filter problem

hi,

if i have a transfer function for my filter in frequency domain, how m
able to convert this transfer function to filter coefficients in
matlab(m-file)?

any help will be greatly appreciated.

regards

		
This message was sent using the Comp.DSP web interface on
www.DSPRelated.com

I suspect that you will get more of a response by posting this question
to a MATLAB group.

"elbib00" <elbib00@hotmail.com> wrote in message 
news:Y7idnbOSVrrjfuffRVn-oQ@giganews.com...
>
> hi,
>
> if i have a transfer function for my filter in frequency domain, how m
> able to convert this transfer function to filter coefficients in
> matlab(m-file)?
>
> any help will be greatly appreciated.
>
> regards

To create *any* file of coefficients first requires that you have the 
coefficients.
So, how do you convert a frequency response to a set of filter coefficients?
Assuming it is an idealized response or one of unknown orgin:
A filter design procedure perhaps?

If you have the identical frequency response of an actual filter and it is 
periodic.  Then an IFFT of the filter frequency response would give you the 
coefficients of a FIR filter

Fred 

>
>hi,
>
>if i have a transfer function for my filter in frequency domain, how m
>able to convert this transfer function to filter coefficients in
>matlab(m-file)?
>
>any help will be greatly appreciated.
>
>regards
>
>		
>This message was sent using the Comp.DSP web interface on
>www.DSPRelated.com
>

One simple way is to perform an ifft (inverse fft).  Next you will need to
window the result by a dot product of your window of choice in the time
domain.  This will get you close to the answer you want.  The windowing is
key and may want to try a hanning window.



		
This message was sent using the Comp.DSP web interface on
www.DSPRelated.com

michaelkre wrote:
> >
> >hi,
> >
> >if i have a transfer function for my filter in frequency domain, how
m
> >able to convert this transfer function to filter coefficients in
> >matlab(m-file)?
> >
> >any help will be greatly appreciated.
> >
> >regards
> >
> >
> >This message was sent using the Comp.DSP web interface on
> >www.DSPRelated.com
> >
>
> One simple way is to perform an ifft (inverse fft).  Next you will
need to
> window the result by a dot product of your window of choice in the
time
> domain.  This will get you close to the answer you want.  The
windowing is
> key and may want to try a hanning window.
>
>
>
>
> This message was sent using the Comp.DSP web interface on
> www.DSPRelated.com

In general you will get complex coefficients if you do that. If a real
valued filter is desired, the input to the IFFT must be
conjugate-symmetric around DC.

John