ICSI Helps Collect Linguistic Data
March 18, 2004
For the month of April, ICSI is one of three sites hosting a data collection effort for the Mixer Project being conducted by the Linguistic Data Consortium (LDC) at the University of Pennsylvania. LDC is collecting switchboard-style conversations of participants talking on the phone with one another on given topics for use in developing speech recognition and speaker recognition technologies. While most of the data collection involves participants making calls on their own phones, at ICSI Madelaine Plauche has set up a recording studio in order to collect data recorded in front of eight different microphones at varying distances from the speaker's mouth. The corpus of data collected for the Mixer Project will include conversations on cellular phones and traditional land lines in addition to the special microphone data collected at ICSI.