Re: [RTTY] A repository of (poor) RTTY recordings?

[George Johnson <w1zt@comcast.net>]

Chen,
RITTY does a form of what you are suggesting.  It will attempt to decode a 
callsign and upon a second occurrence of the pattern it will try to 
reconstruct a "correct" callsign of the two different versions.  I think 
Brian uses a maximum likelyhood algorithm test of the received 
symbols.  What appears on the screen is a slightly delayed "estimate" of 
the corrected callsign which I have found to be accurate more than half of 
the time.  The corrected text comes out in a different displayed color on 
the RITTY screen but that distinct display is not used when RITTY is 
coupled to WriteLog.  I just get a somewhat delayed print which I know to 
be the estimate.

This all makes for interesting signal processing but I also observe that 
there are some pretty vulnerable characteristics to our two frequency 
Baudot encoding and modulation system.  Polar flutter and backscatter 
multipath distortion are pretty rugged conditions.  The best signal 
processing minds have played with these problems and Hal equipment has had 
some of the best solutions ... at a cost.  The software solutions like 
RITTY have worked pretty well for me at this point.  But there have been 
frustrating times that I just wished the DX station had just sent a quick 
ID in CW so I could confirm the printed callsign gibberish...  Hi Hi...
73, George .. W1ZT

At 11:22 PM 4/20/2004, Kok Chen wrote:
> On Apr 20, 2004, at 8:10 PM, Dave Bernstein wrote:
>
> > I have a KAM '98, which I use with MMTTY for RTTY Dxing. WinWarbler lets
> > me run both simultaneously, displaying the received characters from each
> > in adjacent (over/under) panes.
>
> Does MMTTY allow running three or five different parameter sets and then 
> do a vote on which is the best estimate of the character sent and 
> displaying that character?    (The method would even have worked in 
> Florida, just let each person vote three times and then figure out they 
> really meant.  Is that why they vote multiple times where W9OL lives? :-)
>
> Using a human op to do such a mundane task seems such a waste.  Modern 
> computers are plenty fast to do simultaneous decoding.  "Call recognition" 
> is already handled by many contest software, for example.
>
> 73
> Chen, W7AY
>
> _______________________________________________
> RTTY mailing list
> RTTY@contesting.com
> http://lists.contesting.com/mailman/listinfo/rtty


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