A summary of the size and content of the corpus is given below: number of speakers 150 speakers males 75 females 75 range of speaker age 10 yrs. to 70 yrs. number of items per speaker 323 items isolated digits 15 four digit sequences 35 city names 100 monosyllables 110 control words (set A) 13 control words (set B) 24 control words (set C) 26 number of repetitions per item 4 repetitions total number of utterances 193,763 utterances (per channel) sample frequency 16 kHz sample type 16-bit linear number of microphones 2 (dynamic and condenser) --===----=====------=======--------=========--------=======------=====----===-- Distributions of nonstandard orthographic items for the isolated digit subset of the JCSD Corpus. The vast majority of the data is fairly clean, with approximately 10% showing some type of nonstandard behavior. The statistics for other segments of the corpus are comparable. Description Frequency (17,390 items) "clean data": nominal pronunciation/ 15,364 (88.4%) no non-speech markers with a non-speech marker 2,026 (11.6%) and/or an alternate pronunciation with an alternate pronunciation 1,161 (6.7%) distribution: only an alternate pronunciation 1,121 (96.6%) both an alternate pronunciation 40 (3.4%) and a non-speech marker with a non-speech marker: 905 (5.2%) distribution: only a non-speech marker 865 (95.6%) both an alternate pronunciation 40 (4.4%) and a non-speech marker non-speech markers 979 (5.2%) distribution: {mouth noise} 543 (55.5%) {breath noise} 344 (35.2%) {paper rustle} 48 (4.9%) {non-speech noise} 20 (2.0%) {throat clear} 7 (0.7%) {background noise} 7 (0.7%) {cough} 6 (0.6%) {sniff} 2 (0.2%) {mouth_noise} 2 (0.2%) --===----=====------=======--------=========--------=======------=====----===-- An overview of the geographic distribution by region of the speaker population. Geographic Combined Males Females Region (150) (75) (75) Chubu 16 10 6 Chugoku 5 3 2 Hokkaido 1 1 0 Kanto 90 40 50 Kinki 16 10 6 Kyushu 9 5 4 Tohoku 10 6 4 Unknown 1 0 1 --===----=====------=======--------=========--------=======------=====----===-- Distribution of speaker age in the JCSD Corpus. Age Combined Males Females (150) (75) (75) 10-19 1 0 1 20-29 50 25 25 30-39 40 20 20 40-49 32 15 17 50-59 22 11 11 60-69 5 4 1 --===----=====------=======--------=========--------=======------=====----===-- An overview of the prompting material used in the JCSD Corpus. Description Number of items Control Words: Banking Services 13 Word Processors 24 Home Electronic Equipment 26 Digits: Isolated Digits 15 Four Digit Sequences 35 City Names: 100 a phonetically-rich subset of common Japanese city names Monosyllables: 110 all Japanese monosyllables plus several used to pronounce foreign words