[Elecraft] KDSP2 - artifact

[Siu Johnny]

Hi Peter,

Thanks to all the replies including yours.

I experienced exactly the same in the recent CQww SSb contest.  I really 
want to know how I should adjust the beta and decay time to minimise the 
artifact.  I don't mind having much less aggressiveness in NR during SSB 
operation.

In some extreme situation during the contest, I simply turned off the KDSP2

Could any of you in the list point me to the right direction?

TNX & 73

Johnny Siu VR2XMC


From: "PE1E" <PE1E at chello.nl>
To: "Siu Johnny" <jcpsiu at hotmail.com>,<elecraft at mailman.qth.net>
Subject: Re: [Elecraft] KDSP2 - artifact
Date: Tue, 31 Oct 2006 15:52:24 +0100

Johny,

My 0.02.
Underneath I review my K2 DSP SSB speech NR experiences and not the CW ones
( CW NR works much better ).

I also encountered considerable K2 DSP drawback problems ( I guess everyone
does.. ).
( I know about the free lunch, but even then.. ).
I have been tweaking the beta and decay for hours but whatever I tweaked,
problems still bother the most where you need the NR most, i.e. with weak
signals.

Actually, I encountered three kinds of K2 DSP drawbacks :

1. bothering bath tub song sound
2. seriously decreased intelligibility
3. audio artifacts ( distorted audio and splatter like crack sounds etc. )

The audio sample of the K2 SSB speech NR on Elecrafts website ( which
seduced me to buy the K2 DSP after having asked a question on this to the
Elecraft people, which question was not answered properly ) is kinda
misleading ( sorry Wayne ) in that way that it obviously offers a pretty
strong signal where the NR indeed does not suffer from decreased
intelligibility.
Not quite fair of the Elecraft people, since, in contrary, the NR is needed
most where the sigs are weakest.
I would advocate that the Elecraft people would also offer on their website
an audio sample of a weak signal, where the potential DSP buyer can get 
some
idea how the intelligibility suffers by the NR as the sigs get weaker.
DSP NR of ( other :-) commercial tranceivers suffers from the very same (
part of the NR concept ) but I feel it wouldn't be a shame for a high
quality concept like our K2 to show these K2 DSP SSB NR limitations to a
potential buyer before he buys.
Some would then decide not to buy this DSP NR.

BTW, the real audio artifacts ( not the bath tub sounds and not the
decreased intelligibility, but the real audio distortion ) are not really
kinda problem ( at least for me ) since they could be easily avoided by
decreasing the DSP gain setting ( drawback remains that e.g. changing 
filter
settings, changing bands et al requires changing of the gain setting every
time as well ).

The bath tub sounds, critisized repeatedly on this list, are no real 
problem
for me.
My ears/brain quickly adapt to this and though not pleasant indeed this 
does
not decrease intelligibility for me.
Funny to let my inmates think I am talking with little green men from Mars.

On the other hand, concluding and to be honest, I must frankly state that
even with weak signals ( be it not the really weakest ones ) the DSP does
improve intelligibility somewhat if ( and only if for me ) the IF filter is
( default ) widest and the audio filter is ( default ) 2 at maximum or even
better at 1.
Disadvantage is that in contest ( as I noticed in the very very crowded cq
www dx contest last weekend ) that selectivity lacks heavily then and
neighbour qrm is hardly endurable.
That's why ( in spite of some different opinions on list some weeks ago in
the IF shift discussions ) I still do appreciate so much my IF shift in my
FT-100 radios ( yes, I do know, please no extra conversion in the K2... ).

And, end of all, having said my griefs on the K2 SSB DSP, with some 
tweaking
and some good will, the K2 DSP is still more useable for me than I heard on
some commercial radio DSP's.

Do I regret having bought the DSP ?
No, but with some thoughts as exposed above.

Peter, PE1E

_________________________________________________________________
ÓëÁª»úµÄÅóÓѽøÐн»Á÷£¬ÇëʹÓà  Live Messenger; 
http://get.live.com/messenger/overview