TXSoundGen

WIP! This code is not yet fully functional. Low level components are relatively complete, but

there is not a functional client for usage.

Sound pack generator for EdgeTX/OpenTX radios.

Unhappy with downloading sound packs for your radio only to find they are incomplete? TXSoundgen uses self-hosted and cloud text-to-speech generation to create comprehensive voice packs to suit your needs, giving you a consistent voice pack that covers default sounds as well as any custom lines you would like to add.

Technical details

All audio files are generated as 16kHz single-channel .wav files, ensuring compatibility with all EdgeTX / OpenTX radios.

Currently supported speech engines:

• Amazon Polly
• Piper

Requirements

• Python 3.14
• Poetry
• ffmpeg

Planned UX

Create a custom voicepack configuration:

> txsoundgen init --provider piper

Voicepack config created from template at 'my-voicepack.yaml'

Build your custom voicepack:

> txsoudgen build my-voicepack.yaml

Building voicepack 'my-voicepack' using Piper TTS.
[####################-------------------------------------] [57/236]

Completed building voicepacks.
my-voicepack: ./voicepacks/my-voicepack.zip

Future capabilities

• CLI tool for building voicepacks.
• Packaging as docker container.
• Remove ffmpeg requirement and use native Python method for wave conversion.
• Proper webpage with audio examples.
• Web repository of existing sound packs.
• Generate default 'system' data from EdgeTX repository automatically.
• Cache provider responses, speeding up generation where sound data already exists locally.
• Multiprocessing.
• Support for more TTS services.
• Multilingual support.

Notes

https://pypi.org/project/diff-cover/ https://docs.astral.sh/ruff/integrations/ https://github.com/EdgeTX/edgetx-sdcard-sounds https://docs.astral.sh/uv/guides/integration/pre-commit/