JNV (Japanese Nonverbal Vocalizations) corpus

JNV数据集是由东京大学信息科学与技术研究生院创建的，包含420个日本非言语发声（NVs）音频片段，涵盖6种情绪。该数据集通过众包收集NVs短语，并通过情感场景刺激演讲者录制，旨在解决现有日本NVs语料库在短语和情感多样性方面的不足。JNV数据集的应用领域包括情感识别和语音合成，是目前在短语和情感多样性方面最大的日本NVs语料库。

The JNV Dataset was developed by the Graduate School of Information Science and Technology, The University of Tokyo. It consists of 420 audio clips of Japanese non-verbal vocalizations (NVs), covering 6 emotion categories. NV phrases were collected via crowdsourcing, and speakers were invited to record the clips under emotional scene stimulation. Its core purpose is to address the shortcomings of existing Japanese NV corpora in terms of phrase and emotional diversity. The JNV Dataset has applications in emotion recognition and speech synthesis, and is currently the largest Japanese NV corpus in terms of phrase and emotional diversity.

搜集汇总

背景与挑战

背景概述

JNV数据集是由东京大学创建的日本非言语发声语料库，包含420个音频片段，涵盖6种情绪，通过众包和情感场景刺激录制，旨在提升短语和情感多样性。该数据集是当前最大的日本NVs语料库，主要应用于情感识别和语音合成领域。