Been looking at objects like [ej.vdb~] and [fft~]. both mention window size or windowing. I searched around the forum and documentation but couldn’t find anything that would explain exactly what windowing is. Would anyone mind explaining what it is or what the usage for this is? Or point me in the direction of an explanation? Thanks
Fourier analysis for example analyzes audio by taking chunks of it. A succession of chunks can be understood as a window (showing a part, covering the rest) that moves over the audio stream. Windowing is also used as a name for smoothing the edges of the chunk, by applying a short fade-in and fade-out, therefor you will see things like a hanning-window which is a name for a specific fade method. The fft works with a fixed window size that has to be a power of two samples long, eg 512 samples. Hope this brief introduction helps.
And, the larger the window size, the greater the frequency resolution, but smaller the time resolution, so widow size is an important consideration regarding the type of sound you’re analyzing. Short windows are best for time-critical applications like percussion sounds, but you can get better spectral resolution with a larger window on slowly changing sounds like held notes on wind instruments etc.