Anf - totally useless

Giulio

Dear all, I have tried almost everything but I don't hear any difference between ANF on or off. It doesn't change anything. (same thing with WNB) Is it only my impression? I came from a Ft1000 MKV... And I know how a NF works. Thank you for your help. Giulio IZ3EAW

Geir/LA5ZO

Did a comparison using my ELAD FDM-2 RX / SDRRadioV3 software.
Both used on USB bandwidth variable with carrier in the center.

On Flex6300/Smart SDR a carrier is attenuated (not meassured value) but still well audible.
On the ELAD combo it disapears completely.

So there is definately room for improvement here.

Geir/LA5ZO

Walt - KZ1F

When I was in second grade, for lunch we had table monitors.As it was elementary school, these table monitors were only sixth graders. One day one of the table monitors convinced me to take my little milk carton, flip it upside down, and tap on it three times, explaining that if I did that the milk would taste like chocolate milk. Power of suggestion, it actually did! For me ANF merely badly distorts the signal, WNB is useless although I believe I saw it work twice. To be honest, given the power of suggestion, I can't honestly be sure it actually did work, and NB is also useless. It is hoped that when FRS gets done with tablets and linear they will return to noise mitigation. Power of suggestion!

David

This topic has been discussed a lot here. It seems the WNB is for a unique environment and most people don't experience a significant value in using. Tim has commented that if you don't experience any value with WNB enabled at any adjustment level you should not use since it adds an unnecessary load on the radio. I also have found ANF not to improve intelligibility of a signal or reduce the static white noise in way that is helpful. I had hopes that it would be similar to BHI/GAP's Hear-It  product that filter at the audio level vs RF and does a great job of removing the white noise similar to noise cancelling headphones and letting the voice come through more clearly.

The AGC and NR are the most effective in my environment for noise quieting, listening fatigue reduction, improved audio intelligibility. The NB helps too depending on what is going on locally.

As Walt commented hopefully noise mitigation at the RF signal and Audio signal will

Dan -- KC4GO

@ David:"I also have found ANF not to improve intelligibility of a signal or reduce the static white noise in way that is helpful"  The ANF is not a Noise filter it's a Notch Filter so it will not reduce static or white noise, but will notch out a carrier. (Still needs some work I use TNF 90% of the time.

Gene - K3GC

When there is the inevitable tuner-upper in the band a quick click on ANF and he is gone.  I consider that quite useful. 

Lawrence Gray

I find the the ANF does notch out a carrier, but also introduces significant audio distortion.  I rarely use it--I use TNF regularly.  The WNB does nothing for me.  I'm not too sure why effort would be spent on WNB if it is only for special cases, while the ANF and NB are still not up to the standards of many non-sdr rigs.  The NR works reasonably well and I use it most of the time.

mikeatthebeach .

WNB works great for me

https://www.youtube.com/watch?v=9YPfYFx_2I8

73 mike

Mark_WS7M

Does it still taste like chocolate milk Walt!?  :-)

Kevin

I've watched the video linked above and can't reproduce the dramatic results. It might be because the noise level at my location is a lot less than what is displayed on the video (-120 dB vs -100 dB). There are times when some machine in the neighborhood turns on and my noise level jumps up but WNB doesn't seem to help and other noise abatement tools seem to contribute as much to distortion as noise reduction.

As for the AGC adjustment procedure in the video (something I'm always trying to understand on the Flex), is this really a good way to adjust it? Most instructions I see say to max AGC-T then start bringing it down until just before you notice a drop in noise level (the knee I think). Dropping it by 3 dB using Flex Meter as in the video seems to indicate I'd actually be cutting the received audio in half, both noise and signal.

I would like to see more concentration by Flex in the area of noise abatement (including AGC-T). I believe there can be improvements both in their effectiveness and in GUI aids to adjustment. I think Flex would benefit by putting as much emphasis on ragchewing and DXing as they do on contesting. I also think Flex would be well-served by producing instructional videos concerning the radio's features. While the various manuals seem pretty good a video tutorial solidifies the technique and provides Flex a chance to showcase techniques.

73,
Kevin K4VD

Mark_WS7M

I just tested on my 6300 and frankly only the NB and NR do anything.  In fact the WNB raises my noise level just a bit. 

I setup things very much like "mikeatthebeach" and when I turn on WNB things go up.

David

Yes, I keep forgetting that it in a Notch Filter. It would great if they had popup help or expanded description when you hover over an item to remind the function and use of items. Either way I have found using ANF to have more of a negative to the signal quality reducing the intelligibility than removing offending carriers. The TNF is the real value.

Walt - KZ1F

Mark, that actually happened, believe it or not. I hope that table monitor got beat up once he went into Jr High... the other thing he did was tell this girl, I don't recall if I did actually like her, that I liked her. I was so embarrassed (remember this is 2nd grade) that I hid under the table. It's odd for two reasons, one, I remember it and 2, it was like I was in the middle of a call center. It was as if there were a thousand conversations going on at once in my head. I believe, now, one of them was demanding I get up as I was sent to the Principals office in tears. I suspect my mother was called in to 'bail' me out.

Walt - KZ1F

@David, yes, I agree with your prior assessment, AGC-T and NR work fine.

mikeatthebeach .

Do not set WNB at 100%, set at  95% and use the Flexmeter to set level by -3dB 
and works great on man-made impulse noise 

Norm - W7CK

I am in total agreement that the noise mitigation needs more work.  The TNF is excellent though.

On a good note, the other evening I was on 160 meters and there were several storms between 100-200 miles from me.  The static was pretty bad so for grins I decided to try the NR but found that it didn't help much.  I next tried the NB and couldn't believe how much it reduced the static from the storms.  A few days later I was once again on 160, same situation and tried the NB.  No matter how I set things up, it made no difference at all.  Go figure!  I can only guess that the difference might have been the frequency and duration of the static crashes that had changed slightly.

Norm - W7CK

Mark_WS7M

Hi Mike,

ANY amount of WNB causes an increase.  Not sure why.

Giulio

No notch filter on a 2500€ radio? Come on are you joking? Sorry but this radio, actually is not for contesters. NR? it is an artifact, only distortion. WNB is TOTALLY USELESS, maybe there was an empty space in a button. I hope some new enlightening updates. Best 73, Giulio IZ3EAW

mikeatthebeach .

On my Flex6700 the WNB only works on Fast Impulse Noise for Power Line Zap,
Brush Motors that have a Wideband Noise Signature much greater that the Filter Passband.

If I set WNB < 92% and in conjunction with with NB on,
it cleans up a Waterfall that is polluted with Noise.

The Flex WNB is the Waterfall "Wideband" Noise Blanker mathematically removes 
the Noise from the Visual Waterfall. The K3S/P3 and IC-7851 can not do this.

If the Fast Impulse Noise changes in Magnitude, WNB blinks which means it is 
recomputing the Transfer Function of the Fast Noise Pulse in Real Time and applies
the new Mathematical Noise Impulse Function to subtract it from the Signal that has the Noise.

Signal just seem to jump out of the Noise Floor, with a reduction in the Noise Floor
You will not see that on PSDR, and ANAN series, or any other HF radio. This is a 
feature only seen in Commercial Hi-Rel radio's for Harsh RFI Environments 
that have nearby interference caused by Fast Impulse Noise.

Tried a Flex6300 with SSDR 1.8.4, along with a Flex6500 and finally with my Flex6700 with 1.9.3.12 SSDR and all work great

See my interference go away with WNB wit the link below with the Flex6300
https://www.youtube.com/watch?v=z-IRH9KSBbQ

The Fast Impulse Noise is man-made, it can not work on Atmosheric Band Noise,
lightning Crash Noise.

If you have Two Slices, like a Flex6700, like to see a DNB, Diversity Noise 
Blanker to Null Out Noise that does not follow a Fast Impulse Noise Signature similar to the that on the ANAN200D with a Polar/Mag/Phase Display. Maybe an algorithm to auto-adjust between the two to null out interference similar to WNB could be implemented by FlexRadio for the Flex6700.

The ANF is totally different than the WNB feature and yes ANF needs work

Norm - W7CK

It would have been nice if you were actually tuned to a signal that was buried in the noise and then showed how the noise reduction not only lowers the noise floor but also made the signal copyable and without introducing a lot of distortion.

Tim - W4TME

We will be addressing the audio distortion in ANF with the upcoming SmartSDR v1.10 release.

Tim - W4TME

Mark - Because you do not have an well correlated noise for it to work on.

Kevin

Can you put "well correlated noise" into ham radio terms? I found this in a Google search and not sure if it applies:

Correlation is somewhat similar: if your wife starts taking sewing lessons at 11 pm twice a week, and at the same time your best friend is in business meetings, you may think that the two events share some properties.

Tim - W4TME

Noise is a very complex waveform and can be generalized as having various degrees of randomness or periodicity. For example, Ignition type noise is primarily periodic in nature.  Band noise is very random.

If the noise is primarily periodic, it is referred to as correlated.  If the frequency of the periodicity is fairly stable, it is well correlated. 

Kevin

Thanks! So examples of correlated noise (including fast rise time pulse-type noise) include:

• ignition noise
• electric motor noise
• power line noise
• switching power supply hash (noise)?
• carriers

Radio controls pertinent to correlated noise would be WNB, NB and ANF.

Uncorrelated could be considered white (or pink) noise and include:

• solar storms
• lightning storms
• atmospheric noise

Radio controls pertinent to uncorrelated noise would be NR and AGC.

Is this a reasonable summary?

One thing I've found with WNB is that I can't just flip the slider around and expect to see immediate results. WNB takes seconds to actually begin working. According to the manual "When the WNB detects a significant change in signal level or bandwidth, it attempts to normalize the slider value. During this time, blanking will be momentarily bypassed, and the WNB indicator in the Panadapter will start to flash (blink)." This seems to me that it would cause distortion by momentarily bypassing blanking. It seems to me (as a non-engineer) that WNB should remain where it is while it attempts to normalize. Otherwise, wouldn't you be adding the effects of the significant change in signal level or bandwidth with blanking flipping on and off?

[fixed formatting]

Jeffrey Kerber, N3VE

Tim, thank you!

Jeff N3VE

mikeatthebeach .

I get no distortion with WNB turned on with strong S9 * signals 
Mike

W9OY has video's with WNB On versus off, with signals and also no Distortion

https://www.youtube.com/watch?v=9YPfYFx_2I8

David

The distortion is with ANF not WNB. Tim says is being addressed in the next release.

The comments on WNB is that it seems to be effective with an unique environment and many have commented they don't experience any value in the use of it.

EA4GLI

ANF works for me in silencing carriers but I agree that it adds distortion. I find it works better when I have the slider closer to 0, a value of 2 to 10 is plenty in many cases.

I used it on and off this weekend during the contest and it did work. But I quickly turned it off once the tuner decided to go elsewhere.

Any improvements in ANF and any other noise mitigating tools will definitely be very welcome by many users. I am very glad to hear they are being considered for 1.10.

Ken - NM9P

ANF works, sort of, depending upon how it is adjusted, and the band conditions.  It does tend to add some distortion and to "pump" with strong signals, making the offending carrier come and go along with the strength of the modulating signal.

But I don't get to chuffed about it because there is an alternate solution:

Currently, the best way i have found to notch out carriers and "tuners" is to use the TNF (Tracking Notch Filter).  It is easiest to adjust when Zoom is high.  Right-click on the spot and it will give you the option to open a TNF, then just drag it to fine tune it.  You can adjust the width by dragging up or down on the filter line, and adjust the depth by right-clicking on the filter and selecting Normal, Deep, or Very Deep.  You can also make it "remember" the position, or let it be erased when you close the program.

The advantages of TNF are: 1) it doesn't add distortion, 2) you can have more than one in the same passband, 3) You can see where it is going to filter.  4) If it is a persistent carrier, it can remember the same spot always.

The disadvantages are:  1) that it isn't automatic, 2) it doesn't track moving carriers, 3) you must remember to delete it when it isn't needed.

I wish that there was a button that, when pressed, would scan the audio passband for the offending carrier and then slap a TNF on it.  

I also am glad to hear that some of the DSP functions will receive additional attention in an upcoming release.

Ken - NM9P

Here is a short assessment of the current DSP noise abatement routines....

AGC-T is your first line of defense against regular band noise.  Learning how to adjust it carefully solves a lot of problems.  Start with the control all the way to the right and adjust to the left until you hear the noise just begin to drop.  This is the "knee" or "sweet spot"  that lets the noise drop off a little, but weaker signals will begin to pop out of the background noise.  You will need to adjust this up or down a little whenever you change bandwidth, or if band conditions change.  Sometimes even when turning your antenna, which makes your noise profile change.  You may want to set it a little above or below the "knee" depending upon band conditions, or operating situation.  (e.g. in a rag chew session on 80, you may want it to be set for less noise to provide more quiet operation with the stronger signals.  When looking for weak ones on 10 or 6 meters, you may want it a little above the knee to allow more access to weaker signals.)  Experience will teach you what you need, if you listen carefully to the effects as you adjust it.   I usually use FAST for CW, Digital, & AM modes, and MEDIUM or SLOW for SSB in order to control the background noise.  Listening to SSB with FAST AGC gets tiring for me as the noise always returns to high in between syllables....

NR (Noise Reduction) - works pretty well for normal hiss and random white-noise type static.  You don't usually need to turn it up past 5-10.  It sometimes helps if you readjust AGC-T a little bit afterwards.  I also sometimes have improved response if I also tweak RX EQ a little to make up a little bit of highs that are lost with NR...But not too much, or it defeats the purpose of the NR routine.

WNB (Wideband Noise Blanker) - Works for the whole panadapter.  And it works to different degrees depending upon noise type and band conditions.  Sometimes the result is spectacular, sometimes imperceptable.  It is designed for use on "highly correlated" noise patterns - regular patterns of high impulse noise like ignition noise, AC line arcing, other discharge noise.  But it isn't designed to work on wideband hash, storm static, and other random types of noise.  You can see its effect as you adjust the control slider.  But you must remember 1) it takes a second or two for the WNB to "learn" the pattern and take effect.  2) If you turn it up too high, it can be fooled by sharp modulation peaks on strong stations sometimes far removed in frequency.  3) After you see it take effect on the panadapter, you may want to Tweak AGC-T to bring the newly-revealed weak signals out of the noise floor.  4) If the noise pattern or band conditions change, you may need to readjust WNB level and/or AGC-T settings.

NB (This is a slice-level Noise Blanker which works only on the slice for which it is enabled.)  Works on sharp impulse noise.  But it works for impulse noise that is not as well correlated as WNB is designed for.  Use WNB FIRST and adjust it to see if there is any reduction.  But keep WNB to the point that you are not getting any pumping of the panadapter display.  THEN add NB if needed and adjust until the noise pulses are affected.  If you set it too high, then strong nearby stations will make it distort, just like any other noise blanker.   NB is NOT designed for general white-noise and random "hash" or atmospherics, lightning static, etc.

ANF & TNF - Notch filters that I have described elsewhere in this post.

APF (Audio Peak Filter) - Useful to peak the audio frequency for CW operation.  The control adjusts the sharpness of the filter.  Response is more dramatic when using the lower latency versions of the CW filter, and with the wider filter settings.  But it does help to bring the intended CW signal out a little bit.  I seldom use more than half-scale adjustment of the control.  Again...you may want to tweak the AGC-T after adjusting this control.

All of these features have their usefulness, but require some time to learn how to adjust them carefully and how they interact with each other.  It is also important to learn what type of interference each one is designed to fight.  Using the wrong tool for the job will lead to great frustration.  

There is room for improvement for some of these functions.  I am glad to hear that it is scheduled for some attention in the not-too-far future.