[Stk] suggestions for bell modeling

[rurik leffanta]

i am creating an affordable/free series of physically modeled instruments in VST format.

i'm modeling aerophone bell flares with a banded waveguide like structure, filtering and applying a short delay derived from the partial's frequency.

presently the delay applied to each band is simply
(frequency of partial / nyquist) * bell delay coefficient (~3-7ms)

so that the partials have an increasingly larger delay in ascending order. i also tried an "octave unit/linear" spacing of band delay times which sounded less convincing (a hard call given the abstract nature of the method).

do any improvements seem obvious to this technique to those who have reviewed more data?

the behaviours beyond my knowledge reduce the tuning beyond the appreciability for those inexpert users who lack functional continuous pitch controllers, i like to think modeled instrumentation is informationally sensitising to the composer.

are there other elementary methods of bell modeling i should be exploring instead? i lack education above trigonometry.. my experimentations have encountered cascade behaviours that elicits the utility of short periodic switching julius describes for stabilising TIIRs, i was not able to estimate which IIR filter he intended (my erudition has only recently surmised the calculation of damping filter phase to pitch correction in a karplus-strong circuit..).

rurik.
http://www.xoxos.net
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://ccrma-mail.stanford.edu/pipermail/stk/attachments/20090528/b188957d/attachment.html