Forgive me as this has probably been discussed already, but I've seen some 
examples (mr pinkston's rachet, for instance) where the panning is scaled 
with a sqrt function, like:
;p4 = pan (0=left; .5=center; 1=right)
iright=sqrt(p4)
ileft=sqrt(1-p4)
....
out asig*ileft,asig*iright

instead of just using:
iright=p4
ileft=1-p4
....
out asig*ileft,asig*iright

So, what is the reason for using this "geometric" panning, instead of the 
linear one?
This also seems to increase the gain as we approach the center position, 
because for a value of .5 we get a global gain of sqrt(.5)+sqrt(.5) = 1.4.. 
while using the linear way, the added pan factor is always 1
thanks in advance

pedro