The following documents the making of the first
CADSD didgeridoo. I was interested in a D# with an easy first toot
on F (i.e. the toot is an octave + one note above the drone). The
back-pressure was to be chosen so that a fast, percussive playing
technique was possible. Also, when the drone was played, at least
one singing tone should be clearly audible.
The starting material was a hawthorn log about 1.80
metres long with a peculiar fold in the future bell-end.
After about 75 different computer simulations I had an inner form that
should lead to the desired sound and playing characteristics.
Simulation of the desired internal didgeridoo form
with a D# drone (1st white peak 77,8 Hz) and easily playable first
toot on F (2nd white peak 174,6 Hz). The strengthened 4th and 5th
overtones in the blue sound spectrum of the drone are clearly visible
and should be audible as "singing" tones. The orange-green sound
spectrum comes when the first toot is played.
Following the building instructions described by
Kay, I then shaped the approximate external form, sealed the
outside with wood glue and had a cabinet maker I know well cut it
down the middle with a bandsaw.
Also following the method described in the new book
and using criteria determined by the simulations and the unique shape
of the hawthorn workpiece, building templates were made of paper and
the measurements transferred to the two halves.
The most complex work step was the millimetre-precise
working of the calculated internal form on the basis of the templates
with the aid of woodcarver, chisels and power file.
The two halves were then doweled, glued under
controlled conditions, and after final drying, sanded down. The inner
surfaces were treated several times with natural oils.
I chose a part of the simulated drone sound spectrum as the motif for
the painting. So, at least the construction of this instrument had
been successfully completed.
Now came the moment of truth!
Could the didge be played and would it sound as planned?
Playability and back-pressure were fully up to expectations at least.
A reverse-check by means of acoustically analysed FFT measurements of
the simulated sound spectra would show everything else.
Soundexample:
Comparison of the actual FTT measurements
with the simulations
The most difficult measurement to make is the acoustic input impedance
spectrum. Unfortunately I lack the experimental facilities for this.
As a compromise one can, however, slap a flat hand on the mouthpiece
to encourage all the inherent resonances of the air column in the
didgeridoo to appear. (After slapping, the flat of the hand must keep
the mouthpiece closed.) The frequency spectrum analysable by this
technique indicates at least the location and frequencies of the toots.
But because the height of the impedance peaks expresses something
about the acoustic back-pressure, this can only be tested subjectively
by playing.
Analysed FFT spectrum of inherent
resonances of the air column stimulated by slapping the flat of the
hand on the mouthpiece.
White peaks: simulated input impedance spectrum
(drone and toot sequence)
Blue-violet spectrum: simulated sound spectrum of the drone
As can be seen, a relatively good match with the frequencies of the
black simulated impedance peaks was achieved. The following
measurement of the sound spectrum when playing the drone also shows a
satisfactory match. When one considers that in the simulation I
deliberately strengthened the height differences between the peaks (for
easier recognition of influences of the projection), the match is
impressive. Anyone who’s done FFT spectra will know how sensitively
the definition of the forms react to changes in the behaviour of lip
vibrations.
Analysed FFT sound spectrum when
playing the drone
In addition, the measurement of the FFT sound spectrum when playing
the first toot shows an astounding match with the simulation.
White peaks: simulated input
impedance spectrum (drone and toot sequence)
Blue-violet spectrum: simulated sound spectrum of the 1st toot
Analysed FFT sound spectrum when playing the 1st
toot.
Summary
I keep being amazed at how relatively precisely
the playing and sound characteristics of more complex didgeridoo inner
shapes can be simulated and calculated.
==> Computer Aided Didge Sound Design
(CADSD) is no dry theory, but can really be used successfully in
practice.
