You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

MNV-17: A High-Quality Performative Mandarin Dataset for Nonverbal Vocalization Recognition

Dataset Description

MNV-17 是一个7.55小时的高质量表演式中文语音数据集,专为非语言发声(NV)识别而设计。该数据集解决了NV感知ASR缺乏高质量、标注良好数据集的问题,提供了一个包含17个不同且平衡良好的常见NV类别的语料库。数据集由来自不同地区的中文母语者录制。通过使用脚本化方法,确保每个NV实例都是有意的且清晰的。该数据集旨在促进表达性ASR的未来研究。

MNV-17 is a 7.55-hour high-quality performative Mandarin speech dataset designed for nonverbal vocalization (NV) recognition. It addresses the lack of high-quality, well-annotated datasets for NV-aware ASR by providing a corpus with 17 distinct and well-balanced classes of common NVs. The dataset was recorded by native Mandarin speakers from various regions. By using a scripted approach, it ensures that each NV instance is intentional and well-articulated. The dataset is intended to facilitate future research in expressive ASR.

Dataset Summary

MNV-17数据集包含2,444个高质量音频样本,涵盖49名中文母语者(25女性,24男性),总时长7.55小时。数据集专门设计用于:

  • 自动语音识别 (ASR):训练能够识别和处理非语言发声的ASR模型
  • 声音事件检测:识别和分类17种常见的副语言声音事件
  • 表达性语音处理:开发更自然、更全面的语音理解系统

主要特点

  • 17种副语言事件:包括拍手、咳嗽、笑声、叹气、清嗓子等常见声音
  • 说话人独立划分:训练集(41人)、验证集(4人)、测试集(4人)严格分离
  • 高质量标注:每个音频文件都有详细的文本标注
  • 平衡的数据分布:确保各类声音事件的充分代表性

使用场景

  1. NV感知ASR模型训练
  2. 多模态语音理解研究
  3. 人机交互中的自然语音处理
  4. 声音事件分类和检测

论文引用

本数据集来源于以下论文:

MNV-17: A High-Quality Performative Mandarin Dataset for Nonverbal Vocalization Recognition in Speech

如果您在研究中使用了本数据集,请引用以下论文:

@misc{mai2025mnv17highqualityperformativemandarin,
      title={MNV-17: A High-Quality Performative Mandarin Dataset for Nonverbal Vocalization Recognition in Speech}, 
      author={Jialong Mai and Jinxin Ji and Xiaofen Xing and Chen Yang and Weidong Chen and Jingyuan Xing and Xiangmin Xu},
      year={2025},
      eprint={2509.18196},
      archivePrefix={arXiv},
      primaryClass={cs.SD},
      url={https://arxiv.org/abs/2509.18196}
}

数据集使用

from datasets import load_dataset

# 加载完整数据集
dataset = load_dataset("maimai11/MNV_17")

# 查看数据集信息
print(dataset)

# 获取训练集第一个样本
sample = dataset['train'][0]
print(f"说话人: {sample['speaker_id']}")
print(f"文本: {sample['text']}")
print(f"音频采样率: {sample['audio']['sampling_rate']}")
Downloads last month
410