Difference between revisions of "Decoders"
From The RadioReference Wiki
m (updated links after target pages were renamed)
m (updated categories)
|Line 96:||Line 96:|
Revision as of 13:23, 16 August 2015
There are many kinds of data decoders available ranging from public domain packages to professional dedicated units. Prices vary from free up to very expensive and price is dependent on how much you want to be able to decode and what tools are available for signal analysis and identification. Public domain packages, while good, can not compete with the capability provided by the more expensive decoding software. It is safe to say that price goes up with increased capability in this market - be prepared to spend some big money if you want to cover a lot of modes.
The big push these days is in the field of soundcard-based software, or those that requires a USB dongle to work (The Hoka 3-32P and 300-32P are examples). There are many price levels as well as operating systems – from Java-based to Windows, even a few for the IPad. Many are free, but if you get into the USD1000 market and up, you find Wavecom's W-Code, Krypto500, gotoMonitor and GX430 from Rohde-Schwartz.
What kind of tools are available on these kilobuck-level decoders? These include
- Signal Identification (FSK, MFSK, PSK)
- Accurate baud rate measurement
- Correlation Bit Analysis (ACF)
- Variety in modes decoded/identified
- Ability to save captured text (disk and printer)
- Tools for analysis, signal decomposition and demodulation
You can't beat a good Signal Identification Mode; Wavecom and Hoka products, along with Krypto500 include this option. A good Signal Identification mode simplifies the task of figuring out what mode is currently tuned, but keep in mind that even the best identification mode is not always 100% correct.
A common problem affecting these automated tools is that some keying systems share common idle characteristics; active traffic is needed to correctly identify the exact mode. Also the presence of local interference, various propagation effects, or a noisy signal can make it difficult to correctly or reliably identify.
Somewhat surprisingly, experienced listeners of digital systems will largely be able to outperform these tools, at least in the initial classification of signal type (FSK, MFSK, PSK, ARQ, FEC etc), speed and even some other characteristics of the signal.
Baud Rate Measurement
Accurate measurement of baud rate is another vital capability. Many modes can be accurately identified on baud rate alone because many rates are unique to a keying system. It also provides the opportunity to "fingerprint" a signal, system or the user. For example, the Hoka decoders can measure baudrate accurately to 3 decimal places in the presence of a quality signal but also do well on marginal signals, eventually settling down on a reasonable measurement. If your signal is full of noise you might not see 3 decimal places but at least on Hoka decoders you will have displayed those decimal places that make sense - a very nice feature.
Autocorrelation Bit and ACF Analysis
Autocorrelation Bit is a technique that samples the incoming digitized bit stream and presents the data as a graph of bit occurrences plotted against time. This will show when patterns occur within a signal, allowing you to determine the number of bits in a character frame (this is commonly referred to as the ACF), giving you another piece of information when determining an unidentified system. This kind of analysis tool reveals cycle period and shows when there are NO patterns in a signal indicating an encrypted or psuedo-random bit-masked signal, allowing you to move quickly onto more productive signals. Hoka, Krypto500, go2 and Wavecom decoders include autocorrelation bit modules.
There is nothing more frustrating than being able to receive a clean signal and then not being able to identify or decode it (ignoring the problem of encrypted signals for the moment). The more expensive the decoder, the more likely it will have a large selection of modes.
Bear in mind however, that the huge selection of modes offered by the professional decoders is now merely a relic of times gone by, many systems having been long replaced by faster, more modern systems. On some rare occasions, organizations may test old, now antiquated backup systems but most of these modes have long disappeared into the history books.
Ability to save captured text (disk and printer)
The ability to save decoded output to a file and/or the printer should be considered a very important feature of any decoder. Having some form of hard copy, on disk preferably, allows for archiving for later reference or later analysis and independent printing and editing.
Tools for Analysis
If you are interested in going beyond the Identification and decoding of signals heard on the air you are going to need tools. Tools such as Spectrum Analyzers, Character Analysis and Phase Analysis modules are some of the necessary tools needed to analyze today's modern systems. This is, for obvious reasons, not for everyone.
Analysis on the Cheap
Up to a few years ago, there were dedicated hobby-level hardware-based terminal decoders, but they've largely been abandoned. Most all of the modes covered by these units nowadays are either amateur radio specific (i.e. Kantronics) or support has been discontinued by the manufacturer (the Universal M-7000 and M-8000, for example). Some of these units had the ability to 'guess' at the mode being used; the stronger the signal, the more the chance that the guess was correct. Unfortunately this method hasn't kept up with the development of newer modes. There's a link to a PDF file on the Universal website that lists all the digital terminal units that were sold by Universal (made by DES and Info-Tech), along with some technical specifications, in the HF Appendix.
In some cases, the frequencies used by modes like HFDL and, to a certain extent, ALE, are either well known and/or widely published. This is, of course, but one clue in identifying a user. It’s worth also mentioning that for many organziations using HF, old habits die hard. Operating procedures, frequencies and other aspects of a network remain as equipment is upgraded. This was seen in the 1990s for example, as many networks moved from manually-managed RTTY modes to MIL-STD-188-141A/B-based automated networks. In many cases, station identifiers and channels were preserved.
A few programs (notably fldigi and MultiPSK) use a method called Reed Solomon Identification, or RS ID. Briefly, it attempts to identify a transmission based on a series of tones that equate to an unique identifier. See this page for a more detailed description. This method is pretty limited, simply because not every mode we can find in the non-amateur bands has been assigned an identifier.
Many digital decoding packages have the ability to display an audio representation of the signal being received. This ability comes by many names; sonogram, waterfall display or audio spectrum display, just to name a few. This article (PDF) describes in some detail how to use a sonogram and how to record the signal. Both are important - you may need to play the signal back several times before you get a clear picture of what the signal looks like. The accuracy of the sonogram is going to be greatly influences by many factors; the quality of the signal being received, your receiving setup and the PC you are using will all play a role here.
Now what about identifying the patterns shown in the sonogram?
The links found in the HF Appendix have pictures of sonograms (and in some cases audio samples) of numerous kinds of digital signals. You can use these as a guide, but remember, it's another clue as to the mode (and by extension, sometimes the potential user) of the signal.
What About the Other Less Expensive Software?
There are some software packages that contain tools that can do at least some of what the more expensive decoders can in terms of analysis. This information comes directly from that package's website;
- RS ID (as described above), FFT Scan and Waterfall
- MultiMode (for Macs)
- FFT / Spectrum Analyzer and Bit Analysis
- Spectrum Analysis, Oscilloscope, Audio Analyzer
- Audio Display and Waterfall
- DSP Filters and Analysis, RS ID (as described above)
- Demodulated Bit Stream to ASCII file, screen or hexadecimal output
- WinRadio's Advanced Digital Suite
- Signal Classifier, Audio Oscilloscope, and Spectrum Analyzer with Waterfall
- WinRadio's Universal FSK decoder
- Spectrum analyzer, character analysis, can estimate baud rates (with some modes) and a bit inversion parser.
In addition, a package called Signal Analyzer has the ability to analyze digital signals with great complexity and variety. At under USD100, it looks like it would be a good option if one doesn't want to be paying for the multi kilobuck packages. There's an introduction article (PDF) on how to set up this software on the UDXF website. It’s sole downside is that it relies on offline analysis and cannot be used on live audio streams.
There are a number of packages that display audio waveforms, and sometimes, manipulate the incoming signal. Programs like Audacity, Baudline and even SR5 (which, despite the description on the website, works just fine on receivers other than the AOR-5000) all can do varying levels of analysis. You're not going to get all the complex functionality such as autocorrelation, but you might find these tools to be useful. It's up to the user to decide how deep into the analysis game s/he wants to go.
These programs, and several others, are linked on the Utility Monitoring Central software page (see the HF Appendix for the link).
Many of these packages do, however, enable a user to measure the frequency placement, spacing and number of tones or channels that make up an unknown signal and determine offsets and center frequencies. These are some of the most important signal parameters used when reporting signals so that others may accurately tune to and decode them in future.