[hpsdr] Steady Waterfall

[Roger Rehr]

If you try the 'large FFT' variation of PowerSDR that I posted last
year and again last month, it has a modified Linrad waterfall display
color scale [in addition to the standard Linrad color scale display]
that automatically adjusts the signal strength range over which the
color scale is applied, based on the noise floor and the maximal
signal strength detected.  It automatically compensates for changes in
the noise floor.

The code [both source and binaries] is at:

http://www.nitehawk.com/w3sz/HPSDR-PowerSDR-Metis-W3SZ.zip

The zipped file is large, approx 100 MB

Unfortunately, I did not change the code to keep the center of the
waterfall frequency scale at a fixed frequency, which I agree would be
EXTREMELY desirable.

73,

Roger Rehr
W3SZ
http://www.nitehawk.com/w3sz

> On Wed, Apr 27, 2011 at 11:46 AM, Ben Witvliet <pa5bw at xs4all.nl> wrote:
>> ***** High Performance Software Defined Radio Discussion List *****
>>
>>
>> Hello all,
>> (2) To get maximum sensitivity and best indication of weak signals, I set the (Setup - Display - Waterfall) LowLevel just below the noise level. I then set the HighLevel some 25 or 30 dB higher when the band is quiet. Weak signals are then showing up very clearly, and even short transmission are caught straight away.
>>
>> I found that adjusting the LowLevel is needed quite often, depending on the time of day, the band and the antenna used. And the HighLevel must be adjusted also to get best results. A band full of strong signals requires other settings than a very quiet band.
>>
>>
>>
>> Would it be an idea to set the Lowlevel automatically to the average noise level within the full receiver bandwidth (200 kHz)? The Waterfall-HighLevel could than be replaced by Waterfall-DR (Dynamic Range). Dr would be the present ´HighLevel´ minus ´LowLevel.  DR could be set automatically also, calculated from the strongest signals within full bandwidth or manually. Setting it manually to 30 dB would result in the DX setting I described above. If helpful I would be happy to devise a suitable formula for this.
>>
>>
>>
>> Kindest regards,
>>
>> Ben PA5BW
>>

 1303926798.0