[CM] clm formant gain bug

[Bill Schottstaedt]

I noticed yesterday (actually I woke up in the middle of the night 
thinking "oh !%#@#") that the CLM formant generator's gain
calculation was a mixture of the Samson box version from around
1978 and the shiny new one that Ken Steiglitz came up with
a few years later.  I don't know how it got that messed up, 
or why I never noticed, but since I had to make an incompatible
change anyway, I took the opportunity to change the order of
the arguments in make-formant, remove mus-formant-radius,
and so on -- enough changes  that I've called it CLM 4.  There's
a note in sndclm.html about how to get the old behavior from 
the new generator, and snd9.scm has old-make-formant.