Re: More on A.pellucida

Hello, Rene and GR:

I am getting headache from reading this part of the thread. I am convinced
that both of you know how to succeed in photomicrography, but this part of
the thread entered the state of verbal jousting and looping.

I can only hope that both agree on the following basic summary:

1) Smallest resolvable distance 'd' (diameter of Airy disk) is given by d =
0.61*lambda/NA(obj) assuming we are using a NA(cond) of 1.4 and hence get no
limitation due to insufficient illumination.

2) The size of the smallest resolvable distance 'D' in the INTERMEDIATE
image (also referred to as real image) is given by D = d * magnification of
objective

3) To resolve a feature of the size 'D', one needs a pixel size of D/2
(Nyquist theorem [1]) to avoid undersampling. But remember, this is the
ACTIVE pixel size and not the PHYSICAL pixel size in case of a Bayer filter
(see point 4). Let's call this the required pixel size 'r'.

4) The required pixel size for a Bayer filter depends now on the wavelength.
If blue light is used the active pixel size is 2 times the physical pixel
size. For green, it is square-root 2 times the physical pixel size. Let's
call a = D/2 to be the active pixel size.

5) To be able to resolve the entire image, the active pixel size must be
smaller or equal to the required pixel size: a <= r

Now let's look at an example:

For a Nikon D70 (I don't know the Pentax that GR is using), the active pixel
size 'a' is a HUGE 12um when we use primarily blue light (no green). This is
much too big to resolve the fine structure with the CCD of a D70 located in
the intermediate image plane (direct projection). For an objective with NA
1.40 with lambda = 425nm, d = 0.185 and D = 18.5 um. This gives r = 9.25 um,
which is smaller than the 12um provided by the D70 for blue light. When
using a 2x projection ocular (or relay lens) the image information is spread
out onto a larger area and hence we can multiply 18.5 um by two = 37 um.
Now, r = 18.5um, which is easily captured by the image sensor's pixels with
a PHYSICAL size of 6um and using blue light.

Now that would explain one aspect of the fact that GR cannot resolve AP when
using direct projection but by using a relay lens. Another important aspect
is, of course, the fact that using the appropriate relay lens, one gets a
complete correction for aberrations when using older gears.

Cheers,

Gregor

[1] http://en.wikipedia.org/wiki/Nyquist-Shannon_sampling_theorem

"rene" <renevanwezel@xxxxxxxxxxx> wrote in message
news:1170089222.158026.197660@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
GR, I still find you lacking on all points except for the Baertierchen
info.

You seem confused by the issue of contrast. I will say it
one more, final time: contrast increases when the target's
image spatial frquencies move to the left on the MTF curve.
The only way for this to happen at a fixed pixel or grain size is
to decrease the spatial frequency, that is to increase the size of
the image.

So what you are suggesting is that if we enlarge the diatom, then we
are able to see more. Duh.

René.


.

Relevant Pages

  • Re: More on A.pellucida
    ... The pixel size for a Bayer filter depends now on the wavelength. ... If blue light is used, the active pixel size would be 2 times the physical ... it is square-root of 2 times the physical pixel size. ...
    (sci.techniques.microscopy)
  • Re: Silly question....
    ... >> the pixel level. ... >> planetary camera, was selected to just about meet the critical sampling ... >> resolvable by the optics. ... The point is a question of what you mean by 'resolve'. ...
    (uk.sci.astronomy)
  • Re: High Definition TV UK FAQ
    ... calculating what pixel size is needed on a certain distance. ... I don't want to be able to resolve the very sharp pixels in an LCD ... My own tests give me a maximum wanted pixel size around 0.85 arc ... It sounds like you have better than 20/20 vision. ...
    (alt.tv.tech.hdtv)
  • Re: High Definition TV UK FAQ
    ... calculating what pixel size is needed on a certain distance. ... I don't want to be able to resolve the very sharp pixels in an LCD ... My own tests give me a maximum wanted pixel size around 0.85 arc ... minute to get a smooth picture on my LCD-TV (and based on the native ...
    (alt.tv.tech.hdtv)
  • Re: Need to convert from Percent Width to Pixel Width
    ... The problem here is that the physical pixel width will only be determined at ... the client side. ... The server is not aware of the physical rendering on the ...
    (microsoft.public.dotnet.framework.aspnet)