Microsoft has announced the introduction of four new AI-powered neural voices for text-to-speech (TTS) applications that can be used in Azure OpenAI GPT starting today to help create speech-based chatbots, voice assistants, or conversational agents.

The four voices, named en-US-AndrewNeural, en-US-BrianNeural, en-US-EmmaNeural (all US English) and zh-CH-YunjieNeural (Chinese), are “adapted to conversational scenarios” And they are now available for public preview in three regions, Eastern US, Southeast Asia, and Western Europe.

Microsoft has Showed some examples of the new voices compared to other neural voices currently available to show the improvements made in making speech more natural and fluid.

The voices themselves can be integrated into existing apps that use Azure OpenAI, by using the Azure Speech SDK or REST API, as well as using the Azure Bot Framework to develop smart bots capable of using new neural TTS voices.

Microsoft goes on to say the following:

“We started by creating the persona of each voice as if it were a real person who is friendly and optimistic about life, always eager to help others and share intriguing or practical knowledge. The speaking style of the voice is similar to a conversation with an acquaintance over a cup of tea, maintaining a natural and unexaggerated tone.”

Furthermore, we are continuously improving our text-to-speech (TTS) modeling techniques to improve the quality of our AI voices. Our latest projects, such as DelightfulTTS 2 and MuLanTTS, have significantly narrowed the quality gap between AI voices and professional human recordings, producing more natural and realistic voices than ever before. These technological advances serve as the foundation upon which these new AI voices are built.

The four new voices sat alongside the existing offering of over 400 neural voices, covering more than 140 languages ​​and locales.