FW: Re:spectran,dspfil, please clarify (Leif Åsbrink Oct 02 2002)

[Leif Åsbrink <leif.asbrink@xxxxxxxxxxxxxxxxxxxx]

-----Original Message-----
From: owner-linrad@xxxxxxxxxxxxxxxxxxxxx
[mailto:owner-linrad@xxxxxxxxxxxxxxxxxxxxx]On Behalf Of Leif Åsbrink
Sent: Wednesday, October 02, 2002 12:05 PM
To: linrad@xxxxxxxxxxxxxxxxxxxxx
Subject: RE: [linrad] Re:spectran,dspfil, please clarify


Hi Arie, Dom and others,
 
Just a comment to:

> Yes that could be interesting. But..... till now the waterfalls on my
> windows dsp programmes are much much faster than on linrad with 5000, 8000
> or 16000 hz sampling. Higher rates are not possible with my linrad. But in
> Windows I easily can run 48 kHz sampling.
> I currently of course only can view the 3 kHz audio as I have not made any
> convertor to audio from RF. 

The waterfall speed does not depend at all on the sampling speed.
Each line in any waterfall graph is the power spectrum vs frequency
averaged N times where N is anything from 1 to some very large number.

The time between two sucessive power spectra is 1/bw where bw is the 
bin bandwidth. In Linrad you can set N and bw to whatever value you
like and there is no particular waterfall speed associated with Linrad.

You may set bw to 200Hz and N to 1 to get 200 lines per second which means
that the waterfall fills half the screen (1024 x 768) in about 2 seconds.
You may also set bw to 1Hz and N to 3600 to get a new line once an hour.

The visibility for signals staying in the same frequency bin for
a time T will be optimum if the time for one waterfall line is T/n
where nis somewhere between 1 and 5 and if the waterfall is zoomed
for at least one pixel per fft bin. 

Linrad does not allow you to set bw accurately because bw is affected 
by the window function you have selected. The fft size has to be a power 
of two so when you ask for 200Hz, the actual bin bandwidth may be 
anywhere between about 130Hz and 330Hz.

73

Leif  /  SM5BSZ 



LINRADDARNIL