fft function returns NaN from non-NaN input

16 visualizaciones (últimos 30 días)
Yuwei Zhu
Yuwei Zhu el 2 de Mayo de 2017
Comentada: Yuwei Zhu el 2 de Mayo de 2017
Hello, I have a one dimension MATLAB vector as fft input. The plot of the vector looks like this:
Zoom it in:
I used isnan() function checked that there is no NaN in this vector. I thought this vector is totally legal to be used in MATLAB fft function.
What is wired is that when I put this vector into the function, all the returns are NaN.
If you want to try out the vector yourself, I have attached it. just load('fmu_q.mat'), and fft(fmu_q). You will see everything is NaN.
Could anyone shed some light on what happened?

Respuesta aceptada

David Goodmanson
David Goodmanson el 2 de Mayo de 2017
Editada: David Goodmanson el 2 de Mayo de 2017
Hi Yuwei, you did check for NaNs but not for any values that equal +-infinity, and there are three of them:
>> find(isinf(fmu_q))
ans = 2959 2960 2961

Más respuestas (0)

Categorías

Más información sobre Fourier Analysis and Filtering en Help Center y File Exchange.

Etiquetas

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by