--I miss it particularly in the middle C area. (In other words, does hitting middle C cause the strings for the D beside it to sound as D? Seems to, a very little, at the transient stage of the sound envelope and a little after that. Maybe that sound is just part of the transient sound, but it's not just noise.)
4. Similarly, is this relationship (of proximity to amount of sympathetic vibration) modeled in the Full sustain resonance parameter--the parameter on the Options menu?
Am I in already explored territory, here?
Last edited by Jake Johnson (01-10-2008 02:54)