Dear C.H.
You wrote:
"check the lower right-hand corner "E-10 All in one lens story"
The normal lens has smaller rear element, the light beam is divergent -
cause problem.
The digital lens has larger than CCD diameter rear concave element that
give
perpendicular beam - right design."
Many thanks for the link.
Ok I've had a look but I still don't believe what they are saying is
correct. However I'm going to check with some colleagues next week. So
I'll get back to you on this.
In the mean time I've got some physicists disagreements with their
statements. I'm bearing in mind here that this is written by advertising
people, not necessarily people who would question the science. No
offence meant to yourself if you happen to be in advertising!!
In 1) they say the film is flat so it can capture light from a variety
of angles. This is not strictly true since at very shallow angles the
light will reflect off even a film surface. More importantly the diagram
they show of a lens imaging onto film isn't strictly accurate. Imagine a
point source in the object plane say a star or a car headlight. Now it's
light rays fan out in many directions, and the front element of our
camera lens gets in their way. The whole of the disc which is the front
element of the lens is illuminated with light rays. These pass through
the inner elements of the lens, being refracted, focussed and defocussed
and emerge from the whole disc which is the rear element of our lens.
But this is not what they draw. They show the rays emanating from the
centre of the lens. This is not accurate.
Now in 2) they show parallel rays coming out of the whole of the rear
element. Well this is also not accurate and is misleading. Yes I can
arrange for rays to come out of the rear element parallel like that. But
they won't form a focussed image on either a lens or a CCD. Parallel
rays like that don't come to a focus (or strictly they meet at
infininity).
What they should really be drawing is a set of rays coming from all
parts of the rear element, and meeting at one of the pixels. This is the
basis for the calculations I made. The angles I quote are the most
oblique rays the lens will produce and will will meet the CCD at its
furthest corner.
Now they talk about the pixel wells being so deep that oblique rays wont
be captured. We can calculate this for our extreme case again, of the
50mm lens, which gives an angle of 26 degrees with the CCD plane.
Looking at the size (2/3") and 4M pixels gives me a pixel size of 8
microns. For this ray to 'miss' the CCD well I can calculate the depth
the well would need to be. It is Tan(26)*8microns, or about 4 microns
deep. I'm going to check with my colleagues (who design CCDs for our
applications) but I think they are going to say this a most unlikely
structure.
Having seen it the page I can also find other arguments to disagree with
(sorry but I'm just that kind of a guy 8^) ). They talk about
the dust problem leading them to a fixed lens solution (i.e. one sealed
against the dust. Yes I can see it could be a problem, but I'm sure a
thin multicoated window in front of the CCD would easily solve it.
Then they talk about needing to fix the lens because an interchangeable
lens would flex too much. I find this hard to believe. I don't think the
olympus mount 'flexes' much during even a relatively long exposure.
They are right that to acheive 35mm format resolution on a 2/3" CCD the
lens needs to be of higher basic resolution. but they haven't acheived
this level with their CCD yet. A good zuiko can image 100 lines/mm,
which is a feature size of 10micron. Their CCD pixel is nealy as big, at
8 microns. To acheive equivalent resolution they would need to be down
to less than 5micron. So, who are they kidding!
If 'flexing' is such a problem then so will be temperature changes,
which can eisily cause flexing of the order of 5microns.
As I said, many thanks for directing me to it. I will be checking some
of their statements with my colleagues. But on my first viewing I would
say they get full marks for marketing hype and very poor marks for their
science.
regards
Chris Barrett
< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >
|