Thursday, 19 January 2017

JTDX - DXing with JT digimode

JTDX means "JT modes for DXing", it is being developed with main focus on the sensitivity and decoding efficiency, both, in overcrowded and half empty HF band conditions.

It is open source software distributed under the GPL v3 license, and is based on the WSJT-X r6462 source code. Optimal candidate selection logic, four/five pass decoding and decoders based on the matched filters making JTDX performance quite different from WSJT-X software for operation on the HF bands. Almost all work goes around JT65 mode, decoding efficienсy of JT9 mode is the same as in WSJT-X and might be addressed in the future JTDX versions. Initially based on the enthusiasm, now in average it taking close to 12 hours of daily work and your donations would be the only way to let me proceed with the further JTDX software development.
Many thanks for your support,
73, Igor UA3DJY

When watching PSK reporter lately I noticed most stations are using JTDX instead of WSJT-X or other software these days. So I thought, what could be the reason? This software has to be the top if so many use it. I installed the software latest version and first impression is that it looks a bit different from WSJT-X although it is based now on the latest WSJT-X 1.7. After some playing and configuring I was courageous enough to dare a TX cycle and suprisingly 3 calls came back, all 3 with only 1Hz apart from each other. So, I actually saw one signal in the waterfall but it was at least a layer of 3!!! I have never seen this in WSJT-X. Wow! I picked the weakest station (-20dB) while the other stations were at least -10dB and -6dB. After a few days I noticed that even weak traces half behind strong signals are decoded, something I hardly experienced with WSJT-X.

Some features not seen in WSJT-X:

To work stations faster I notice more and more stations are using the feature "skip locator" in JTDX, with other words they come back with the report directly in response on a CQ. This is especially useful for DX stations with large piles of calls waiting to make a QSO as fast as possible.

A SWL button increases the amount of decoding attempts. Realize that it will take a faster computer to decode everything before your transmission session will begin!

There is the "hint" decoder button, although I probabely will not use it. It is interesting development but I personally think it is cheating. The "hint" decoder looks for a hint in a file with calls and displays the best possible call it can find, it is to the operator to use it or not for a QSO. It actually is the same as the hints you can get in a contestlogger like N1MM(+). That is actually cheating as well although it helped me numerous times. Anyway, you don't have to use the "hint" decoder of course and actually most of the time JTDX is receiving well enough to do without it.

The filter button is another nice one. A passband filter for the desired signal especially usefull in crowded band situations.

While I was writing this article the software is still developed and another feature has been implemented which is a button for AGCcompensation. Not shure what it does, I've been switching it off/on but am not shure what the difference is? This is what the official release notes tell: AGCc button: use it only if AGC being triggered in receiver by JT signals at beginning of the RX interval. Noise level change in the waterfall might be used as criterion showing that AGC is triggered in the receiver. 


It is recommended to read the JTDX documentation and release notifications before installation.
I had trouble with the performance of my computer with this software package as I still work with my 2008 Intel dual core 2,6GHz. Sounds fast, well not for todays standards. The problem was that my transmitting audio started 3-6 seconds later after the actual transmission (in time) start. First I thought it was JT-Alert in combination with JTDX but without JT-Alert running I had the same problems. The original settings are for up-to-date not too old computers. I use both attempts and passes two steps less as set by the software when you install it. Till now with good results. Below something you have to consider when you got a slower computer.

Advanced settings tab:

- Number of decoding attempts: set it as per required decoding time. It might be set to 1 if CPU frequency is 2GHz, and set to 3 if CPU frequency is equal to 3GHz. Higher values make better decoding performance.
- Number of decoding passes = 4 (may be set to 3 or 2 for CPU frequency less than 2GHz if decoding attempts=1 is not enough to get into 10 second decoding time). For VHF usage optimal number of decoding passes = 2, and number of decoding attempts may be increased to improve decoding efficiency.
- Hinted decoding DT range = 1

I've been using JTDX now for a few times/days last couple of weeks and notice that it is very addictive. It simply detects so many signals. There is always something interesting to see in my opinion. Have you ever seen 3-5 different stations coming back on your call all at once? With this software you have to actually choose who you make QSO with first. Well, I hope I've helped some of my blog readers to get a better JT digimode experience. Now, back to the radio and make contacts...


  1. Thanks for this info, Bas - I am just now starting to use JT9/65 but had not heard of this software. I will try it this weekend.
    73 - John AE5X

    1. Hello John, I saw your call for some time ago on hamspots calling for stations. I was just too late. Last few weeks I'm primairely active on 60m (5357 dail) in JT65/9. This program is certainly worth trying. Hope to work you some time. 73, Bas

  2. Thanks for the infos, may be I'll give a try. "hint decoder" option looks like can cause arguments in the future hihi.
    73 de SV1GRN

    1. I agree with you Panagiotis so I probabely will not use it. But at least you can try this option some time if you have a fast computer. 73, Bas

  3. Hallo Bas, ik had er nog niet van gehoord. Inmiddels wel geïnstalleerd, gelijktijdig laten draaien met WSJT 1.7.0 maar ik zie geen verschil in ontvangst tot nu toe. De rapporten met WSJT zijn zelfs iets beter. ;-) afijn, verder uitproberen. 73 Paul PC4T

    1. Ik heb nog geen vergelijk gedaan. Of JTDX echt beter ontvangt? Tja, ik ben geen expert. Maar ik was wel onder de indruk van het decoderen van stations die maar 1 Hz van elkaar af zaten. Daarnaast heeft JTDX natuurlijk een paar extra opties. Gewoon proberen wat jou het beste bevalt. 73, Bas

    2. Ja, dat ga ik doen. Ik beide op de PC staan en zal wat meer gaan experimenteren. 73 Paul

  4. Tnx for sharing this Bas.
    Running now in my QTH with a half size K9AY and G11 SDR. We shall see how it does.

    73 de Angel M0HDF / M8A

  5. Thanks for the rundown on the software and JTDX. One day this week I will have to check out the software and as you said read the manual as well.

    1. Hello Mike, just like Paul mentioned the decoder isn't able to be more sensitive as WSJT-X as it uses the same WSJT-X source code. But since it decodes in more passes and with the add of filters you have chance to decode signals that would otherwise not be decoded. Good luck with trying...
      73, Bas


Thanks for your comment. Bedankt voor je reactie. 73, Bas