Hi Joe, Thanks for those results. Yes, SF behaves well on those conditions and FT8 may be close to the edge that is obvious in the HM model. I was eager to see whether the 10 ms Doppler spread starts to affect to the high SNR performance, it may or may not. It's not an important information, just curios.
73, Reino OH3mA > -----Original Message----- > From: Joe Taylor via wsjt-devel <wsjt- > de...@lists.sourceforge.net>h > Sent: Saturday, September 7, 2024 2:51 PM > To: WSJT software development <wsjt- > de...@lists.sourceforge.net> > Cc: Joe Taylor <j...@princeton.edu> > Subject: Re: [wsjt-devel] SuperFox and FT8: Weak-Signal > Performance > > Hi Reino, > > We could do simulations for the "High-latitude > Moderate" (HM) channel, of course, but have not done > do. We do know that FT8 does not do very well under > such conditions (Doppler spread = 10 Hz, Differential > delay 3 > ms): see Figure 7 and Table 6 of our QEX paper on FT4 > and FT8, here: > https://wsjt.sourceforge.io/FT4_FT8_QEX.pdf > > Steve did simulations for a path fairly close to the HM > parameters: > Doppler spread = 5 Hz, Differential delay 3 ms. Curves > for these results are included in the plot I've posted > here: > https://wsjt.sourceforge.io/fig_superfox_ft8b.pdf > > As you can see from the curves with triangle points, > sensitivities for SuperFox and FT8 are nearly the same > (within about 1 dB) on this path. > We also know (Fig 7, QEX paper) that the FT8 curve for > HM fails to approach a decode probability of 1.0 at high > SNRs. > > -- 73, Joe, K1JT > > On 9/7/2024 5:41 AM, Reino Talarmo via wsjt-devel > wrote: > > Joe, > > > > Thank you about the excellent document. It does > clarify where we are > > at the fight against noise and signal spreading. > > > > I am interested to see also results for the "High-latitude > Moderate" > > propagation model for SF. FT8 have a lot of difficulties > on that kind > > path as presented in Figure 7 of The FT4 and FT8 > Communication > > Protocols. FT4 in that figure may give a hint. I assume > that a part of > > N5J QSO difficulties was the inferior decoding of FT8 > over the polar > > path not the SF signal. > > There was sometimes also unintended FT8 QRM on > the even timeslots > > adding repeats from fox. > > > > 73, Reino OH3mA > > > >> -----Original Message----- > >> From: Joe Taylor via wsjt-devel <wsjt- > de...@lists.sourceforge.net> > >> Sent: Friday, September 6, 2024 10:54 PM > >> To: wsjt-devel@lists.sourceforge.net > >> Cc: Joe Taylor <j...@princeton.edu> > >> Subject: [wsjt-devel] SuperFox and FT8: Weak-Signal > Performance > >> > >> Hi all, > >> > >> WSJT-related lists have recently seen a lot of > > discussion > >> about FT8 and SuperFox decoding thresholds. Some > of it has been > >> incorrect or misleading. > >> > >> I have posted a short document comparing measured > sensitivities of > >> SuperFox and FT8 (including multi- > >> streaming) here: > >> > https://wsjt.sourceforge.io/SuperFox_Performance.pdf . > >> > >> -- 73, Joe, K1JT > >> > >> > >> > ______________________________________________ > >> _ > >> wsjt-devel mailing list > >> wsjt-devel@lists.sourceforge.net > >> > > https://lists.sourceforge.net/lists/listinfo/wsjt-devel > > > > > > > > > ______________________________________________ > _ > > wsjt-devel mailing list > > wsjt-devel@lists.sourceforge.net > > https://lists.sourceforge.net/lists/listinfo/wsjt-devel > > > ______________________________________________ > _ > wsjt-devel mailing list > wsjt-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/wsjt-devel _______________________________________________ wsjt-devel mailing list wsjt-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/wsjt-devel
