How Do You Choose a Sliding Window?

Am, the first systems, you take a window and you take a furiar transform. You look at the spectral content of that window. And so i might take a very small window, which means that i know that this is only one speeth sound in this window. But then you've got a lot less data to kind of figure ad to kind of estimate the content ofthat window. It's something that you don't just see in speech processing. You see at in any type of cana, a signal processing an task.