posted 22 April 2007 13:28            

I think that other samplers that are dedicated to that job have an easy time doing a multi sample piano. Kyma can do it but making a patch matrix of multi samples for both velocity and pitch range is possible but is more laborious to construct without the user interface that a sampler offers. The crossfilter could be used to add the resonance of the piano body but then we would have to have a set of sampled notes that had no body resonances in the first place. I don't know how one could make such a set of samples.

That said one thing that sampled pianos normally miss out on is the fact that when the sustain pedal is pressed all the unplayed strings resonate in sympathy with the played notes. A crossfilter with an impulse of all the piano notes, all being played at the same time (with the pedal down) would give this effect very well.

On the other hand what would be wonderful would be a physical modeled piano that sounds realistic. This is what I was trying to achieve with the Resosynth but failed. I notice (by listening carefully) that as a piano string dies away you can hear harmonics moving up and down in level. This happens even for the single string notes at the bottom of the keyboard, so it's not caused by the chorusing of three slightly de-tuned strings. This was what the 2D effect was supposed to achieve as it was supposed to emulate the fact that the string can vibrate not only in and out but also left and right and that the level would change as it swapped it's oscillating direction. But when the 2D effect is used it didn't make harmonic change in level a bit but instead made them appear and disappear completely and then appear again. In other words it didn't work.

Maybe the reason the changing harmonics happens is because the dampened strings are still adding a lot to the sound, in which case a crossfilter with impulse of a piano being hit with the sustain pedal not pressed could give this effect. I don't know? The Resosynth is quite good at making guitar type sounds (where there are not so many near by strings) so maybe this is the missing link?

If I had a piano I'd try hitting it and recording the result. Most samples of pianos being hit are with the sustain pedal down which is not what we want.

A physically modeled piano would have so much more variety than a sample based one but I just haven't got there yet.

Any ideas from the physics experts or piano acoustic experts would be appreciated.

Thanks