25. Online Gaming Voice Chat Corpus with Emotional Label (OGVC)

Producer, Project

Yoshiko Arimoto, Hiromi Kawatsu, Tokyo University of Technology

Contents

Natural speech and acted speech for emotional research.

Speaker

Natural speech: 9 males and 4 females

Acted speech: 2 males and 2 females

Recording environment

Meeting space of laboratory office

Head-set microphone

Speech file format

Natural speech: WAV format (48 kHz, 16 bit, Mono)

Acted speech: WAV format (44.1 kHz, 16 bit, Mono)

Distribution media

2 DVD

Licensing

For research purpose only

Price

No fee

Further information

https://sites.google.com/site/ogcorpus/ (in Japanese)

Note

All documents are written in Japanese.

Speech sample for test listening

Natural speech

A: 
えーっとショートカット―キー。じゃあ3。うおっと何何何?何?ちょっちょちょちょっこ、これ付いてこうぜ。これ付いてこうぜ
   
(Uhmm, the short cut key. OK, `3'. Wow, wh- what? What? Heeeeeeeeeey, le- let's follow 'em. Let's follow 'em.)
B: 
うん?
   
(Uh?)
B: 
何?何があった?
   
(What? What's going on?)
A: 
{ 笑 } ねえ、あ、あれ右右右、すごいすごいすごいすごい
   
([laughs] Hey, th- that on your right, on the right, on the right, wow wow wow wow...)
A: 
{ 笑 }
   
([laughs])
B: 
あらー
   
(Gee.)
A: 
これ低レベルパーティだよね?
   
(What a terrible party this is?)
B: 
すげえ。なんか懐かしいな
   
(Wow. It takes you back, doesn't it!)


Acted speech

「あらー」

(Gee.)

「すげえ。なんか懐かしいな」

(Wow. It takes you back, doesn't it!)

Go to corpora list