benson

To try out SuperSDR, the installation instructions are at https://github.com/mcogoni/supersdr/

Hi Paul,

I am assuming that you have Python installed correctly and already unzipped the kiwiclient package.Depending on how far you got with trying to run kiwirecorder and using Python on your windows machine you can try the following steps.

1. With explorer go to the directory where you unzipped the kiwiclient repository into.

2. Check Python is properly installed by opening CMD and typing python or perhaps python3 in your installation. To exit python type in exit()

3. For the remote kiwi(s) you want to use, have the ip adresses and port numbers ready. 

4. As example here Shannon Volmet parallel transmission is used on 5505 and 8957 and a kiwi receiver in Finland. Provided the remote Kiwi has 2 slots available it will record both transmissions which then show up in your above mentioned kiwiclient directory. In the CMD window type:

python kiwirecorder.py -k 30 -s kiwisdr.oh6ai.fi,kiwisdr.oh6ai.fi -p 8073,8073 -f 5505,8957 -m usb -L 200 -H 2700

Stop the recording after a minute with the CTRL + C keys combination.

5. Once you got this working, have a look at kiwirecorder.py --help to see what else you want to add, like limiting the recording time, add squelch parameters, append a station name or adding a directory where you want to save those wav files.

Best regards, Ben

Well, not really abandoned as you can see in this post from April this year: https://priyom.org/blog/priyom-releases-rivet-build-91

You could also have a look at Rivet (https://github.com/IanWraith/Rivet/wiki/Introduction). This is an open source decoder written in java that can work with kiwi wav recordings or directly from the audio when using a virtual soundcard. The Rivet code does not do automatic classification however.

It looked quite promising as a FSK type of signal decoder with features as automatic speed and shift adjustment, but it appears the development has stopped unfortunately.

What you will find is that some of those signals such as CIS36-50 (BEE) can indeeed be decoded with it, but as expected seldom deciphered. Those that are in plain text (NAVTEX, DSC, MSI, FSK meteo) can already be explored with the existing Kiwi extensions.

brgds, Ben

Indeed an option for the Kiwi admin that removes the callsign field such as Martin suggested would be helpful to address the issues with fake or false callsigns and take care of some of the points Philippe brought up.

brgds, Ben.

Hi Jim,

Thank you for the hint. I found all needed info in https://wiki.archlinux.org/title/ConnMan .

Next I used the BBAI cloud9 IDE interface and deleted the older wifi directories from var/lib/connman. A Beagle reboot seemed necessary for the changes to take hold.

Best regards,

Ben

VK5QI created some good looking rrd plots as shown in his posts above. Also linkfanel's page with signal-to-noise ratio based scores is useful, but in order to visualize how these SNR values change over time and location I have been working on creating SNR plots in Python for single and multiple Kiwi stations. Now using jks recent addition of up to a week of hourly SNR data this has become a lot easier and quicker than previously where waterfall data had to be retrieved and processed for each Kiwi node.

The multi station plots are for comparison of hourly data. Kiwi nodes have been selected for 4 longitude bands trying to get some reasonable geographic coverage and the list has been limited to 20 of the subjectively "better" stations.

Please note that sometimes the SNR json files cannot be retrieved due internet connectivity time-outs and at times will possibly have missing hourly data points when all channels are in use. In addition the individual Kiwis can be offline for quite long periods as well.These plots are generated with 'bokeh', a versatile Python library for interactive visualizations that produces javascript for use in html web pages .

The plots spanning a whole week do look crowded, but by clicking on the legends any or all bands can be turned off. On the top right corner you will find some tools including (wheel) zoom which make it easy to navigate the plots.

Currently running as a trial on a RP3, you can check out the alpha version on http://sparks.epizy.com/SNR/MultiSNR_plots.html

A pdf screenshot is attached.

Note : on many browser F5 refreshes the browser cache.

About changing the SNR measurement details and referring to LA3RK's post about diurnal effects.

Is there a place where the administrator can change "utc":0,"ihr":6 ?

For longer time monitoring hourly SNR snapshots would be useful, if the rolling data history retained on the Kiwi nodes can be expanded to say cover one week of data. This was also the time frame Marco -IS0KYB previously used for his quite well working Dynamic SNR map.

Thanks, Ben

An interesting attempt to classify demodulated signals by using audio fingerprinting of known samples is published by Jose Rueda (http://jcrueda.com/?p=916). It is based on a Python program called Dejavue which is meant to recognise songs from just "listening" to a small fragment. A good explanation is in the program description at ( http://willdrevo.com/fingerprinting-and-audio-recognition-with-python ).

This looks like a great idea, but trying this out on live signals from a Kiwi SDR gives rather low accuracy and a lot of false captures (mis-identification). This perhaps is not surprising because of the large amount of QRM and QRN mixed in with the audio signals. In addition filter bandwidth, gain and center frequency chosen for demodulation can make identification more difficult. It may be worth an attempt to use live recordings from the Kiwi as "known samples" to replace the mostly high SNR sample data that Jose extracted from the sigidwiki.com database.

Another route would be to use deep learning based signal classification and train a Convolutional Neural Network (CNN) using a much larger set of sampled data that would covering a wide range of SNR's and other variations. This would involve a lot of work though.
Aside from demodulated data as input the IQ datastream could be used to also classify the modulation type. Using either Tensorflow or Python Keras deep learning networks by Randaller shows some promising results ( https://github.com/randaller/cnn-rtlsdr ).

It is still early days for BB-AI adaptation into the Kiwi project but it seems the AI would be quite capable for this type of classification job.
Probably some folks on this forum have already looked at this and are willing to contribute ideas, code and results.
Anyway I just bring up auto signal classification as one of the many possible further development directions that would be unique.

Best regards, Ben

One of those less well known but excellent alternatives is directKiwi created by llinkz.
(https://github.com/llinkz/directKiwi)

The program allows you to select any of the available Kiwi receivers on a map interface GUI and get the streaming audio signal, with a mouse click you can obtain the current SNR estimate as used to be available on Marco's IS0KYB website using his micro -kiwi-waterfall routine and it shows the not so frequently updated SNR measurements from rxfanel page.

I have been using this for quite some time to find the better performing Kiwi's bypassing the obviously manipulated voting method that occurred at sdr.hu

73, Ben