hmmph, I used both fftn and fft2, they both produce the same result. Is
there a restriction on the dimension of the input? power of 2 or some such?
On 12/29/2011 07:21 AM, Torgil Svensson wrote:
This is because fft computes one-dimensional transforms (on each row).
Try fft2 instead.
//Torgil
Sorry, i should have looked at your image. A few test you can do is
1) does ifft2 give you back the original image? (allclose returned
True for a little test I did here)
2) does scipy.fftpack.fft2 yield the same result?
//Torgil
On Thu, Dec 29, 2011 at 6:32 PM, Burlen Loring
This is because fft computes one-dimensional transforms (on each row).
Try fft2 instead.
//Torgil
fft(a, n=None, axis=-1)
Compute the one-dimensional discrete Fourier Transform.
fft2(a, s=None, axes=(-2, -1))
Compute the 2-dimensional discrete Fourier Transform
fftn(a, s=None,
there seems to be some undocumented restriction on dimensions as when I
work with 512x512 data things work as expected.
On 12/29/2011 09:43 AM, Torgil Svensson wrote:
Sorry, i should have looked at your image. A few test you can do is
1) does ifft2 give you back the original image? (allclose
Hi
I have an image I need to do an fft on, I tried numpy.fft but results are
not what I expected, and differ from matlab.
My input image is a weird size, 5118x1279, I think numpy fft is not liking it.
In
numpy the fft appears to be computed multiple times and tiled across the
output image. In