Topband: Top Band and JT65

[Mike Waters]

Ok, per your request, here's Joe Taylor's first reply (thank you, Joe! :-).

My second question --and Joe's reply-- follows, below the first reply.

73, Mike
www.w0btu.com

On Mon, May 15, 2017 at 7:58 AM, Joe Taylor <joe at princeton.edu> wrote:

> On 5/14/2017 11:36 PM, Mike Waters wrote:
>
>> Ok, what is the truth here? Is JT9 better than JT65 on 160m, or is it
>> inferior??
>>
>
As you should expect, the correct answer is "it depends".  Depends on
conditions of propagation, QRM, etc.

On an ideal AWGN (additive white gaussian noise) channel JT9 has a 2 dB
advantage over JT65.  But the JT65 code has more redundancy than that in
JT9, and the 2 dB advantage tends to disappear on a fading channel.

JT65 is more robust than JT9 in the presence of QRM (and possibly QRN?).

Overlapping JT65 signals are readily decoded.  Not so much for JT9. Savvy
operators using JT9 often respond to a CQ or tail-end "up" or "down" by 20
Hz or so, thereby avoiding calling on top of another caller.

I don't subscribe to topband at contesting.com or TBD160 at yahoogroups.com, but
you may re-post this message there if you choose.

        -- 73, Joe, K1JT


On Sat, May 13, 2017 at 5:50 PM, Brian D G3VGZ
<topband at planet3.freeuk.co.uk <mailto:topband at planet3.freeuk.co.uk>> wrote:

    ... I find the better ability for JT65 to decode co-channel signals
with its two pass decoding makes up for any 2dB improvement in decoding
JT9. I also find a single static crash can take out JT9 decodes, more so
than JT65.*
    *
That certainly got my attention! Has anyone else experienced this? *And if
so, what filter were you using: the wide SSB filter or a narrower CW
filter?? *I'm inclined to think that JT9 is superior to JT65 on 160m. /But
I have an open mind.
[snip]

=====================================

*Okay, here is another question, and Joe's reply:*
=====================================

On Tue, May 16, 2017 at 9:46 AM, Joe Taylor <joe at princeton.edu> wrote:
Hi Mike,

... I would like to know if this is true regardless of what the bandwidth
> and shape factor is in the receiver itself. The state of the art in SDR can
> create an almost perfect filter with zero ringing /no matter what the
> bandwidth is.
>
> I realize that the SSB filter is great for browsing the JT* signals, but
> what about narrowing the signal after establishing a JT9 contact, when
> there are no overlapping signals?
>

An SSB receiver is a linear device: typically a series of amplifiers,
filters, frequency mixers, etc.

As long as things remain linear, it doesn't matter in what order these
things are done.  There is NO advantage to using a narrow filter ahead of
WSJT-X, because the program already uses digital filters tailored exactly
to the needs of the protocol.  Final detection and measurement bandwidths
for the JT65 and JT9 tones are equal to the baud rates, 2.692 and 1.736 Hz
respectively.

        -- 73, Joe, K1JT