The KiwiSDR 2 online store is open for orders! Please visit kiwisdr.nz

Continuous automatic scaling of waterfall display?

jksjks
edited February 2020 in KiwiSDR Discussion
People are always posting pictures showing the Kiwi waterfall in a horrible washed-out state due to improper scaling of the colormap via adjustment of the "WF min/max" controls (or use of the "Auto Scale" button). I.e. nothing but greens and yellows with no differentiation across the display.

What are people's experiences with other SDRs / SDR software? Do they have continuous automatic scaling of the waterfall display so this is not an issue? Or some other mechanism like a better colormap scheme?

Comments

  • edited February 2020
    All of the SDR software I've used, HDSDR, SDR#, SDR Console, GQRX, and others, use manual waterfall adjustment. I generally find the KiwiSDR software auto scale to "hot" for my visual liking. I prefer a "cooler" waterfall using maximums between 50 and 30, and minimums between 140 and 90, depending on the band conditions. This usually gives a red carrier, green/yellow sidebands, blue background with minimal noise sparkles. So maybe, beauty is in the eye of the user, eh?
    Manually Scaled to my viewing preference:



    Auto Scaled:



    Ron
    KA7U
  • @Ron: This is how I manually adjust my displays as well. I agree that the Auto Scale button doesn't work optimally (except possibly at zoom 0 where the low noise floor > 25 MHz helps to set the display noise floor low enough to get a dark blue color).

    The other problem is that I think most people don't even know adjusting the WF min/max is a thing. They just leave it at the default values, either the factory ones or the values set by the admin.
  • I agree with Ron
  • edited February 2020
    Yes, I had noted this.
    My current workaround is to tap shift+"s" once to set auto scale, then tap shift+"w" between 5 and 10 times quickly until satisfied with the appearance of the waterfall. I'm usually tending to go 10 steps so suit my aesthetic requirements.
    When I'm using the sqrt=2 parameter in the URL I'm picking 8 steps.

    I also think that people are forgetting to redo the auto scaling when changing zoom level, and that can have a less than desired effect on the view.
    I wonder, would a solution to this be to mark the current scaling numbers relative to what the auto-scale values are for that view, and when changing zoom level that the same delta between auto and user be applied to a new autoscale in the new zoom level?

    As an example, the user is in WF6, and the autoscale gives -49 and -101, but the user has chosen -55 and -93; that's a delta of -6 and +8.
    Then the user zooms to WF9, and the autoscale for that view when the zoom has completed is -43 and -110; add the delta of -6 and +8 to give -49 and -102
    That seems to work for the scenarios I've tried manually here at least.

    John, to better answer your question:
    My own experience is that the waterfall in SDRConsole has been the most aesthetically pleasing for me, and the "auto" button there does a pretty good job of getting the range generally right. Their choice of colour scheme really does help that though with the navy/blue/grey/white/yellow scheme is nice for contrast for radio signals. The refresh rate of SDRConsole waterfall has also really helped my visualisation of noise generators in the area, but that's a rabbit-hole for another time.
    I've done a fair amount of personal work in visualisation of astronomical images and using false-colour for gaining more info from monochromatic images I've taken, and I've found that the most useful is something similar to the SDRConsole scheme, with nice contrast and colour only really used for the off-scale areas.

    Hope that helps.
    -Cathal.
    Powernumpty
  • I aim for a waterfall color mapping similar to KA7U's. Autoscaling is tricky, along with auto-anything else. It can get you close, then touch things up with some manual adjustments. What I end up with depends a lot on what particular signals I am looking at, it varies.
  • Something I came across recently - Google group massages standard false colour spectrum gradients for decent improvements in differentiation:
    https://ai.googleblog.com/2019/08/turbo-improved-rainbow-colormap-for.html

    The comparisons shown appear to be pretty good, and appears to make sense.

    If there's a suggested choice for a change to the colours currently used for the waterfall, it's entirely possible that a lot of the reason for this thread may go away if there's more contrast available to the viewer without having to massage the waterfall limits too much.
Sign In or Register to comment.