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

JT9 decode r3496 vs r3673

Expand Messages
  • k3wyc
    I stuck with r3496 for a long while as other versions were fielded. I decided to upgrade to r3673 and have been using it for a bit over a week. Since I
    Message 1 of 9 , Feb 16, 2014
    • 0 Attachment

      I stuck with r3496 for a long while as other versions were fielded.  I decided to "upgrade" to r3673 and have been using it for a bit over a week.  Since I started using the new version I have had many failed JT9 decodes for signals that looked ok on the waterfall.  The number of failed decodes seems to have increased significantly since I left r3496.


      Was there any change in the JT9 decoder between r3496 and r3673?


      Anyone else seeing JT9 decode issues with r3673?


      73,


      Andy k3wyc.

    • chas cartmel
      Sorry no name to address you. I have had no issues with 3673. You don’t say what machine spec you are running it on so I wonder if this version needs more
      Message 2 of 9 , Feb 17, 2014
      • 0 Attachment

        Sorry no name to address you.

         

        I have had no issues with 3673. You don’t say what machine spec you are running it on so I wonder if this version needs more CPU due to the flattening algorithm and the PC can’t keep up? Have flattening switched off here.

         

        Hope it’s sorted.

         

        73

        Charlie

        www.G4EST.me.uk

         

        From: wsjtgroup@yahoogroups.com [mailto:wsjtgroup@yahoogroups.com] On Behalf Of a.durbin@...
        Sent: 17 February 2014 01:24
        To: wsjtgroup@yahoogroups.com
        Subject: [wsjtgroup] JT9 decode r3496 vs r3673

         

         

        I stuck with r3496 for a long while as other versions were fielded.  I decided to "upgrade" to r3673 and have been using it for a bit over a week.  Since I started using the new version I have had many failed JT9 decodes for signals that looked ok on the waterfall.  The number of failed decodes seems to have increased significantly since I left r3496.

         

        Was there any change in the JT9 decoder between r3496 and r3673?

         

        Anyone else seeing JT9 decode issues with r3673?

         

        73,

         

        Andy k3wyc.

      • chas cartmel
        Sorry no name to address you. I have had no issues with 3673. You don’t say what machine spec you are running it on so I wonder if this version needs more
        Message 3 of 9 , Feb 17, 2014
        • 0 Attachment

          Sorry no name to address you.

           

          I have had no issues with 3673. You don’t say what machine spec you are running it on so I wonder if this version needs more CPU due to the flattening algorithm and the PC can’t keep up? Have flattening switched off here.

           

          Hope it’s sorted.

           

          73

          Charlie

          www.G4EST.me.uk

           

          From: wsjtgroup@yahoogroups.com [mailto:wsjtgroup@yahoogroups.com] On Behalf Of a.durbin@...
          Sent: 17 February 2014 01:24
          To: wsjtgroup@yahoogroups.com
          Subject: [wsjtgroup] JT9 decode r3496 vs r3673

           

           

          I stuck with r3496 for a long while as other versions were fielded.  I decided to "upgrade" to r3673 and have been using it for a bit over a week.  Since I started using the new version I have had many failed JT9 decodes for signals that looked ok on the waterfall.  The number of failed decodes seems to have increased significantly since I left r3496.

           

          Was there any change in the JT9 decoder between r3496 and r3673?

           

          Anyone else seeing JT9 decode issues with r3673?

           

          73,

           

          Andy k3wyc.

        • k3wyc
          I signed my post with my name and call so don t understand the no name to address you comment. I have no shortage of CPU power (2.3 gig quad core A10). In
          Message 4 of 9 , Feb 17, 2014
          • 0 Attachment

            I signed my post with my name and call so don't understand the "no name to address you"  comment.

            I have no shortage of CPU power (2.3 gig quad core A10).  In any event, to the best of my knowledge, CPU power only influences the speed of the decode and not the probability of a successful decode.


            I started wsjt modes with JT65-HF and then moved to wsjt-x to be able to use JT9.   It has always been my understanding that the waterfall display parameters had no influence on decode, and that the signal received by the decoder was completely independent of any waterfall tailoring.  Did something change?


            73,


            Andy k3wyc




          • Sam Morgan
            it s a yahoo groups thing if you receive and answer any yahoo group email from your email program, such as thunderbird and have a signature line like this for
            Message 5 of 9 , Feb 17, 2014
            • 0 Attachment
              it's a yahoo groups thing
              if you receive and answer any yahoo group email from your email program,
              such as thunderbird
              and have a signature line
              like this for example:
              "--
              GB & 73
              K5OAI
              Sam Morgan"

              it turns out, the -- and all below the --
              is not included when yahoo groups
              posts to their site and/or send out your email

              I started getting around it by simply typing in my signature line
              without the -- and under my comment

              here:
              K5OAI Sam

              On 2/17/2014 6:50 AM, a.durbin@... wrote:
              >
              > I signed my post with my name and call so don't understand the "no name
              > to address you" comment.
              >

              just for grins, I'll include it here also to see if it gets through

              GB & 73
              K5OAI
              Sam Morgan
            • chas cartmel
              Apologies Andy For some reason I did not see it – must have scrolled off the screen due to the added line-feeds I receive. Me being lazy and not scrolling
              Message 6 of 9 , Feb 17, 2014
              • 0 Attachment

                Apologies Andy

                For some reason I did not see it – must have scrolled off the screen due to the added line-feeds I receive. Me being lazy and not scrolling down – mea  culpa.

                 

                I agree that the waterfall settings have no influence on the decoding, however the processing of the flatten will take up CPU and on marginal systems may drain CPU resource. Your system though should cope OK I would have thought. I have noticed at times that I get (for example) 4 decodes appearing in JTAlert-X with a few seconds break after which several others appear. Total time elapsed around 5 seconds even 4 12 or so decodes, and I assume that this is due to a delay in writing the information to the file from which JTAlert gets this data from, my logical assumption is that should this time then stretch into the next TX period then decoding may be affected.

                 

                Are you using split mode or just JT9? And have you tried both?

                Is it OK on the sample audio files?

                Is the audio source set to DVD quality?

                 

                Again clutching at straws perhaps, but possible lines of investigation.

                 

                73

                Charlie

                www.G4EST.me.uk

                 

                From: wsjtgroup@yahoogroups.com [mailto:wsjtgroup@yahoogroups.com] On Behalf Of a.durbin@...
                Sent: 17 February 2014 12:50
                To: wsjtgroup@yahoogroups.com
                Subject: RE: [wsjtgroup] JT9 decode r3496 vs r3673

                 

                 

                I signed my post with my name and call so don't understand the "no name to address you"  comment.

                I have no shortage of CPU power (2.3 gig quad core A10).  In any event, to the best of my knowledge, CPU power only influences the speed of the decode and not the probability of a successful decode.

                 

                I started wsjt modes with JT65-HF and then moved to wsjt-x to be able to use JT9.   It has always been my understanding that the waterfall display parameters had no influence on decode, and that the signal received by the decoder was completely independent of any waterfall tailoring.  Did something change?

                 

                73,

                 

                Andy k3wyc

                 

                 

                 

              • Joe Taylor
                Hi Andy and all, ... The decoder incorporates some necessary trade-offs that balance decoding performance against speed. There s a straightforward way to
                Message 7 of 9 , Feb 17, 2014
                • 0 Attachment
                  Hi Andy and all,

                  K3WYC wrote:
                  > I stuck with r3496 for a long while as other versions
                  > were fielded. I decided to "upgrade" to r3673 and
                  > have been using it for a bit over a week. Since I
                  > started using the new version I have had many failed
                  > JT9 decodes for signals that looked ok on the waterfall.
                  > The number of failed decodes seems to have increased
                  > significantly since I left r3496.

                  The decoder incorporates some necessary trade-offs that balance decoding
                  performance against speed.

                  There's a straightforward way to investigate your suspicion. You and
                  other users can easily do this, and at the same time make a contribution
                  toward future improvement of WSJT-X.

                  1. Install both versions of WSJT-X. Download links for Version 1.1,
                  r3496 and Version 1.3, r3673 are on the WSJT web site. They must go
                  into separate directories, say C:\WSJTX1.1 and C:\WSJT1.3 .


                  2. Start v1.3, check "Save All" on the Save menu.

                  3. Let it run for a while, making QSOs if you wish, and make a note of
                  the UTC for any instance where a JT9 signal looked OK but failed to decode.

                  4. After you have accumulated several instances of apparent decoding
                  failures: terminate WSJT-X v1.3, start v1.1, go to "File | Open", and
                  navigate to the "Save" folder under in the v1.3 installation directory.

                  5. Open the file(s) that you thought may have had improper failures to
                  decode, and see if v1.1 does any better. If you find any such examples,
                  send the *.wav files to me.

                  -- 73, Joe, K1JT
                • k3wyc
                  Thanks for the reply Joe. I have all the waves files saved and I also record the no decode events in my log remarks e.g -12,-14,ND,-13.. Now I have to dig
                  Message 8 of 9 , Feb 19, 2014
                  • 0 Attachment

                    Thanks for the reply Joe.  I have all the waves files saved and I also record the no decode events in my log remarks e.g -12,-14,ND,-13..    Now I have to dig through the data.  I'll send files if I find anything useful.


                    73,


                    Andy k3wyc

                  • k3wyc
                    Well that effort failed. I found 10 candidate QSO in my log for which I had recorded no decode events. Unfortunately I have no wave files saved. I had set
                    Message 9 of 9 , Feb 19, 2014
                    • 0 Attachment

                      Well that effort failed.  I found 10 candidate QSO in my log for which I had recorded no decode events.  Unfortunately I have no wave files saved.   I had set the SAVE option to SAVE DECODED when I installed the new version.   Either I don't understand what that selection does, or it's not working right.  The number of files being saved is far fewer than I expected and there are no files near the QSO times for the 10 events of interest.


                      I changed the SAVE option to ALL and that is saving a file every minute.  Maybe I'll be able to gather some no decode data in the future.


                      73,

                      Andy k3wyc

                    Your message has been successfully submitted and would be delivered to recipients shortly.