[sdiy] Thoughts on drum synth

[Peter Grenader]

I am very impressed with these figures, especially the mode information from
Tim.  It really is great that you guys have such knowledge. Very impressive.

I think though that the answer, the sucessful implimentation will lie in a
balance of the left and right brain lobe. The 'left 'in this analogy being
the synthesis of the physics behind the sound, the 'right' being in the end
it will come down to tweeks by ear.

My suspicion is if even if the harmonic/amplitude ratios could be executed
fathfully via analog, that it still might not sound like a kick bass because
there is an aweful lot that goes into it outside of these overtones.  In the
very least, while you may be able to nail one sound, the normal use of a
bass drum within a single track contains many different sounds even though
only one drum and one drummer is present.  What will be difficult to
synthesize is the human qualities of the playing, which as you know is a
real big part of the the whole effect of  a drummer drumming.

It has been my experience that while it's possible to get close via analog
or even digital, it's usually impossible to get spot-on.  In this instance
though I think this is a job for digits, not volts.

If you look into those who have tried this with other models, the human
voice (Charles Dodge, Paul Lansky and Jon Appleton) in the end it was
impossible to accomplish via analog and they all had to go to computers (at
Columbia Uni, Bell Labs and Dartmouth College respectfully).   It's silly
however for me to compare the two sounds, as the human voice is by far more
complex than a kick. One who has come close is Wendy Carlos  but be aware
that most of her drums sounds were played at half speed in their final form,
the effects of which you just can't replicate through real-time analog
means.

And while I am not a real big fan of the the technology, the subject will
undoubtedly come up "why don't you just sample a bass drum", although I
understand that this is not the thrust behind the experiement.

Again, I am not trying to quell this idea because it is quite admirable.  At
the very least, I am trying to suggest that a very careful analysis of the
entire sound event needs to be taken into account, which goes much beyond
the harmonic structure of the instrument.

very respectfully submitted -

P



on 3/31/02 7:11 PM, Tim Ressel at madhun2001 at yahoo.com wrote:

> Christian,
> 
>> Fascinating...
>> is this EVERY drum with a circular head?  I'm
>> assuming this is for a head
>> stretched to a specific tension all the way
>> around... or does the tension of
>> the different parts of the drum just change the
>> relationship of the mode
>> amplitudes to the fundamental?
> 
> Assumes equal tension across the head.
> Formula for freq of stretched membrane:
> 
> Fo = (0.382/R) * sqrt(T/m)
> 
> R=Radius of membrane in centimeters
> T=tension in dynes per centimeter
> m=mass of membrane in grams per square centimeter
> 
> 
>> this would be interesting.  Modeling this seems
>> intriging... I've been
>> studying physical modeling for a while and wonder if
>> we this could me
>> modeled with a delay line representing the 'height'
>> of the drum with a
>> scattering junction at end.  Sort of like a Physical
>> model for a horn or
>> flute, but with a really different excitation
>> signal...
> 
> Because the radius is much larger than the length, the
> end effects kinda swamp the tube effect. That's why I
> said the filter would have a low Q. Physical modelling
> would be fun, if a bit overkill...
> 
> 
>> oh, and the Walsh thing looks super cool.  I was
>> thinking of trying to do
>> this on my DSP too.
> 
> I'm still confused over how one would do this. And i'm
> wondering why one would do this, when Fourier
> synthesis is so much more simple.
> 
> --tr
> 
> 
> __________________________________________________
> Do You Yahoo!?
> Yahoo! Greetings - send holiday greetings for Easter, Passover
> http://greetings.yahoo.com/