Question regarding synthetic voice consent for academic research

#4
by rec1219 - opened

Hello,

I'm a master's student conducting a cybersecurity capstone on detecting synthetic Spanish-English code-switched speech.

I'm using the CS-FLEURS READ dataset and would like to know whether the participants consented to having their recordings used as voice references for synthetic speech generation with XTTS-v2.

My proposed use is strictly noncommercial academic research, comparing genuine recordings with synthetic recordings using the same speaker identity and transcript.

Does the dataset's consent and licensing permit this type of voice cloning, or would separate permission be required?

Thank you for your time!

Owner

Hi Rocio - Thank you for your interest in our dataset. Please feel free to use it for your non-commercial purposes. At this time there is not a separate consent required for particular usages, but note that individual contributors may ask for their data to be withdrawn at their discretion.

Sign up or log in to comment