RTL-SDR LSM trunking via OP25: audio sample

[KA1RBI]

The recording (linked below) was taken from a local P25 LSM system using OP25 with a $20 RTL SDR stick. No other hardware (besides the PC itself) was used. OP25 is Free Software.

It is interesting to note that the stock vertical antenna supplied with the RTL was used, and that it was just randomly placed on top of the PC, and wasn't moved around to find a "sweet spot".

The trunking features of OP25 were utilised; specifically one TGID was whitelisted, suppressing all other activity in the system. The selected talkgroup was for a competition being held this weekend at a local ski area.

As in the previous sample the empty gaps between conversations are skipped in this WAV file. The file is approx. 1MB.

The WAV file is at
http://www.lightlink.com/mhp/rtlsdr.wav

 

[mtindor]

If I ever get approved in op25-dev [which you suggested I join, even though I'm not a dev], I'll be asking some questions regarding how to get this going. I can get the OP25 block running, but I don't think the standard OP25 block is all I need. Unfortunately, I live in a rural area and there is no source of near-constant P25 traffic let along a constant source of LSM-specific traffic. Makes it every difficult to even know if OP25 is working properly since P25 transmissions are few and far between and spread out among a couple dozen frequencies].

Hopefully one day I'll be able to provide my own sample audio out from monitoring an LSM system.

Mike

 

[br0adband]

So now the fun begins, so to speak. Guess I'll muck around with OP25 and see how I fare in the coming days. For those that aren't aware, I had sent KA1RBI a rather lengthy PM (sorry!) asking if it would be worth the time and effort to get OP25 functional using just an RTL stick for monitoring a trunked P25 system. As I've stated before, I have hardly any P25 systems here in the Las Vegas area but the Nellis AFB system is a candidate and I was looking for a way to do it with an RTL stick specifically. I hadn't seen much info about GNURadio and monitoring analog trunking systems (still can't find anything of any use so I suppose that's not going to happen, guess using Windows is the only solution at this time) so I asked about some aspects of OP25 that could prove appealing, primarily the ability to hold on a given talkgroup specifically.

Being able to scan a trunked system is fine but if you can't actually follow the trunks or hold on a talkgroup then it's like having a traditional bank/channel scanner and putting all the voice channels in and hitting Scan - sure you hear some comms but in terms of following a particular conversation, forget it.

Since OP25 is P25 specific, I'm still somewhat stuck in most respects for monitoring a trunked analog system without a lot of "stuff" - it's entirely possible to do, sure, but it requires both my RTL sticks to get it done and I tend to not use both at the same time (one for Unitrunker to handle the control channel, the other for tuning the voice channels per info supplied by the decoding of that control channel).

I'm simply going to guess that we'll probably never see a "single app solution" for doing analog trunk tracking with SDR and a single RTL stick - it seems like it's possible but given the need to handle not only the control channel itself but then decoding the sub-channel audio data when on a voice channel so as to keep things functioning smoothly I don't think it's going to come about anytime soon if ever.

Would be damned nice, however...

Thanks for the audio sample, KA1RBI, will put some time into OP25 in the next few days and see what comes of it.

 

[SCPD]

Hi BB;

Here are two things to consider - one economic and one technical.

I'm simply going to guess that we'll probably never see a "single app solution" for doing analog trunk tracking with SDR and a single RTL stick - it seems like it's possible but given the need to handle not only the control channel itself but then decoding the sub-channel audio data when on a voice channel so as to keep things functioning smoothly I don't think it's going to come about anytime soon if ever.

Economic: the desire for two VFOs from the same SDR makes sense when that SDR costs hundreds or thousands of dollars and you have the beefy hardware to run it. The RTL sticks go for $10 USD if you don't mind waiting three weeks. If you bought one RTL stick, you should have bought two (or three).

Technical: many systems' channel set spans more than 2.5 Mhz. An RTL supporting two VFOs (one voice and one control) must run at a higher sample rate. This requires more CPU. You'll get better CPU performance with two sticks instead of one because you can run each RTL stick at a lower sample rate.

 

[br0adband]

You're right, Rick, completely dead on accurate and even in spite of me actually having two RTL sticks (as you recommend and I've said since I got into this RTL/SDR kick about 2 months ago) I still don't use them both at the same time with any consistency. I suppose I'm still hoping to be able to "scan" and monitor trunked systems more efficiently (and yes I'm aware that even the simplest trunk tracking scanner can do it effortlessly, I just don't own one presently), that's all.

I'm not knocking any software that exists currently, especially not Unitrunker since there's pretty much absolutely nothing even remotely as capable as that is and you're still making progress in most every respect which is damned awesome and again I thank you for all your contributions to this hobby over the years and into the future.

SDR# (my preferred program) doesn't "hit" hard on performance on this older laptop of mine (Core 2 Duo P8700 @ 2.53 GHz) even with two copies of it + two copies of Unitrunker + DSD+ doing its thing all working at the same time I barely see 50% CPU usage most of the time, occasionally spikes higher but a steady ~50% or so.

There are times when I really had spent a better part of my past experience with computers focusing on actual programming and software development aka being a coder myself; if I had the knowledge and the expertise I think I'd focus on creating something more akin to what I'm looking for myself, and then if I did manage to pull it off I'd see if anyone else might find it useful too.

Well, let's see how this experience with OP25 works out...

 

[SCPD]

Whether they realize it or not, this is the direction op25 is headed ...

Super Trunking Scanner

She's using a USRP to record everything - with just enough control channel decoding to tag the conversations. That's one VFO for every channel.

 

[KA1RBI]

Whether they realize it or not, this is the direction op25 is headed ...

Super Trunking Scanner

She's using a USRP to record everything - with just enough control channel decoding to tag the conversations. That's one VFO for every channel.

Direction op25 is headed ?

Simultaneous multi-channel RX has been in OP25 for years-

RepeaterPage

 

[br0adband]

Yah, I saw that several weeks ago thinking "Ok, here's what I've been looking for, a proper trunking capable SDR program suite..." but after looking at what it does it turns out that no, it's not what I want.

That basically just records an entire trunking system at once, every frequency, every transmission, etc and yes tagged as well if necessary with talkgroup/RID info, date/time/etc. Crazy serious stuff there that's way beyond what I would like to do but pretty amazing considering.

But the thing is, it's not using a VFO for every channel, it's capturing the entire baseband audio of the entire spectrum the SDR hardware is capable of at the same time, pulling out the particular frequencies (control and voice) for content, decoding as required, tagging, then saving the info as audio files (since it's not digital capable yet, it's strictly for analog systems as I understand from reading the documentation).

I honestly admit that till about a week ago I had never used the option in SDR# to record the baseband spectrum - it's not the same as sending baseband audio to Unitrunker or DSD+ or DSD, it's vastly different: it records the entire spectrum at the same time however wide that happens to be. Since I use 1 MHz for my spectrum (1,024 MSPS) I did a 5 minute recording a few days ago that ended up being over a gig (about 1.3GB) in size but when I load that recording (an *.iq file) back into SDR# I can then "replay" everything that happened during that 5 minute period of time by tuning on the activity.

I felt stupid never realizing this capability in the ~2 months I've been at this, and now I use it quite often and it's pretty amazing that it works as well as it does. It's not just a recording of a given frequency, it's everything in that 1 MHz (or however large your spectrum sampling is set for) that happens for as long as the recording is in progress. Pretty damned amazing.

Wonder how large such recordings would be on devices like Airspy with a 10 MHz spectrum bandwidth or even larger ones. The recordings for using that Super Trunking Scanner setup must get into the multi-terabyte range pretty fast.

Edit:

Well, based on what KA1RBI just posted, and after reading that documention for RepeaterPage it appears you specify frequencies for capturing/decoding/recording purposes, not everything in the entire spectrum which would obviously be a lot less space. 10-20 frequencies instead of the whole spectrum, yeah, that'll cut down the recording sizes considerably...

 

[PiccoIntegra]

Super Trunking Scanner

She's using a USRP to record everything - with just enough control channel decoding to tag the conversations. That's one VFO for every channel.

That's all fine and dandy for analog, but wait until they convert to digital. She's already mentioned on her blog they were moving to Phase 2 in the future. It won't be quite so easy to accomplish. She's real smart though, and has the right hardware, I can't wait to see what she comes up with.

 

[rachelbythebay]

The recordings for using that Super Trunking Scanner setup must get into the multi-terabyte range pretty fast.

Hi br0adband,

You might be pleasantly surprised. There's no multi-TB-anything in the system in terms of storage space. It's a flood of I/Q data from my USRP, sure, but it doesn't wind up being stored that way. Figure the input data is about 120 Mbps (64000000 / 18 sample rate, 2 channels, 2 bytes per channel). It's all sliced and diced and turns into a series of much smaller streams: the control channel turns into a series of OSW strings via gr-smartnet and is parsed by my own home-brewed controller.

The audio channels on the system are all analog, so they go through a FM demod stage to get WAV type data, but then immediately go straight into libmp3lame to become MP3s. Only then are they written to disk. Given the audio itself is pretty limited, the resulting files are pretty small. The whole set of audio files is only about 49 GB at the moment.

I've had it running for about two and a half years now, and it's recorded just under 2.5 million calls. The production database for all of that metadata makes about a 150 MB SQL dump -- hardly anything.

Random bit of trivia: the loopback device on my recording system has its counters up around 333 TB as I write this, and it's been up for about 150 days. That's because part of my system uses a split-process model: one talks to the USRP and serves up the data over TCP, and the other consumes it and does all of the above control channel and audio channel stuff. It makes it much easier to try other things without affecting regular operations.

 

[br0adband]

Wow, that's some impressive stuff. And I take it you do this just for yourself, 'cause you love knowing what's going on after it's happened?

 

[tylerwatt12]

This is really awesome. I've never even seen this site. But it has a striking resemblance to mine

SDRSCAN.COM

She has features in her site, I wish I had in mine.

 

[SCPD]

Simultaneous multi-channel RX has been in OP25 for years

Multi-channel, yes - but that was conventional only. Support for control channel information is relatively new.

Now - if anyone wants to argue over who built the first trunked radio logging recorder - I've got you all beat. But that's another story.

She's already mentioned on her blog they were moving to Phase 2 in the future. It won't be quite so easy to accomplish. She's real smart though, and has the right hardware, I can't wait to see what she comes up with.

She's already familiar with Gnu Radio. She can download op25 and hack/slash away to make it do what she wants. Me thinks she'll do just fine.

 

[rachelbythebay]

Wow, that's some impressive stuff. And I take it you do this just for yourself, 'cause you love knowing what's going on after it's happened?

I got into it after watching a helicopter go around and around and around one night while SCPD units swarmed the area, and I had no idea what was going on. I owned an ordinary scanner which would more or less work (modulo rebanding), but of course by the time I hear the chopper, the context is long gone. You have to back up to the original dispatches or calls for fills or whatever to get some idea of why the entire police force just descended on your neighborhood.

It took a long time to go from just an idea to something that other people could enjoy via the web site. It wouldn't have been possible without GNU Radio or gr-smartnet for that matter.

 

[rachelbythebay]

She can download op25 and hack/slash away to make it do what she wants. Me thinks she'll do just fine.

I've been trying my best to help the op25 folks nail down the whole phase II thing over the past year or so. Things moved up a fair amount after I got in touch with someone who both owned a USRP and lived within range of a real production system (something up in Maryland, if I remember correctly). They graciously recorded nice long raw I/Q data captures in the middle of the band with sufficient bandwidth to capture the entire system. The bigger capture is 6 MHz and 643 MB in bzip2 format, so you can imagine how big it is in raw uncompressed form.

I've been offering up this data to anyone who wants to try their hand at hacking up phase II support. I'd love to see a bunch of software appear which can support the new symbol rate and codec and all of that stuff, and that's my way of trying to help.

That goes for anyone reading this who wants to try: send me a private message and I'll send you the URLs and details of the capture files.

 

[br0adband]

I got into it after watching a helicopter go around and around and around one night while SCPD units swarmed the area, and I had no idea what was going on. I owned an ordinary scanner which would more or less work (modulo rebanding), but of course by the time I hear the chopper, the context is long gone. You have to back up to the original dispatches or calls for fills or whatever to get some idea of why the entire police force just descended on your neighborhood.

It took a long time to go from just an idea to something that other people could enjoy via the web site. It wouldn't have been possible without GNU Radio or gr-smartnet for that matter.

Was being a bit facetious with my comment so I apologize. The work you've done is truly impressive stuff, I wish I had even a small inkling of the talent put into it, and if I did I'd be all over that Phase II side of things myself trying to help out in any way possible.

I'm sure in time we'll have decoders that can handle Phase II just as readily as we can handle Phase I and X2-TDMA, on all platforms. I'm still considering a go with Linux for GNURadio and OP25 here just so I can learn something in the process which I'm sure will be a lot of fun (ok, now I'm being facetious).

Regardless, thanks for your efforts and for sharing them with everyone.