K3 NR

[Bill K9YEQ]

I changed my audio equalizer from where it was and that makes a huge
difference for me. I tried to remove some of the noise with the equalizer
and that caused the sound to be very hollow.


73,

Bill
K9YEQ
K2 #35; KX1 #35; K3 #1744; mini mods
ATS-3B


-----Original Message-----
From: [hidden email]
[mailto:[hidden email]] On Behalf Of N2TK

 Yeah, I agree with what you are saying. I like the NB on CW, but on SSB
even at F1-1 it sounds somewhat like being in a tunnel. I would like to see
a NB setting that is less aggressive and that doesn't change the audio as
much.
Also it would be nice if the audio didn't reduce in volume when turning on
the NB.
 
Will any of the AGC settings like SLP, THR, PLS, HLD, DCY have an impact on
what SSB sounds like with the NB?

N2TK, Tony


______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[John E. Reiser]

I'm running version 3.27/2.25 on K3 #384.  I live in a high noise suburb of
NYC.  When I tap the NR button, the noise just about disappears, and the
signal remains, SSB or CW.  Magic!

The difference is more noticeable on SSB, of course, because I usually have
the pass band set to 500 Hz for CW.

My favorite NR setting is F1-3.  I seldom change from that setting anymore.
I don't use EQ.

I operate 90% CW and RTTY.  When I do operate SSB, I'm not too concerned
about audio frequency response.  If I can understand what the other guy is
saying, I conclude that I've got him tuned in OK.

When I want hi-fi, I get out the IPod_ hi!  But, don't be mad at me.  That's
just me.  I respect all differing opinions.  Maybe I'm just not as
discerning as some folks.

73,  John, W2GW


----- Original Message -----
From: "N2TK" <[hidden email]>
To: <[hidden email]>
Sent: Friday, August 28, 2009 6:10 PM
Subject: Re: [Elecraft] K3 NR

> Yeah, I agree with what you are saying. I like the NB on CW, but on SSB
> even at F1-1 it sounds somewhat like being in a tunnel. I would like to
> see
> a NB setting that is less aggressive and that doesn't change the audio as
> much.
> Also it would be nice if the audio didn't reduce in volume when turning on
> the NB.
>
> Will any of the AGC settings like SLP, THR, PLS, HLD, DCY have an impact
> on
> what SSB sounds like with the NB?
>
> N2TK, Tony
>
> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of Paul - WW2PT
> Sent: Wednesday, August 26, 2009 1:17 PM
> To: [hidden email]
> Subject: Re: [Elecraft] K3 NR
>
>
> I found the NR in v3.25 to be excellent to my ears and a big improvement
> to
> previous versions, but I'm having a lot of trouble finding a NR setting
> with
> v3.27 that I'm comfortable with -- everything sounds like it's coming
> through a tunnel, even the x-1 settings. I generally prefer minimal DPS
> noise reduction for SSB and have rarely used the x-4 settings. I generally
> stuck to 4-1 or 4-2 in previous versions and rode the RF Gain to further
> reduce background noise when needed. Just can't find a happy setting with
> v3.27.  It might sound OK on one signal but awful on another, resulting in
> far more dial tweaking than I usually like to do.
>
> RX EQ is flat as has been suggested, just doesn't help much. The tunnel
> effect is a bit less bothersome when using headphones, but I generally
> prefer using a pair of powered desktop speakers with AFX (Bin) turned on.
> Haven't tried NR in CW mode yet, my comments apply only to SSB.
>
> I've reverted to 3.25 and then back to 3.27 just to make sure I'm hearing
> what I think I'm hearing. I think I can live with the new NR settings, but
> sure would like to have them *in addition to* the 3.25 filters which were
> damn near perfect for my ears. Maybe some additional less aggressive NR
> settings (5-1 through 5-4, 6-1 through 6-4, etc.) will make everyone
> happy?
>
> Just more free advice... ;-)
>
> 73,
> Paul WW2PT
>
>
>
> Lyle Johnson wrote:
>>
>>
>> The 3.27 NR F3-1 is *identical* to 3.25 NR F1-4.
>> The 3.27 NR F3-2 is *identical* to 3.25 NR F2-4.
>> The 3.27 NR F3-3 is *identical* to 3.25 NR F3-4.
>> The 3.27 NR F3-4 is *identical* to 3.25 NR F4-4.
>>
>> The 3.27 NR F1- is "lighter" than the 3.25 NR.
>> The 3.27 NR F2- is a little lighter than the 3.25 NR.
>> The 3.27 NR F4- is "heavier" than the 3.25 NR.
>>
>> I am an SSB op, not a CW op, and the NR now works much better for me.
>>
>>
>
> --
> View this message in context:
> http://n2.nabble.com/K3-NR-tp3515659p3518140.html
> Sent from the Elecraft mailing list archive at Nabble.com.
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Don Wilhelm-4]

Tony,

Did you mean "NR" when you stated "NB"?  I find no audio reduction when
turning on the NB unless I have the IF NB cranked up to maximum.

I first thought the new NR was reducing the audio, but I found it was
not so when one is tuned to a signal. The noise reduction is so
effective that in my case it almost behaves like a squelch because there
is not band noise at all (my noise level is not usually very high
anyway) - so there is a natural tendency to turn up the audio gain
because there is no noise to hear - but when a signal is present, the
audio is then LOUD.
As I have stated in prior posts, I *do* find a bit of AF reduction on
SSB if I have the NR set to F3-x or F4-x, but I find NR on SSB quite
effective at the F1-x or F2-x settings.

73,
Don W3FPR

N2TK wrote:
> Also it would be nice if the audio didn't reduce in volume when turning on
> the NB.
>  
>
>  
>
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[David Woolley (E.L)]

The main source of published research on real time noise reduction of
audible signals seems to be the hearing aid industry.  I just came
across this article suggesting that hearing aid noise reduction
strategies make people think that the noise is less sever, but don't
actually make the signal any more intelligible.

Trends in Amplification, Volume 10, No. 2, June 2006: Acceptance of
Background Noise, Mueller et al.
<http://tia.sagepub.com/cgi/reprint/10/2/83.pdf>

Obviously being perceptually more desirable is good for marketing, which
is largely based on feeling, rather than fact, and it may also be of
value in that, over longer periods, reduced fatigue improves
performance.  In the hearing aid world, it increases compliance (the
number of people who actually wear their aids).

On the other hand, if anyone comes up with a algorithm that actually
increases intelligibility in the "cocktail party" context, I think the
hearing aid industry would love to know!

--
David Woolley
"we do not overly restrict the subject matter on the list, and we
encourage postings on a wide range of amateur radio related topics"
List Guidelines <http://www.elecraft.com/elecraft_list_guidelines.htm>
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Bill W4ZV]

David Woolley (E.L) wrote

The main source of published research on real time noise reduction of
audible signals seems to be the hearing aid industry.  I just came
across this article suggesting that hearing aid noise reduction
strategies make people think that the noise is less sever, but don't
actually make the signal any more intelligible.

Trends in Amplification, Volume 10, No. 2, June 2006: Acceptance of
Background Noise, Mueller et al.
<http://tia.sagepub.com/cgi/reprint/10/2/83.pdf>

That's an interesting article David!  It agrees with my feelings over the many years I've tried various types of noise reduction.  Sometimes I think we're fooled by the level changes introduced by NR, when in fact we could probably do as well simply by turning AF Gain down a little.  

A few years ago when NQ5T and I both had Orions, I challenged Grant to measure the actual S/N with NR on and off.  At narrow bandwidths, there was no difference in measured S/N.  Of course this was for CW where NR simply builds a narrow filter around a discrete signal.  

I believe a filter is a filter is a filter...whether crystal, DSP or NR.  There is no magic.

73,  Bill

[Bill W4ZV]

Bill W4ZV wrote

David Woolley (E.L) wrote

The main source of published research on real time noise reduction of
audible signals seems to be the hearing aid industry.  I just came
across this article suggesting that hearing aid noise reduction
strategies make people think that the noise is less sever, but don't
actually make the signal any more intelligible.

Trends in Amplification, Volume 10, No. 2, June 2006: Acceptance of
Background Noise, Mueller et al.
<http://tia.sagepub.com/cgi/reprint/10/2/83.pdf>

That's an interesting article David!  It agrees with my feelings over the many years I've tried various types of noise reduction.  Sometimes I think we're fooled by the level changes introduced by NR, when in fact we could probably do as well simply by turning AF Gain down a little.  

A few years ago when NQ5T and I both had Orions, I challenged Grant to measure the actual S/N with NR on and off.  At narrow bandwidths, there was no difference in measured S/N.  Of course this was for CW where NR simply builds a narrow filter around a discrete signal.  

This was for Orion II,  but for anyone interested, here's the summary by Grant NQ5T:

http://lists.contesting.com/archives//html/TenTec/2006-03/msg01118.html

On March 25, 2006, NQ5T wrote:

[TenTec] Orion II NR Performance Measurements

I won't bore you with the setup here, but will be happy to provide details
to anyone who is interested.  Consistent results were obtained by two
independent methods: (1) graphical computation of SNNR, and (2) spectral
analysis software that directly computes an estimate of SNR.

The results are as follows (LCW, 1000Hz spot tone, NR=9)  

BW=3000Hz: SNNR improves by approx 1dB with NR=9.

BW=500Hz:  SNNR degrades by approx 2dB with NR=9.

There is improvement at 3 Khz bandwidth, but it's negligible.  At 500 Hz
(and anything below that as well) you're better off without NR at all.  The
distortion created by NR at any bandwidth in both CW tone and SSB voice is
very unpleasant compared to typical noise reduction products.

Even without having a v1 Orion to compare with I'm basically moving from the
"uncertain" bench to the "put it back the way it was" bench.

Grant/NQ5T

[Vic K2VCO]

Bill W4ZV wrote:

[about NR and SNR]

I've been watching this thread with interest. For a few years and several different
radios, I've been repeating the same test:

1) Tune around with a bandwidth of about 400 Hz. Find a weak CW signal close to the noise
which I can't copy 100%.

2) Try all the possible techniques to improve intelligibility, looking for the ones that
improve the percentage of copy.

Here is what I've learned:

1) If there is a kind of noise that a NB will reduce, that helps.

2) If there is not too much noise, reducing the bandwidth further helps. I sometimes go
down to 50 Hz. on the K3. But on a noisy band this makes it worse.

3) AFX doesn't matter one way or the other.

4) Dual-diversity reception (polarization diversity) *REALLY* helps. IMHO this is one of
the K3's greatest features.

5) NR doesn't help with the weakest signals.

6) I can't decide what pitch is best. Usually I use around 500 Hz. but that's because I
like the sound of it.
--
73,
Vic, K2VCO
Fresno CA
http://www.qsl.net/k2vco
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Guy, K2AV]

That is really very good reading, if highly technical. Helps to make a
little index card with all the abbreviations used.

One thing clearly suggested by the article is that immense efforts to tailor
the K3 NR based upon complaints/praise, or measurable improvement of
intelligibility, are, stated simply, doomed to failure absent a breakthrough
invention in DNR that would make the patent holder rich.

I'd just as soon see Lyle get rich as someone I don't know, so keep at it
:>)

I have done some amount of testing on NR. My earlier conclusions, fuzzy
compared to the article, are similar.  For a weak CW signal in the noise, NR
is more likely to DEGRADE intelligibility. The article attributes this kind
of thing to a masking effect.

NR makes listening to signals in noise more tolerable, until the signal
becomes masked, then it gets in the way.

Using NR, all signals will degrade vs. no NR at and below the masking point.
Note that for contesting, these are the QSO's which make the winners. Those
who can hear these difficult signals will get stuff others can't.

There are similar issues with what some mistakenly call ringing. The narrow
CW bandwidths in a noisy band situation will produce a constant "narrow
noise" which competes for attention if the listener finds it irritating. It
seems to be particularly irritating for those who tend to mentally
demodulate the signal in a wide bandwidth.

My own explanation for those who hear this way is that the wide bandwidth
allows them TO IDENTIFY THE NOISE in their mind, assigning it a
pseudo-diversity, ignore it and thus separate the CW.  I surmise that for
them, in the narrow situation, THE CW SOUNDS LIKE THE NOISE.

Identifying the CW by throwing away the noise seems the case in diversity
RX, where a fairly real sounding pseudo-spatial diversity spreads around
non-discrete signals, but leaves the desired signal focused. Diversity RX is
the only setup where narrow bandwidths have a way to spread the noise
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[N2TK]

Typo. Tnx Don for picking that up. Yes, I mean NR. I like the way the NB
works.

I do see a reduction in audio when turning on the NR with SSB. As an
example, I usually have the AF control around 9-9:30 position using an
external speaker or my Heil headphones. When I turn on the NR I have to use
around 10:30-11:00 position for the same audio. This is using F1-1. As I
increase (Fx-y) X from one to 4 the audio level reduces. As I increase Y
from one to 4 the audio level also decreases. So if I go to F4-4 I have to
crank the AF control fully CW and I still am not at the audio level I was
with the NR OFF and the AF control at 9:30.  Also I can definitely pull out
weak SSB signals better with the NR OFF. Right now the noise level is low,
so not sure what impact that is having.
With CW I do not need to change the AF level. I do like the way the NR works
on CW.
Typically I have the RF gain control fully CW or close to this position.
When running I don't want to be fiddling with knobs or get my ears blown
off.
Both $1435 and #311 behave the same.

I still like the Icom 756Pro III's NR better. I find I can leave the NR ON
at a low level to just take the background noise away without causing issues
or artifacts with received signals.

Because  of the  varying comments on how the latest NR is behaving I am
wondering if there are any other parameters, especially with the AGC that
would account for these wide variations?

Work in progress.

73,
N2TK, Tony


-----Original Message-----
From: Don Wilhelm [mailto:[hidden email]]
Sent: Friday, August 28, 2009 7:48 PM
To: N2TK
Cc: [hidden email]
Subject: Re: [Elecraft] K3 NR

Tony,

Did you mean "NR" when you stated "NB"?  I find no audio reduction when
turning on the NB unless I have the IF NB cranked up to maximum.

I first thought the new NR was reducing the audio, but I found it was
not so when one is tuned to a signal. The noise reduction is so
effective that in my case it almost behaves like a squelch because there
is not band noise at all (my noise level is not usually very high
anyway) - so there is a natural tendency to turn up the audio gain
because there is no noise to hear - but when a signal is present, the
audio is then LOUD.
As I have stated in prior posts, I *do* find a bit of AF reduction on
SSB if I have the NR set to F3-x or F4-x, but I find NR on SSB quite
effective at the F1-x or F2-x settings.

73,
Don W3FPR

N2TK wrote:
> Also it would be nice if the audio didn't reduce in volume when turning on
> the NB.
>  
>
>  
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Bob Cunnings NW8L]

I'm running 3.27/2.25 on K3 #569, mostly CW. No EQ, default AGC settings.

The NR is so improved (for CW at least) with this version that for the
first time I'm tempted to actually use it!. Normally I simply reduce
DSP BW to overcome noise when working the weaker CW signals, but like
to use a relatively wide BW when tuning around or copying the stronger
signals. Yes, a filter is a filter but the NR artifacts are so minimal
now that I can leave the BW at 800 Hz or so (I have the 1kHz roofing
filter) and just turn on the NR ( I also like F1-3) if the band noise
becomes too great. Give it a couple of seconds to converge and the
noise vanishes yet all but the very weakest signals in the passband
remain, with minimal distortion.

Good work!

Bob NW8L

On Fri, Aug 28, 2009 at 5:26 PM, John E. Reiser<[hidden email]> wrote:

> I'm running version 3.27/2.25 on K3 #384.  I live in a high noise suburb of
> NYC.  When I tap the NR button, the noise just about disappears, and the
> signal remains, SSB or CW.  Magic!
>
> The difference is more noticeable on SSB, of course, because I usually have
> the pass band set to 500 Hz for CW.
>
> My favorite NR setting is F1-3.  I seldom change from that setting anymore.
> I don't use EQ.
>
> I operate 90% CW and RTTY.  When I do operate SSB, I'm not too concerned
> about audio frequency response.  If I can understand what the other guy is
> saying, I conclude that I've got him tuned in OK.
>
> When I want hi-fi, I get out the IPod_ hi!  But, don't be mad at me.  That's
> just me.  I respect all differing opinions.  Maybe I'm just not as
> discerning as some folks.
>
> 73,  John, W2GW
>
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[.k8dd.-2]

Tony .....

Yeah, I couldn't have said it better about the NR as it is now - you
nailed it.  See what the next few DSP f/w's have!

I had an IC 746 (not PRO) - the NR on it was great .... and it was just
that - Noise Reduction, pure and simple.  We took it to V47 about 6
years ago and it was almost as good as the K2/100 w/o the KDSP2.  Wish
I'd have never sold the 746!

73    Hank    K8DD



N2TK wrote:

> I do see a reduction in audio when turning on the NR with SSB. As an
> example, I usually have the AF control around 9-9:30 position using an
> external speaker or my Heil headphones. When I turn on the NR I have to use
> around 10:30-11:00 position for the same audio. This is using F1-1. As I
> increase (Fx-y) X from one to 4 the audio level reduces. As I increase Y
> from one to 4 the audio level also decreases. So if I go to F4-4 I have to
> crank the AF control fully CW and I still am not at the audio level I was
> with the NR OFF and the AF control at 9:30.  Also I can definitely pull out
> weak SSB signals better with the NR OFF. Right now the noise level is low,
> so not sure what impact that is having.
> With CW I do not need to change the AF level. I do like the way the NR works
> on CW.
> Typically I have the RF gain control fully CW or close to this position.
> When running I don't want to be fiddling with knobs or get my ears blown
> off.
> Both $1435 and #311 behave the same.
>
> I still like the Icom 756Pro III's NR better. I find I can leave the NR ON
> at a low level to just take the background noise away without causing issues
> or artifacts with received signals.
>
> Because  of the  varying comments on how the latest NR is behaving I am
> wondering if there are any other parameters, especially with the AGC that
> would account for these wide variations?
>
> Work in progress.
>
> 73,
> N2TK, Tony
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html