Wavelet transform question

[mystique_unbound]

WAVELET TRANSFORM

I found from net that Fourier analysis fails when non stationarity is introduced into the signal . When the signal has randomness
say a signal with sinusoidal frequency of 100K for 10ms and then it has a frequency of 50K for next 10ms fourier analysis will show both the frequencies .But i am having a doubt here .

the signal can be viewed as apulsed signal
sin(100hz)*pulse(10ms) + sin(50K)*pulse(time shifed 10ms)

For this the fourier transform will be a sinc function at 100K and an exponentially decaying (decay constant is 10w)SINC at 50K .

Which cuses the problem ? We are using normal fourier analysis for FSK,BPSK which are also of pulsed nature ? how can we employ Wavelet there .What is the signigficance

 

[Resistance]

Re: WAVELET TRANSFORM

hello friend,

See impulses are only there is theoritical cconcepts.. mathematically speakinging yes u are right they will be different but See matlab simulations and practically observing things in the analyzer u wont able to distinguish the non stationary and stationary spectrum.

regards.

 

[darush_a_m]

WAVELET TRANSFORM

i need a good toturial for wavelet transform, source code (matlab) and its application, please help me
my mail : darush_a_m@noavar.com

 

[redsk_y]

Re: WAVELET TRANSFORM

This is a nice and a very clear tutorial pn wavelet transform:

h**p://users.rowan.edu/~polikar/WAVELETS/WTtutorial.html

Try it.

Regards,

--rs