what is an effective size of a window

i want to know that what is an effective window size of window and how can i apply that .........why actually we are using this effective window size in hamming and hanning window???pls help me out guys!

[Geoff]

For what?

A window size of 4 is great for smoothing temperature readings taken at half-hour intervals, but is pretty hopeless for smoothing audio data sampled at 96kHz.

Please provide some context.

[newbie traveller]

thank you sir .....for showing your interest!!sir actually ,i wanted to know ........why we actually use hamming and hanning and other windows????for smoothing???is that so how does it works???and while working i heard of effective window size what that suppose to mean???

i am expecting an answer from you sir!

[Jan]

Please explain the problem you want to solve. Currently your questions are to general to be answered efficiently.

Answer 1

[Image Analyst]

An effective window size is one that does what you want. Wouldn't you say that is true? If you're taking the FFT of a signal and don't have the same amplitude of your signal at the left and right edge of your range, then you'll get high frequency artifacts due to the fact that the FFT expects a periodic signal. So the left edge and right edge, when matched up, exhibit a big mismatch - an edge. For example you have a signal that is a ramp, and it's 10 at the left side and ramps up to 200 at the right side, so you have an "edge" of 190. And you know edges are comprised of a lot of high frequencies. So what those windows do, and I'm not telling you anything you won't find via Google, is to bring down the edges to zero smoothly, thus getting rid of any edge and getting rid of artifacts. However you're also eating away into your valid signal when drop down the edges so that improvement comes at a cost. How much artifact removal you want (that's good) versus how much of your signal you're willing to lose (that's bad) is a judgment call.