数据堂— 205人噪音环境口音普通话手机采集语音数据_引导

205人噪音环境口音普通话手机采集语音数据_引导，该数据发音人用带口音的普通话在多种日常场景下的噪音环境中录音，录音内容涉及车载场景、智能家居、智能语音助手领域。噪音环境口音普通话手机采集语音数据可用于语音识别声学的语言模型训练和算法研究、机器翻译的语料库建设、声纹识别的模型训练和算法研究

Guided Accented Mandarin Speech Dataset Collected via Mobile Phones in Noisy Environments (205 Speakers): This dataset contains speech recordings from 205 speakers who spoke accented Mandarin in noisy environments across multiple daily scenarios, including in-vehicle scenarios, smart home environments and intelligent voice assistant-related fields. It can be applied to the training of acoustic language models and algorithm research for speech recognition, the construction of corpora for machine translation, as well as the training of models and algorithm research for voiceprint recognition.

搜集汇总

背景与挑战

背景概述

该数据集包含205人在噪音环境下通过手机采集的带口音普通话语音，用于测试普通话识别模型。数据覆盖车载、智能家居和语音助手等场景，格式为16kHz的WAV文件及文本转录，包含205位参与者（其中58%为男性）。