Clarification on speaker consent and commercial TTS use — Kikuyu dataset
#27
by bobbybob2026 - opened
Hello WAXAL Team,
Thank you for making the WAXAL dataset available, particularly the Kikuyu (Gĩkũyũ) TTS recordings.
I'm working on a project to develop a natural-sounding Gĩkũyũ text-to-speech voice, primarily for narrating traditional stories, proverbs and educational content on a YouTube channel that may be monetised.
I've reviewed the dataset documentation and understand that the Kikuyu recordings, provided by Loud and Clear, are licensed under CC BY-SA 4.0.
Before proceeding, I'd appreciate clarification on the following:
- Did the original speakers provide informed consent for their recordings to be used to train synthetic speech or voice-cloning models?
- Does that consent extend to generating new speech in a recognisable speaker's voice, including for commercial or monetised content?
- Are there any additional restrictions, attribution requirements or permissions required from Loud and Clear or the individual speakers?
- Is there a contact person at Loud and Clear who could provide written authorisation if necessary?
I want to ensure that the project respects the speakers' rights and complies fully with the applicable licensing and consent requirements.
Many thanks for your guidance and for supporting African language technologies.
Bob