Microsoft announces Text To Speech Avatar AI tool that can create virtual 3D digital humans

Microsoft released an AI tool called "Azure AI Speech text to speech (TTS) avatar" at the Ignite conference, claiming to be able to generate realistic virtual characters. This tool is now available for public preview

Microsoft said that users can use Azure AI Speech text to speech (TTS) avatar to create a virtual avatar based on "inputting text to speak content" and combine it with real-person photo training to build an "interactive avatar" based on real people. "Chatbot" can be used in corporate marketing, business or customer service scenarios.

According to reports, Azure AI Speech text-to-speech (TTS) avatar is mainly composed of three modules, namely text analyzer, TTS sound synthesizer and TTS virtual avatar synthesizer

The text analyzer will first analyze the text content entered by the user and generate a phoneme sequence. Then, the TTS speech model in the TTS sound synthesizer will predict the acoustic characteristics of the user's input text and then synthesize the sound. Finally, the neural network sound synthesis model Avatar will predict the character's lip image based on the above acoustic characteristics, thereby ultimately forming a virtual avatar image

Microsoft explained that traditional avatar production is time-consuming and labor-intensive, requiring the establishment of a dedicated shooting environment, and the post-production process of shooting and editing is also quite costly. Currently, Microsoft's latest Azure AI Speech text to speech (TTS) avatar service is used. After establishing the model for the first time, users can create various product introductions, interactive videos, etc. just by inputting text. Together with Microsoft Azure OpenAI Service and neural network TTS functions, it can also present a more natural interactive experience.

IT House reported that Microsoft gave an example that users can use Azure AI Speech TTS avatar to batch produce various video content, such as corporate culture videos, product introductions, or the CEO's digital avatar at a conference. In addition, you can also create virtual live broadcast digital humans, chat robots, business robots, and AI teachers for online teaching, etc.

Microsoft announced that Azure AI Speech Synthesis (TTS) avatars are now available to Azure subscribers and support multiple languages. Users can select the desired role from preset avatar options or customize the avatar

If users want to customize their own virtual avatars, they need to upload a batch of character video clips, and the Azure platform will process these videos online to generate virtual avatars. The character itself is separated from the sound source. Users can choose the official default sound source or upload their own training sound source.

The above is the detailed content of Microsoft announces Text To Speech Avatar AI tool that can create virtual 3D digital humans. For more information, please follow other related articles on the PHP Chinese website!