Japanese Multi-speaker Speech Synthesis Corpus (Multi-emotion)
The database recorded 10,504 sentences (161,520 words) from 3 voice talents (1 female and 2 males). The total audio duration is about 9.77 hours, including the original silence at the beginning and ending (about 350 ms each).
The recorded content is organized into 24 texts, including multiple fields and emotions, such as news, technology, dialog, entertainment and angry, sad, happy and surprise. We used ja-jp_hepburn_intro phone set for labeling.
The database recorded 10233 sentences (132550 words) from a female voice talent. The total audio duration is about 10.84 hours, including the clear silence at the beginning and ending (about 500 ms each).
The recorded content is organized into 35 texts, including multiple fields, such as news, dialog, song, etc. We used Ko-kr_rr_introl phone set for labeling.
Korea Korean Female Speech Synthesis Corpus (Multi-emotion)
The database recorded 1,782 sentences (27,266 words) from a female voice talent. The total audio duration is about 2.2 hours, including the zero silence at the beginning and ending (about 400 ms each).
The recorded content is organized into 2 texts, including multiple emotions, such as happy, and sad. We used ko-kr_rr phone set for labeling.
The database recorded 10233 sentences (132558 words) from a male voice talent. The total audio duration is about 10.9 hours, including the clear silence at the beginning and ending (about 350 ms each).
The recorded content is organized into 35 texts, including multiple fields, such as news, dialog, song, etc. We used Ko-kr_rr_introl phone set for labeling.
Korea Korean Multi-speaker Speech Synthesis Corpus
The database recorded 26,664 sentences (432,872 words) from 29 voice talents. The total audio duration is about 26.84 hours, including the original silence at the beginning and ending (about 300 ms each).
The recorded content is organized into48 texts, including multiple fields, such as news, biography, dialog, etc.
Korean Multi-speaker Speech Synthesis Corpus (Multi-emotion)
The database recorded 6,904 sentences (165,209 words) from a female and 2 male voice talents. The total audio duration is about 9.66 hours, including the original silence at the beginning and ending (about 350 ms each).
The recorded content is organized into 30 texts, including multiple fields, such as news, encyclopedia, dialog, narration, etc. and 4 emotions, such as happy, sad, angry and surprise. We used ko-kr_rr.phset phone set for labeling.
The database recorded 1,165 sentences (13,157words) from a female voice talent. The total audio duration is about 2.16 hours, including the original silence at the beginning and ending (about 350 ms each). The recorded content is news text.
The voice talent was born and raised in Uganda, and was 22 years old when recording the database. She has a standard Luganda pronunciation and is a professional broadcaster. The recording has a warm timbre and even speech rate.
The database recorded 637 sentences (12,989 words) from a male voice talent. The total audio duration is about 2.09 hours, including the original silence at the beginning and ending (about 350 ms each). The recorded content is news text.
The voice talent was born and raised in Uganda, and was 23 years old when recording the database. He has a standard Luganda pronunciation and is a professional broadcaster. The recording has a deep timbre and even speech rate.
