RE: [Leica] Nyquist again (was scanning)

["Austin Franklin" <darkroom@ix.netcom.com>]

> > The problem is "virtually any alignment".
>
> Not really.  Only a handful of very precise alignments will
> produce artifacts,
> and those alignments are extremely improbable in practice.

You just agreed that I am correct.  If even one case fails, your premise is
wrong.  Like I said, you obviously haven't really done any real design work.

> > Of course I disagree, because your claim is
> > wrong.  It would be a range, equal to 53 lp/mm
> > down to 1/2 that.
>
> I didn't ask for a range, I asked for a specific figure.  If you
> are going to
> give a range, then I expect to see the MTF curves themselves.

The information I gave was stated correctly.

> > This is really really simple to understand.
>
> Perhaps, but since there is so much more to resolution, image capture, and
> visual perception, it may also be simplistic in consequence.
>
> > 1/2 that and I've explained why.
>
> No.  You've mentioned it, but you haven't explained why.

Go back and read the thread.  It is clearly explained, that is, unless you
just want to play games, which you are doing.

> > You haven't explained why, nor have you
> > explained why my correction to your mistaken
> > assertion is not correct.
>
> In my previous post, I explained a great deal about why 53 lp/mm,
> apart from
> being quite achievable in the scanner, is also more than adequate
> for virtually
> all purposes.

I looked back, and find no such "explanation", I only find figures and
incorrect assumptions stated.  Please provide a correct explanation that is
not just faulty and repeated assumptions.

> > Yeah...and isn't 44.1kHz slightly more than
> > 22kHz?
>
> Yes ... but you said 20 kHz.

So?  Can't 20kHz be sampled by a 44.1kHz sampling rate?  This is clearly a
diversion on your part to avoid answering the question that obviously shows
that your claims are wrong.

> > Reliable is hardly amorphous.
>
> Well, provide some figures that define it, then.

Go look it up in a dictionary for your self.  Again, you're just playing
games.

> > I have given you a very clear definition of it.
>
> No.  A clear definition is something like "50% modulation
> transfer at 40 lp/mm."
> Something with numbers, that is.

Oh, well, you said "No", so I must be wrong!  If your claim fails in ONE
case, it is a faulty claim.  Your claim DOES fail in at least one case, so
it IS a faulty claim.  Reliably means does not fail in ANY case.

> > As I stated, it is a standard term in signal
> > processing, and if you have any experience in
> > signal processing, as you claim you do, you would
> > understand what "reliable" means, and that it
> > is a VERY specific term.
>
> I'll understand it when you put numbers behind it, as I have been
> doing for my
> assertions all along.

You apparently can't or don't want to understand anything, so no matter what
I "put", you'll just play games.

> > You made the original claim, and you still fail
> > to substantiate it.
>
> I gave the numbers and sources in my previous post.

The numbers don't substantiate squat.  Just claiming 2700PPI is sufficient
to sample 53 lp/mm in EVERY case is not substantiation, it is a claim, and a
faulty one at that.