Skip to content

NOTICES.md

Latest commit

 

History

History
77 lines (58 loc) · 3.32 KB

NOTICES.md

File metadata and controls

77 lines (58 loc) · 3.32 KB
Raw

Third-Party Notices

GMRS-TTY is licensed under the MIT License (see LICENSE). It depends on and/or redistributes the following third-party components, each under its own license. This file is provided to satisfy attribution requirements and to make redistribution terms transparent.

Runtime dependencies (installed via pip, not redistributed)

Package License Project
PySide6 LGPL v3 / commercial https://wiki.qt.io/Qt_for_Python
sounddevice MIT https://github.com/spatialaudio/python-sounddevice
soundfile BSD-3-Clause https://github.com/bastibe/python-soundfile
numpy BSD-3-Clause https://numpy.org/
piper-tts MIT https://github.com/rhasspy/piper
faster-whisper MIT https://github.com/SYSTRAN/faster-whisper
silero-vad MIT https://github.com/snakers4/silero-vad
noisereduce MIT https://github.com/timsainb/noisereduce
scipy BSD-3-Clause https://scipy.org/
jiwer Apache-2.0 https://github.com/jitsi/jiwer
pyserial BSD-3-Clause https://github.com/pyserial/pyserial

These are installed at the user's site via pip and are not bundled in this repository. Each retains its own license terms.

Build / script dependencies (not runtime, not bundled)

Package License Project
reportlab BSD-3-Clause https://www.reportlab.com/

reportlab is used only by scripts/build_user_manual.py to generate the PDF user manual. It is not required to run the application.

Bundled voice models (Voices/)

The .onnx voice models in Voices/ are pre-trained Piper TTS voices redistributed from the rhasspy/piper-voices collection on Hugging Face. Each voice has its own model card and license — please consult the upstream source for authoritative terms.

Upstream: https://huggingface.co/rhasspy/piper-voices

Voice file Dataset Upstream license (as published) Attribution required
en_US-amy-medium.onnx Amy MIT No
en_US-arctic-medium.onnx CMU Arctic Public domain / BSD-style (CMU Arctic terms) No
en_US-lessac-high.onnx Blizzard Challenge 2013 (Lessac) BSD-3-Clause (per upstream model card) Recommended
en_US-libritts-high.onnx LibriTTS CC BY 4.0 Yes
en_US-ryan-high.onnx Ryan MIT No

Attribution for CC BY 4.0 voices

The LibriTTS-derived voice (en_US-libritts-high) is distributed under the Creative Commons Attribution 4.0 International license. If you redistribute or use this voice, you must:

  • Credit the original creators of the LibriTTS dataset (Zen et al., "LibriTTS: A Corpus Derived from LibriSpeech for Text-to-Speech")
  • Indicate that the voice model was trained on LibriTTS
  • Link to the license: https://creativecommons.org/licenses/by/4.0/
  • Indicate if changes were made

The Piper project provides per-voice MODEL_CARD files at: https://huggingface.co/rhasspy/piper-voices/tree/main/en/en_US

Models downloaded at runtime

faster-whisper downloads OpenAI Whisper model weights on first run (default: small.en). Whisper model weights are released by OpenAI under the MIT License. See: https://github.com/openai/whisper

silero-vad downloads its VAD model on first run, released under the MIT License. See: https://github.com/snakers4/silero-vad

These models are not redistributed by this repository.