Skip to main content
Topic: FFT window question (Read 1429 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

FFT window question

A sliding window FFT analysis equation:

Y(lamda,k) = \sum_{mu =0}^{L-1} [y(lamda*R + mu) * h(mu) * exp (-j*2pi*k*mu/L)].    (A)

Here,R is frame shift and  h refer to the window.

Rainer Martin said:

Furthermore,to facilitate our notation and to avoid unnecessary normalization
factors we assume

\sum_{mu =0}^{L-1}( (h(mu))^2) = 1 ;                                                  (B)


I wonder: how the equation B can reach the idea " to facilitate our notation and to avoid unnecessary normalization
factors " ?

some maybe useful hint: the subsequent work is to optimally smooth the spectra ,and then to get noise estimate in terms of mimimun statistics.

Regards
HyeeWang

 
SimplePortal 1.0.0 RC1 © 2008-2019