Classic FM Demodulation, using Matlab

>Perhaps you can tell us why, when 99.9% of all radio receivers don't use a
=
>Hilbert Transform, why you feel you need one here. I assume you are trying
=
>to recover phase. The usual method is to use a PLL on the carrier or IF
sig=
>nal - or a discriminator - quad detectors are normally the best though due
=
>to phase noise of PLLs. Also, if you want to recover phase then FFTs can
be=
> used using complex cepstrum instead of Hilbert Transforms.
>

I measure the FM modulated carrier in the lab using bench instruments. From
that, I need to determine the modulation amplitude using any valid
technique via post-processing in Matlab. I'd don't prefer Hilbert Transform
over any other technique. However, in reading about how to perform FM
demodulation, the Hilbert Transform is often mentioned. Thus, my interest
to use it. If there are other mathematical techniques to perform FM
demodulation, they'd work too. So far I've summarized the two methods I'm
aware of above. 	 

_____________________________		
Posted through www.DSPRelated.com

On Sunday, January 25, 2015 at 3:22:37 PM UTC+13, all4dsp wrote:
> >Perhaps you can tell us why, when 99.9% of all radio receivers don't use a
> =
> >Hilbert Transform, why you feel you need one here. I assume you are trying
> =
> >to recover phase. The usual method is to use a PLL on the carrier or IF
> sig=
> >nal - or a discriminator - quad detectors are normally the best though due
> =
> >to phase noise of PLLs. Also, if you want to recover phase then FFTs can
> be=
> > used using complex cepstrum instead of Hilbert Transforms.
> >
> 
> I measure the FM modulated carrier in the lab using bench instruments. From
> that, I need to determine the modulation amplitude using any valid
> technique via post-processing in Matlab. I'd don't prefer Hilbert Transform
> over any other technique. However, in reading about how to perform FM
> demodulation, the Hilbert Transform is often mentioned. Thus, my interest
> to use it. If there are other mathematical techniques to perform FM
> demodulation, they'd work too. So far I've summarized the two methods I'm
> aware of above. 	 
> 
> _____________________________		
> Posted through www.DSPRelated.com

I would have put the Hilbert Transform at the bottom of any list on FM demodulation. Quad detector being the most common. For ordinary FM demodulation is quite simple, just like AM after you defferentiate - that's a discriminator. Most of the methods work by a form of band limited differentiation. The PLL is 3dB better but as I said, making a VCO with low enough phase noise can be a problem. If I had x+jy I would use 1/(1+(x/y)^2)X differential of x/y found using a band limited differentiator of some sort as used in software radio.