Corpus of Spoken Mandarin (2005-2025)
Description
A corpus of spoken Mandarin spanning 20 years from 2005 to 2025. The corpus consists of four sub-corpora (2005-2009, 2010-2014, 2015-2020 and 2021-2025), each one containing between 170,000 and 180,000 characters of unscripted spoken Mandarin, for a total of approximately 700,000 characters. The spoken utterances in the first three sub-corpora were sourced from data available freely in the public domain, mostly transcripts of interviews and popular Chinese chat shows, such as "Luyu you yue" (‘A Date with Luyu’), broadcast originally on television or radio. The fourth sub-corpus was built by sourcing audio-visual data online and using free speech-to-text software to create the transcripts of the broadcasts. The transcripts were checked for accuracy against the recordings using spot checks.