freq mult

[Hairy Harry]

But really really seriously folks...

This is a great idea and I've been working on it too...
OTOH I have NOTHING that works, but maybe can steer you
from things that don't work...

The guitar wave is a bastard.  The harmonics are shard because
of the stiffness of the string, so the waveshape will change
constantly... It is almost impossible to get a square wave out
from it. The duty cycle will never be, or stay at 50%.

Check out guitar synthesizer patents, especially Akimatu, Merriman,
and Gold for some insight as to the problems involved. I like (and use) the 
Gold method myself, it trades an overwhelming tendency for
the square wave to "octave hop" for having the wave reverse phase...
which is much less noticible.

If you can't get a 50% square, integration is out of the question...
this is probably where your DC offsets are comming from.

If you have a triangle (pure) you can use a wave folder to get higher 
harmonics... but this needs a very stable amplitude.

If you can get a sine wave you might use frequency shifter (sideband -
like the Bode or similar), or do a squaring function... the same sine
multiplied with itself will give a 2F sine wave... but again any distortion 
will kill this idea.

I wanted to multiply the guitar wave so that I could do pitch to
voltage conversion, without the horrible delay....

PLLs will probably be as bad as any other method here... better than some of 
them...

How about digital pitch shifting ??? (i said a bad word...:)

Let me know if you find the holy grail anyway.

H^)  harry


>From: Toby Graves <carpet8 at mac.com>
>To: <synth-diy at node12b53.a2000.nl>
>Subject: freq mult
>Date: Wed, 29 Nov 2000 01:20:16 +0800
>
>but seriously, folks...
>
>Here's the deal.  What i'm making is a freq. multiplier for guitar that
>multiplies by fractions.  What i need to do is multiply a square wave by 4
>or 8 and then divide it by an odd number or some integer+1/2.  i can get
>nice square waves with 50% duty cycle and multiply them once (thus getting
>me back to original frequency because i divided it by 2 to even the 
>pulses),
>but things fall apart after that because of the different reactance of the
>integrator at different frequencies.  There's got to be a nicer way to
>double frequencies without synthesis or whatever else is used.
>
>
>toby
>

_____________________________________________________________________________________
Get more from the Web.  FREE MSN Explorer download : http://explorer.msn.com