SHACHI - Language Resource Metadata Database

CIAIR constructed the corpus of simultaneous interpretation between Japanese and English for five years (from 1999 to 2003). CIAIR had already completed the transcription and the visualization of speech data and spoken language analysis parts of the corpus with 182 hours of speech data recorded. The transcribed speech data size of CIAIR simultaneous interpretation corpus reaches about one million words (morphemes). The corpus is interactive and bilingual between Japanese and English, containing spoken language data of lectures of daily topics and conversations in travel-related settings. See the heading of database for further information.

CUDIGIT, a collection of spoken Cantonese digit strings of length from single digit to 16 digit, is a part of CUCopora, a large scale Cantonese spoken language corpora. Manually verified phonemic transcription is contained in the corpus. The package comes with CUCMD. (Version 1.0).

CUCMD, a collection of Cantonese command words simulating a navigation control scenario, is a part of CUCopora, a large scale Cantonese spoken language corpora. It is ideal for the development of word based command and control products or application systems. The package also contains CUDIGIT. (Version 1.0).

CUCall Sentences, a collection of continuous Cantonese sentences designed to be phonetically rich, is a part of CUCall, a systematic collection of Cantonese/Putonghua speech data over the telephone networks. The sentences were extracted from local newspapers. The corpus also contains corresponding phonemic transcriptions.

CUCall Words is a collection of word utterances (names of listed companies, foreign currencies and places in Hong Kong, and navigations commands), and a part of CUCall, a systematic collection of Cantonese/Putonghua speech data over the telephone networks. Materials were partitioned into small sections and read by hundreds of speakers. The corpus also contains corresponding phonemic transcriptions. The package includes CUCall Cantonese Digits Version 1.0.

CUCall Digits is a collection of digit-strings (from single to 7, 8 and 16 digits long) utterances and a part of CUCall, a systematic collection of Cantonese/Putonghua speech data over the telephone networks. It covers commonly used digits types such as ID numbers, telephone numbers and credit card numbers. The corpus also contains corresponding phonemic transcriptions. The package includes CUCall Cantonese Words Version 1.0.

CUCall Paragraphs is a read speech data of Cantonese sentences and paragraphs randomly selected from newspapers and a part of CUCall, a systematic collection of Cantonese/Putonghua speech data over the telephone networks. It is designed for capturing various speaking behavior in long utterances. The corpus contains corresponding phonemic transcriptions. The package also includes CUCall Cantonese Spontaneous Speech Version 1.0.

CUCall Spontaneous Speech is a speech (answers to prompted short questions) database of Cantonese collected from large number of speakers, and a part of CUCall, a systematic collection of Cantonese/Putonghua speech data over the telephone networks. It is designed for capturing spontaneous speaking behavior. The corpus contains corresponding phonemic transcriptions. The package also includes CUCall Cantonese Paragraphs 1.0.

The Putonghua Speech data is a part of CUCall, a systematic collection of Cantonese/Putonghua speech data over the telephone networks. It contains utterances of Mandarin words and sentences covering travel and financial domains. There are also spontaneous Mandarin answers to prompted short questions. The corpus contains corresponding phonemic transcription in pinyin transcription scheme.

CUCall is a systematic collection of Cantonese/Putonghua speech data over the telephone networks, composed of two parts; phonetically oriented continuous speech data and application-oriented short phrases and digit strings. The corpus contains 6 sub-corpora; Sentences, Digits, Words, Sentences and Paragraphs, Spontaneous Speech, and Putonghua Speech. The reading materials of the corpora are designed with phonetic as well as application specific consideration.