How to increase windows size in sound classification

Hi, I am using the TinyML board and trying to classify various sounds based on their source. I am using edge impulse and the recommended window size in 986ms. Is there a way to increase this windows size?