Dream NR:
Very similar to Ron's thoughts.
I would love if the O's had both DE and a leaky adaptive predictor filter and we
could chose one or the other or both and be able to vary each separately. As Ken
Brown mentioned, I think any dramtic "pull the signal from below the noise
floor" implementation would end up taking too much time (latency) and DST
horsepower to be practical (but I don't know this definitively).
I had a short e-mail exchange with Doug Smith recently, because I was wondering
if the leaky part of the leaky adaptive predictor could be increased with a
corresponding increase in adaption rate, so one gets both spectral filtering and
DE noise gate action from a single implementation. He felt you could, but only
to a point. Here's part of what he said
"...the speed of adaptation is also proportional to the
amplitude of the input signal. In other words, you have
to make it more aggressive when the signal is small--
just when you need noise reduction the most. That can be
a bit tricky because of certain overflow problems and other
instabilities in software."
So it's the same old story, DSP isn't magic.
But I sure like being able to reduce the noise on most signals! I don't
experience the noise fatigue I had with my old Heath equipment.
Lin
WB1AIW
Ron Castro wrote:
Downward-expansion only for SSB, and DE plus bandpass filtering for CW. I
have found that in SSB, anything that cuts into the high-end of the bandpass
reduces the ability to copy.
In my experience, NR seems to work for CW, but I have never heard an NR
scheme on any radio that helped me copy an SSB signal that I couldn't copy
without NR. I have a feeling that anything that would work better on SSB
would require massive processing power that is not now available or
affordable, and/or introduce unacceptable latency issues.
Ron N6AHA
----- Original Message -----
From: "Douglas Shock" <douglas.shock@gmail.com>
To: "Discussion of Ten-Tec Equipment" <tentec@contesting.com>
Sent: Thursday, May 04, 2006 11:26 AM
Subject: Re: [TenTec] 2.033
Let me pose a question to you guys:
Regarding the NR function in the Orion
What approach would be best suited if TenTec allowed you to choose your
poison?
On 5/4/06, Gary Hoffman <ghoffman@spacetech.com> wrote:
Grant,
Actually I've now gotten several off-line requests for just such a
reference. I used to have several books and magazines that talked about
it.
I've told everyone that I will go and search through the "stuff" I've got
in
the house and see if I can find any of those documents. I used to have a
lot of these references.
Regards,
Gary, AA2IZ
----- Original Message -----
From: "Grant Youngman" <nq5t@comcast.net>
To: "'Discussion of Ten-Tec Equipment'" <tentec@contesting.com>
Sent: Thursday, May 04, 2006 2:03 PM
Subject: Re: [TenTec] 2.033
As to what I am trying to prove - nothing. This is just a
way in which DSP can be done, and it does not work by using
any sort of spectral information or filters. I've noticed
that many here are not familiar with this approach.
It actually sounds familiar as you describe it, and I'm reasonably sure
I've
come across some material that may be what you're describing. Of
course,
today I can't find it.
Any chance you can point to a web source with some information? I
purchased
the book that Jerry referenced in an earlier email -- maybe it will be
discussed in there.
Thanks ... Grant/NQ5T
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
|