SA: Automatical Parameters Detection. Update to Version 6.0.7.2

Signals Ananlyzer: Autodetection of Parameters.

Version 6.0.7.2

In version 6.0.7.2 function ADP - Autodetection of Parameters of a signal is announced . We added this function for the several reasons. In the first, it is well-known to that presence of this function will strongly please and attract some part of users:-). In the second, the task is extremely difficult and therefore is very interesting, well and in the third, there is the big desire to release an analyst from routine operations at the analysis.

At once should tell that the automation will not substitute and cannot make a competition to the skilled analyst. In this sense it is not necessary to create the big illusions about automation, there always will be the signals, which will demand an absolutely not standard approach and which will stump any automation. As well, since we only started development in the field of autodetection of parametres, it is not necessary to expect at once very amazing results, it is very serious sphere, and we have lot of work in this sphere. We have decided to start with small: From definition of speed of manipulation, as the speed of manipulation is one of the key moments in analysis.

Current implementation of ADP, works primitively enough, at the level of an average-skilled analyst, and solves only tasks of definition of the speed of manipulation. Further we plan to well expand quantity of auto-detected parametres and the reliability of definition. With this start implementation of ADP we designate one of main directions, in which we are going to work and develop SA.

It is simple enough to launch ADP: you only need select signal's area with horizontal and vertical markers. Horizontal markers set/limit the signal's bar by frequency, and vertical markers set the time-segment.

Certainly the interval of time should not be very short and or too long. In the first case it is not possible to receive exact value of frequency of manipulation or even impossible to measure it at all, in the second case too much time can be demanded. Standardly it is desirable to have the time-segment not less than 1 second and not more than 10-20sec, but depends of sampling rate and features of the signal, there is a lot of nuances.

Examples:

Some minor problems detected by users and beta-testers are solved. We would like to express much thanks for it to users and beta-testers.

I would like to remind once again, that ADP don't give 100% guarantees yet. Alhought ADP alghorytm successfully solves it's tasks in most cases on rather simple signals.

Good luck!

Signal : Clover 2000

Clover 2000
Clover 16p4a, mixed PSK-16 manipulation and four-lever AM at the same time

Author: SergUA6 6.0
Band Width ~ 2000 Hz
Low Range 500 Hz
Baud Rate 62.5 Гц, in the channel
n-Ary (PSK/MPSK) PSK-16
Count of Carriers 8 spectral bands with not crossed spectrum
Step between carriers 250 Hz between carriers of the channels
RX mode SSB

The very qualitative sample of Clover 2000

Sonograms:

pic.1 General view

Diagrams:

pic.2 Graphical spectrum

Pictures:

pic.3 Phase constellation

pic.4 Crossings in constellation

This signal is from Clover family of signals. It is rare enough signal on the air. For definition of parametres of the channel, it it is necessary preliminary to select it. The constellaions, which are represented in the description, concerned to the lower channel by the spectrum.

Usage of the mixed modulation, allows to increase speed in the channel, for this variant it turns out (if I'm not mistaken) 6*62.5=375 bit per second on the channel, the common speed is 3000 bps. That is not bad characteristic, at the 2000Hz band on HF.

UDXF Forum discussion:Unidentified signal on 8600.9kHz USB (France)‏

Unidentified signal on 8600.9kHz USB (France)‏.

http://f1jek.ns5-wistee.fr/mp3/8600.9khz.mp3
http://f1jek.ns5-wistee.fr/mp3/8600.9khz.mp3

After analysis we identifed that signal. It is Globe Wireless OFDM Waveform.

The quantity of channels is well visible from the picture, there are 16 channeles. Not each analyzer is able to show it on such Signals. The matter is that GW has the very short CP and short enough body of the signal, it does not allow to distinguish channels by usual analyzers. Manipulation mode in the channels for GW is standardly PSK-4 (QPSK). Speed of manipulation approximately is 55-56 Hz.

The record is in mp3, althought *.wav format is the best for analysis. For the ones who want to know more about how to make a good record for analysis i post the link on our short article with easy steps "Record for the following analysis".

It will be usefull to read.

Thanks to UDXF forum for interesting discussion and UDXF members for the signals.

PSK: Arcotel Farcos 1800 EADS/RACOMS psk-4, ALE

Arcotel Farcos 1800 EADS/RACOMS psk-4, ALE.

Author: SergUA6 6.0
Band Width ~2200-2400 Hz
Low Range 500-600 Hz
Baud Rate 1800 Hz
n-Ary PSK/MPSK PSK-4
Carrier frequency 1600 Hz is standard, in this example carrier frequency is 1648 Hz
RX mode SSB

The signal is here.

Sonograms:

pic.1 General view

pic.2 Speed

Pictures:
pic.3 Phase constellation

ARCOTEL FARCOS 1800 EADS/RACOMS - this signal is from the same series as ARCOTEL MAHRS 2400 EADS/RACOMS, but it has other PSK mode, a little bit shorter bursts and slower speed. The presence of header with psk-2 duration about 70 msec is characteristic.

PSK: Arcotel MAHRS 2400 EADS/RACOMS PSK-8, ALE

Arcotel MAHRS 2400 EADS/RACOMS
PSK-8, ALE

Author: SergUA6 6.0
Band Width ~2800-2900 Гц
Low Range 300-400 Гц
Baud Rate 2400 Гц
n-Ary PSK/MPSK PSK-8
Carrier frequency 1800 Гц
RX mode SSB

The signal is here

Sonograms:

pic.1 General view

pic.2 Manipulation speed

Pictures

pic.3 Phase constellation

ARCOTEL MAHRS 2400 EADS/RACOMS - It Sounds similar to one of LINK-11 variants, so it is possible to confuse these two signals.
At attentive studying it is visible that it is psk-8 serial. on such short signals it is difficult to receive the qualitative picture on the phase plane, but if to run whole file, that all is easily clarified. The presence of the short PSK-2 fragments in the begining of each pachcage with duration 60 m, is is typical.

SA updating to version 6.0.7.0

Version 6.0.7.0

In version 6.0.7.0 the number of radical updates is realized. One of the new feature is the ability of reffer as "read only" to resource, where the analyzed file is located.

All SA versions work only with a copy of the original record, it was made with the purpose to guarantee safity of original record, in case of force majeures. But, in old versions the copy had been created in the same place, where the original record was located. Thus, the direct work with the records, which are located on CD, flash disk or network resource was too slow or even impossible.

The problem is solved in the current update, now the copy is created in the own SA "Temp" directory.
Since, sometimes it is extremely desirable to see the full path to initial file, this path is specified in the Status Bar of the main window. The path to the initial file is specified for active window, or, for the window, over which the cursor is placed. In the working windows, the relative path to copies, with which SA works, is specified.

Another serious enough, but outwardly not very noticable changes:

• the algorithm of the graphics displaing/mapping in WF module is optimised: speed is increased in several times, lthought on the fast computers it may be not so noticable, but on the weak ones, especially in a fullscreen mode the difference is very essential.
• the operation/work of the semi-automatic corrector is optimised: the algorithm is refreshed. Earlier on the fragments, where the signal is absent the proofer simply stopped, now it is fixed. Precision of correction is increased.
• the various collisions which could be called by operations of user are removed. For example, earlier, in WF module or VMW the automatic increasing/reduction of the graphic's scale at moving of the cursor and pressed on mouse button at the same time, would led to uncontrolled process.
• some minor moments are solved.

Good luck.

MT-63 and some other non standard OFDM signals.

Non standard OFDM signals.

Being engaged into analysis of MT-63, there was found out several similar signals. As it has appeared, it is not such rarity. Before we will start to consider MT-63 in details, we will try to understand, basically, what are specifics and features of these signals .

Probably we wouldn't pay attention to this class of co -called pseudo OFDM signals, but examination of the developed module of the analysis of OFDM, has shown that there are signals, which are generated obviously not by standard technology, at least the analyzer is "slips" a little bit on the records, which quality seems to be exellent.

Besides, some discussions on our forum were linked to this question.

There are three signals which do not fit into the standard scheme of OFDM creation.

Probably, there is nothing unusual can be visible at once, but attempts to simulate Codan-16 by the standard technology do not lead to reception/getting of adequate model. The difference will be more obvious if to compare Codan-16 with Well-known CIS-12 tone, from the point of view of the main relations of space between channels and speed of manipulation, these signals have approximaly identical structure, and thus we can expect approximaly identical spectral allocations. However distinctions are very essential.

It is well visible that Codan-16 has too obvious almost complete partition of channels, while the classical methods do not provide it. CIS-12 tone has the typical result of the classical OFDM creation , and the overwhelming majority of cases OFDM signals look same.

It is known that, sometimes the difficult enought methods are used at creation of OFDM.
For exmaple the methods of :

• reducing of out-of-band radiations
• increasing of the spectral density
• reducing of the peak-factor

But all these methods are not used untill complete partition of the channels, besides all, these methods do not distort the signals to that point, then the reverse demodulation by the classical ways becomes impossible, on the contrary, these methods are aimed to provide compatibility.

An obvious conclusion - these signals are created by other way. It is not the big secret or strong innovation, but nevertheless, it can cause complexities at the analysis, and basically, it can even to lead to incorrect results of the analysis.

Signals, which are similar to DIGTRX and Codan-16 are formed by FDM technology. Let's not go deep into details, but in general, the sense of FDM technology is: that the prototype of the one channel with the obviously established form is created, and then on the basis of this prototype, the grid of working channels is formed.

Practically the spacing can be any, orthogonality is not necessary.

It allows to reach more density of usage of the spectrum, in comparison with the classical multi-channel (not OFDM) systems.

We consider the task of analysis, not only as definition of parametres of a signal for the purpose of its identification, but also with the purpose to understand what is necessary for the signal's demodulation.

In this sense, the conclusions for Codan-16 and DIGTRX that these signals are the classical OFDM signals certainly is not true. These signals do not have CP, which "is easily calculated" from the relations of speed and spacing, they have not orthogonal grid of frequencies, and it is impossible to demodulate these signals by the classical OFDM approaches. From the point of view of analysis there should not be vaguenesses in the signal at the analysis. It is necessary to know all features of the analyzed signal and understand why these features are placed in the signal. Ofcourse it is not always possible, but one of the main goals of analysis is the full detalization of the signal and it's nuances.

Let's come back now to MT-63. Principles of creation, which were described above, had no success in case with MT-63. The signal is obviously formed by the standard OFDM technology, but the result is not standard. Fast enough we have ascertain that the signal is divided/separated on two groups of the channels: even and odd. These groups have shift/displacement relatively each other on half-clock/tact of frequency of manipulation.

That is, as the variant, by the standard resources of OFDM technology, two independent streams of even channels and odd accordingly, are forming. CP is added in each stream, and both these streams are merged into one. It is possible to provide the shift/displacement on half-clock/tact both at creation of streams and at their merging, as the result we receive pseudo OFDM, where according to all signs/characteristics, it is classical OFDM, but all attempts to demodulate it by the standard for OFDM methods are failing.

It is not easy to get to know how it is possible to devide/separate these streams at reception/recieveing by simple methods. Ofcourse there are solutions, but it turns out difficult enough.

Besides, the sense of such creation is not really obvious, because there are no particular advantages before the classics methods, at least we had not found them.

That would be interesting to get to know: why is there the necessity of such shift/displacement of even and odd channels? And that it gives basically? Except doubtful reducing of the peak-factor and obvious, considerable complicating of the demodulator.