Linrad 3.34 report : spur. removal feature

[ON4KHG]

Leif, Dears,


Yesterday I have installed and tried Linrad 3.34 but I experience one
issue with the spur. removal :

- Computer with XP and Linrad 3.27 => spur removal works fine
- Same XP computer and Linrad 3.34 with all par_... files copied from
3.27 (so, same params as in 3.27) => spur removal doesn't work : when
spur removal activated via the top right "C" in the wideband
waterfall, there is a kind of notch of a few dB's and which lasts
maximum 1 sec on the spur. but that is all.

What kind of recording can I send you to demonstrate the problem
(a .wav recorded through pressing "W" on keyboard or some other kind
of recording ?)

Thanks & 73,

Gaëtan, ON4KHG

[Leif Asbrink]

Hello Gaëtan,

Oooh! There are many things that can go wrong.....

The spur removal in Linrad-03.34 does not work because of
the same error that causes the second fft to produce
transforms in groups rather than one by one.

Now, that is not really the error although fixing it
will restore spur removal functionality.

The code is intended to allow for occasional overload
from other programs. That could make the second FFT output
come in a group now and then. Provided that the user
has set an adequate output delay margin to accomodate
for such interrupts in the Linrad processing everything
should work properly.

It is obvious (now) that the spur removal algorithm
is in error. I will look into this.

In the meantime, if you use the spur removal together
with the second FFT, do this correction:

To fix the error, edit line 379 in wcw.c:
if( ((timf2_pn2-local_timf2_px+timf2_totbytes)&timf2_mask) >=

Change it to:
if( ((timf2_pn2-timf2_px+timf2_totbytes)&timf2_mask) >=

NOTE THAT THE PREVIOUS MAIL ABOUT THIS CORRECTION GAVE
THE WRONG FILE. Look in wcw.c and not fft2.c

73

Leif / SM5BSZ

[ON4KHG]

Thanks for the feed-back Leif.
I have made the correction in the wcw.c file but it remains the same (no
spur removal or very much time limited and more an attenuation than a true
removal, as it was the case in 3.27), both from an audio point of view but
also (which I didn't mention in my previous message) from a graphical point
of view. Indeed, activating (yellow) or deactivating (blue) the "C" at the
top right of the wideband waterfall, no spur is removed on the waterfall (as
it was the case in 3.27) even though the number above the "C" indicates an
amount of spurs supposed to be removed.

Same behaviour in SSB.

73,

Gaëtan

-----Message d'origine-----
De : lin...@googlegroups.com [mailto:lin...@googlegroups.com] De la part de
Leif Asbrink
Envoyé : jeudi 9 février 2012 18:38
À : lin...@googlegroups.com
Objet : [Linrad] Re: Linrad 3.34 report : spur. removal feature