In your message dated: Sat, 05 Sep 1998 19:45:00 EDT you write:
>"Why would you build a camera with CCD to take superb OM lenses when cheap
>zooms on point and shoot digital cameras exceed the image quality that can
>be delivered by a current commercially available CCD?" - Winsor Crosby, Long
>Beach, California
>
>Er... Current CCD technology uses 10-micron pixels, i.e., 100 pixels to the
>millimeter = 50 lines per millimeter. Although some Zuiko lenses are
>appreciably better than that, it's not an unreasonable match.
True, but that is *only* for black and white. If you want color, you
have to use 3 pixels each with a microfilter (red, green, blue) in front
of them, so you've cut that 50 lines/mm down to a Polaroid-like 17 or
so.
>
>Did you think CCDs were inherently 640x480 just because the cheap ones are?
>Even those have 10-micron pixels -- on a 6.4 x 4.8-millimeter chip!
>
>The digital OM-5 should of course have a 2000 x 3000-pixel array, like
>Nikon's best.
I went thru a *long* argument why it will be a *long* time before ccds
will be as good as film for normal color photography a few months back.
Basically, you need 4kx4k CCD's to get to the resolution of 35mm in
color. Below is a copy of that post.
Cheers,
--Lee
(Message outolympus:514)
Return-Path: lhawkins
To: olympus@xxxxxxxxxxxxxxx
Cc: lhawkins
Subject: CCD's vs. Film --- an astronomer's perspective....
Date: Thu, 19 Mar 1998 11:33:13 -0500
From: "R. Lee Hawkins" <lhawkins>
Hi all:
I thought I'd chime in on the film vs. CCD issue, as a few important
points have not been discussed yet.
First, it's important to keep in mind that film grains are *binary*
detectors. A film grain is either exposed or it isn't. Thus, saying
the resolution of film is equal to film grain size isn't exactly
correct, since it takes just as many film grains to get a given dynamic
range as it does bits in a CCD a/d converter. This all works out to
mean that a good CCD pixel size to match film is 9 microns on a side.
What a coincidence that most Kodak CCD pixels have exactly this size :)
Ok, now that we know that, let's figure out what size CCD we need (in
pixels):
36/0.009 = 4000 pixels (H)
24/0.009 = 2667 pixels (V)
This gives a total number of pixels for a 35mm frame of:
10,668,000
Now, some problems:
1) Dark noise and read noise: Older generation chips had to be
read out slowly to minimize read noise. This, however, meant
that they had to be cooled, since they created so many thermal
electrons at room temperature that even a 10 second readout
would saturate them. Kodak seems to have solved both of these
problems in their newer chips, so they can read out fast with
low read noise. The dark current for these chips is also much
lower than that for earlier chips.
2) Image storage: This has already been discussed as a problem,
but with 32 bit/pixel, this would translate to an image of
~40 Mb. Pretty large to put in a handheld camera.
3) Device yield: The cost of a CCD chip is directly related to
how many you can successfully make out of given wafer of
silicon. Likewise, the more complex (read: the more pixels) a
chip is, the more likely you are to have some bad chips.
While the situation has gotten better in recent years, and
continues to improve, I think chips that are ~4k x ~3k will
still be in the $10,000 range for a number of years to come.
I don't see mom & dad buying a $10,000 camera. (Actually, this
is assuming a lot... 2k x 2k SITe chips cost $75,000 right
now).
4) Color photography: All of the above assumes that you want
to do only B&W photography. If you want to do color, you have
3 choices:
a) Use "on chip" filter arrays, and cut your
resolution by 3.
b) Make your pixels 3x smaller, and use on chip
filter arrays, which means you need 9x more pixels.
This probably makes your CCD far too expensive.
In fact, the state of the art is 9k x 9k, and that's
not up to the 12k x 9k we need.
c) Use three separate CCD's, one for each color.
This works well, and only increases the cost
of your camera by 3x. So just the CCD's cost
$30,000 :)
As you can see, I don't think digital photography will replace film in
the near future. However, I think it will kill polaroid if it gets
cheap enough. No one likes the smelly polaroid chemicals.
With all that said, we had the next to best Oly digital camera with us
on my recent eclipse expedition, and it was *great* for documenting the
experiments, and we even got some images of totality. You can see one
of these images at:
ftp://pao.gsfc.nasa.gov/newsmedia/eclipse/TWO_VIEWS.JPG (case sensitive)
Disregard the caption --- it was not taken through a telescope, but with
the zoom on the Oly. Note that this image on has the inner corona, and
the dynamic range isn't anything to write home about. However, all in
all we found it to be a very useful camera.
Cheers,
--Lee
________________________________________________________________________
R. Lee Hawkins lhawkins@xxxxxxxxxxxxxxxxxxx
Department of Astronomy lhawkins@xxxxxxxxxxxxx
Whitin Observatory http://www.astro.wellesley.edu/lhawkins/lee.html
Wellesley College Ph. 781-283-2708
Wellesley, MA 02181 FAX 781-283-3667
________________________________________________________________________
< 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 >
|
