AwezaMed automatic speech recognition (ASR) test data

The corpus contains orthographically transcribed broadband speech in four official languages of South Africa: Afrikaans, English, isiXhosa and isiZulu. Respondents read a number of ASR prompts (10 or 20) in a real-world environment. The dataset includes 1 hour of test data.The data is not housed in this repository and can be accessed and downloaded via https://hdl.handle.net/20.500.12185/688.

本语料库包含南非四种官方语言的正字法转录宽带语音数据，涉及语言包括南非荷兰语（Afrikaans）、英语（English）、科萨语（isiXhosa）以及祖鲁语（isiZulu）。受访对象在真实环境中朗读了若干自动语音识别（Automatic Speech Recognition，ASR）提示句，单人次朗读量为10或20条。该数据集包含1小时的测试数据。本数据集未存储于此代码仓库，可通过链接 https://hdl.handle.net/20.500.12185/688 进行访问与下载。