[RTTY] Decoder performance on crowded bands

Mon Sep 28 15:48:07 EDT 2015

Don,

I ran Alex Skimmer Server on a local QS1R in the WPX earlier this year and 
the results were hard to believe.  Performance was exceptional.  And the 
benefit is that 100% of the spots are workable because they come off your 
local system.  I had almost zero in the way of S&P mis-spots.  The downside 
to this configuration is that you need the QS1R ($$) and if you want to skim 
multiple bands in their extents (96 Khz x 3 active bands), you literally 
need the best Intel chip available (more $$).

73/jeff/ac0c
www.ac0c.com
alpha-charlie-zero-charlie

-----Original Message----- 
From: Don AA5AU
Sent: Monday, September 28, 2015 2:42 PM
To: Tim Shoppa ; rtty at contesting.com
Subject: Re: [RTTY] Decoder performance on crowded bands

Tim wrote:
"The frequency resolution and accuracy seems to have improved in the 
skimmers in the past few years and often times I would click on a spot and 
be tuned within 10Hz, that is great!"
Yes, indeed. I found the frequency accuracy of skimmer spots to be 
excellent. I had ALL the RTTY skimmer spots coming in this weekend and it 
was amazing to see the spots flying in at a pace that seemed like well over 
a hundred spots a minute.
Unfortunately, IMHO, the RTTY skimmer still needs a lot of work. I would 
prefer to see less skimmer spots so that there are less S&P stations are 
being spotted and less busted calls spotted. It seems I got spotted quite a 
bit when I was S&P. This means that other S&P stations are being spotted as 
well. This means that a MAJORITY of the spots are not valid run stations. 
This is not good. I understand there will always be some busted calls but 
hopefully work is being continued to make this better. There was one 
particular skimmer that seemed to bust the start of calls and this caused a 
LOT of spots to show up as new multipliers. For the most part, I knew the 
calls were busted but it seemed I was continuously clearing bogus calls from 
my new multiplier list. I nearly blacklisted this one skimmer, and that's a 
solution, but I was afraid I'd miss a new multiplier! HI
I am grateful for the time, effort and resources from the entire crew of 
skimmer ops. They are doing great work and need to be commended. This was 
the first time I've used the RTTY skimmer spots for a full contest and it 
was very useful. It will take some thinking to determine the best options 
for using this new tool.
When using non-skimmer spots, you can nearly rest assured that a station 
spotted is a run station and you can chase them. The only real problem with 
that is knowing whether or not the spotter was using FSK or AFSK and if the 
spotted station is on the frequency of my radio when I go there. With the 
skimmers, that problem appears to be eliminated (when using fFSK) and it's a 
great thing.
But now you can't just click on spots because most spots are NOT run 
stations. My solution, for now, is to run two separate packet windows (not 
sure this is possible with WriteLog on a single PC but I did it on networked 
PCs this weekend since I was single band). One packet window will receive 
skimmer spots and the other non-skimmer spots. Use the skimmer spots to 
populate new multipliers to a Packet Spots window and use the non-skimmer 
spots to populate the bandmap. It may take some work from software authors 
or there may be solution already out there I don't know about?
73, Don AA5AU

      From: Tim Shoppa <tshoppa at gmail.com>
To: rtty at contesting.com
Sent: Monday, September 28, 2015 9:25 AM
Subject: [RTTY] Decoder performance on crowded bands

Wow, there was a lot of activity for CQ WW RTTY! 20M and 15M were filled
completely between 080-150kc above band edge,

The "bottom ends" of the band were particularly crowded.

I used 3 decoders simultaneously - MMTTY, 2Tone, and Gritty. I don't always
run Gritty for the more minor contests, but I did fire it up for CQ WW RTTY.

MMTTY is still my "main decoder". It decodes just a little faster than the
other decoders and handles the broadest range of senders.

2Tone did spectacularly well in decoding in the presence of strong adjacent
signals. This was particularly important on 40M and 80M where all the DX
usually had a strong adjacent local CQ'ing. It did not do so well with some
of the "slow RTTY" signals. 2Tone AFC always did "what I wanted" and helped
with a couple of off-frequency replies.

90% of the time I was looking between MMTTY and 2Tone.

Gritty did very well decoding in the presence of flutter and other unusual
effects, pulling out callsigns from flutter on 10M and the odd stuff that
happens on 40M at sundown.

Very shortly after the start of the contest I set my receiver bandwidth to
about 1200Hz because 2Tone and Gritty work best like this. If I set my
bandwidth too narrow it seems that Gritty just stops working. I also turned
IF gain way down so AGC wouldn't pump. Only a few times did I narrow up my
receiver bandwidth.

While I was not looking at the Gritty decoder all the time - I was always
looking at the Gritty waterfall. The Gritty waterfall is very very useful.
The color-intensity range and scroll rates work very well.

I happened to enter assisted as well. I started with VE7CC cluster and its
default skimmer filtering but it was not letting all the skimmer spots I
wanted through. Even though the calls were being spotted by 3 or 4
skimmers, lots of time they were not getting through whatever additional
checks were enabled in VE7CC cluster. So I went straight to
reversebeacon.net and took the full brunt of the firehose. It was obvious
not all the skimmers covered all the RTTY band - some (especially EU)
seemed to only be skimming in the 080-100 segments. But others were
skimming all the way up to 150kc above the band edge and those proved very
useful because activity was just huge this weekend! The frequency
resolution and accuracy seems to have improved in the skimmers in the past
few years and often times I would click on a spot and be tuned within 10Hz,
that is great!

Again, thanks to the authors of all these wonderful decoders, and the
skimmer guys.

Tim N3QE
_______________________________________________
RTTY mailing list
RTTY at contesting.com
http://lists.contesting.com/mailman/listinfo/rtty

_______________________________________________
RTTY mailing list
RTTY at contesting.com
http://lists.contesting.com/mailman/listinfo/rtty 