In signal processing, windowing is the process of splitting the input signal into temporal segments where the signal can be considered quasi-stationary. This process is also referred to as [[short-time processing]]. The [[window function]] is typically a mathematical function that is defined to be zero outside a certain range and non-zero inside that range. Each signal segment is called a [[frame]]. For example the m-th frame of the signal $x(n)$ is obtained by: $ x_{m}(n) = x(n-mM)w(n) $ where $M$ is the number of samples between each frame, commonly referred as the [[hop length]]. The length of each frame depends on the length of the window function $w(n)$. ![[signal-rect-win.png]] ![[signal-hann-win.png]] [[Backstrom 2022]] The most common approach to re-construct windowed signals is called [[overlap-add]].