Radio chatter mod for Kerbal Space Program that reacts dynamically to your flight and EVA.
Your crew talks to Mission Control and to each other about what is actually going on during the flight or EVA. A quiet coast in orbit stays mostly quiet. A capsule tumbling through space gets a crew shouting over each other while Mission Control tries to keep them calm. Probes get their own voice too in the form of beeping noises. This way your crew will react dynamically to your flight and EVA!
loñapsE is the language of the kerbals. Under the hood it is Spanish with every word spelled backwards ("Separación" becomes "nóicarapeS"), rendered with a text to speech voice and pitched up. It sounds like a real language without being one you can follow. If you are not fluent in loñapsE, the English subtitles show what your Kerbals are saying. They are on by default.
Each kerbal has their own voice based on their gender, and their own personality too! The personality is based on the Kerbal's stats as seen in the Astronaut Complex. For example, a Kerbal with high stupidity would make more jokes in flight, while one that has a low stupidity would make fewer jokes. The voice is generated based on a seed, and the seed stays the same even after restart. That means that the Kerbals will keep the voice that is unique to them.
This update focuses mostly on impoving the ambience with new sound effects. Additionally, it changes the way science is communicated by running the lines at runtime and puts new checks for docking in place.
Fall lines wait until the kerbal is calm. After a fall, the talk starts when the kerbal has not been tumbling for 1.5 s, and at the latest 12 s after the first fall. Before, it could start in the middle of a second fall.
Kerbals no longer sometimes sound very deep and slow. Each kerbal plays at their own pitch, pushed apart from the rest of the crew and lowered a little under G. For a low-voiced kerbal this could undo the whole Kerbal shift. It happened most on stations with many kerbals, in intercom chats and under high G. The pitch now has a floor, so every kerbal keeps the Kerbal sound. On a big station the voices sit a bit closer together because of this. G strain lowers the voice by at most 6% instead of 12%.
Tense and panicking kerbals no longer sound like chipmunks. Under stress the crew voice went up in pitch, in the recordings and again in game. That is gone. A tense or panicking kerbal now has their normal voice and only talks a bit faster.
Reset data gets no report. The talk waits until the results dialog is closed. An experiment you reset or discard there is left out. Data you keep or transmit still gets its report.
undock, dock_close and dock_callout, and three new events for on=, undock, dockclose and dockcallout. The lines are in Phrases/Runtime/undock.tsv and dock_close.tsv.science_talk: a kerbal and Mission Control after an experiment, in Phrases/Runtime/science_talk.tsv. New values {temp_c}, {press_kpa}, {grav_g}, {sci_flight} and {exp_runs}. See docs/API.md.Install
OR
That's all. Lonapse has no dependencies.
To update, delete the old GameData/Lonapse folder first, but keep GameData/Lonapse/PluginData if you want to keep your settings.
Settings
Click the antenna button on the toolbar in flight. There you can set how chatty the radio is, the volumes, which events Mission Control reacts to, the probe behaviour, the subtitles, time warp chatter and a bunch more just to your liking. Most of the new lines need the runtime voice, so keep "Speak names and new lines at runtime" in the General tab on.
Hotkeys:
You can change both keys in the window.
Settings are saved in GameData/Lonapse/PluginData/settings.cfg.
Compatibility
Known limits
Credits and license
Inspired by Chatterer. Lonapse contains none of its code or sounds.
The voices are spoken by Piper, a neural text to speech engine, with three Spanish voice models. Their credits are in the GitHub in audio/CREDITS.md. The beeps and static are synthesized.
I wrote the mod with help from Claude (an AI assistant) for code and documentation. The design, the testing in game and the decisions are my own.
The code is MIT licensed. The sounds and the phrase text are CC BY 4.0. The runtime voice keeps its own licences (MIT, GPL-3.0 for espeak-ng, and CC0, Apache-2.0 and the Unlicense for the voices). See LICENSE on GitHub.
Raw stats are from the beginning of time until now. Each follower and download entry represents one hour of data. Uneventful hours are omitted.