Microsoft Releases Custom Neural Voice for Azure With Strict Misuse Controls
Microsoft launches Custom Neural Voice, an Azure AI tool that generates highly realistic synthetic voices while enforcing strict ethical guidelines and consent requirements.
Microsoft announces the general availability of Custom Neural Voice, an Azure Cognitive Service that allows developers to create highly realistic synthetic voices using neural text-to-speech technology. The service operates under a limited access model, meaning customers must apply and receive approval before using it in production across most Azure cloud regions.
The technology uses machine learning models to accurately predict and replicate prosody, which encompasses the tone and duration of specific sounds, resulting in highly natural-sounding speech. This advancement provides significant benefits for voice actors and brands by reducing the need for time-consuming recording pickups, thereby increasing productivity and allowing actors to focus on creative work while earning residuals.
To prevent misuse, Microsoft implements strict safeguards that require explicit consent from voice talent and speaker verification to ensure the submitted recording matches the training data. Additionally, Microsoft reviews each use case, mandates a customer code of conduct, and requires clients to clearly disclose to users that they are hearing a synthetic voice.