Loading ...
Sorry, an error occurred while loading the content.

Re: [linuxham] fldigi waterfall shape

Expand Messages
  • w1hkj
    Here are some side-by-side comparisons of the gmfsk-hkj and fldigi waterfall displays: gmfsk - scale / speed processing load fldigi - scale / speed processing
    Message 1 of 12 , Sep 13, 2006
    • 0 Attachment
      Here are some side-by-side comparisons of the gmfsk-hkj and fldigi waterfall displays:

      gmfsk - scale / speed
      processing load
      fldigi - scale / speed
      processing load
      gmfsk waterfall
      fldigi waterfall
      x4 / Normal

      Xfree - 19%
      gMFSK - 4 %
      x1 / NORM

      Xfree - 22 %
      fldigi - 6 %


      x2 / Normal

      Xfree - 18
      gMFSK - 5 %
      x2 / NORM

      Xfree - 23 %
      fldigi - 6 %


      x1 / Normal

      Xfree - 21%
      gMFSK - 5 %
      x4 / NORM

      Xfree - 21 %
      fldigi - 5 %


      x1 / Double

      Xfree - 21%
      gMFSK - 7 %
      x1 / FAST

      Xfree - 33 %
      fldigi - 11%


      x2 / Double

      Xfree - 21%
      gMFSK - 7 %
      x2 / FAST

      Xfree - 35 %
      fldigi - 12 %


      x4 / Double

      Xfree - 21%
      gMFSK - 7 %
      x4 / FAST

      Xfree - 33 %
      fldigi - 9 %



      The sampling rates and drop rates are the same in both programs.  The primary difference is in the fft processing.

      gMFSK computes a different fast fourier transform size depending on the scale (x1/x2/x4).  The FFT routine used is an N point complex fourier transform.  The frequency resolution and span for the three zoom levels are:
      • x1- 3.906... Hz / 1024 points / 4000 Hz
      • x2 - 1.953... Hz / 2048 points / 4000 Hz
      • x3 - 0.9756... Hz / 4096 points / 4000 Hz
      The waterfall display blanks and restarts with each change in FFT length.

      fldigi computes the same size fast fourier transform for all zoom levels and then decimates in frequency as the display zoom is changed.  The FFT routine is a 2N real fast fourier transform
      • x1 - 0.9756... Hz x 4 = 3.906 / 4096 points / 4000 Hz
      • x2 - 0.9756... Hz x 2 = 1.953 / 4096 points / 4000 Hz
      • x4 - 0.9756... Hz x 1 = 0.9756 / 4096 points /4000 Hz
      Decimating in frequency after the FFT allows the waterfall display to transparently change zoom level with no blanking and restarting of the display.  The fuzziness at the x1 and to a lesser degree the x2 zoom level is due to the way that I am decimating (sub-sampling) the fft output.

      I will try some different decimation routines to see if the x1 and x2 zoom levels can be made a crisper in appearance.

      You will note from the above that the highest resolution displays are nearly identical in appearance.


      Hope this answers the question on waterfall appearance.

      Dave, W1HKJ

    • w1hkj
      I checked you web page Leigh but did not see any reference to the Visual CQ. Can you provide me your code? Dave
      Message 2 of 12 , Sep 13, 2006
      • 0 Attachment
        I checked you web page Leigh but did not see any reference to the Visual
        CQ. Can you provide me your code?

        Dave
      • w1hkj
        I checked your web page Leigh but did not see any reference to the Visual CQ. Can you provide me your code? Dave, W1HKJ
        Message 3 of 12 , Sep 13, 2006
        • 0 Attachment
          I checked your web page Leigh but did not see any reference to the
          Visual CQ. Can you provide me your code?

          Dave, W1HKJ
        • Tim Seed
          Dave, fldigi installed and running very well Red Hat FC5 in the ever sunny Middle East - great work. Will FeldHell be making an appearance in the near future ?
          Message 4 of 12 , Sep 13, 2006
          • 0 Attachment
            Dave,
            fldigi installed and running very well Red Hat FC5 in the
            ever sunny Middle East - great work.

            Will FeldHell be making an appearance in the near future ?

            Good decode in all modes heard - don's suppose you could send me
            some propagation - the bands are very dead here at the moment.

            73s
            Tim
            A45WG
          • Rick Kunath
            ... I really like this behavior, i.e. not refreshing the waterfall and losing data with every resolution change. So often you might hear something out of the
            Message 5 of 12 , Sep 13, 2006
            • 0 Attachment
              w1hkj wrote:

              > Decimating in frequency after the FFT allows the waterfall display to
              > transparently change zoom level with no blanking and restarting of the
              > display.

              I really like this behavior, i.e. not refreshing the waterfall and
              losing data with every resolution change. So often you might hear
              something out of the display area and now the data is there in a
              resolution change. You don't have to wait for a refresh from the top.

              Rick Kunath, k9ao
            • w1hkj
              Glad to hear it works well on RH/FC-5 Tim. FeldHell and Throb are both on the drawing board. We could all use some help from the ionsphere. Seems like we have
              Message 6 of 12 , Sep 13, 2006
              • 0 Attachment
                Glad to hear it works well on RH/FC-5 Tim.

                FeldHell and Throb are both on the drawing board.

                We could all use some help from the ionsphere. Seems like we have a
                great big rf sponge in the sky.

                Dave
              • Leigh L Klotz, Jr.
                Maybe you need a LPF before the decimation?
                Message 7 of 12 , Sep 13, 2006
                • 0 Attachment
                  Maybe you need a LPF before the decimation?
                  > I will try some different decimation routines to see if the x1 and x2
                  > zoom levels can be made a crisper in appearance.
                  >
                  > You will note from the above that the highest resolution displays are
                  > nearly identical in appearance.
                Your message has been successfully submitted and would be delivered to recipients shortly.