Prev: Erorr:Subscript indices must either be real positive integers
Next: Image pointer problem
From: madhwendra tripathi on 10 May 2010 08:49
"Jaime Delgado Saa" <delgado(a)sabanciuniv.edu> wrote in message <hs2207$4ao$1(a)fred.mathworks.com>...
> Hi, just call me Jaime!
>
> I just upload a function that make the calculation of the Marginal Hilbert Spectrum (mhs), it takes as input the amplitude/energy of the intrinsic mode functions and the instantaneous frequency for each of those imf's (wich are the output of the hilbert spectrum), then if you have 10 imf´s the input are 2 matrices of size (length of data,10) , the outputs are 2 vectors, one containing the mhs and other the frequencies values for each element of the first vector, the you can use this 2 vectors to plot the mhs.
>
> I just upload the file, so in next days (maybe tomorrow) will be availabale, let me know if work for you, if there is some bugs also letme know to fix it.
>
> I hope be useful for you.
>
> see you!
>
> Jaime












hi jaime,

its pleasure to me that u replied me again..
but as u told in your last message that u have uploaded a function file for calculating "Marginal hilbert spectrum".. but, till now i have not received that file...
please send me that file again.. i'll be very thankful to you..

u can also send that file at my e-mail id.

madhwendra.711(a)gmail.com


thank you very much !!

madhwendra
From: Jaime Delgado Saa on 10 May 2010 12:17
Hi, just a couple of minuts ago the file was acepted and appear in my profile.
this shoud work.

This function calculate the contribution (sum) of each frequency in the instantaneous frequency vectors of each imf.

I am reviewing the teory againg in orther to make this comprable with the furier transform, where there are bins for the frequency, in hilbert-huang transform, the frequency is continuos, If I decide that some improvement should be done, I will let you know.

I am using this package to calculate the hilbert huang transform (emd, imf, hilbertspectrum) :
http://perso.ens-lyon.fr/patrick.flandrin/emd.html, my function use the output of the gilbert spectrum of this package, but should work in general independently.

let me know any question

Jaime
From: madhwendra tripathi on 12 May 2010 13:28
"Jaime Delgado Saa" <delgado(a)sabanciuniv.edu> wrote in message <hs9bim$l5p$1(a)fred.mathworks.com>...
> Hi, just a couple of minuts ago the file was acepted and appear in my profile.
> this shoud work.
>
> This function calculate the contribution (sum) of each frequency in the instantaneous frequency vectors of each imf.
>
> I am reviewing the teory againg in orther to make this comprable with the furier transform, where there are bins for the frequency, in hilbert-huang transform, the frequency is continuos, If I decide that some improvement should be done, I will let you know.
>
> I am using this package to calculate the hilbert huang transform (emd, imf, hilbertspectrum) :
> http://perso.ens-lyon.fr/patrick.flandrin/emd.html, my function use the output of the gilbert spectrum of this package, but should work in general independently.
>
> let me know any question
>
> Jaime








hi Jaime,

thanx for sending me the program file for calculating mhs. but could not got the desired results... perhaps due to wrong implementation of signal etc.

i'm sending you text file of original heart sound signal signal (on which, i'm working for) this signal is sampled at 5000 samples/sec. please find its IMFs and then marginal hilbert spectrum.

after applying empirical mode decomposition on this signal, i got total 13 IMFs.
but after applying mhs (from your sent programs) on it i got an amplitude-frequency plot. but its not according to my desired results.

bcoz it should have frequency varying from 0 to 500 hz (maximum), but its varying from 0 to 3500 hz.

and it should show the plot like ; very first , strongest frequency, the lesser strong frquency....

but its not showing in this manner....... its showing the zig-zag variation...

please download my signal file and try to get the desired result...

is there need to apply any window, like hanning, hamming etc. (actually i dont have any idea to apply window technique on signal )

one thing more....... after obtaining marginal hilbert spectrum of my signal, i have to get the fourier spectrum (perhaps fft) of original signal. and then i have to compare the both spectrum (means mhs is compared with fourier spectrum).
so , i request you to send a program file to calculate fourier spectrum of same signal.

(from the whole data file, i have selected starting 7000 samples for complete HHT analysis)
First  |  Prev  | 
Pages: 1 2
Prev: Erorr:Subscript indices must either be real positive integers
Next: Image pointer problem