Tetra decoding

[henrygb]

i made a minor update to osmo-tetra-sq5bpf to show the LA (Location Area) parameter, which is useful to distingiush cells. you can see it in tetra-rx (receiver1 etc) output, and telive.log if you enable alldump.

after enabling alldump in telive.log you will see something like:
20141227 22:48:09 FUNC:NETINFO1 CCODE:02 MCC:0104 MNC:xxxx DLF:39xxxxxxx ULF:38xxxxxxx LA:2 RX:1
which means that the first receiver (the one that reads from /tmp/fifo1) got a broadcast from network with colour code 2 and LA 2

What is the best procedure to install this update ? just copy the new files in the directory structure ?
Not everyone is good friends with linux . I try to be my best friend linux

 

[sq5bpf]

What is the best procedure to install this update ? just copy the new files in the directory structure ?
Not everyone is good friends with linux . I try to be my best friend linux

this has nothing to do with linux, but with git (which is used also on other operating systems, including windows). read the manual to see how to use git pull (like git pull https://gitbub.com/REPOSITORY master).

or just delete the old tree, clone and recompile again

 

[henrygb]

Thanks for your answer, delete the directory was my first option, but when reinstall, so many errors.
Now using git pull in the directory, and just update the project. So easy when you know.
In the past I did run telive_1ch, did copy the information and group the frequencies with same DL freq, CCCode, and MNC. Now with this update we can see this information in multichannel mode

Still learning and reading

 

[giannit]

Hello , in your opinion, could the system work on asus netbook with Atom processor that does not use more but given the size it would be handy to have below. Has anyone tried already tried to install on atom , possibly linux distro which I could use to exploit it even with only one channel at a time ?

 

[racingfan360]

>could the system work on asus netbook with Atom processor
I have the 1ch configuration running on a 2core asus 1001 atom with the max 2GB memory and Mint 17.1. My main driver is portability to survey other TETRA systems away from base. Performance is nowhere near as good as a modern desktop machine which easily copes with 4ch confg.

It also looks like hansheesbeen is running on an Atom: interested to know how this performs?
http://forums.radioreference.com/di...ware/302347-tetra-decoding-3.html#post2295721

Also interested to know of ways of maximizing the decode performance from anyone (eg I don't run the scope in 1ch simple).

 

[RMW1010]

look at some graphics acceleration settings etc, and try to disable them

No luck, abandoned the Virtual Machine for the time being.

try to disable the spectrum display, this way it eats less cpu. if it is still bad, then mute the audio, and play the resulting ogg files automatically (this isn't exactly realtime, but the delay is a few minutes) - i've posted a script to do this a few posts ago.

also look at memory usage, maybe you could kill a few unneeded processes to free some?

Yeah, the Centrino is definitely to slow. It's running all the time at 100% CPU load. Stopping the flowgraph is a bit better, but not much.

So I tried on a laptop with an i5 processor and that works great with live audio!

 

[sq5bpf]

one last thing that you could try is to lower the sample rate. below 2M the antialiasing filter in the rtl2832 chip doesn't work well, but it might work well enough for you (is you have some aliasing problems, then maybe you can pick an offset to take the alias signal out of the output).

i know people have tried 1000k, 1024k, 256k with some succes, check some groups regarding rtl-sdr for suggestions what might be a good sample rate.

 

[woodpecker]

Happy new year all

Just starting to try and build, initially I tried Ubuntu 14.04 in VMware workstation 7.1.5 under Windows7 but had many problems just getting Ubuntu installed.

Now switched to Virtualbox, Ubuntu 14.04 32 bit installed ok, currently building gnuradio.

Would also like to try and get this running using a VM on a windows 8 tablet, which is the best route to go?

 

[woodpecker]

Up and running in Virtualbox with Ubuntu 32 14.04, seeing lots of data but voice is all broken, have very strong tetra signals here so not a signal issue, how do I debug garbled audio?

 

[firefather_sdr]

Garbled audio could be encrypted tetra. Look into receiver-terminal, if Air encryption 1 shows up.
Also you could test if the acelp-file play the word tetra.

 

[woodpecker]

Garbled audio could be encrypted tetra. Look into receiver-terminal, if Air encryption 1 shows up.
Also you could test if the acelp-file play the word tetra.

Looks like Air Encryption is showing 1, I just assumed it would mute if encrypted, everything may be encrypted here, do networks mix encrypted and non encrypted traffic or should I tune to something else?

How do I test the acelp-file?

 

[firefather_sdr]

Looks like Air Encryption is showing 1, I just assumed it would mute if encrypted, everything may be encrypted here, do networks mix encrypted and non encrypted traffic or should I tune to something else?

How do I test the acelp-file?

Try tovlisten along time with activate mutessi and record. It could be possible that traffic is mixed, but i don't belive.
For test acelp-file, please read the Dokumentation. I mean it works with tetraplay.

 

[woodpecker]

Try tovlisten along time with activate mutessi and record. It could be possible that traffic is mixed, but i don't belive.
For test acelp-file, please read the Dokumentation. I mean it works with tetraplay.

The testfile plays ok so I guess its the encryption, maybe nothing to listen to...

 

[racingfan360]

>How do I test the acelp-file?

/tetra/bin/tplay < testfile.acelp
The testfile.acelp is available from here: https://github.com/sq5bpf/telive

>everything may be encrypted here
I'm guessing you are trying to listen to Airwave.....in which case all users are encrypted. You'll also see that the SSI data is suppressed. You should get all the network info though using 'a' - alldump

 

[sq5bpf]

It could be possible that traffic is mixed, but i don't belive.

actually there are many networks with mixed encrypted and unencrypted traffic

 

[woodpecker]

>How do I test the acelp-file?

/tetra/bin/tplay < testfile.acelp
The testfile.acelp is available from here: https://github.com/sq5bpf/telive

>everything may be encrypted here
I'm guessing you are trying to listen to Airwave.....in which case all users are encrypted. You'll also see that the SSI data is suppressed. You should get all the network info though using 'a' - alldump

The airwave network is the strongest, someone here reported they had received unencrypted fire comms though, are all airwave users encrypted or just police?

I can also see a much weaker network around 450.650 MHz, struggling to get decode though, lots of 'could not find successive burst training sequence' messages.

When it does decode it still appears it may be encrypted though

 

[racingfan360]

>are all airwave users encrypted or just police?
They all are - blue lights and the numerous other gov users too

>received unencrypted fire comms though
That was the Northern Ireland Fire Service: TETRA radios but not Airwave

>a much weaker network around 450.650 MHz
On that freq I'd guess it was a prison. It won;t be Airwave, but don't know about what encryption is used. I'd be interested to see what network data you can decode though (post here or maybe PM me direct?)

 

[woodpecker]

On that freq I'd guess it was a prison. It won;t be Airwave, but don't know about what encryption is used. I'd be interested to see what network data you can decode though (post here or maybe PM me direct?)

It could be Bedford prison, its strongest with the beam pointing that direction.

I can see 4 tetra signals, 450.550, 450.650, slightly weaker 450.675 and 450.725

 

[sm0vec]

Now I've also tested with NUC with Lubuntu and it worked a lot better than with Ubuntu thansk to the lighter GUI. I can't decode 4 channels, but now I can decode 2 channels without being on the limit as before. With two channels and no graph running, the CPU load is about 73% and the CPU temperature is just under 50C which is almost no difference to idle. The fan have to do a bit more work, but running at 3200 rpm it's not easy to detect.

One thing I noticed was that the "Enhanced Intel SpeedStep" should be ticked in the bios. Without this the decoding lose sync sometimes. When this is ticked it might looks like the CPU is running slower than the spec when it's probed by cpu info, however I believe that the presented value is not fully true, and maybe the application creates a lot of bursts which the SpeedStep (including Turbo Boost) handles better (up to 2,4 GHz) than when running the CPU on the steady specified speed (2,16 GHz)

So the cheapest Intel NUC BOXDN2820FYKH with Lubuntu is a good box for decoding two channel systems.

It remains to compare the power consumption between the different NUC's. I'll get back to that in a couple of weeks. Now it's time for Christmas!

SQ5BPF; Weso&#322;ych &#346;wi&#261;t.

A small update, I've now played more with the gnuradio settings and by lowering the FFT refresh rate to 20 and only running 1,4 MHz sample rate (which fits all 4 channels of the 4 channels sites - spread over 1,125 MHz), and stopping the FFT, I can also use this cheap NUC without problems (even though being quite close to the limit).

So with NUC worth 30 USD and a cheap RTL stick it's possible to cover 4 channel sites within 1,125 MHz.

 

[arroja]

Meanwhile, on Spain´s Ebay, some douchbag is trying to sell something very similar to Jacek´s work for a botload of euros:

TETRA Prototipo receptor experimental decodifica voz | eBay

I wonder if anyone is stupid enough to place a bid.

Doesnt look like Jackec´s software and it´s based on the old Funcube, which by itself, it´s a pain in the neck to use.

Dig?