about computing fft in wavenumber domain

I have signal distributed in space domain (x,y,z). and I want to calculate its wavenumber doain for (x,ky,z) domain. how can I carry out this task(without using a do loop over all of y values)?

Answer 1

[David Young]

Use

fft(data, [], 2)

[smabtahi]

Thanks David

Do you mean that with the mentioned command , second column of "data" assumed as "y" and the "[]" contains array of wavenumbers?

[Walter Roberson]

No, the [] indicates that the default number of fft points is to be used, and the 2 means to work across rows rather than down columns.

This syntax is the third one listed in the possibilities in the fft reference documentation, http://www.mathworks.com/help/techdoc/ref/fft.html

[smabtahi]

Thanks Walter

I have read the document, but it says only that :

Y = fft(X,[],dim) and Y = fft(X,n,dim) applies the FFT operation across the dimension dim.

then it means that "Dim" refers to a determined coordination direction, and even may be more than 2 and it doesn't refer to specify working with rows rather than columns.

please tell me if I got the point true or not.

[Walter Roberson]

The dimension refers to the matrix dimension index in standard matlab order -- columns first, rows second, "pages" third, unnamed entities for the higher dimensions. It is logically equivalent to using ":" as the dim'th element in indexing the vector with the other elements held constant for any one operation.

It is the value 2 that indicates working across rows, rather than just the presence of a parameter at that location.

[smabtahi]

Thank you very much.

your explanation was really helpful.