Microsoft introduce the latest voice technology with support of “multi-emotional level”

Microsoft released the latest voice technology, which supports the easy adjustment of “emotional level”, making the emotional expression of intelligent voice more delicate and controllable. Moreover, Human emotions are largely reflected in subtle changes in voice and tone. For example, a “goodbye” sentence is sometimes calm and reserved, sometimes happy and relaxed, and sometimes decisive and angry.

Microsoft Smart Voice can distinguish Happy, Sad, Angry, Fearful, Disgruntled, Serious, Affectionate, Gentle, Depressed, Embarrassed, calm, and other emotions, with one percent as a quantitative unit, with a calm tone as zero points, so that virtual characters have thousands of emotions at once, making content creation more flesh-and-blood.

Furthermore, Microsoft’s artificial intelligence Chinese voices such as Xiaoxiao, Yunxi, Yunye, Xiaohan, Xiaoxuan, Xiaomo, and Xiaorui all support the “emotional level” adjustment technology, and they have different ages, genders, and personalities. It is based on an adaptive neural network. Developers can use SSML tags (Speech Synthesis Markup Language) to easily control the degree of emotion.

At the same time, mass users without any programming or SSML labeling experience can also use this feature through the audio content creation platform. By combining with automatic text sentiment analysis technology, Microsoft’s intelligent voice technology can automatically predict emotion categories and intelligently interpret works full of emotional changes.

In addition, Microsoft’s intelligent voice emotional level adjustable technology makes audio creation just like a director’s casting, using the most suitable voice and the most appropriate emotion to interpret better works. It is suitable for chat robots, audiobook reading, automatic film, and television dubbing, and games Wait for many scenarios.

|VIA|