Algorithm for Sound Recognition and Classification of Ethnic Musical Instruments Based on Deep Learning

In music, voice recognition technology can not only assist music creation and automatically mark music metadata, but also provide new tools for music education and cultural heritage protection. This article proposes the voice recognition and classification algorithm of national musical instruments based on deep learning to realize accurate recognition and efficient classification of different national musical instruments. In view of the uniqueness and complexity of the sound of national musical instruments, this study collected high-quality datasets of national musical instruments and carried out preprocessing and feature extraction. Then, the deep learning model is iteratively trained using the processed dataset, and the model's performance is assessed by comparing the loss curves in different training stages and the classification accuracy of different musical instruments. Results show that the model has good optimization and convergence characteristics in the training process and has achieved high accuracy in the voice recognition and classification tasks of typical national musical instruments such as guzheng, erhu, and pipa. This achievement verifies the practicability of the voice recognition and classification algorithm of national musical instruments based on deep learning.