On 17/11/2018 14:29, Jesús Gutiérrez Rodríguez wrote:
I've been reading messages about problems, improvements, all referred
to the current beta version rc4.
Really many of the improvement suggestions or problems with the new
version that we are testing, some of them, we have alternatives like
- Distinctive colors, secondary issue if we operate with JTAlert that
informs us about this topic in an appropriate way.
-worked or new countries, also secondary because we want it JTAlert
-links on Eqsl or others, we have, for example, Log40M, which gives us
many links, statistics, etc.
I want to express with this that there are many improvement issues
that we are contributing, which is very good, but, from my experience
in FT8 digital systems, I see as worrying that, now the program, I
think it has lost sensitivity and this should be to address it is the
essence of decoding in FT8
HI Jesus,
although several testers of WSJT-X v2.0.0 RC4 have reported a loss of
sensitivity, no one has given good evidence of that actually being the
case, recorded .WAV files for example. There are, for sure,
complications when sharing frequencies with incompatible versions of
WSJT-X, particularly with prior RC versions that had a dual decode
capability that has be dropped from RC4 onwards. These have lead to
situations where RC3 operators. for example, have set up sending 75-bit
FT8 messages, quite rightly, because they are operating on the normal
sub-bands (14074, 7074 etc.) and then not being able to inter-operate
with RC4 testers above 2000Hz.
The simple answer here is that *no one should still be running WSJT-X
v2.0.0 RC3, RC2, or RC1*. These are beta test release candidate versions
and testing feedback, which is why you are running them, must be from
the latest iteration to be useful to the development team. If anyone is
still running RC3 or earlier, and wishes to continue with beta testing,
then please upgrade to RC4; otherwise revert back to v1.9.1 and leave
the testing program for now.
With respect to your comments on worked before highlighting and JTAlert
overlap, please remember that many WSJT-X users do not use MS Windows,
others do not want to run multiple applications or do not use a station
logging application that interacts smoothly with JTAlert. Also please
consider that one of the advantages of the new FT8 and MSK144 77-bit
message payloads is the ability to partake in digital modes during some
contests. Contest may not be preference but for those that do enjoy that
aspect of amateur radio, the ability to flag duplicates and multipliers
is essential these days. The enhancements to highlight some decoded
messages are a step towards such facilities.
As a further aside, a basic communications theory analysis of the 75-bit
and 77-bit FT8 and MSK144 modes will show that, all other things being
equal, there would be a loss of sensitivity due to the slightly higher
code rate necessary to convey the extra bits in the same time frame and
bandwidth with the same modulation scheme. But all other things are not
equal, in fact diligent and thorough work by Steve K9AN has yielded a
better decoder sensitivity that offsets the loss of sensitivity due to
the higher code rate. This has been thoroughly verified in simulations
and we are confident that this new FT8 mode in WSJT-X v2.0.0 is at least
as sensitive as the 75-bit payload version in most usage situations.
There is also better immunity from undetected errors (false decodes) in
this latest decoder iteration, even though there are more possible
message combinations, thus improving it's practical usability on a
crowded HF band.
73
Bill
G4WJS.
_______________________________________________
wsjt-devel mailing list
wsjt-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/wsjt-devel
