AGC Misunderstanding

  • 1
  • Question
  • Updated 3 years ago
  • Answered
I noticed AGC has no affect when using a program like WJST-X. I switched AGC off and brought the slider all the way to zero. No audio from the speakers but there was no impact on WSJT-X's waterfall nor on its ability to decode. I'm guessing AGC is really just for listening?

While doing the pre-post search on this question I came across two videos from Steve N4LQ. One of the two showed how he adjusts AGC and I think it is really helpful. The other kind of complained about AGC on the Flex and I found that one confusing. The links have been posted here before but I'll repeat them.

Flex-6000 VS FTDX-3000 Demo For AGC-T

Flex-Radio 6000 Series AGC-T Adjustment Procedure
Photo of Kevin N4TT, Elroy

Kevin N4TT, Elroy

  • 924 Posts
  • 266 Reply Likes

Posted 3 years ago

  • 1
Photo of George Molnar, KF2T

George Molnar, KF2T, Elmer

  • 1626 Posts
  • 589 Reply Likes
Perhaps an operator error? AGC makes a profound difference in the waterfall, no matter what setting (high/med/low/off). Could you be adjusting one slice and using another in WSJT-X?
Photo of Kevin N4TT, Elroy

Kevin N4TT, Elroy

  • 924 Posts
  • 266 Reply Likes
Hi George. I'm on the correct slice. When I back down AGC to zero with AGC off I hear no audio which I expect but I'm still seeing traces in WSJT-X's waterfall and I'm still decoding JT65 and JT9. This wasn't expected.

Here's what I see when AGC is set to zero:


When I set AGC to max I'm seeing:


If anything, the trace looks even brighter when AGC is zero. I could still be looking at things wrong. Would you expect the trace on WSJT-X to show no or minimal signals when AGC is set to zero? I also tried when AGC is set to FAST and still don't see a noticeable change on WSJT-S waterfall.
Photo of George Molnar, KF2T

George Molnar, KF2T, Elmer

  • 1626 Posts
  • 589 Reply Likes
Odd one. I duplicated your settings as best I could - works as expected. Are you sure you're using the right DAX channel (Maybe using DAX IQ)?
Photo of KM4CQG

KM4CQG

  • 231 Posts
  • 38 Reply Likes
Kevin

I ran similar test with JT65 and originally came to the same conclusion. However in furthur testing I did in fact see a diffetence when using AGC I suggest you continue to evaluate it.

I would screen shot you my setting but am traveling on business.

Ian
Photo of Kevin N4TT, Elroy

Kevin N4TT, Elroy

  • 924 Posts
  • 266 Reply Likes
HI Ian. Yes, looking closer, the AGC slider does make a difference. Watching the signal strength reported by WSJT-X in the adjustment at the bottom left I can see that when AGC is set to max and signals are present WSJT-X is reporting about 55dB. When AGC is set to minimum the WSJT-X reading drops to -29dB. Big difference no doubt. Oddly, the WSJT-X spectrum display does not change as I would expect it to and copy remains fine at max and min on the AGC slider. At this point it probably isn't a big deal. It works. I was just wondering why I was seeing this behavior. Here's a screen shot of the WSJT-X waterfall. At 15:50 the AGC is off and the slider is at minimum. At 15:51 AGC is off and the slider is at maximum. I wonder if I'm overdriving the waterfall? It looks pretty clean.

Photo of Kevin N4TT, Elroy

Kevin N4TT, Elroy

  • 924 Posts
  • 266 Reply Likes
By the way, sound through the speakers goes from silence (agc slider min) to ear splitting (agc slider max). This is why I don't understand why the WSJT-X waterfall doesn't follow suit. I can understand the SmartSDR waterfall not reacting to the AGC setting (though I wish it would).
Photo of EA4GLI - 8P9EH - Salvador

EA4GLI - 8P9EH - Salvador

  • 1773 Posts
  • 543 Reply Likes
This is where the suggestion to have a visual cue to the AGC on the screen makes a lot of sense. When you are working phone on SSB you can find the sweet spot by ear... but when working digital modes, the horizontal line for AGC would be great.
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Kevin

Pay attention to the time stamps showing when they were published on YouTube.
the first was Feb/2015, the second Aug/2015. The AGC-T and the Flex were both the same, but maybe Steve changed. The second video is demonstrating how the AGC-T works when it's properly set. The first shows how not to set it up. For everything but signals buried down in the noise floor the second video's method works. When a signal is in the noise then you have to accept that noise will be heard. When I've adjusted the AGC-T correctly, and I start hearing noise coming out of the speakers, I know it's time to touch up my AGC-T adjustment, usually dropping it a few units\percent  will reduce the noise or make it inaudible. When I suspect that I'm not hearing a weak signal, I bump up the adjustment until I start to barely hear the band noise again, and maybe then dropping it back down a unit, so that the AGC-T's affect, is riding just barely out of the noise. This of course doesn't work when things like thunderstorms are bouncing the noise floor rapidly. Maybe Flex can add a software defined lightning eliminator button to SmartSDR as a new feature.

Remember also, that DAX controls the audio levels being fed to WSJT-X. thru it's own separate ethernet connection to the radio. The AGC-T processing is being done in the radio, SmartSDR, thru it's own separate ethernet connection is controlling the radio and the AGC-T in the radio, that's being fed thru DAX. WSJT-X has it's own internal AGC running on the audio going to the decoders, while WSJT-X's waterfall is mostly just a pretty picture to look at, and has no effect on what get's decoded. The  can be adjusted separately so you see the colors you like. Were you still seeing some signal on the DAX channels receive level when you had the AGC level zeroed?

Only Flexadio's, have an AGC-T and it's not really comparable to another brands AGC. 

 SDRgadgets

73, Jay - NO5J

Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Kevin
 
Could you take screenshots , including all three windows SmartSDR, DAX and WSJT-X with there settings sliders visible?

It might help us figure it out.
  
 SDRgadgets

73, Jay - NO5J

Photo of Kevin N4TT, Elroy

Kevin N4TT, Elroy

  • 924 Posts
  • 266 Reply Likes
Here ya go Jay. Let me know if you see anything. Kev

Photo of John-K3MA

John-K3MA

  • 102 Posts
  • 29 Reply Likes
And please post them here as I am interested in this topic also.

John - K3MA
Photo of John-K3MA

John-K3MA

  • 102 Posts
  • 29 Reply Likes
You were posting while I was typing.  Thanks for the posts.
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Kevin

I just tried AGC==off and AGC-T==0 and I was still seeing one or two signals in the WSJT-X waterfall but the Audio level temp guage in WSJT-X was still indicating a 30 dB level. DAX was still showing about a quarter scale red bar though. Weird! I can reproduce part of what your seeing, still don't understand it though. Not what I was expecting to see. Let me ponder on this a while. Before I reach any conclusions.

 SDRgadgets

73, Jay - NO5J

(Edited)
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Kevin 

After duplicating the WSJT-X waterfall settings from your screenshot, First I really wish I had your antenna, second I'm still seeing similar, not near as strong, but very nice looking signals in WSJT-X waterfall, with the temp gauge reading 20 dB when signals are present. and DAX channel level==50 AGC-T==0, AGC==off, 6500\Preamp==0. The WSJT-X temp slider==default\midscale\0 gain, With WSJT-X decoding signals all the while, but not in proportion to whats visible in it's waterfall.
Like 4-5 visible, resulting in 12+ decodes

Still Weird!
 
 SDRgadgets

73, Jay - NO5J

(Edited)
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Kevin 

I just noticed you've got Flatten turned on now my waterfall looks like your screenshot. I never turn flatten on, it's not needed, and tends to overdrive the WSJT-X waterfall. But that's just a visual setting.

 SDRgadgets

73, Jay - NO5J

Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Kevin

Another observation, I turned the speakers back on and I'm hearing those signals too.

 SDRgadgets

73, Jay - NO5J

Photo of Kevin N4TT, Elroy

Kevin N4TT, Elroy

  • 924 Posts
  • 266 Reply Likes
Hey Jay... thanks for proving I'm not going nuts! Not a big deal but it wasn't expected so I thought I would ask here. Looking over past posts in the community, AGC is a very popular topic. By the way, did you happen to check out the two videos I linked above? If so, curious what you thought of them. Being a visual person I have to admit they help me a lot more than just reading about knees and such.
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Kevin 

Which version of WSJT-X are you using?
I using v1.7.0-devel r6913, which is a tad old.
I need to recompile a fresh copy.

We need apples to apples for a good comparison.

 SDRgadgets

73, Jay - NO5J

Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Kevin

I re-watched both videos, the second one is better, it sticks pretty close to the adjustment procedure in the Flex user guide. I'd set it 1 or 2 unit's lower though. But that's just me, I tend to listen to signals without any noise. 
Watching and hearing Steve N4LQ, adjust it right help's me to visualize it too.
It makes me trust the settings I use.
Confirmation is good!

 SDRgadgets

73, Jay - NO5J

Photo of Kevin N4TT, Elroy

Kevin N4TT, Elroy

  • 924 Posts
  • 266 Reply Likes
1.6.0 r6263. This appears to be the latest version.

But... would this behavior have anything to do with WSJT-X? It can't do anything unless SmartSDR provides it some audio.
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Kevin
Yeah that's the current stable release. I'm running the Alpha build it yourself version. It might have something to do with what I'm seeing, Joe Taylor likes to change things. There's a new release, very close on the horizon.
Even set like this, DAX is still sending audio, to WSJT-X until you turn off the DAX channel.
 AGC-T==0 does not equal off.
Yeah, I did need to recompile it, it's building v1.7.0-devel r7034 right now.
 SDRgadgets

73, Jay - NO5J

(Edited)
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Kevin 

Ok, I've got the current source code, built and installed now, You just might like the 1.7.0-rc1 release when it happens, and only Joe Taylor knows when that will happen. The one thing I just realized is that when adjusting back to my normal settings in WSJT-X. They have changed. I can run the AGC-T up to 30 \ Fast now. before i usually ran the AGC-T @15 with either Off or Fast set. My Preamp is still set to 0. but a couple of weeks ago AGC-T==15 DAX channel==50 resulted in a 30dB reading on the temp gauge in WSJT-X during the quiet periods. now DAX==50, AGC-T==30 equals @30dB on the gauge. So did somebody calibrate audio levels to better match the slider position? It wasn't me! Maybe last nights thunderstorm did a little magic. 

Something weird "is" going on, But for now I like the 30%==30dB result though.
So my old tweaks are going to need an overhaul. It's a Brand New Day! 
If I'm turning up the AGC-T to get 30dB, why is setting it to 0 still sending nearly enough audio to WSJT-X? Still Weird!
 
 SDRgadgets

73, Jay - NO5J

(Edited)
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
All
 
BTW, WSJT-X v1.7.0-rc1 now seems to work correctly, using DXlabs Commander for the CAT control choice. I hadn't checked recently but it seems to work ok. Something to look forward to, maybe.

 SDRgadgets

73, Jay - NO5J

Photo of Ernest - W4EG

Ernest - W4EG

  • 608 Posts
  • 117 Reply Likes
Jay,
Thanks for the information you have provided on this subject.  I will start applying your settings to WJST-x and now looking forward for K1JT release of the new version. 73, W4EG
Photo of Dan -- KC4GO

Dan -- KC4GO

  • 339 Posts
  • 69 Reply Likes
It is true if you feed the WSPR from I/Q stream it makes no difference to the levels on the WSPR Water fall.  You should be capturing your WSPR via one of the audio DAX channels.
for one thing your filter with is not taken into consideration when you use I/Q streaming. So your reports of S/R with a 2.3 band with will not be correct.  Try changing it to DAX channel 1
and give it a try.
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Dan
I somehow missed the discussion regarding I/Q streaming here.
All of my comments have been regarding WSJT-X fed with regular DAX channels.

Who were you speaking about\to?

 SDRgadgets

73, Jay - NO5J

Photo of Kevin N4TT, Elroy

Kevin N4TT, Elroy

  • 924 Posts
  • 266 Reply Likes
Hi Dan:

I'm pretty sure my audio is set up correctly. When I change the filter bandwidth on SmartSDR I can see the changes on the WSJT-X graph. Hare are the settings I'm using for audio.
Photo of Dan -- KC4GO

Dan -- KC4GO

  • 339 Posts
  • 69 Reply Likes
Jay, Sorry for delay I was under the weather a bit... Here is what I'm seeing here. AGC-T adjustment does work. You can watch the DAX output and see the level go down. I'm using WSJT-X for WSPR and if you adjust the AGC-T you can see the level meter go up and down on it's input, but the DAX level has to go down to almost nothing (so input shows 0 on the WSJT-X software) before the Waterfall stop showing signal.  So I think that AGC-T is working, but that the input to the WSJT-X is so sensitive it displays any small amount of audio all at the same level. In order to make the display go away I had to turn everything down to 0 and it didn't fade away it just ended.  I do find this very interesting.  I'm looking at the Flatten as George described below.  73's Dan
Photo of Dan -- KC4GO

Dan -- KC4GO

  • 339 Posts
  • 69 Reply Likes
As George said turn off the Flatten and the AGC-T will adjust the Waterfall down to nothing.. Its all to do with the WSJT-X software and what it does with Flatten... Thanks George...
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Dan
I'm in total agreement about the sensitivity of WSJT-X's inputs. "Weak Signal" modes ought to be incredibly sensitive. I'm guessing WSJT-X's internal AGC just makes it even more likely to decode signals out of nothing. I mentioned "Flatten" first, a @2-3 days ago. Hello is this thing on, anybody hear me.? 
Maybe, I just didn't say it clearly, or loudly  enough. 

 SDRgadgets

73, Jay - NO5J

Photo of Kevin N4TT, Elroy

Kevin N4TT, Elroy

  • 924 Posts
  • 266 Reply Likes
I heard you Jay. I just ignored you. ;)

Flatten doesn't seem to do anything bad. I have a few hundred (930 according to the log) contacts on JT and I've always had Flatten turned on. My interpretation was (still is) that it basically normalizes the band. It may be that the Flex doesn't require any normalization but it doesn't appear it had any negative impact. The only noticeable affect I see/hear/copy with Flatten on or off is in the WSJT-X waterfall but not in the ability of WSJT-X to decode JT65 or JT9.

Flatten is off now and the AGC action of the Flex is clearly visible in WSJT-X's waterfall particularly when I have it adjusted properly. When I set AGC to slow and bring down the slider as in the video above I can clearly see the background become quiet when signals are present and then noise coming back up for the 12 seconds no signals are present. It's an excellent demonstration of AGC I think. It also makes me wonder, as I mentioned above, why there is no AGC indication/tool on SmartSDR's panafall display. Maybe that needs its own "idea" thread to get attention.

Next time, I won't ignore you!

73,
Kevin K4VD
Photo of EA4GLI - 8P9EH - Salvador

EA4GLI - 8P9EH - Salvador

  • 1773 Posts
  • 543 Reply Likes
I am pretty sure it is Kevin. Al keeps better track of these ideas.
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Kevin 

When you behave like I do, you get used to being ignored.

I seem to remember several feature requests, here for a visual indication of where the AGC action was active.

I just set it by ear, and ride it with my FlexControl. tweak it up until static is barely heard, then backing it down 1-3 units, usually keeps the audio clean. The Maestro AGC-T controls make it even easier to ride.
 
 SDRgadgets

73, Jay - NO5J

Photo of George Molnar, KF2T

George Molnar, KF2T, Elmer

  • 1626 Posts
  • 589 Reply Likes
Official Response
Aha moment - in the WSJT-X waterfall, you have "flatten" checked, which will cause the display to try to compensate for uneven levels in the passband. It's like an AGC for the waterfall. However, decoding doesn't take place from the waterfall values, so the "thermometer" gauge is your best bet - adjust rig levels for 30 dB there. AGC to the noise "knee" and volume to get the level.
Photo of Ken - NM9P

Ken - NM9P, Elmer

  • 4148 Posts
  • 1324 Reply Likes
Correct,
If I understood it when I read it, the "Flatten" switch was created for use with some analog rigs whose filter response dropped off at the upper and lower edges so that they could send and receive at the edges.  If I remember correctly, it boosts not only the receive, but the transmit audio in order to compensate.

It isn't needed with the current Flex DIGI filters.  (BTW, I usually use 5K width for WSJT-X unless I have some really strong signals taking over the AGC)
(Edited)
Photo of Kevin N4TT, Elroy

Kevin N4TT, Elroy

  • 924 Posts
  • 266 Reply Likes
I'm pretty sure transmit is not affected by the Flatten button but I incorrectly thought it only affected the spectrum display. This is not so: "Check Flatten if you want WSJT-X to compensate for a sloping or uneven response across the received passband." I haven't found an explanation of how WSJT-X compensates for an uneven response - whether it applies an assumed formula to compensate for the average radio's response or it actually calculates the response and adjusts dynamically.

I now have it off. And yes! When I vary the AGC I go from blackness (min) to red (max) on the WSJT-X display. So... it looks like Flatten was almost acting like an AGC for WSJT-X? It was keeping the same relative audio signal strength for WSJT-X regardless of the input signal strength. 

Thanks Ken and George. I have to keep reminding myself that different programs implement fixes for the average radio and that the 6500 is not average.
(Edited)
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Kevin

We probably need to ask Joe Taylor, to get a definitive answer about Flatten, and what it affects. 

I've ASSuMed it affects the waterfall window "only" because that's where the control is at, and Joe Taylor has stated numerous times that waterfall controls don't have any effect on decoding.

Long ago, in PowerSDR / SDR1000 days the WSJT-X waterfall noise floor sometimes  appeared slightly tilted horizontally I don't remember which end appeared lower. but selecting Flatten would flatten it out so that all the signal levels across the waterfall would balance. it was as if WSJT-X was applying an internal eq boost to a portion of the audio stream. DAX is of course a digital stream of audio with out any analog eq emphasis/de-emphasis added. DAX streams are Flat to begin with.

The SDR1000 used external sound cards and a rat's nest of audio cables, ground loop eliminating transformers, firewire cables and firewire soundcards, etc. to feed the audio into the PC. where PowerSDR processed it, and then connected it thru VAC to WSJT-X. Sometimes that would result in the audio being slightly colored by all the stuff it passed thru. Flatten was a way of compensating for that but it only affected the waterfall.

The recommendation was that if you didn't like the tilted waterfall then turn Flatten on to flatten it out.

Flatten sometimes not only Balanced the audio it sometimes boosted it, which would then require readjusting the waterfall gain and zero controls, to clean up the waterfall display.

But that was then, The 6xxx models process all that audio digitally, inside the radio, and output it to DAX as digital stream over ethernet. DAX then passes that digital audio stream to WSJT-X. Eliminating all the analog audio portions.
Which keeps the audio unadulterated.

Flatten is still useful for the folks stuck with using analog PC soundcard connections with WSJT-X.
 
 SDRgadgets

73, Jay - NO5J

(Edited)
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
For WSJT-X I tend to run with either AGC==Off or AGC==Fast, with it off My hope is that a really strong JT9 signal doesn't drive the weaker JT65 / JT9 back down into the noise floor, and out of the decoders. I also customized a WSJT-X palette to make it a little simpler to adjust the AGC-T visually. It mimics the color choices of the default SmartSDR waterfall palette. With my palette the background color is Black, and the overload color is Pink with Blue\Green shades for everything in between. when the Noise Floor bounces or twitches, I get a horizontal line all the way across the WSJT-X waterfall. I try to ride the AGC-T control, so that the background between signals stays Black and I rarely see any horizontal lines. This usually works better with AGC set to off. When I'm busy or lazy, I get it close, and then set AGC on Fast. This give's me what I need for a Visual AGC-T adjustment. Better antennas might require a bit more palette tweaking.
   
 SDRgadgets

73, Jay - NO5J

Photo of Kevin N4TT, Elroy

Kevin N4TT, Elroy

  • 924 Posts
  • 266 Reply Likes
There are two AGC Idea threads. I think both have merit and are worth consideration for your vote.

https://community.flexradio.com/flexradio/topics/option-to-automate-the-agc-t-setting
2 years old, 46 votes proposes a setting to automate the AGC. I hadn't considered this idea but as an option (still allowing me to manually set it if I like) it makes sense. I gave it a vote.


https://community.flexradio.com/flexradio/topics/horizontal-line-for-agc-t
4 months old, 39 votes. I like this one a little better. It's more manual but possibly more instructional. It proposes adding a line to indicate the AGC-T level. I voted for this one also.

Kev
Photo of Al / NN4ZZ

Al / NN4ZZ

  • 1806 Posts
  • 641 Reply Likes
Hi Kevin,
Adding both the line and offering the automated option would be great. Automating the setting would likely be very effective for many of us that don't always remember to tweak the setting as conditions change, or as we rotate our beams, change bands , etc. And having the visual indicator on the screen would serve as a reminder, an aid to manually adjusting , and feedback on how well the automated option is working.

Although these requests are ranked highly and have been on the list for several years, I don't think we will get any official feedback until/unless it is part of a release. While both are a SMOP, they don't seem to be a high priority compared to the other big ticket items. Maybe in V2, or V3, or V4.

Regards, Al / NN4ZZ
Photo of EA4GLI - 8P9EH - Salvador

EA4GLI - 8P9EH - Salvador

  • 1773 Posts
  • 543 Reply Likes
Al, do you happen to have a list of the current list of requested features? I am always amazed at how you keep track of it!
Photo of Al / NN4ZZ

Al / NN4ZZ

  • 1806 Posts
  • 641 Reply Likes
Hi Salvador,
Here is my trick.....

On the lower right side of the main page is a place to "Share Your Expertise."    By default it shows "Problems."   You can change it to show "Ideas" and order them by the number of VOTES.  This is how I have it set up and it is sticky and stays that way.   I find that more useful than list the problems.     Here are some screen shots to show the steps to set it up.


Look for this section on the lower right of the main community page......notice in my case it is listing the top 3 ideas.  
 



Click on "VIEW all.......conversations."   You will see all of the posts with the search selection box at the top of the list.  All of the posts will be listed based on the criteria you select, which again is for problems by default. 





Click on "REFINE" to set up the search parameters you prefer.   Here are the ones I use to see active ideas ordered by number of votes. 



Unless you come back and change the parameters later, you will now see the ideas listed here when ever you log in.  I



Regards, Al / NN4ZZ  
al (at) nn4zz (dot) com
6700 - HW.................. V 1.8.4.84
SSDR / DAX / CAT...... V 1.8.4.168
Win10
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Thanks for sharing that Al, I've often wondered how you keep such a good track of Feature requests, over time.
 
 SDRgadgets

73, Jay - NO5J

Photo of EA4GLI - 8P9EH - Salvador

EA4GLI - 8P9EH - Salvador

  • 1773 Posts
  • 543 Reply Likes
LOL He should have kept it a secret :) to keep us in awe of his memory!!!

Thanks Al!

I foresee a bunch of new VOTES in the near future! 
(Edited)
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Kevin 

I voted for the second idea, about the horizontal line. Automated as an option might be Ok too. My guess is a horizontal line AGC-T indicator might be both easier to code into SmartSDR, Maestro, etc, so it might get done faster, and likely needed as part of the Automated AGC-T option, If that ever gets added to the client software. Visually it seems easy enough to do, but letting the software decide on it's own, how close to ride the noise floor will need a lot more magic. But as long as it's functional magic, go ahead and cast the spell. But leave me a manual option, too.

 SDRgadgets

73, Jay - NO5J

Photo of EA4GLI - 8P9EH - Salvador

EA4GLI - 8P9EH - Salvador

  • 1773 Posts
  • 543 Reply Likes
The horizontal line serves two purposes in my opinion, one the obvious of providing us with a direct visual indication of AGC... the other is educating us in what the AGC does.... it took me a while of trial error and reading here to understand AGC.... when I did it made the Flex a so much better radio. That is the first thing I always explain to new users of Flexes when I help them by Teamviewer.
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9154 Posts
  • 3506 Reply Likes
FYI: These follow-up comments regarding the AGC visual indicator Idea post will not be considered in the evaluation process since they were not associated with the Idea post.
Photo of EA4GLI - 8P9EH - Salvador

EA4GLI - 8P9EH - Salvador

  • 1773 Posts
  • 543 Reply Likes
??
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9154 Posts
  • 3506 Reply Likes
When people start commenting about the Idea post in the problem or question post, theses comments are not associated with the Idea post, so when our product managers review the Idea posts, theses comments are never seen.
Photo of EA4GLI - 8P9EH - Salvador

EA4GLI - 8P9EH - Salvador

  • 1773 Posts
  • 543 Reply Likes
Crystal clear now. Thanks!
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9154 Posts
  • 3506 Reply Likes
De nada.
Photo of Kevin N4TT, Elroy

Kevin N4TT, Elroy

  • 924 Posts
  • 266 Reply Likes
Official Response
Just a quick follow-up. Many thanks for helping me to understand what was going on. I've been working with AGC setting recently and think I can properly set it for any mode. I've noticed, when set right, strong signals can pretty much wipe out noise but leave weaker signals intact. 


In this screen shot of WSJT-X displaying JT-65 the first minute (lower portion) there's a strong signal around 600 Hz. Background noise dropped off almost totally and the other signals within the display sounded and decoded fine.

You can see the next 12 seconds with no signal. Noise came up on the display and in the audio output. Then the next minute starts with no particularly strong signals. Noise dropped a bit but again, copy of all signals was perfect. Relative volume in each case was easy listening. The strong signal was hardly any louder than any other signal.

I'm wondering (and I'll experiment), if the second minute trace indicates I should drop my AGC threshold down a little bit more.

On CW with a decently strong signal I get the same results. The noise just drops out and I'm left with a pleasant sounding CW note. I practiced this on the QRP Fox Hunt last Thursday and got both foxes within 25 minutes. Generally I only get one or no fox.

I kind of wonder why this AGC action wouldn't be made visible in the panafall of SmartSDR. I suppose it has to do with where in the signal flow WJST and SmartSDR is getting its data to display.

Again... thanks for the help. This has been fun.

73,
Kevin 
Photo of Jay / NO5J

Jay / NO5J

  • 1647 Posts
  • 249 Reply Likes
Kevin

Have you tried using the "Negative" WSJT-X waterfall palette as an aid to seeing where the AGC-T effect is riding above the noise floor. It can help. Watch the quiet periods and lower the AGC-T until you aren't seeing a noise increase when the signals are gone. You dont have to get the background totally white, but you don't want to see the quiet periods getting darker either. Which would indicate that you can lower it another percentage point if they are.

 SDRgadgets

73, Jay - NO5J

(Edited)