Which software do you use to measure repetition-rate and other stats?
I have been looking for a suitable spectrum-analyzer, but so far I have only found software oriented at music-use, but none that were more 'signal-agnostic' or oriented towards technical-signals.
I use a variety of softwares to measure and display sweeps and rates.
You can often use audio focused software for this purpose, after you demodulate the signal using a radio, then everything is at audio frequencies. So things like Audacity, Spectrogram16, Goldwave, Spectrum Lab, Soundcard Scope, etc, all allow you to display and measure once the signal is demodulated to audio. And of course, there is an entire market of high end programs to do detailed analysis of signals, like Go2MONITOR, Go2ANALYSE, Hoka Code 300, Krypto500, Wavecom, etc.
However, ideally to do so (use audio measurement and display software) you need to be able to have the entire waveform inside the audio passband. For example, a 14 kHz wide sweep would need to be demodulated in USB (or LSB) with a 14+ kHz audio bandwidth, and that is not always possible. You can still take some measurements even if you can't get the entire thing in the passband, but it is best if you can use a wide enough bandwidth.
It is sometimes better if you can do measurements and such at the RF level. For that programs like Signal Analyzer, SpectroVue, etc.
And just a random piece of speculation here;
If these are 'early-warning' radars... then why aren't they running 24/7?
OK, who said these radars don't run 24 hours a day? Right now I can look at the FullDay display on the University of Twente WebSDR and see that the Russian 29B6 has been on, one frequency or another, for over 24 hours. And anytime I go look for it, or the British PLUTO, I can generally find it, unless conditions are very bad.
There is no doubt, these are radars, no doubt at all. You can question the exact source, but the operations, habits, and waveforms are easily correlated to the task.
A quick primer on HF radars, particularly these kinds of long range OTHRs (British PLUTO, Russian 29B6, French Nostrodamus, Australian JORN, Chinese, Iranian, etc).
The basics of any radar apply, the system emits a signal, the signal reflects off a target back to the radar receiver, and the radar processing measures various aspects of the signal to detect / plot / track the target.
Some of the things radars may do: The radar may measure time of flight (the time it takes the signal to leave the radar until the time the signal returns to the radar) to determine range to the target. The radar may measure Doppler shift to determine radial velocity. The radar may look at specifics in the return, such as JEM lines (Jet Engine Modulation) to identify aspects of the target parameters. The radar may attempt to measure RCS (Radar Cross Section) to estimate target type. It may plot the track to determine target speed. And the radar uses a steered beam or angle of arrival to measure the azimuth angle to the target.
However, while these radars often can steer the beam in azimuth, they typically cannot really control the take off angle of the beam in elevation (other than by frequency selection). And then they rely on the ionosphere to reflect the beam back down to the Earths surface in the desired target area. To make the beam "land" in the right geographic areas they dynamically select frequencies. They intentionally pick a frequency that will reflect off the ionosphere at the right geometry to illuminate the desired region. And this "correct" frequency changes constantly.
So these radars typically step up and down the bands all day long. Sometimes they set on a given frequency for many hours, sometimes they are on a frequency for just a few minutes. The point is you generally see them following the propagation trends. During the morning hours they typically move up in frequency, during the evening hours they move down.
And... and this is most likely a wild one... since most of their number-stations always broadcast numbers of 5-digit lengths (I got a video of 'the english man' spitting out a +20 minutes long sequence of 5-digit numbers recently), is it remotely possible they could be using the frequencies as a kind of 'number-station'?
In your video that would be these 8 groups of 5-digit numbers:
10454, 10852, 10940, 13932, 14740, 15073, 15881, 15957
It would beat the vox-transmissions for precision, where some individual spoken numbers may drown in noise and not be copied by the recipient. With such a method the 'number' would simply fade in and out but always be the same, and the signals are relatively easy to distinguish from simple narrow-band empty carriers so the chance of mistake is low I would think.
Yeah...no. First, that is a big signal to handle such a small bit of data. While an SDR shows it pretty clearly, using something like a portable receiver it can be hard to get an exact center frequency.
Some beams will set on a frequency for hours each day, others move every fewn minutes. And if you do not know what "number" to look for, how do you find the next and be sure it is the correct sequence number? Also, since the 29B6 works between 6000 kHz and 19999 kHz, the 5 digit groups would always start with a 0 or a 1.
No, there is nothing in the frequency selection habits to suggest the selection of frequencies has anything to do with a message.
T!
) and went 'channel-surfing' when I fell over this thing which fascinated me enough to record this video 
