Ratten Chatter | Nuclear Option

Ambient radio chatter for Nuclear Option, so the cockpit never feels dead quiet.

Around eighty short conversations play quietly in the background while you fly — check-ins, intercepts, SEAD calls, tanker joins, carrier approaches, a nine-line from a forward air controller, a wingman who will not hold formation, and the odd release authorisation you would rather not have overheard.

Every one is two people and two or three lines: two aircraft, or an aircraft and a tower, or an aircraft and mission control. One call, one answer, maybe a closing word. Short enough that you catch it without stopping what you were doing, and never a crowd scene that pulls your attention out of the cockpit.

Every callsign is its own person — nearly fifty of them, each a different real voice: American, Canadian, English, Scottish, Irish, Welsh, Australian, South African and Indian, men and women, cloned locally with Chatterbox from the studio recordings of the CSTR VCTK Corpus. They breathe, hesitate and sound like somebody keying a mic rather than a screen reader. Every voice goes through a full radio treatment: band-limited to a comms channel, compressed flat by the transmitter's level control, dirtied with clipping, and buried in Rauschen — a steady open-carrier hiss with a crackle of pops and ticks riding on top, opening a fraction of a second before the mic click and closing with a squelch tail. An AWACS with a real antenna sounds nothing like a forward air controller on a handheld or a wingman at the edge of the map.

It is mixed to sit behind the aircraft rather than in front of it. Turn Volume up if you want the net closer.

Pacing comes from the script rather than a fixed timer, so a snap "Two." lands straight away while "stand by" leaves an actual gap. Conversations are drawn from a shuffle bag — you hear all of them before any repeats — and now and then the net stays busy and a second exchange starts on top of the last one instead of after the usual silence, which is what a genuinely busy net sounds like.

Chatter holds off while something else is talking on the interface channel, so it never steps on a voice warning from the game or from another mod, and it pauses partway through an exchange rather than talking over one.

Extracted from MKMods into its own plugin so you can run the chatter without the rest of that mod.

Installation

1. Install BepInEx 5 into the Nuclear Option game folder and launch the game once so it generates its folders.

2. Copy the RattenChatter folder from the release into Nuclear Option/BepInEx/plugins/, so you end up with:

   BepInEx/plugins/RattenChatter/RattenChatter.dll
   BepInEx/plugins/RattenChatter/Assets/chatter/*.mp3
   BepInEx/plugins/RattenChatter/Assets/chatter/manifest.txt
   

3. Launch the game. Config is written to BepInEx/config/dev.sewerlabs.rattenchatter.cfg.

Clips load in the background over the first few seconds after startup rather than blocking the loading screen, so the net may be quiet for a moment on the first flight.

Configuration

Settings live under 1. General (master switch, volume) and 2. Timing (pauses between conversations) and are editable in-game with ConfigurationManager.

Setting	Default	Description
Enable mod	true	Master switch. Turning it on mid-game needs a restart; turning it off works immediately.
Volume	1.2	How loud the chatter is. 1 matches the game's other cockpit sounds; the clips are already mixed low, so raise this to bring the net forward.
Shortest pause (seconds)	12	Minimum silence between conversations.
Longest pause (seconds)	45	Maximum silence between conversations.
Busy radio chance	0.2	Chance the next conversation starts within a few seconds instead of after the usual pause, so traffic comes in waves. Set to 0 for evenly spaced chatter.

Adding your own chatter

The shipped voices are generated, and the whole pipeline is in tools/ — scripts, cast, text-to-speech and the radio filter chain. See tools/README.md. Adding a conversation is a handful of lines of dialogue plus one command, and that is also where the knobs for loudness and how much Rauschen sits under the voice live.

If you would rather drop in your own recordings, the plugin reads clips two ways.

With a manifest (what the generator writes) — Assets/chatter/manifest.txt lists one transmission per row as conversation|file|gap, where gap is the seconds of dead air before that line. Row order is playback order:

chatter07|chatter07a.mp3|0.00
chatter07|chatter07b.mp3|1.40
chatter07|chatter07c.mp3|0.70

Without one — delete manifest.txt and clips are grouped by file name: everything but the last character is the conversation key, the last character orders the lines, and gaps are randomised between 0.8 and 1.8 seconds.

chatter13a.mp3   ->  conversation "chatter13", line 1
chatter13b.mp3   ->  conversation "chatter13", line 2

Either way the plugin goes quiet for a random time between Shortest pause and Longest pause between conversations and works through all of them before repeating.

Building

Requires the .NET SDK and a local Nuclear Option install (the game assemblies are referenced directly).

dotnet build RattenChatter.sln -c Release

The default game directory is Y:\SteamLibrary\steamapps\common\Nuclear Option (set in RattenChatter/RattenChatter.csproj). Override it with:

dotnet build RattenChatter.sln -c Release -p:GameDirectory="D:\Steam\steamapps\common\Nuclear Option"

A successful build deploys the DLL and assets straight into <GameDirectory>/BepInEx/plugins/RattenChatter.

Credits

Radio chatter code, scripts and audio pipeline by Lukas Struckl. Voices are generated with Chatterbox by Resemble AI (MIT licence), cloning speakers of the CSTR VCTK Corpus (CC BY 4.0): Yamagishi, J., Veaux, C., MacDonald, K. (2019), CSTR VCTK Corpus: English Multi-speaker Corpus for CSTR Voice Cloning Toolkit (version 0.92), University of Edinburgh, The Centre for Speech Technology Research, doi:10.7488/ds/2645. The lines are synthetic — none of the VCTK speakers recorded them — and every clip carries Resemble AI's inaudible Perth watermark. The edge-tts fallback uses Microsoft Edge neural voices via edge-tts. All of it is put through an ffmpeg radio chain. Built on BepInEx and HarmonyX. Originally part of MKMods, a fork of mkualquiera/nuclear_option_modding.