Weighted average

DrTeeth

The 'weighted average' button in the display settings does not seem to do much at all in 1.4.0 whilst it had a noticeable effect in 1.3.8. I have also noticed that I have to have the AVG slider at 100 in 1.4.0 when I had it at 50 in 1.3.8 to get my desired display. I'm afraid I do not recall if I had weighted average set or not in 1.3.8.

Has anybody else noticed this?

Steve N4LQ

I tried the weighted average and it makes the cw band look like popcorn popping. I don't really know what it's for.

W4YXU

As best I can tell the button just didn't get "hooked up" in the software and appears to be stuck in "weighted average" mode.

I prefer the plain average and normally ran 1.3.8 at AVG 50, FPS 10.

Jim
W4YXU

Ken - NM9P

It seems to me that the Weighted Average has more effect when using a higher averaging number (slider to the right).  What it does when ON is to allow the rising peaks to move up faster while maintaining the average decay time of the trace.  When turned OFF, both the rise time and decay time seem to be averaged more equally.  When using lower averaging numbers, there isn't as much difference because fewer samples are being averaged to compare to.

I prefer the control ON  in order to see more of the frequency peaks in a SSB signal, and it allows weaker CW signals to peak higher and stand out better from the background.  But I use a setting of 30 for software V. 1.3.8 and have moved to between 50 & 60 for V. 1.4.  (With FPS of 15.)
That gives the nicest display for my.  Your own preference may be far different.  That's the beauty of the adjustments.  NICE!

Ken - NM9P

Steve-N5AC

The weighted average function is an average that favors (weights) an increasing signal more than a decreasing one.  For a dramatic demonstration, use a signal generator and place a tone in the panadapter and then tune the generator.  With weighted average on, the peak will remain almost constant.  With weighted average off and averaging turned up, the signal will have a decreased amplitude and will have more of a smearing effect with reduced amplitude as it is moved across the panadapter.  Without weighted average, it takes just as long for a signal that is suddenly turned on to rise in the panadapter as it does for them to decay (fall).  

The purpose of weighted average is to more quickly signal where something that has just arrived in the band is and to keep it present in the display as the modulation stops and starts.  There are few visual artifacts to this also.  Averaging without the weighting looks more natural to most people. With the averaging, it tends to look like ants crawling up a wall and falling off the wall (sorry, it's hard to unsee this for some).  The net-net is that there are people on both sides of the weighted average discussion -- some like it one way while others like it the other way.  It is just another tool in your toolbox to represent what is going on in the spectrum.  Neither is more "accurate" necessarily, but simply show the signals in a different way.

While we're on the topic, part of what we did for v1.4 is to completely decouple the panadapter and the waterfall controls.  Prior to v1.4, the waterfall used the base data from the panadapter to create the waterfall display.  Altering the speed of the panadapter display also had an impact on the waterfall.  Now each has their own source data with independent controls (thanks again to the big FPGA in your radio).  As a result of this change, there is no doubt that things will behave differently, but this change gives us more flexibility to do some cool things in the future and is a good long-term change.  We worked to make the controls mostly transparent so that similar settings would get you the same basic result as best as we could.