公开(公告)号：US12094484B2

公开(公告)日：2024-09-17

申请号：US18360838

申请日：2023-07-28

Applicant: ZHEJIANG LAB

Inventor： Jingsong Li ,  Zhenchuan Zhang ,  Tianshu Zhou ,  Yu Tian

IPC: G10L21/0232 ,  G10L17/02 ,  G10L17/04 ,  G10L25/30

CPC classification number: G10L21/0232 ,  G10L17/02 ,  G10L17/04 ,  G10L25/30

Abstract: The present disclosure discloses a general speech enhancement method and apparatus using multi-source auxiliary information. The method includes following steps: S1: building a training data set; S2: using the training data set to learn network parameters of a model, and building a speech enhancement model; S3: building a sound source information database in a pre-collection or on-site collection mode; S4: acquiring an input of the speech enhancement model; and S5: taking a noisy original signal as a main input of the speech enhancement model, taking auxiliary sound signals of a target source group and auxiliary sound signals of an interference source group as side inputs of the speech enhancement model for speech enhancement, and obtaining an enhanced speech signal.