759小时印度印地语语音数据_朗读（手机）

印度印地语语音数据_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1425名来自不同地域和文化背景的印度本土人参与录制，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。

Hindi Mobile-recorded Read Speech Dataset: Collected by reading pre-defined scripts via mobile devices, this dataset covers multiple domains including general daily scenarios, conversational interactions, in-vehicle environments and smart home scenarios, with rich and diversified content. It is annotated with multiple attributes such as the corresponding transcribed text. Recorded by 1425 native Indian speakers from diverse regional and cultural backgrounds, the dataset features authentic regional accents. The transcribed texts have undergone manual proofreading to ensure high annotation accuracy, thus providing abundant resources for research and practical applications in the field of speech recognition. Verified by multiple AI enterprises, this dataset enables models to achieve robust performance when coping with real-world diversities. We strictly comply with data protection laws and privacy regulations to safeguard users' privacy and legitimate rights and interests throughout the entire process of data collection, storage and utilization. All dataset materials comply with the requirements of GDPR, CCPA and PIPL.

搜集汇总

背景与挑战

背景概述

该数据集包含759小时的印度印地语朗读语音，通过手机采集，覆盖通用、交互、车载、家居等多个类别。由1425名印度本土人录制，口音正宗且文本经过人工校对，准确率高，适用于语音识别研究与应用，并严格遵循数据保护法规。