Project description: Chinese Speech Corpus (Chinese04)

◦ Chinese names, command words for cell phones, and 11-digit telephone 
numbers uttered by native Mandarin speakers 
◦ A total of 100 speakers (45 males and 45 females for training data; 5 
males and 5 females for testing data) 
◦ Office environment
◦ Microphone
 - SENNHEISER E-835S
◦ Sampling and data format
 - 16,000 Hz, 16 bit Windows wave format
◦ Prompts
 - Chinese names 
 150 items for training data (50 items in consideration of distribution of 
Chinese names and 100 items in consideration of syllable balanced words); 120 
items for testing data (80 items in consideration of distribution of Chinese 
names and 40 items in consideration of syllable balanced words)
- command words
 57 command words (divided into 36 crucial words (A) and 21 non-crucial words 
(B)): 36 A items and 7 B items for training data; 36 A items and 21 B items for 
testing data 
- 11-digit telephone numbers 
 Chinese telephone numbers including cell phone numbers generated by random 
sampling. 50 items for training data; 36 A items and 73 items for testing data. 
In the prompts “i”was used for the pronunciation “yi”and 
“1”was used for the pronunciation“yao”in order to elicit 
two variable pronunciations for one (1) (speakers were informed of this fact.) 
 
◦ Prompts sheet
 - 100 sets

◦ Per speaker
 - 1 set (250 tokens)