Information about WSJT experimental Modes JT2 and JT4 from Joe Taylor K1JT

To: Users of WSJT
From: Joe Taylor, K1JT
Subject: Experimental Modes JT2 and JT4
Date: October 11, 2007

Thanks to all who have tested the experimental modes JT2 and
JT4 in WSJT 5.9.8 r558 ... and special thanks if you reported your observations to me or on one of the relevant reflectors, or sent me representative wave files. I can report that over the past few days QSOs have now been made with both JT2 and JT4 on amateur bands and propagation modes ranging from trans-Atlantic nighttime DX on 1.8 MHz to 144 MHz EME. In addition, JT4 has been used to send and receive messages on 10 GHz EME.

If you have not sent comments or a report, it's not too late! Please note that some of the most useful wave files are those containing received signals that you think "should" have decoded, but did not. I am interested in reports of all types of on-the-air tests, from HF to microwaves.

As described in the announcement posted several days ago at

The new modes are basically functional but do not yet have certain amenities that JT65 users are accustomed to. The new modes show considerable promise, but for a number of reasons they are not yet ready for general use:

1. Decoding performance is overly dependent on the average audio level at the sound card input, and perhaps other factors.

2. Software AFC should be included, to correct small frequency drifts.

3. Options for message averaging and matched filtering should be included in the decoders.

These improvements should not be very difficult to make, but my own schedule is such that they will probably take at least some weeks to complete.

I have not heard much from others about the appearance of
JT2 and JT4 signals on the SpecJT waterfall, or the perceived operational importance of the narrow bandwidths of these modes compared to JT65. Some relevant facts are these:

1. JT65B, the dominant mode for 144 MHz EME, has bandwidth
B=355 Hz. JT65A, mostly used on HF and 50 MHz, had B=177 Hz. These modes have a narrow "sync tone" only 2.7 Hz wide, which is easy to recognize on a waterfall display.

2. JT2 has bandwidth B=8.75 Hz and JT4A B=17.5 Hz. In these modes sync information is embedded in the data stream.
Waterfall traces of JT2 and JT4 signals are therefore much narrower than the full extent of JT65 signals, but appear "fuzzy" compared to the JT65 sync tone.

3. Many JT2 or JT4A signals can fit into the bandpass of a standard SSB transceiver without interfering. This means that random digital EME operation could become easy if, for example, JT2 activity were concentrated in a relatively small (5 or 10 kHz?) portion of each band. It would then be possible for WSJT to acquire many of the advantages of the
MAP65 software, which at present are available only to those with wide-band receiving systems like Linrad.

4. The probability for collisions between JT65 signals or between a JT65 signal and a birdie is much larger than that for the much narrower JT2 or JT4A signals. On the other hand, the JT65 decoder is relatively mature and has been made remarkably immune to QRM. The performance of JT2 and
JT4 in comparable situations is not yet known.

Community input on these matters is needed! Assuming that the present deficiencies of JT2/JT4 can be corrected, so that S/N performance is essentially equal to that of JT65, do you consider the narrow bandwidth of JT2/JT4 to be a significant advantage? Or should we forget about JT2/JT4 (at least on bands where JT65 is now widely used), and stick with JT65? Please share your views on these and any other questions you think relevant.

-- 73, Joe, K1JT

PS: I will be travelling for much of the next week, so any responses will likely be delayed.

For reference:

The original announcement about JT2 and JT4 is at

The experimental version of WSJT with these modes can be downloaded from


Back to news of my latest activity

Back to homepage