Joe Taylor -- would you share your thoughts on the fix for this problem?
Dave...You'll find it interesting to do a scatter plot of that data.
Only 16,838 entries for almost a full month? I have 34,920 just for yesterday
but I run about 16 hours a day.So this is 24 hours for FT8 here and my history
of FT4 except for the test from my location here. My delays are due to using
ANAN100 and Thetis so all the data processing in my computer and adds to the
latency and I need buffer time to avoid dropouts on my Intel Core I7. There's
some variance in here due to my playing with buffer sizes but the moving
average is pretty consistent.
FT8 mean = 0.4Mode = 0.4Average = 0.391 Sigma = 0.41Skew = -.03
You can see the 300ms loss here for me in the mean and mode -- and the
reduction in standard deviation is worrisome as there's no reason the dt values
in FT4 should be any different than FT8 except for the difference in the time
tolerance -- which means quite a few decodes being missed. If everybody's'
time was the same we'd have mean=0 and sigma would be equal in both
modesthat might be asking too much though.
FT4 mean = 0.1Mode = 0.1Average = 0.181 Sigma = 0.15Skew = .11
Mike W9MDB
You note a little negative excursion around 17,000 -- that's when I adjusted my
computer clock to -300ms.You'll also see a negative stepping trend here. I
think some people were abandoning FT4 that were having decoding problems.
| 0.1 |
| 0.175057 |
| 0.146898 |
| 0.111946 |
| 0.1 |
On Friday, May 17, 2019, 4:09:08 AM CDT, David Birnbaum
wrote:
I took a slice out of my ALL.TXT file from 4/12/19 through 5/9/19 and did some
simple statistics on DT. The results show:
FT8: Mean = 0.1830 Variance = 0.1674 for 16838 entriesFT4: Mean = 0.055707
Variance = 0.031910 for 4249 entries
DaveK2lyv___
wsjt-devel mailing list
wsjt-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/wsjt-devel
___
wsjt-devel mailing list
wsjt-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/wsjt-devel