[Elecraft] eHam.net forum message

[Tom W8JI]

>FWIW I have a Perseus and two other receivers which both use H-Mode mixers,
> one of which uses a version of PA3AKE's H-Mode front end.
>
> With the Perseus used as a stand alone receiver, I have not experienced 
> this
> "noise problem" in the presence of mild or quite strong static or manmade
> pulse QRN, UNLESS I wind up the Perseus' NR  too far, then indeed there is
> some tunnel/ reverberation effect, which can at times be annoying.

I think people are talking about two different things here.

What I am talking about, and what I understand some others to be talking 
about, is a problem with DSP systems processing noise floor signal, 
especially when the noise is a bit rough, without adding artifacts that make 
copy or quality worse.

This is not anything to do with noise blanker use, which would have limited 
use with a weak noise-floor signal.

Processing with a noise blanker punches a hole or reduces gain for fast 
level increases, and noise blanking in a DSP is a different process. I'm not 
sure what goes on now, but years ago when I worked on some DSP NR systems it 
was necessary to compare many samples of signal and look for repeating 
patterns that would be a steady signal and try to subtract out things that 
did not repeat. This always caused a delay or echo or distortion, depending 
on the aggressiveness of the noise reduction.

I fully expect any kind of blanker or processor to screw up the signal 
characteristics. There is no way around it. :-) My concern is when the 
processing is off and there should be a faithful accurate reproduction of 
signal, even when it is below noise floor.

I have no problem at all with how the K3 blanker and noise reduction works, 
and actually IMO it does a good job for me. I just don't think any DSP radio 
is as good as a full analog system when signal levels are near noise levels, 
especially when the noise floor is a bit rough.

73 Tom