美声民族唱法数据集 Bel Conto & Chinese Folk Song Singing Technique Dataset

美声民族唱法数据集专注于区分美声与民族唱法，所有音频均由专业歌唱家演唱。该数据集包含四个明确定义的类别，分别为男美声、男民族、女美声和女民族。经过处理的数据是通过对演唱录音进行去静音、合并、频谱转换以及切片而获得，总计包含9603个数据样本。该数据集的建立旨在为研究者提供一个具有代表性的音频数据集，以便支持对美声和民族唱法声学特征的深入研究，并为相关领域的深度学习模型开发提供基础。

The Bel Canto and Folk Singing Dataset focuses on distinguishing between bel canto and folk singing styles, with all audio recordings performed by professional singers. This dataset contains four clearly defined categories: male bel canto, male folk singing, female bel canto, and female folk singing. The processed data is obtained by conducting silence removal, audio merging, spectral transformation and audio slicing on the original singing recordings, with a total of 9603 data samples. This dataset is established to provide researchers with a representative audio dataset, so as to support in-depth research on the acoustic characteristics of bel canto and folk singing styles, and lay a foundation for the development of deep learning models in related fields.

搜集汇总

背景与挑战

背景概述

该数据集包含203段由专业声乐家录制的美声与中国民族唱法无伴奏音频，总时长达5.08小时。数据集经过集成处理，提供了音频、梅尔频谱图及唱法与性别等多类标签，旨在填补唱法数据集的空白并支持跨文化声乐研究。