Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site nyit.UUCP
Path: utzoo!linus!philabs!nyit!jpl
From: jpl@nyit.UUCP (John Lewis)
Newsgroups: net.music.synth
Subject: more about FM
Message-ID: <137@nyit.UUCP>
Date: Sat, 5-Oct-85 02:26:05 EDT
Article-I.D.: nyit.137
Posted: Sat Oct  5 02:26:05 1985
Date-Received: Sun, 6-Oct-85 08:40:15 EDT
Distribution: net
Organization: NYIT Computer Graphics Lab., Old Westbury, N.Y.
Lines: 21

The relation underlying FM is (in the most basic form)
	cos(I*sin(m)) = sum J<n>(I)*cos(n*m)
J<n> being a Bessel function fo order n.  The significance of this
is that one can generate an infinite Fourier series weighted by Bessel
functions using only two oscillators, and the effective bandwidth 
(to a greater extent than its character) can be changed by changing
the parameter I.

I think the idea that FM sounds are quite good got started in the
early days of computer music when it was one of the only affordable
methods known.  To my ears this idea is way off--Fm covers the
range of electronic bell sounds, from doorbells to menacing underwater
bells... quite suitable for spacey muzak, or the carefully crafted
bleeps of some serious, modern composers.   The popularity of the
Dx synthesizers suggests otherwise, but it may have more to do with
what you get for the money than the actual quality of the sound
(and of course one doesn't have to count on the timbre as
much if other instruments are available)

	zilla
	NYIT graphics lab