Writing

Feed Software, technology, sysadmin war stories, and more.

Saturday, June 8, 2013

Want to do some codec hacking? Does P25 TDMA sound tasty?

While I wait to see whether fred will be funded or not through Kickstarter, I've turned my attention to the P25 radio project. Every indication is that the South Bay's new radio system is going through tests at the manufacturer and will start being installed here next month. This pretty much means I have to figure out how to make sense of a trunking control channel.

It also means there's the matter of figuring out the new "phase II" TDMA scheme which keeps showing up in all of the documents. Unlike the other agencies which have rolled out ordinary "phase I" P25 where it's just one call per channel, this one chops the channel into two slots. First it blasts some data for slot 0, then slot 1, and so on.

There are tools which will handle some modulations which are similar. DSD in particular will make short work of things like DMR. The problem is that it doesn't support this new mode, probably because there is a lack of sample data.

Well, I'm trying to change that. I've gotten friends to plug in those $20 RTLSDR sticks in areas where agencies are supposedly running newer systems with this modulation. I've talked to people with USRPs who live in areas which supposedly have all-phase-II systems. After a few months of trying I think we finally have an actual call captured as raw data. DSD actually recognizes it as P25 (probably because of the same 5575... sync), but can't make sense of the rest of it. It shows up as some "unknown DUID".

So here's the thing. If there is anyone out there who is hacking on DSD or similar projects and who needs a phase II TDMA I/Q sample, I think I have a couple now. I can also cook them down to a WAV file if you'd rather deal with something which looks more like a discriminator tap output.

FFT of probable TDMA signal

Maybe you're the sort of person who's looking for a new sort of challenge. How about turning a blob of random data into voices? In theory, it's all documented. Just go to the IHS online web store and prepare to pay through the nose for each piece. Whatever your motivation, if you want a copy of this stuff, just let me know. Be sure to give me some way to respond to you. I can't reply without an e-mail address!

(Hey, SIM card person in SF, I need your contact info!)

I'll be plugging away at it in the meantime, but I make no guarantees as to whether I'll be able to make any headway. I'd much rather have a whole bunch of people trying to make sense of this instead of going it alone.

The world really needs a free decoder for this stuff.