Re: FFT1 Resolution Bandwidth - back to basics

[Leif Asbrink <sm5bsz.com; leif@xxxxxxxxxxxxxxxx>]

Hi Guy,

> I quite understand that provided FFT2 is enabled, the main function of
> FFT1 is to enable the removal of strong signals before the Noise
> Blankers. This does not need a particularly narrow resolution
> bandwidth, as Leif has often written, e.g.
> > For fft1 it is a good idea to set a high bandwidth
> > like 200 Hz with a sin 3 window.
> > It will give a rather wide peak but with very good shape factor .....
> 
> Now if you choose a particular window power (like 2 or 3), and input a
> particular bandwidth, you will actually get a rather lower value of
> BW, often much lower, because the FFT size has to be an exact power of
> 2.
> I take it that Leif's comment really refers to the bandwidth you
> actually get rather than the one you ask for.
If you ask for 200 Hz you should get a bandwidth 
around 200 Hz. Usually it will be a bit smaller, but sometimes
it could be a bit larger. Quite often another power of two
would give a bandwidth that is closer to 200Hz. 

> With 96 kHz sampling, as with the Delta-44, we find that:
> FFT size 1024 gives a BW of 187Hz for a sine^2 window, and a BW of
> 225Hz for a sine^3 window.
> No other FFT size comes even close to 200 Hz, and either of these
> seems ok, though I think Leif would choose the higher power window.
When asking for a large bandwidth like this one has to use a sine^3
window to get enough suppression at something like 1 kHz away from
a strong signal. Another alternative would be to go for something
like 20 Hz bandwidth with a sine^2 window. It would also have
adequate attenuation 1 kHz away but time delay would be larger
and it would also cost more CPU time.

> With 111.111 kHz sampling, as I am currently using with my SDR-IQ, we
> find that:
> FFT size 1024 gives a BW of 217Hz with a sine^2 window, and a BW of
> 260Hz with a sine^3 window.
> FFT size 2048 gives a BW of 130Hz with a sine^3 window, and a BW of
> 149Hz with a sine^4 window.
> Now which one would Leif recommend?
> The sine^2 window with 217Hz, sine^3 window with 260Hz or 130Hz, or
> the sine^4 window with 149Hz?
> My guess is the sine^3 window with 260Hz! Leif?
It depencs on what you want to achieve.
If you really have a neighbour with an extremely good transmitter
that would allow you to listen to weak stations, more than 100dB
weaker than him 1 kHz away while you also need the noise blanker
to remove strong QRN, these things become important. Under normal
circumstances it does not matter what you choose.

Look here:
http://www.sm5bsz.com/linuxdsp/blanker/fp/fpblank.htm
http://www.sm5bsz.com/linuxdsp/rxiq/demo/blanker.htm
http://www.sm5bsz.com/linuxdsp/install/dlevel.htm
Some info about how the blankers work. I think I have written
a page about why a sine^3 is much better than sine^2
when fft1 bandwidth is large, but I could not find it.
 
> Footnote to Leif:
> Thanks for the C-coding suggestion re the attenuator in the SDR-IQ.
> I understand the change, I would have tried this already myself, but I
> have no C-compiler!
> You may remember that I am running under Windows!
Oooh! Look here:
http://www.sm5bsz.com/linuxdsp/linrad.htm
the line: "To compile linrad.exe from source code under Windows,"

The compilers can be downloaded from my site (15 megabytes).
Follow the simple instructions and you will have linrad.exe
compiled from source code under Windows:-)


73

Leif / SM5BSZ

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "Linrad" group.
To post to this group, send email to linrad@xxxxxxxxxxxxxxxx
To unsubscribe from this group, send email to linrad-unsubscribe@xxxxxxxxxxxxxxxx
For more options, visit this group at http://groups.google.com/group/linrad
-~----------~----~----~----~------~----~------~--~---


LINRADDARNIL