I know mtf is a curve, but to specify "resolution"
one but also specify at what ration of black to white
contrast is diminished to.
JCO
----------------------------------------------------------------------------
----
J.C. O'Connell mailto:[EMAIL PROTECTED] http://jcoconnell.com
----------------------------------------------------------------------------
----
-----Original Message-----
From: Bob Blakely [mailto:[EMAIL PROTECTED]
Sent: Monday, September 15, 2003 4:38 PM
To: [EMAIL PROTECTED]
Subject: RE: Lens resolution
wrong!
You must have at least *two* black lines separated by white space or at
least two white lines separated by a black space. The edge of resolution is
defined as the boundary distance where the two lines can just (almost) be
discerned as separate vs. one line or where the two lines just (almost)
merge to one line or where a series of equally spaced & sized black lines
and white spaces merge to 50% gray without spatial ripple. It's the same
number and it's not the same as the MTF or Modulation Transfer Function. It
is only one point on the function.
> From: J. C. O'Connell [mailto:[EMAIL PROTECTED]
>
> I dont agree with the final conclusion in this post.
> resolution is detemined by 1 black line next to
> one white line and determining a reference of how thin they
> can be made before the black and white fade into a middle grey
> value, MTF to be more specific...
> JCO
>
> Bob Blakely wrote:
> >
> > Logically, because for one line to exist, it must have a
> background. If it
> > is thought of as a black line then it has a white background. If it is
> > thought of as a white line then it has a black background. By
> saying that
> > the "pair" is a black line and a white line, then it is not possible to
> draw
> > one line pair!
> >
> > No one draws a black line on a white piece of paper and claims
> it's a line
> > pair!
> >
> > No one draws two black lines parallel together on a sheet of paper and
> > claims it's four lines!
> >
> > No one draws 100 black lines parallel together on a sheet of paper and
> > claims it's 200 lines!
>
> Thank you Bob!
> That says it, doesn't it?
>
> > I'm starting to become convinced that the idea of "line pairs per mm"
> > started as the result of some anal retentive who, for the sake
> of argument
> > alone, claimed that one could think of the resolution target as white
> lines
> > on a black background just as well as black lines on a white background
> > (true enough). Either he or others decided that it was somehow "more
> > scientific, more intellectual or whatever" to start calling
> both the black
> > area and the white area as lines. To me, this leap is nonsense. I am
> quickly
> > beginning to believe that "line pairs per millimeter" is
> nothing more than
> a
> > pseudo intellectual attempt to sound more knowledgeable about the topic,
> and
> > (in fact) is really the same as lines per millimeter. Indeed, a 10^100
> > search of the web shows that the terms are often used
> interchangeably and
> > refer to the same number.
>
> [snipped]
>
> Resolution as spoken of by astronomers works with points of light
> against a very black background, and they talk about separating (or
> "resolving") known twin stars, which have known separations, and it's a
> test of their system to see if they can separate the two -- see two
> distinct stars.
> This brings up talk of airy discs and such, but there's no talk about
> white stars and black stars, there's only two close together white stars
> against the black firmament.
>
> So it seems camera lens resolution should be just the lines themselves.
> If you prefer, black lines drawn on white or gray paper.
> And like astronomers do, it's a test to see how close together those
> parallel lines can be and still be resolved into a distinct pair of
> black lines, or whether it's a bar-shaped smudge.
>
> That sort of says it all, far as I'm concerned.
>
> keith whaley
>
>
>
