Regular smart speakers often have arrays of many microphones, but as they are separated ... their voice will arrive there first,” explained by paper co-author Tuochao Chen.