hi peter,

i guess butterworth should be good because of maximally flat passband with reasonable phase response/groupdelay. If attenuation is adequate i would definately use it as they are much easier to design on ICs (unit circle poles makes for easy choice of well matched components R,gm,C etc,)

the thing that would make me thing abit more about inverse cheby is that it has transmission zeros which can make the design a bit tougher. i guess if you are doing bandpass, this doesn;t matter, but for low pass, i think its easier to do a slightly higher order butterworth than an inv cheby.

cheers.
aw