Digital Signal Processing Reference
In-Depth Information
collected at Nagoya University [7], which contains data from 800 speak-
ers. They include isolated word utterances, phonetically balanced sen-
tences and dialogues recorded while driving. The data collection is per-
formed using a specially designed data collection vehicle that has multiple
data acquisition capabilities of up to 16 channels of audio signals, three
channels of video and other driving-related information, i.e., car position,
vehicle speed, engine speed, brake and acceleration pedals and steering
handle.
Five microphones are placed around the driver's seat, as shown in Fig-
ure 19-3, where the top and the side views of the driver's seat are depicted.
Microphone positions are marked by the black dots. While microphones
#3 and #4 are located on the dashboard; #5, #6 and #7 are attached to
the ceiling. Microphone #6 is closest to the speaker. In addition to these
distributed microphones, the driver wears a headset with a close-talking
microphone (#1).
Figure 19-3.
Microphone positions for data collection inside the vehicle: Side view
(top) and top view (bottom).
In the majority of the corpus, the speaker is driving in the city traffic
near Nagoya University. Considerable part of the corpus that we use in
