VOZ AI PUBLIC DEMO: MEDIA CREDITS

Original: WIKITONGUES: Omar Speaking English and Jamaican Patois
Creator: Wikitongues
Source: https://commons.wikimedia.org/wiki/File:WIKITONGUES-_Omar_Speaking_English_and_Jamaican_Patois.webm
Original publication: March 2014
License: Creative Commons Attribution 3.0 Unported (CC BY 3.0)
https://creativecommons.org/licenses/by/3.0/

Rubix Labs uses the opening 31.24 seconds. The footage was trimmed, resized
to 1280x720 and re-encoded. Audio levels were adjusted. The English version
retains the original speaker's recorded voice. The Spanish adaptation replaces
the English audio with translated synthetic speech on the same video timeline.
The speaker's appearance, mouth movements and voice were not cloned or altered.

The excerpt, poster, transcript, translated captions and Spanish media adaptation
are provided under CC BY 3.0 with the original attribution retained. This applies
to these media assets, not the website code. No endorsement by Omar or Wikitongues
is implied. The footage is not a Rubix client recording or testimonial.

Spanish speech: Piper (https://github.com/OHF-Voice/piper1-gpl), using
es_ES-davefx-medium. Its model card identifies a CC0 dataset and fine-tuning
from the lessac medium voice:
https://huggingface.co/rhasspy/piper-voices/blob/main/es/es_ES/davefx/medium/MODEL_CARD

Transcription assistance and independent audio checks: faster-whisper,
Systran/faster-whisper-small. Translation and timing prepared for this demo.
The revised Spanish dub uses individually timed phrases and native synthesis
pacing, without slowing down the finished recording. Captions follow the dub.
The browser plays prepared files. It does not upload, transcribe, translate,
clone a voice, lip-sync or render a new video. This is not output from the
original Voz AI models and is not an accuracy or audio-quality benchmark.