Disclaimer: I already know what I'm doing, no tutorials needed, please

Background: I want to monitor only my own town's dispatches. It's hard to do without a way to decode the Plectron two-tone call my town uses. Otherwise I have to listen to the 36 other towns and agencies (not exaggerating!!!) that are dispatched by a commercial dispatch center that all of these towns contract to.

Right now it looks like the go-to solution is to buy a used Uniden BCT15 off of eBay for about $200 and be done with it.

I'm surprised I have not seen something in the RTLSDR hackerspace, but no joy there so far.

Looking for any innovative and cheap methods. I have plenty of old scanners and RTLSDRs I can put to use.

Methods I have already considered and discarded:

- Dredge up enough old Plectron parts to make one work on the right freq's/tones; discarded because it would become a labor of love and require too many hours of searching and repairing.
- Can't use an old Minitor because no reeds for Plectron tones. Same issue with it being a labor of love.
- A Midian decoder board is $125 and some assembly required Although I am familiar with them from projects in days gone by, I might as well just get the scanner for $200.
- Don't know how to code, don't want to learn, so unless somebody has an Arduino project all done, right and tight, not interested.

So, now at the bottom of the barrel, anything down there? Any cheap Chinese handhelds that could be programmed?

Anytone HTs can be setup to break squelch for tones. The 868 was $150 a few months ago.

There's tons of older QCII capable commercial radios.  Get one you can program and have a dedicated receiver.

Radius GM300 or mactrac would get it done.

You could get a basic LMR handheld that has two-tone decoding for under $200.

Example:
https://www.theantennafarm.com/catalog/index.php?main_page=product_info&cPath=89_112_160_176&products_id=3169

Homebrew (or possibly find on ebay or somewhere) a 12v battery eliminator and you'd be set.

As I would expect you to, you've already arrived at the optimal solution. You can spent more on newer Unidens as most will have that feature.

It would be relatively trivial to implement this as a simple filter for a RTL-SDR, but someone has to write it. It's basically just a simple pipe filter with a couple goertzel tone detectors controlling the feed-through logic.

868: good option, under consideration, thanks!

GM300: will it decode Plectron tones, or only QC2? I'm guessing only the latter since it's a Motorola.

F3001: same question. Will it do Plectron tones?

"Plectron" is just a (/\/\) brand name for 2-tone signaling isn't it? I don't have a lot of experience with programming 2-tone stuff.

It will decode QCII which I suspect is equivalent what you are referring to as Plectron.  We do QCII here and back in the day people used these big old plectron receivers.  I don't think "Plectron" is a signaling standard though. I could be wrong.

I had a tube type Plectron that was tuned to 33.90, the fire dispatch frequency. Damn that was a long time ago.

Looked it up, QC2 and Plectron are both two-tone paging "protocols" if you can call them that. They work the same but aren't the same as the possible tone sets are pretty different, though you can find a close-enough QC2 tone to match most but not all Plectron tones.

The tone timings are different as well, QC2 apparrently uses 1/3 (one second tone followed by three seconds of a different tone) while Plectron has a choice of a single 3-second tone, a 3-second tone followed by a 0.25 second tone, or a 0.75 second tone followed by a 0.25 second tone.

I suspect pretty much anything scanner-related will just react to two-tone pair in a short enough time frame.

seek2 is giving you guys the tutorial I didn't need

The problem is that the Plectron tones in question are 2361Hz (first tone in this case) and 701Hz. These are only Plectron tones and not available in the Moto Quick Call II scheme, although you can get close to 701.

You can see the tone charts here.

What's interesting is that these are being transmitted using the QCII timing scheme and not the Plectron timing scheme, so I have that going for me.

I do know from reading both the Uniden and the Anytone manuals that you can set any tone freq's you want in 1Hz increments. Plus both the Uniden and the Anytone implementations are relatively tone duration insensitive. So both of those approaches should work.

What is not clear is whether the GM300 or the F3001 are the same. I'd need somebody familiar with programming them to chime in before pulling the trigger on either. I'm particularly suspicious of a Motorola solution entertaining other than a strict implementation of QCII.

Have you looked at Two Tone Detect?  Not sure where to download the latest software package anymore though.

They want $70 for this. And I have to provision a PC along with the scanner. It didn't seem worth it.

Wow, not a good solution at all.

Also using a sledgehammer in place of an eyeglass screwdriver.

I remember way long ago someone make a speaker with a two-tone paging decoder built in that you could plug into simple FM scanners. I did a quick search and didn't find anything.

I did discover someone sort of copying the flightaware model that offers software called twotonedetect -- it turns around and grabs the tones and audio and pushes it back out on the internet to clients running on phones. Radioreference has an absolutely massive thread on it.

Uniden scanner still wins.

The F3001 allows any tone frequency to the tenth of Hz along with user-defined tone lengths. I think all the Icoms are that way, pretty much all of them support 2-tone and 5-tone signalling. What they won't do (as best I can tell) is support long single tones as some other stuff will.

It's not really the FlightAware model. Those who contribute feeds to FlightAware get a free "Enterprise" level account. I should know, since I'm one of those geeky bastid's Ultimately you must have a $300/year subscription to use twotonedetect.

Uniden still wins!

Great info, this now becomes another viable option, thanks!

Holy crap. Getting paid too much while other people do your heavy lifting, must be nice.

Yeah, Uniden wins! You'd be $100 ahead in just the first year with no PC or software hassles.

That didn't use to be the case with TTD but then again it has been a few years since I have used it.

Anything Motorola is going to be limited to QCII. No one offs.

So if Plectron is unique and it sounds like it is my idea is off the table.  Sorry.

I ended up with a Unication pager at work... let me look into what it can do after the kids go to bed here...

$400, plus $60 for the programmer...

It is certainly not the cheapest option but it works quite well.  Realistically if you are only looking for tone out alert, a uniden base scanner will probably be your best option.

Definitely a step above the old Minitors!

Also Scanner Master has the BCT15X for $169.95!

About par for that scanner right now -- must be getting closed out?
HRO has it for the same price.

To aa's goal, it does support two-tone for under $200 and doesn't require dealing with ebay, so I'd say it's a winner.

Can you get me some IQ recordings of what it is you're after.

Why would you need IQ data? This is a simple, well understood, audio tone recognition problem. Links to relevant (but expensive) software are given above that demonstrate the desired functionality.

Well because I write dsp software for sdrs and was going to see if I could knock out a decoder for use with something like and RTL.

I have referenced the links you posted, that's why I ask. This shouldn't be too hard of a problem to solve. Buffer samples, apply fm demod, window the ttl of the first possible tone, slide it in the frequency domain, if tone 1 is hit, chuck the next time of samples for tone 2, if tone 2 is hit, perform FM demod of remaining samples and return to user.

This could be done with scipy/pandas/numpy.

* eta, to answer the "why" is because there many standards referenced and would have like to confirm it suites your desired work case.

This processing can occur at the baseband audio stage, IF stage processing really isn't required.

There's plenty of exemplar audio available. If that would help I can PM you the specific dispatch center link on broadcastify.com.

Interestingly, this would not even need to be an SDR app per se, just a simple audio processing app. If you built an app that worked with a broadcastify feed that would be brilliant. People would flock to it in droves. You could be rich and famous Here's how it would work:

1. Open the .asx stream file that broadcastify provides (an example of this would be to open it in VLC player, or Media Player, etc.).
2. Monitor for tones (preferentially a list of tones or tone pairs so that more than one alert could be recognized, but maybe just start with a single pair of tones).
3. Upon recognizing a valid tone or tone pair, open the squelch for a user-programmed length of time (e.g. 30 sec, 60 sec, etc.) and perhaps have an option to leave squelch open until manually reset.
4. Optionally record audio for that same length of time, make the filename as a time-date group (e.g. 191406JUN21, etc.), or perhaps provide some fancier method of labeling.
5. Optionally have some user readable log of detections (ultimately a database, perhaps).

If you did not want to handle the feed part, you could do this as a VLC plug-in, or perhaps as an Audacity plug-in (in this latter case it would be up to the user to feed Audacity an audio stream, but that's trivial, in fact that's how I identified the Plectron tone pair I needed).

The only tricky part is that when you look at the tone charts you will see that you need to implement a tone detector that has a very, very high Q because the tones are closely spaced.

Yeah the audio streams could work you just loose sensitivity. As far as high q goes Ideally you just run a high FFT on the demod audio and if you get what you think is a detection you can decimate that chunk in the frequency domain, run a higher FFT and get as accurate detection as you have CPU. Talking sub 24khz about anything should be able to run a 128k fft.

I'll look at the streams and see what I see. Obviously good, fast, cheap (any 2) come into play.

Since you only care about two tones and have a 0.25s detection window, I'd just run Goertzel filters rather than a full FFT.

Your IM inbox is full. Sent you an email.

Awesome??????

Yeah I got your email. I guess my inbox didn't shirk since I let my membership expire.

I was poking around github and saw this.

https://github.com/syastrov/twotonedecoder


I might fork it and update it to py3 and qt5.

Good find!

I need this to run on Windows, although I don't mind having to install Python if that's required.