[Elecraft] K2 DSP - Standardised Input for Comparison Testing

[Lyle Johnson]

Hello Mike!

> I've just read the article describing the design philosophy behind the new
> K2DSP.  The ability to adapt many of the filtering parameters is sure to
> result in a large amount of dialogue on the reflector about what settings
> are best for SSB, CW, RTTY etc.

As you hint in your posting, band conditions vary widely.  Noise conditions
vary widely.  Signal con...  you get the idea.  The ability to adjust the
settings *while you operate* is important for those who like to tweak
things.

I can adjust the settings to make certain cases sound really great -- but in
the real word of dynamically changing conditions, the resulting performance
might be very poor.

We picked default settings that seemed to give good results under typical
conditions.  But some people have very critical ears and are sensitive to
certain kids of distortions, while others don't and aren't.  Sort of like
allergies, I suppose!

In general, if you set the decay to very low values, you'll get better noise
or carrier rejection - but at the expense of more pronounced artifacts.
Similar with very high settings for beta.

> However, individual assessment of settings
> will be subjective unless there is a standardised input signal.

Exactly!  This whole topic is subjective.  Personally, I think it should be.

> What sort of empirical testing did the design team do?  What if
> anything did they use/do to standardise the input conditions?

During Field Test, folks were asked to play..er..test and report back what
they liked.  We then averaged things, added a weighting factor (my vote
counts for 3, yours for 1/2 :-) then got out the dartboard and the
chimpanzees.  But we selected the darts very carefully...

73,

Lyle KK7P (ex VK6GXD)

> AC7MZ  (also VK1KCK)