Is each voice in the voice gallery based on a clone of one specific natural person or is it synthetic?

Question

I would like to understand whether:

1. Each voice in the voice gallery is based on a clone of one specific natural person?
2. Voices are synthetic (similar to those from 11Labs Voice Design) that cannot be traced back to an individual person?

Thank you!

Answer 1

Hi @mpsb,

Thank you for reaching out to Microsoft Q&A forum!

Voices in the Azure Voice Gallery are primarily synthetic, created using advanced machine learning and AI techniques. While some voices may be based on recordings of a specific natural person's voice to achieve a particular tone or style, the resulting voice models are typically synthetic.

Some voices are created from recordings of a specific person's voice, capturing their unique characteristics but often modified during synthesis.

Other voices are entirely AI-generated, not based on any real person, designed to sound natural but without a real-world counterpart.

In practice, Azure may use both methods depending on the specific requirements and design goals of each voice. Some voices might be designed to offer a high degree of naturalness by starting with recordings of a specific person, while others might be generated entirely synthetically for broader flexibility and privacy considerations.

For more info see: Code of conduct for Azure AI Speech.

Hope this helps. Do let us know if you any further queries.

---

If this answers your query, do click Accept Answer and Yes for was this answer helpful.