Josep M Comajuncosas wrote:
> 
> Wow congratulations! The gong sounds pretty cool and allows a nice variety of
> textures. I see the orchestra more as a phenomenological model of the instrument
> than as a true physical model of it. But you could try to better approximate the
> instrument behaviour, for example using a much more complex waveguide mesh to model
> the plate morphology, and applying more PNF´s in parallell at different points of
> the mesh, as suggested in Duyne´s.

Thanks! I based the goto statements of my PNF on your Karplus-Strong orchestra
with the PNF. I have recently worked on a gong with 8 delay lines joined by a
"scattering" junction, similar to the JOS physical modelling (but in a feedback
delay network format, as opposed to bidirectional waveguides). It sounds pretty
good, but the original gong based on the Puckette/Stautner reverb sounds a bit
better to my ears. The feedback matrix makes a HUGE difference - when I changed
the feedback matrix in the Puckette/Stautner reverb to a feedback matrix based
on a scattering junction with equal impedences, it sounded HORRIBLE. Too much
beating of modes with the scattering junction feedback matrix. Of course, I need
to do some tweaking of the PNFs with the new gong sounds, to find the "sweet
spots." I have also incorporated some subtle modulation of the delay lengths
with sub-audio random functions, and this seems to smooth out much of the
beating between modes.

I plan on working on the 2-D waveguide meshes suggested by Scott Van Duyne in
the next week or so. Unfortunately, most of the shortcuts that he uses to gain
speed (i.e. binary shifting instead of division) won't be achievable in a Csound
orchestra. Still, I do plan on writing an orchestra that implements a waveguide
mesh. I might have to write a small program in C that writes the bulk of the
orchestra, as the number of repeated statements should be huge...