[PD] Avian syrinx revisited

Andy Farnell padawan12 at obiwannabe.co.uk
Thu Feb 21 06:23:05 CET 2008

Avian syrinx revisited 

Synthetic birds have been done quite a bit, by Tamara Smyth (pressure impedance),
Seppo Fagerlund (a three port waveguide) and Hans Mikelson ( a csound FM
version ) [BTW if anybody else has got published models I don't of please let
me know.]


I first put a tutorial on the syrinx up on my site 2 years ago
after copying Hans FM/AM version from Csound to Pd and then messed
with waveguide models and adding different beaks and trachea with some


But the one thing kept evading me, squawking birds. Oscines are singing birds
with a highly developed syrinx, and those are the ones everybody studies and
tries to emulate. Turns out oscines are the easier ones to create. But "working
class birds", crows, pigeons, ducks and seagulls are a different problem. 
I live by the sea and when they're not stealing my chips, seagulls fly by 
my room taunting me saying "come on then.. make a seagull sound!", but I could
never work it out. Seagulls 1: Andy 0

Well I fixed that, I think. 
Thanks to some papers by these chaps, Suthers, Fee and Fletcher
it was time to have another look at the syrinx.

Starting with Tamaras pressure analysis again I decided to start making each 
bronchus separately as a  pulse source then feed an FM stage with 
feedback. I guess this is the same as using FM with a complex modulator, but 
the feedback gets a semi chaotic effect which is what Fletcher suggests is 
behind the squawk of birds with bigger lungs and smaller syrinx.

Here's the results (with background to situate them - all
synthetic scenery as usual)

First the regular oscines in a forest (with some recent insects)

Now two examples from the new model
The star of this one is the duck...

And here's the seagulls (gotchya you feathered freaks!)

The same model can do both, so it's quite flexible. All (pure) Puredata The Pd
code is still a little messy to release but it will be on the site soon or in
the book maybe.


Use the source

More information about the Pd-list mailing list