Synthesis-based clear hearing method in noisy conditions

• Main Author: WENG FULIANG
• Format: Patent
• Language: Chinese; English
• Published: 16.07.2024
• Subjects: DEAF-AID SETS; ELECTRIC COMMUNICATION TECHNIQUE; ELECTRICITY; LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKEACOUSTIC ELECTROMECHANICAL TRANSDUCERS; PUBLIC ADDRESS SYSTEMS

The present invention provides a novel and improved hearing aid system that employs a high quality noise cancellation method and apparatus to overcome the limitations and difficulties encountered in the prior art. Technical limitations of noise uncertainty and speech distortion in the hearing aid field are solved by restoring high quality speech by converting speech content into intermediate language characterization and by synthesizing the speech of the same speaker with a pre-trained artificial intelligence (Al) module. In the present invention, due to the Lombard effect, it can be assumed that the speech of a target speaker is a dominant speech, and noise uncertainty can be avoided by selecting corresponding settings to focus on the target speaker or picking up the dominant speech. 本发明提供一种新型且经改进的助听器系统，该新型且经改进的助听器系统采用高品质噪声消除方法和装置，以克服传统技术中遇到的限制和困难。助听器领域中的噪声不确定性和语音失真的技术限制通过将语音内容转换为中间语言表征来恢复高品质语音，并通过用经过预先训练的人工智能(Al)模块对同一发音者的语音进行合成来解决。在本发明中，因为隆巴德(Lombard)效应，可以假定目标发音者的语音是主导语音，通过选择相应设置来专注于目标发音者或拾取主导语音来规避噪声不确定性。