Hi Paul, all,

SlowJT9 vs WSJT-X: as I mentioned it was pure luck that SlowJT9 performed 
better that night.
But as the only reason I can think of is a small timing difference between both 
it might be worth to time-shift the incoming audio by small steps, then try to 
decode each step nd take a "best off" as result.

Frequency conversion: I will investigate that furher. Anyway, if I implement 
this it will be optional (can be switched off).

For the moment WX is rather bad here, so the antenna is lowered. But in a few 
days, when the  WX permits I would like to try some TA QSO's on 475 kHz using 
the slower JT9 modes.
Any takers at the other side of the pond?
I worked Paul, N1BUG and Eric, NO3M, mid February last year.

73, Rik  ON7YD - OR7T

________________________________________
Van: [email protected] <[email protected]> namens N1BUG 
<[email protected]>
Verzonden: donderdag 31 januari 2019 21:43
Aan: [email protected]; [email protected]; 
[email protected]
Onderwerp: Re: [rsgb-lf-group] SlowJT9 update (v0.9.15.0)

Hello Rik, all,

Two comments, see below.

> Interestingly Rik's SlowJT9 which was running in parallel,
> managed to decode a couple of more transmissions than WSJT-X."

I have seen this work both ways. Sometimes SlowJT9 will decode
something WSJT-X does not. Other times WSJT-X will decode something
SlowJT9 does not. Usually they are equal.

> 2. Add optional internal frequency conversion to allow reception
> at higher frequencies I have done some preliminary tests (based
> on the code snippet Wolf, DL4YHF, sent me) and it seems to work
> fine. But I am not sure that it will be very useful as frequency
> conversion will not increase the usable frequeny range, but just
> shit it. Eg: now JT9-10 can be received from 100 Hz to 415 Hz
> audio. After USB downconversion with a 1 kHz carrier the range
> will be 1100 Hz to 1415 Hz. The only reason I can think of to
> want frequency conversion is if you use a fixed frequency (ofen
> 800Hz) CW filter, so you can fit the JT9-5 or JT9-10 frequency
> range into the filter passband. But this downconversion requires
> a Hilbert transform (90 degrees all pass filter) that never can
> be perfect and this will cause some distortion (in particular at
> the lower and of the passband). So before I start the effort to
> implement this in SlowJT9 I would like to know if there is an
> audience for it.

I think there is justification for it on 2200m if it comes without
too high cost in lost sensitivity. Some people cannot easily tune
antennas down to 136.4 or so. But if we operate at 137.x then
listening stations must adjust their receiver because the usual
136.000 'dial' frequency will not allow receiving JT9-5 or JT9-10.
This is not the usual setting for most who also want to monitor WSPR
or other modes. For pre-arranged QSOs this will not matter, but for
calling CQ it is good to operate where we can be heard without
special receiver settings.

However, if this feature comes with significant loss of decoding
sensitivity, then I would vote to stay as we are.

73,
Paul N1BUG

-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.

View/Reply Online (#8560): https://groups.io/g/rsgb-lf-group/message/8560
Mute This Topic: https://groups.io/mt/29606322/1555778
Group Owner: [email protected]
Unsubscribe: https://groups.io/g/rsgb-lf-group/leave/3440736/168031436/xyzzy  
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-