Voice and Likeness Rights
A voice points to a specific person; get consent you can show before using it
Related
The problem
Imitating a singer's voice with AI is no longer technically hard. The difficulty is not resemblance but reference. A voice identifies a specific person, and most legal systems protect it as a personality interest, while some US states add a right of publicity and Europe tends to work through personality and data protection. The names, scope and duration differ. There is no single line.
How it works
Separate three things. Melody and lyrics are a work, owned by the author. A particular performance belongs to the performer. A recording belongs to its producer. The voice itself is a person's identifier and sits outside the work system, so owning the song is not the same as being allowed to clone the singer. Protection for public figures is often narrower, and the line between imitation, parody and endorsement is drawn by local rules and by the specific use. Risk peaks when a listener is led to believe the wrong person made it.
Practice
When cloning or imitating a real voice, get written consent naming the use, media, territory, term, exclusivity, withdrawal and payment. Keep the original recording and the signed record, with names and dates. Log each version you use and where it was published. Label synthetic output as synthetic, especially in advertising, news and education. Never use a real voice for political messaging, sexual content, fraud or customer-service impersonation. If your own voice is taken, preserve evidence first, screenshots, audio, links and timestamps, then use platform complaints or legal routes. Minors, deceased performers and voice actors under agency contracts each need separate confirmation.
What to listen for
In the Sapphische Ode score you can see the pitches and the piano part, but not the throat, the vocal identity is not in the file. In Deep River, how a tune hangs on a particular way of singing and on a community. In Black Is The Colour, how a song survives through named singers. Synthetic voices fail first, and offend first, in exactly these places.
Listen in the library
All tracks from midicn-lib — inline badge is the license tier (C1 commercial / C2 non-commercial / C3 study only).
Sources
- 声音与肖像的人格利益保护，依公开法律评论归纳，未作结论
- 表演者与录音制作者权利的区分，依公开行业说明核对，未抄原文
Used for fact-checking only; all prose is original.