Simulation of vocal teaching platform based on target speech extraction algorithm and cloud computing e-learning

Abstract

This paper design a platform that conforms to teaching characteristics, starting from independent data, business logic, and relatively independent functions, a system music teaching platform has been established, and a wide range of system architecture designs have been completed. Therefore, this article adopts targeted speech extraction algorithms to design the application of cloud computing databases on music teaching platforms. This article designs a cloud computing database system based on cloud computing technology under the target speech extraction architecture. The system utilizes corresponding technologies to solve the problems that exist in the system. And propose a training process for a directed speech extraction algorithm and multitask learning scheme based on speech separation and mixed speech, which includes a speech separation module, an embedded attention mechanism module, and a speech extraction module. The existence of targeted speech extraction algorithms and cloud computing databases has made music education more three-dimensional and networked, and teachers' teaching methods and practices have begun to develop on a larger scale. This teaching method has had a profound impact on music education, not only improving students' artistic literacy, but also laying the foundation for innovation and reform of similar courses.