is this is a bug for fft program?

i am a student in china.

i get diffrent result when i run the code below:

fft([1,2,3,4],4).'
fft([1,2,3,4],4)'
the first:
  10.0000 + 0.0000i
  -2.0000 + 2.0000i
  -2.0000 + 0.0000i
  -2.0000 - 2.0000i
the 2th
  10.0000 + 0.0000i
  -2.0000 - 2.0000i
  -2.0000 + 0.0000i
  -2.0000 + 2.0000i

but i think it should be the same

could you tell me where is wrong?

Answer 1

[dpb]

See the documentation for

doc transpose % .'
doc ctranspose % '

You're transposing the output of the fft() which is complex; in Matlab the ' operator is the complex transpose whereas the "dot" version .' is non-complex transpose.

So, it is doing exactly what is documented to do--

ADDENDUM

And, to emphasize what SS said, the output of fft() itself is identical as would be expected--it was given identical inputs.

The transpose operators are applied to the output of fft(); it (fft, that is) has actually nothing whatsoever to do with the result at that point; its job was done and over before the interesting action occurred.

Same as if had written

>> complex(-2,2)
ans =
  -2.0000 + 2.0000i
>> complex(-2,2)'
ans =
  -2.0000 - 2.0000i
>> complex(-2,2).'
ans =
  -2.0000 + 2.0000i
>> 

which clearly demonstrates the properties of the transpose operators has nothing to do with the source of the data (and vice versa); it simply has to be a complex value before one can observe the difference since w/o an imaginary component the complex conjugate is a "do nothing" operation.

>> [1:4]
ans =
     1     2     3     4
>> [1:4]'
ans =
     1
     2
     3
     4
>> [1:4].'
ans =
     1
     2
     3
     4
>>