Oh great watchers of grass and wet paint:

This new app, JTDX, seems to be becoming all the rage among JT65 aficionados:

http://www.qrz.lt/ly3bg/JTDX/jtdx.html

and

https://www.qrz.com/db/UA3DJY

Based on WSJT-X but using a multi-pass decoding algorithm, apparently people are decoding down to -30!

Tag

I just upgraded from WSJT-X 1.6 to 1.7 On 1.6 I would decode by the 50th second, now on 1.7 I decode by the 52nd second.  I can only imagine more decoding would hurt me.

FWIW. "it is said" that JTDX is based on the old KSVAD decoder, while the latest WSJT-X releases v1.7 uses some sort of newer/better decoder.  It "is also said" on some less-than-scientific comparisons that JTDX decodes fast, but generally misses some signals that WSJT-X v1.7 rc3 catches.

Just "by eye" (which may or may not be true) WSJTX v1.7 rc3 decodes more signals, and closer together signals.  It's now not uncommon to decode 25 or more signals on a busy band.  It looks like it's about the same speed to me, but decodes quite a few more JT65's than v1.6. It's also fairly common to see 2 or 3 signals decoding on the same Hz.

I also like the meteor scatter mode MSK144 - very cool.

If anyone has a link to a  properly done head-to-head test, I'd be interested.

I'm curious and would like to be able to get down to -30....but I'm also kinda tied to JT alert now as my 'b4' log is a big part of how I've been working stations. If it doesn't integrate with JT alert, or provide some way to easily upload a flat file of my 'b4' call signs I feel like I'd be adding a manual step (checking the call sign against my log) before working stations.

Still...about 5dB of signal range is intriguing.