bronchia - 50 samples
note! delay lengths are in samples, which means pitch will vary considerably at different sample rates!
being unsure of how to apply keyboard tracking, there is no internal response to midi note, however key tracking can be assigned as a mod source.
the pulse clock i used modulates at block rate which can have unexpected results.
oscine tract and synsect
vsti by xoxos
oscine tract
oscine tract is an elementary waveguide model of the songbird voice.
incoming pressure must be met with enough reflection to produce oscillation within the tract. note that pressure does not have a trimmer and must be a modulation target to produce a signal, however this can be an envelope, lfo or step event generator.
the dual-branched structure exhibits frequency modulation at some modes. while this a simple abstraction, i was able to recognise this dynamic in bird song after using this model.
each path is represented by a length coefficient, and by an area coefficient that is used to calculate reflectivity at the junctions.
the syringeal membranes are modeled as reeds using 1st order mass-springs. to provide some tonality, a low pass filter is applied to the mechanical movement of the membranes.
if you wish to relate the delay lengths for each section, the maximum slider settings are:
trachea - 400 samples
beak - 100 samples
this was done to reduce cpu cost with modulation. the presets were patched at 44.1k
the panel includes a send region, where modulation sources can be routed to multiple destinations.
synsect
after researching various analyses of insect sound, i decided a filtered pulse clock would provide a flexible platform for emulation.
the pulse clock triggers a one sample impulse used to drive the filters, which are effectively sub-sample delays, and suitable for modeling minute resonances.
because many insects use two mechanisms at once, the impulse has two instances. random temporal variation is applied to one or both impulses.
the output from the first set of resonator filters is summed and passed through two filters to provide more definition. these two filters blend between a parallel and serial structure.
for fans of reference, here are some numbers:
some cricket:
2.7k or 5k clicks at 28hz
some cicada:
1.3k clicks at 5, 10-15 or 76hz
some brazilian cicada:
4.1-5.3, 9.8-12.3khz output at 400-600hz
some red milkweed beetle:
1.4khz clicks at 46hz
note that some of the test conditions in these studies would make you output 12.3khz at 400hz as well.
the bee and fly patches were incidental :)
given the current development of new gui resources, there will probably be a 2.0 release improving these models.
made with synthedit
http://www.synthedit.com
these vsti make extensive use of technology created by chris kerry
http://www.chriskerry.f9.co.uk
special gratitude goes to julius o. smith among others for documentation of his waveguide synthesis development.