Having been inspired by Giorgio's work on the Solo Strad, I recently reconsidered how I postulated in a past thread how I might approach developing such a virtual instrument. Having no Stadavari at my disposal I fell back to my old crappy (yet still-loved) Holton trombone. (Oh, and no anechoic chamber, either: pity my working conditions!)
I'm attempting to do this through purely additive synthesis, for now, intending to pre-fill wavetables in an interpolating hypercube (blah blah blah tech talk blah). The first step is simply to capture the dynamic range and associated frequency response of the instrument. So trusty bone in hand, I recorded a crescendo from quiet-to-breaking. Here's what a lot of math and coding spat out when I was done massaging the data:
One Note Trombone Solo in Bb (940 kb .wav)
I did take the liberty of attempting to emulate a tonguing sound up front through the use of rapidly varying amplitude: still a bit rough for now, but I think I'm on the right path for not having to rely on samples! Overall I think it's a decent first step.
But that's just my opinion: what do you people think?
I'll admit that proof is in the playability and blending with other instruments, and this is just a single note of a single articulation, but does this humble beginning warrant further investigation? I mean, after this point, it only gets harder and harder...
Of course, the fact that nobody has *ever* asked for a trombone synth has crossed my mind.