3 ms·

Looks great - I'd even be willing to spend some time training it to help it distinguish voices of people I tend to meet with!