rx.kiwisdr.com improvements
A few improvements to rx.kiwisdr.com (aka kiwisdr.com/public)
The search: field at the top of the page now allows search terms to be AND'ed using the comma "," character. For example uk,gps will only show entries containing both uk AND gps. Similarly for uk,loop,gps etc. The vertical bar "|" character allows search terms to be OR'ed. For example uk|germany|france will show entries containing one OR more of those terms. The two operations cannot be combined currently.
The search field can be initially setup from the browser URL. E.g. rx.kiwisdr.com/?s=uk,gps or rx.kiwisdr.com/?search=uk,gps
The URL params field will be applied to each Kiwi link shown. For example if the field is set to f=10Mamz10&sp then each Kiwi link will include the parameter in the URL, e.g. my_kiwi:8073/?f=10Mamz10&sp In this way a "preset" of URL parameters can be conveniently applied to all the Kiwi entries in the list. See the image below.
As with the search field, the URL params field can be initially setup from the browser URL. E.g. rx.kiwisdr.com/?u=f=10Mamz10&sp Instead of u= the sequences url= or p= or params= may be used.
If s= and u= are used together then s= must appear before u= E.g. kiwisdr.com/public/?s=uk,gps&u=f=10Mamz10&sp
So it is now possible to create browser bookmarks that produce very specific rx.kiwisdr.com results.
Comments
Hi John,
This is a useful enhacement, but in order to reject the poorer performing KiWi's, I wonder if it would somehow be possible to sort / filter by SNR score too.
e.g HF>30dB
Regards,
Martin
Thinking about this a bit further, and this is just a wild idea...
The reason I'm interested in the SNR scores is mainly when choosing KiWis fir TDoA runs.
The big problem is the identifiers used on the TDoA map don't always match the names used on the public lists that give SNR scores.
The problem most folks seem to have when using the TDoA extension is picking KiWis that have decent SNR values and are a suitable distance away from the target location.
I wondered if the filters you are providing could also be used with the TDoA map, to remove KiWis that do not meet the required criteria.
It also struck me that maybe if the TDoA map was initially centred in the likely transmitter location and the frequency entered, it may be possible fir the TDoA server to semi automatically select three suitable KiWis with good SNR values, which were also placed at suitable skip distances for the TDoA run at the specified frequency.
In some cases it may then be possible to automatically select three different KiWis to perform a second run, which could be compared against the first as a test for accuracy.
Regards,
Martin
Okay, an additional
snrfilter field should be working. Also, the HF snr value is included in each Kiwi listing (to the right of the user counts).When the
searchandsnrfields are both specified then the result is the logical AND between them.Use a negative
snrfield value to get a less-than-or-equal function.When no snr value is shown in a listing it could be because snr measurement is disabled or the antenna is disconnected. With some changes on the server side these two conditions could be distinguished which might be helpful.
"No antenna connected" detection should be working. Enter
no antinto the search field to find them.There are 8 currently. It doesn't seem to detect any false-positives, but it might be missing some.
Hi John,
Thanks for this most useful.
It would be great if the search filters also worked with the http://rx.linkfanel.net/ map link on the rx.kiwisdr.com page so that only the filtered KiWi's map pins appeared.
Regards,
Martin
You'll have to talk to Pierre about changes to rx.linkfanel.net -- I don't control any of that.
Well the "no antenna connected" detection is going to have to move to the Kiwi server code (v1.602). There were false-positives today. And the fix can only happen with a different analysis of the waterfall bins.
@G8JNJ Yes, adding automation like that to the TDoA extension is a good idea. Lots of work to get something like that working though..
Hi John,
Yes I agree, it was a wild idea, as a lot of work would definitely be required to automate the TDoA selection process.
I guess the accuracy could also be further improved by some sort of machine learning, by analysing the TDoA requests sent to the server, the KiWis that were selected by users and the accuracy of the end result (size of map 'bulls eye').
Just thinking out loud...
Regards,
Martin
The other thing I'd like to see is some automation of connection to the candidate Kiwis (after SNR criteria etc.) with a confirmation dialog so you could more easily weed-out those that don't have adequate reception of the signal-of-interest. Right now it's kind of a tedious process of manually testing and rejecting candidates.
I don't think you could easily differentiate between noise and the desired signal on a single KiWi automatically, but maybe after the TDoA run the KiWi(s) that didn't correlate well could be flagged up with an icon, so that the user knew and could choose another KiWi. Which would save having to look through all the maps to ID the one(s) to be excluded.
The 30 second period is sometimes a bit too long if signals only appear once a minute or if there is other interference. This means there is a lot of other stuff in the recordings that has to be tested.
The stand alone TDoA application allows you to snip the recordings, so that you only use the relevant section for correlation. I sometimes use this for difficult signals, but it is far too complicated for casual use.
That all being said, the current TDoA extension is pretty good, and has made something available to the hobby community that was previously only available to the professionals. It was my original reason for buying my first KiWi and it has not dissapointed.
Regards,
Martin
And there is more...
I wonder if the degree of correlation (or size of map 'bulls eye' ) could be used to provide some sort of percentage confidence figure along with the likely co-ordinates.
Sometimes it is difficult to know which world region to select when initially trying to identify a likely source. It can take several TDoA runs using many different KiWis before being able to narrow it down to the most likely area to concentrate on, all of which is very time consuming. This is particularly true when trying to TDoA HF radars, which seem to be strong just about everywhere :-(
Should have added a link to Direct TDoA Github
https://github.com/llinkz/directTDoA
Hi John,
I have left a request as a Github action, but I won't hold my breath.
https://github.com/priyom
Regards,
Martin