数据堂—197小时韩语手机采集语音数据_朗读

韩语语音数据_朗读（手机），基于给定的脚本朗读并模拟录制，录音人共计291名，来自韩国、朝鲜，录音环境为在安静无回音的环境。录音内容广泛，包括经济，娱乐，新闻，口语，数字，字母等，每人约400句。文本经过人工校对，准确率高，为语音识别相关研究及应用提供了丰富的资源，经多家AI公司验证：有助于模型面对真实世界的多样性时能够表现出色。我们严格遵循数据保护法规和隐私规定，确保数据采集、存储和使用的过程中维护用户的隐私和合法权益，所有数据均遵循GDPR, CCPA, PIPL。

Korean Speech Dataset: Read-Aloud (Mobile Phone)
This dataset is collected by having speakers read aloud from provided scripts with mobile phone-based recording. A total of 291 speakers hailing from the Republic of Korea and the Democratic People's Republic of Korea participated in the data collection, and all recordings were conducted in quiet, non-reverberant environments. The recorded content covers a diverse range of categories including economics, entertainment, news, colloquial speech, numerals, letters, and more, with each speaker contributing approximately 400 utterances. The corresponding transcripts have undergone manual proofreading with high accuracy, providing rich resources for research and applications related to speech recognition. Verified by multiple AI enterprises, this dataset enables models to deliver excellent performance when dealing with real-world diversity. We strictly comply with data protection regulations and privacy rules, ensuring the protection of users' privacy and legitimate rights and interests throughout the entire process of data collection, storage and utilization. All data adheres to GDPR, CCPA and PIPL.

搜集汇总

背景与挑战

背景概述

该数据集是一个197小时的韩语朗读语音数据集，专为语音识别研究与应用测试设计，包含291名韩语母语者通过手机在安静环境下录制的语音，内容覆盖经济、娱乐、新闻等多个主题，每人约400句，转录经人工验证，准确率达95%。数据格式为16kHz WAV，遵循相关隐私法规，适用于商业用途。