Adding DCS

CMakeLists.txt

@@ -256,6 +256,7 @@ list(APPEND trunk_recorder_sources
   trunk-recorder/formatter.cc
   trunk-recorder/source.cc
   trunk-recorder/call_conventional.cc
+  trunk-recorder/tone_manager.cc
   trunk-recorder/systems/p25_trunking.cc
   trunk-recorder/systems/p25_parser.cc
   trunk-recorder/systems/smartnet_parser.cc
@@ -289,6 +290,8 @@ list(APPEND trunk_recorder_sources
   #lib/gr-latency/latency_tagger.cc
   #lib/gr-latency-manager/lib/latency_manager_impl.cc
   #lib/gr-latency-manager/lib/tag_to_msg_impl.cc
+  trunk-recorder/gr_blocks/ctcss_squelch_ff.cc
+  trunk-recorder/gr_blocks/dcs_squelch_ff.cc
   trunk-recorder/gr_blocks/freq_xlating_fft_filter.cc
   trunk-recorder/gr_blocks/transmission_sink.cc
   trunk-recorder/gr_blocks/decoders/fsync_decode.cc

README.md

@@ -64,6 +64,8 @@ RTL-SDR dongles; HackRF; Ettus USRP B200, B210, B205; BladeRF; Airspy; SDRplay
 
 ## Setup
 * [Configuring a system](docs/CONFIGURE.md)
+* [Adjusting Squelch for Conventional Systems](docs/SQUELCH.md)
+* [CTCSS / DCS Tone Squelch (analog conventional)](docs/TONE-SQUELCH.md)
 * [Uploading to OpenMHz](./docs/OpenMHz.md)
 * [FAQ](docs/FAQ.md)
 

docs/CONFIGURE.md

@@ -852,7 +852,7 @@ This file allows for you to specify additional information about conventional ch
 |-------------|----------|-------|
 | TG Number     | ✔️        | The Talkgroup Number formatted as a decimal number. This has to be the first column |
 | Frequency        |  ✔️       | The frequency in MHz or Hz for the channel (decimal point must be used for MHz) |
-| Tone |        | The CTCSS Tone for the talkgroup. |
+| Tone |        | Tone-squelch spec. Empty / `0` = no tone. A positive number = CTCSS frequency in Hz (e.g. `118.8`). `D<code><N\|I>` = DCS, with 3-digit octal code and polarity (e.g. `D023N`, `D131I`). `S` = search mode (record everything, identify the tone in the call metadata). Multiple rows at the same Frequency become one shared recorder routed by detected tone — see [TONE-SQUELCH.md](TONE-SQUELCH.md) for the full feature, multi-row freq groups, and the SKIPPED behavior for non-matching tones. |
 | Alpha Tag |       | A 16 character description that is intended as a shortened display on radio displays |
 | Description |   | A longer description of the talkgroup  |
 | Category |    |  The category for the Talkgroup |
@@ -868,6 +868,11 @@ A **Header Row** is required for the file, with a header provided for each of th
 | --------- | --------- | -------- | ------------- | ---------------------- | ------ | ------ | ------------------- | ---- | ---- |
 | 300       | 462275000 | 94.8  | Town A Police | Town A Police Dispatch | Police | Town A |    |  false |  |
 | 325       | 462275000 | 151.4 | Town B DPW    | Town B Trash Dispatch  | DPW    | Town B | false   |  |  -50 |
+| 410       | 155857500 | D023N | County Fire   | County Fire Ops        | Fire   | County |    |    |  -48 |
+| 510       | 155250000 | S     | FD Paging     | Fire Paging Out        | Fire   | Paging |    |    |  -48 |
+
+Multiple rows can share the same `Frequency` to route one shared recorder to
+different talkgroups by detected tone — see [TONE-SQUELCH.md](TONE-SQUELCH.md).
 
 
 ## unitTagsFile

docs/SQUELCH.md

@@ -18,6 +18,8 @@ The **Signal Detector** automatically tries to find a power threshold to separat
 
 Each Conventional Recorder also has a Squelch setting. Once a Conventional Recorder has been enabled and is receiving samples from the source, it will use a Squelch block and only process samples above a certain power level. This prevent noise from being recorded. The Squelch value needs to be hand tuned, and can be set for the entire Conventional System, or per channel.
 
+For analog channels you can also configure a **sub-audible tone squelch** (CTCSS or DCS) per channel, gate audio on a specific tone, or share one recorder across multiple talkgroups on the same frequency. See [TONE-SQUELCH.md](TONE-SQUELCH.md).
+
 
 ## Configuring a Conventional System
 

docs/TONE-SQUELCH.md

@@ -0,0 +1,192 @@
+# Tone Squelch (CTCSS / DCS)
+
+Analog conventional channels often use a sub-audible tone or digital code to share
+one RF frequency between multiple talkgroups. Only receivers configured for that
+specific tone open audio; everything else stays muted even though they're all
+listening to the same carrier. Trunk-recorder supports the same idea via the
+**Tone** column in the `channelFile` CSV.
+
+Three tone families are supported:
+
+| Family | What it is | Tone column value |
+|--------|-----------|-------------------|
+| **None** | No tone; record any signal that breaks squelch | empty, `0`, or `0.0` |
+| **CTCSS** | Continuous Tone-Coded Squelch System — a sub-audible sine tone (67–254 Hz) | The frequency in Hz, e.g. `118.8`, `151.4` |
+| **DCS** | Digital Coded Squelch (a.k.a. DPL/Motorola, CDCSS) — a 134.4 baud Golay codeword | `D` + 3-digit octal code + polarity, e.g. `D023N`, `D131I` |
+| **Search** | Detect the tone but don't gate on it — report whatever's on air | `S` or `s` |
+
+## Modes in detail
+
+### `0` / empty — no tone
+
+The recorder records everything the squelch lets through, like a stock
+conventional channel before this feature existed. No tone detection runs.
+
+### CTCSS (e.g. `118.8`)
+
+Enter the tone frequency in Hz as a number, e.g. `118.8`. trunk-recorder
+recognizes the full standard PL set from 67.0 Hz up through 254.1 Hz.
+Only transmissions carrying that tone will be recorded.
+
+### DCS (e.g. `D023N`, `D131I`)
+
+Enter the 3-digit code (with the leading zero — `D023N`, not `D23N`)
+followed by `N` or `I` to match what your code list shows. You'll usually
+see `N`; some older radios use `I`. If you're not sure, try `N` first.
+
+trunk-recorder is forgiving on the `N` vs `I` suffix and will record the
+channel regardless of which one the transmitter actually uses.
+
+One thing worth knowing: DCS has a small number of codes that look
+identical over the air — for example, a `D023N` channel will also pick up
+`D047I` transmissions, because the two signals are indistinguishable to
+any receiver. This is a quirk of the DCS protocol itself, not specific to
+trunk-recorder. The pairs are rare and almost never both used in the same
+geographic area, but if you ever see unexpected recordings on a DCS
+channel, this might be why.
+
+### Search (`S`)
+
+The recorder records everything the squelch lets through (no gate), but both
+the CTCSS and DCS detectors run side-chain. At end of call, the
+identified tone is written to the JSON sidecar and the call log — useful for
+discovering what tones are actually in use on a channel before configuring
+specific channels for them.
+
+## End-of-call JSON fields
+
+Every call's `*.json` sidecar gains three flat fields:
+
+```json
+{
+  "tone_mode":       "ctcss",     // "off" | "ctcss" | "dcs" | "search"
+  "tone_detected":   "118.8",     // "" | "<Hz>" | "D<code><N|I>" | "D<a>/D<b>"
+  "tone_confidence": 0.957        // 0.0 – 1.0
+}
+```
+
+- `tone_mode` — what the recorder was *configured* to do for this channel.
+- `tone_detected` — what was actually heard. Empty when nothing was
+  identified. In search mode, when a DCS transmission is one of the
+  ambiguous pairs (see the DCS section above), both possibilities are
+  reported joined with `/`, e.g. `D023N/D047I`.
+- `tone_confidence` — for CTCSS, this is the spectral concentration; for DCS
+  it's phase diversity × match rate. Use it to threshold consumers downstream.
+
+For non-analog (P25/DMR/digital) recorders the fields are still emitted but
+will always be `"off"` / `""` / `0.0`.
+
+## Multi-row freq groups
+
+A common operator scenario: one frequency carries three or more separate
+talkgroup labels, each gated by a different tone. Express this directly by
+writing one CSV row per (TG, tone) pair, *sharing the same Frequency*:
+
+```csv
+TG Number,Frequency,Tone,Squelch,Alpha Tag,Description,Tag,Category,Enable
+203,154325000,D223N,-48,County Fairfield,County Fairfield,County Fire,Fire,true
+204,154325000,151.4,-48,County Eaton,County Eaton,County Fire,Fire,true
+205,154325000,118.8,-48,County North,County North,County Fire,Fire,true
+```
+
+Trunk-recorder collapses all rows at the same frequency into a **single
+analog recorder running in search mode**, then routes each captured call to
+the matching row by detected tone. This is much cheaper than three separate
+recorders sharing an antenna, and it correctly produces three independent
+"call streams" tagged with the right talkgroup, alpha tag, and category.
+
+Setup behavior:
+
+- The first row at a freq is the **primary** — its `Squelch` and
+  `Signal Detector` settings drive the shared recorder.
+- All other rows are **alternates** — only their TG/Alpha/Description/Tag/
+  Category metadata is used (their per-row Squelch values are ignored).
+- A row with `Tone=0`/empty or `Tone=S` in a group acts as a **catch-all**
+  for that frequency: any tone (or no tone) that doesn't match a specific
+  alternate ends up routed to it.
+
+### What happens when no row matches: SKIPPED
+
+If the detected tone doesn't match any row's `Tone` value **and** there's no
+catch-all row in the group, the call is concluded as **SKIPPED**:
+
+- No WAV files are kept.
+- No plugins fire (no upload, no stream, no log emission to external systems).
+- A single log line records what was dropped and why:
+
+  ```
+  [system]  17C   TG: <primary>  Freq: 154.325 MHz  SKIPPED — detected tone '127.3' did not match any allowed row
+  ```
+
+This makes the multi-row config a **strict allow-list**, not a catch-all.
+Use it when you specifically want to ignore traffic on a frequency that
+isn't from one of your configured groups.
+
+If you want the more permissive "record everything, label what we can"
+behavior, add a single `Tone=S` row to the group — it acts as a fallback
+bucket that catches non-matching transmissions.
+
+## Examples
+
+### Single channel with a CTCSS tone
+
+```csv
+TG Number,Frequency,Tone,Alpha Tag,Description,Tag,Category
+100,154265000,151.4,FD Disp,Fire Dispatch,Fire,County
+```
+
+### Single channel with a DCS code
+
+```csv
+TG Number,Frequency,Tone,Alpha Tag,Description,Tag,Category
+106,155857500,D162N,FD South,Fire South Ops,Fire,County
+```
+
+### Search-only channel (discover tones)
+
+```csv
+TG Number,Frequency,Tone,Alpha Tag,Description,Tag,Category
+109,155250000,S,FD Paging,Fire Paging Out,Fire,Paging
+```
+
+Every call on this channel concludes with `tone_detected` populated; no calls
+are dropped.
+
+### Multi-row freq group, strict allow-list
+
+```csv
+TG Number,Frequency,Tone,Alpha Tag,Description,Tag,Category
+203,154325000,D223N,County A,County A Fire,Fire,County
+204,154325000,151.4,County B,County B Fire,Fire,County
+205,154325000,118.8,County C,County C Fire,Fire,County
+```
+
+Three TGs share 154.325 MHz; only D223N, 151.4 Hz CTCSS, and 118.8 Hz CTCSS
+make it to disk. Anything else gets SKIPPED.
+
+### Multi-row freq group, with catch-all
+
+```csv
+TG Number,Frequency,Tone,Alpha Tag,Description,Tag,Category
+203,154325000,D223N,County A,County A Fire,Fire,County
+204,154325000,151.4,County B,County B Fire,Fire,County
+299,154325000,S,County Misc,County Other Traffic,Fire,County
+```
+
+D223N and 151.4 Hz are labelled specifically; everything else lands on
+TG 299 with the detected tone in the sidecar.
+
+## Tips
+
+- The CTCSS detector typically locks within 200–400 ms of carrier; DCS
+  needs at least one full codeword (~170 ms) plus a confirming repeat,
+  so very short keyups (< 250 ms) may still record but with empty
+  `tone_detected`.
+- For DCS, don't sweat the `N`/`I` suffix — pick whichever your code list
+  shows (usually `N`). trunk-recorder will record either polarity.
+- The end-of-call diagnostic verdict from each detector is emitted at
+  `debug` log level; turn it on (`"logLevel": "debug"` in config) when
+  tuning a new channel.
+- The recorder still runs all of trunk-recorder's normal squelch and
+  signal-detection logic. Tone gating is *additional* — it's evaluated
+  on audio that has already passed the squelch.

docs/intro.md

@@ -51,6 +51,8 @@ RTL-SDR dongles; HackRF; Ettus USRP B200, B210, B205; BladeRF; Airspy
 
 ## Setup
 * [Configuring a system](docs/CONFIGURE.md)
+* [Adjusting Squelch for Conventional Systems](docs/SQUELCH.md)
+* [CTCSS / DCS Tone Squelch (analog conventional)](docs/TONE-SQUELCH.md)
 * [FAQ](docs/FAQ.md)
 
 ### Troubleshooting
@@ -62,6 +64,6 @@ For those not familiar, trunking systems allow a large number of user groups to
 
 Most trunking system types (such as SmartNet and P25) set aside one of the radio frequencies as a "control channel" that manages and broadcasts radio frequency assignments. When someone presses the Push to Talk button on their radio, the radio sends a message to the system which then assigns a voice frequency and broadcasts a Channel Grant message about it on the control channel. This lets the radio know what frequency to transmit on and tells other radios set to the same talkgroup to listen.
 
-In order to follow all of the transmissions, Trunk Recorder constantly listens to and decodes the control channel. When a frequency is granted to a talkgroup, Trunk Recorder creates a monitoring process which decodes the portion of the radio spectrum for that frequency from the SDR that is already pulling it in.
+In order to follow all the transmissions, Trunk Recorder constantly listens to and decodes the control channel. When a frequency is granted to a talkgroup, Trunk Recorder creates a monitoring process which decodes the portion of the radio spectrum for that frequency from the SDR that is already pulling it in.
 
 No message is transmitted on the control channel when a conversation on a talkgroup is over. The monitoring process keeps track of transmissions and if there has been no activity for a specified period, it ends the recording.