This corpus comprises 240,296 entries uttered by 201 speakers (100 males and 101 females), recorded over 4 channels (desktop in quiet office). Speech samples are stored as a sequence of 16-bit 44.1kHz for a total of 95.63 hours of speech per channel.