Mandarin Conversational Speech Recognition Corpus (TicHome Mini)
This database is collected over Tichome mini in quiet (home) and noisy (TV shows) environment, which were from 186 speakers, including 71 male and 115 female. The total pure recording time is about 100 hours, including the reasonable leading and trailing silence.
Mandarin Emotional Speech Recognition Corpus (Mobile)
This database is collected over Mobile phones in quiet (office/home) environment, which were from 250 speakers, including 116 male and 134 female. The total pure recording time is about 49.7 hours, including the reasonable leading and trailing silence.
Mandarin Far-field Speech Recognition Corpus (Desktop)
This database is collected over Desktop microphones in quiet (living room) and noisy (living room with stationary noisy/babble noisy from TV and music) environment, which were from 1012 speakers, including 502 male and 510 female. The total pure recording time is about 3989.4 hours, including the reasonable leading and trailing silence.
This database is collected over Mobile phones in noisy (running car) environment, which were from 66 speakers, including 34 male and 32 female. The total pure recording time is about 54 hours, including the reasonable leading and trailing silence.
Mandarin Keywords Speech Recognition Corpus (Desktop)
This database is collected over Desktop microphones in Studio environment, which were from 300 speakers, including 149 male and 151 female. The total pure recording time is about 53.89 hours, including the reasonable leading and trailing silence.
Mandarin Keywords Speech Recognition Corpus (Desktop)
This database is collected over Desktop microphones in quiet (office/home) environment, which were from 150 speakers, including 76 male and 74 female. The total pure recording time is about 16.1 hours, including the reasonable leading and trailing silence.
Mandarin Kids Emotional Speech Recognition Corpus (Mobile)
This database is collected over Mobile phones in quiet (office/home) environment, which were from 250 speakers, including 114 male and 136 female. The total pure recording time is about 55.2 hours, including the reasonable leading and trailing silence.
This database is collected over Desktop microphones in quiet (office) environment, which were from 199 speakers, including 101 male and 98 female. The total pure recording time is about 66.4 hours, including the reasonable leading and trailing silence.
Mandarin Kids Speech Recognition Corpus (LeTV-Desktop-Mobile)
This database is collected over Desktop microphones, Mobile phones and LeTV in quiet (office) environment, which were from 575 speakers, including 283 male and 292 female. The total pure recording time is about 460.2 hours, including the reasonable leading and trailing silence.
