I don't think it will save any bandwidth. ie if you only monitor one
it is turning on and off at the data rate. The other channel (polarization) is
also turning on and off at the the data rate, but the data is complimented.
Each detector still needs the bandwidth to soak up the changes....
However...it is an interesting idea. In a digital sense it may be like
RTTY, ie the output from the two detectors (filters).
On the otherhand, there may be the technical problem of isolation
between channels of say for example 20db. I think this would mean
that the dynamic range of the recovered information, may be limited
to 20 db.
Anyhow it's interesting...I have to think about it some more.....
Over and out....Don wa8qzz