995 local Cantonese speakers participated in the recording, and conducted face-to-face communication in a natural way. They had free discussion on a number of given topics, with a wide range of fields; the voice was natural and fluent, in line with the actual dialogue scene. Text is transcribed manually, with high accuracy. Format：Mobile phone: 16kHz, 16bit, mono channel, .wav;   Voice recorder: 44.1kHz, 16bit, dual channel, .wav; Environment：quiet indoor environment, without echo Recording Content：dozens of topics are specified, and the speakers make dialogue under those topics while the recording is performed Demographics：995 Cantonese; 45% speakers of all are in the age group of 26-45; 504 speakers of them spoke in groups of two speakers, 195 speakers of them spoke in groups of three speakers, 196 speakers of them spoke in groups of four speakers, and the other 100 speakers spoke in groups of five speakers Annotation：annotating for the transcription text, speaker identification and gender Device：mobile phone and voice recorder Language：Cantonese Application Scenario：Voice Recognition, Voice Print Recognition Accuracy rate：95%

Creator(s)

-

Distributor(s)

ELRA

Right Holder(s)

-