NR within AM/SAM mode

k3Tim

SSDR 1.4 is working like a charm. The 6500 is powered up 24/7 and no hiccups!  The change to SAM mode giving more volume (comparable to AM / SSB) is nice touch.

Question: Why is the NR (DSP) filtering disabled for AM/SAM?  
Is there any possibility of NR for the SAM mode in the future releases?
The NB is possible but not helping the line noise I hear in early evening.  Pre-dawn all is quiet therefore no DSP filtering required.

Thanks for any insight (hope).

_..--
 k3Tim

WA2SQQ

I did not realize it was now disabled in AM. 1.38 was enabled as I had power line noise that was somewhat reduced by using it. About the time we got 1.4, my local power company found and fixed the problem, so I've not had the need to use NR.

Since it was available, hopefully it can be reactivated.

Doug Hall

While I don't know the reason NR is not available for AM I can hazard a guess, and (maybe) offer a little insight into the workings of NR. Based on my NCAA tournament picks, however, my guess might be completely wrong.

Based on the way it sounds I suspect that the algorithm for NR is an adaptive FIR filter, and the method used to adapt the coefficients is based on the Least Mean Squares (LMS) algorithm. At the risk of oversimplifying things, the adaptive filter looks for correlated information and builds filters around it. A CW signal (a single tone) is an example of a correlated signal where adaptive filters work well. A speech signal is not highly correlated, but it's more correlated than noise, so with the proper choice of the adaptation constant and the choice of decay constants (to "unbuild" the filter when the signal changes) we can use an adaptive filter for speech signals as well.

My experience, however, has been that an adaptive filter never improves the fidelity of a speech signal, and I would even argue that it also does not improve the intelligibility of a speech signal. It can improve the "listenability" of a single sideband signal by suppressing the noise between words or syllables, although one could argue that proper choice of AGC-T and AGC settings can also do that.

When I choose AM I do so because it usually offers higher fidelity (due to greater bandwidth) and because the carrier keeps the channel quiet between words and syllables. If an AM signal is sufficiently noisy that you need NR then you have probably lost the benefit of AM, and maybe even the benefit of synchronous AM. So I would never likely use NR with AM, and didn't even know it wasn't available. But then I don't use it on SSB either for the reasons listed above.

My 2 cents worth. I have no idea if the above method is what Flex implemented, but that's what it sounds like to me. Won't hurt my feelings to discover that I am wrong because even in that case I can learn something new, and anyway I'm used to it :-)

73,
Doug K4DSP

N7AIG

Very nicely stated! Kudos!!

73 de Dave, N7AIG