Ah. That's interesting. I will repeat my tests over the weekend and 
report back.

Mike
====

> Not true - the way its reported.
> 
> S/N is always (in amateur circles) referred to 2.5kHz.  So although
> OP32 has 6dB lower bandwidth than OP8 (0.12Hz compared with 0.5Hz
> signalling)   by teh time it is normalised to 2500Hz the figures end
> up the same
> 
> 
> Quite why, since WSJT introduced the idea, we chose to refer to S/N in
> 2500Hz is lost in the shrouds of time.  Just becasuse its about the
> same as an SSB filter Why couldn't we have done what the rest of the
> RF world does and use a 1Hz reference?
> 
> Andy
> G4JNT
> 
> 
> > Reports on my signals from PA0A, PA0AM, G4WGT and F6CNI (between 257
> > and 465km) show variable results. The slower speed never produced a
> > worse S/N, though sometimes it was exactly the same. Usually the
> > result was between 4 and 9dB improvement with the slower speed.
> >
> > Intuitively, I would assume a 6dB improvement (though I am no
> > digital expert), but occasionally better iif QRM is present as I
> > would expect this to affect the faster signal more. Is this a
> > reasonable
> >
>