Project description: Multimodal Speech Corpus

---

◦ Multimodal corpus of voice and video of the frontal face captured by 
the camcorder 
◦ A total of 100 speakers speaking the standard dialect (Age: 20 ~ 30)
◦ Studio: 230cm * 230 cm
◦ Lighting: basic lighting + front (one 200W), right/left(two 100W), and 
ground (one 60W)
◦ Background: blue screen
◦ Equipment: Sony DCR-VX2000
◦ Video sampling and data format(avi)
 - resolution: 720*480
 - frame speed: 30 fps
 - condensation: MPEG-4 V1, key frame at maximum 
◦ Speech sampling and data format(wav)
 - 16kHz, 16bits, mono, Microsoft Windows WAVE format
◦ 5 basic vowels 
◦ 12 single digits 
◦ 452 phonetically balanced words (PBW)
 
◦ Per speaker
 - 5 basic vowels 
 - 12 single digits 
 - 90 ~ 92 PBWs