It practices identifying slight differences in sounds (e.g., denki vs. tenki ) that are easily missed at normal speaking speeds.
