Prev: "figure properties" Java exception on Linux
Next: Inverse of Matrix and Signal reconstruction
From: kk KKsingh on 22 Jul 2010 16:38
Which is the best way to compare the result of different interpolation methods on same sample in matlab !

I am using different interpolation techniques with FFT..and Want to know difference between them so what is best you suggest !
From: Walter Roberson on 22 Jul 2010 16:57
kk KKsingh wrote:
> Which is the best way to compare the result of different interpolation
> methods on same sample in matlab !
> I am using different interpolation techniques with FFT..and Want to know
> difference between them so what is best you suggest !

The best method for this will be invented too late for you to implement but
late enough that who-ever will be evaluating your results will devalue your
results because you didn't use the newly invented technique. If you are given
a chance to update your work, then 28 years later someone will prove that that
newly invented best method was disproved by a more general result rigorously
proven by Nikola Tesla in a section of one of his notebooks that had long been
considered to be pure nonsense, and the legacy of your reputation will be ruined.

If you don't have 300 or 400 years to work on this, I suggest that instead of
seeking the *best* method, that you settle for some well-respected and
understood methods.
From: kk KKsingh on 22 Jul 2010 17:25
Walter Roberson <roberson(a)hushmail.com> wrote in message <i2abjr$ppi$1(a)canopus.cc.umanitoba.ca>...
> kk KKsingh wrote:
> > Which is the best way to compare the result of different interpolation
> > methods on same sample in matlab !
> > I am using different interpolation techniques with FFT..and Want to know
> > difference between them so what is best you suggest !
>
> The best method for this will be invented too late for you to implement but
> late enough that who-ever will be evaluating your results will devalue your
> results because you didn't use the newly invented technique. If you are given
> a chance to update your work, then 28 years later someone will prove that that
> newly invented best method was disproved by a more general result rigorously
> proven by Nikola Tesla in a section of one of his notebooks that had long been
> considered to be pure nonsense, and the legacy of your reputation will be ruined.
>
> If you don't have 300 or 400 years to work on this, I suggest that instead of
> seeking the *best* method, that you settle for some well-respected and
> understood methods.

Thanks for reply walter ! It will be great if i can get your views on it question below too

Hi !

I have a matrix which looks like (A*A) where A is Fourier Matrix,It need to apply in this form

(A* WA)^(inverse) * A*Wy;

Where A* is conjugate of A....y is irregular sample signal (So A*Wy; is bandlimited spectra)...W is the weights for samples...

which is the best way to invert is using backslash (\) is making the condition number bad resulting in unstable result. same happens with using inv(A*W*A)...What you suggest about pinv(A*WA)....

Also did any body tried writing this code for signal processing, I wasted so much time on this algorithm still not able to get any results ...Following is my algorithm

1. Take a irregular signal Y
2.Apply A*Y; (Make it band limited , cut off high frequencies)
3. Apply inv(A*WA) to A*Y;
4. Apply IDFT to get samples back

Please let me know, if any one has ever tried this ..I really need some help on this
From: Walter Roberson on 22 Jul 2010 17:33
kk KKsingh wrote:

> Thanks for reply walter ! It will be great if i can get your views on it
> question below too

My view is that it duplicates your earlier thread and did not need to be asked
about here.

From: Jan Simon on 22 Jul 2010 17:51
Dear kk KKsingh,

> Which is the best way to compare the result of different interpolation methods on same sample in matlab !
> I am using different interpolation techniques with FFT..and Want to know difference between them so what is best you suggest !

The "best" method depends on what you call "good". Depending on the actuall job you try to solve, it might be more important to minimize the average absolute difference, the maximum peak difference, the average euclidian distance, the same for blocks, etc.

Under a certain point of view MP3 and JPEG are interpolation methods also, but both have different purposes. As you can imagine, it is impossible to define one "best" method to assess the results of these two methods.
So the best I can suggest is to clear what you want to achieve.

Kind regards, Jan
 |  Next  |  Last
Pages: 1 2
Prev: "figure properties" Java exception on Linux
Next: Inverse of Matrix and Signal reconstruction