> >... I might take SED more seriously if it was shown
> > how increasing noise would improve detection efficiency. Usually
> > detectors are cooled to reduce noise and improve efficiency.
> There was an article (in SciAm? or New Scientist?) fairly recently about
> how adding noise can improve detection efficiency.
seems to have alot in common with some of the methods used in
mu-law is one data compression technique that introduces "noise"
in a sense to pull out correlated values and reject the largely
non-correlated ones for the simple reason to reduce the amount
of transmitted information; which is not exactly the same as squeezed
light, but I seem to recall some DSP methods that introduce noise
for the purpose of increasing some desirable component of signal
strength. Maybe someone in comp.dsp can help ?