Voice generation method and device, intelligent voice interaction method and device and electronic equipment

By analyzing the multimodal semantic features of the text to be pronounced and generating speech that conforms to these characteristics, the problems of rigid, monotonous and wrong pronunciation in the prior art are solved, and more accurate and vivid speech generation is achieved, improving user experience.

CN119832895APending Publication Date: 2025-04-15TAOBAO CHINA SOFTWARE
View PDF 8 Cites 0 Cited by

Patent Information

Application Number
CN202411724364.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-27
Publication Date
2025-04-15

AI Technical Summary

Technical Problem

When the prior art converts answer text into pronunciation, the generated pronunciation is often stiff and monotonous, especially for sentences in long sentences or complex contexts, which can easily lead to errors in pronunciation or breaking sentences, affecting accuracy and user experience.

Method used

By determining the multimodal semantic features of the text to be pronounced, its semantics are represented from multiple dimensions, and a semantic-compliant speech is generated based on these features, ensuring the accuracy and vividness of the speech.

Benefits of technology

It realizes accurate and vivid voice generation of various sentences, and improves the user's experience of using intelligent dialogue, especially when dealing with long sentences or complex contexts, it can generate the pronunciation of various sentence patterns more accurately.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119832895A_ABST
    Figure CN119832895A_ABST
Patent Text Reader

Abstract

The invention discloses a voice generation method and device, an intelligent voice interaction method and device, electronic equipment and a computer readable storage medium. The method comprises the steps of determining a to-be-pronunciated text; analyzing the to-be-pronunciated text to obtain a multi-modal semantic feature corresponding to the to-be-pronunciated text, the multi-modal semantic feature being used for representing semantics of the to-be-pronunciated text from multiple dimensions; and according to the multi-modal semantic features and the to-be-pronunciated text, generating generated speech corresponding to the to-be-pronunciated text and conforming to the multi-modal semantic features. According to the scheme provided by the invention, more accurate and vivid voices can be generated for various statements, and the user experience of using intelligent dialogues by a user is improved.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Model training method and device, dialogue system evaluation method and device, equipment and storage medium

    CN110188331A

  • Speech synthesis method, speech synthesis device and intelligent equipment

    CN112530406A

  • Voice recognition method, device, equipment, system and storage medium

    CN113436614A

  • Voice recognition method and device, medium and equipment

    CN113470619A

  • Speech synthesis method and device, computing equipment, storage medium and program product

    CN114373443A