535小时德国英语语音数据_朗读（手机）【数据堂】

德国英语语音数据_朗读（手机），基于给定的脚本朗读并模拟录制，涵盖通用、交互、车载、家居等多类别，内容丰富。此数据集标注了文本内容等多种属性，由1162名来自不同地域和文化背景的德国人参与录制，口音正宗，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

German-accented English Speech Dataset (Read Speech Recorded via Mobile Phones): This dataset is collected through simulated recording sessions where participants read aloud from standardized scripts, covering a wide range of scenarios including general daily use, human-computer interaction, in-vehicle applications, and smart home environments, with rich and diverse content.

The dataset is annotated with multiple attributes including text content. It was recorded by 1,162 German participants from diverse regional and cultural backgrounds, featuring authentic pronunciation. All source texts have undergone manual proofreading to ensure high accuracy, serving as a valuable resource for research and industrial applications related to speech recognition. Verified by multiple AI enterprises, this dataset helps speech recognition models achieve robust performance when facing the diversity of real-world scenarios.

We strictly comply with global data protection and privacy regulations, ensuring the protection of participants' privacy and legitimate rights throughout the entire lifecycle of data collection, storage, and usage. All data in this dataset fully 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535小时的德国英语朗读语音数据集，由1162名德国人使用手机在安静室内录制，涵盖通用、交互、车载等多类别内容，标注文本准确率达95%，适用于语音识别研究，并遵循隐私法规。数据格式为16kHz、16bit的单声道WAV文件。