Creating the Complex FFT from the Real only

If I have a test instrument that only outputs the real part of the FFT, is
it possible to fabricate the complex FFT from the real part to enable
calculation of the transfer function?  This instrument has more than one
channel.

"nealvb" <neal@soundsciencecat.com> wrote in message 
news:U8adnTKkbMgif5PanZ2dnUVZ_r2nnZ2d@giganews.com...
> If I have a test instrument that only outputs the real part of the FFT, is
> it possible to fabricate the complex FFT from the real part to enable
> calculation of the transfer function?  This instrument has more than one
> channel.
>

I doubt the assertion that an instrument would output "the real part of the 
FFT".
It's much more likely that an instrument would output the (real) magnitude 
of the FFT - or something very close to that.
Otherwise, the output would be pretty worthless - it would be loaded with 
zeros where the FFT is purely imaginary - and that would be misleading, thus 
useless.
I have never seen such a machine.... but plenty of the type I describe.

You might just tell us which machine it is!

I believe there are methods for estimating a transfer function from the 
magnitude response but someone else would have to comment on that.

Fred 

On Oct 12, 8:37 am, "Fred Marshall" <fmarshallx@remove_the_x.acm.org>
wrote:
> "nealvb" <n...@soundsciencecat.com> wrote in message
>
> news:U8adnTKkbMgif5PanZ2dnUVZ_r2nnZ2d@giganews.com...
>
> > If I have a test instrument that only outputs the real part of the FFT, is
> > it possible to fabricate the complex FFT from the real part to enable
> > calculation of the transfer function?  This instrument has more than one
> > channel.
>
> I doubt the assertion that an instrument would output "the real part of the
> FFT".
> It's much more likely that an instrument would output the (real) magnitude
> of the FFT - or something very close to that.
> Otherwise, the output would be pretty worthless - it would be loaded with
> zeros where the FFT is purely imaginary - and that would be misleading, thus
> useless.
> I have never seen such a machine.... but plenty of the type I describe.
>
> You might just tell us which machine it is!

University professors must have lots of such strange machines.  Why
else would they ask about them on exams and homework assignments?

I'd go ask professor Hilbert (maybe offer him a barometer in
exchange :)
about how to transform one of these test instruments into something
more complex.


IMHO. YMMV.