Olympus-OM
[Top] [All Lists]

Re: [OM] RMS Grain Size - slide film

Subject: Re: [OM] RMS Grain Size - slide film
From: Paul Wallich <pw@xxxxxxxxx>
Date: Sat, 12 Feb 2000 11:32:46 -0500
At 1:02 AM -0800 2/12/00, Jim Terazawa <jimt@xxxxxxxxxxx> wrote:

I am aware of a ~6million pixel CCD area imager
soon to be marketed (donno when).  This CCD, as I
understand, has the area of ~2.7um2 per pixel.
Today's 2.5~3million CCDs employ ~3.4um2 pixel.

I think with this high density CCDs I feel there
should be no impact to the image degradation for
still image capture operations.  As you might know,
the high-end/professional digicamcoder have
three CCD units with independent RGB filters.
Of course there will be some image degradations
whenever you start adding any elements in front
of the sensors.  But with improved wafer processing
at the fab and with digital signal processing(DSP)
applications, I think the image degradation due
to what kind of optics used are no longer issue.

What is the pitch for those high-density arrays?
If the pixels are about 1.65 um across, that sounds
like a spacing of 2-3 um (depending on fill factor) which
could give you another kind of optics problem. To get good
information onto those sensors you need a lens capable of
resolving 150-200 line pairs/mm (conversion from lines to
pixels is somewhere between 2 and 3 depending on relative phase).

Although shorter focal lengths may make this kind of resolution
somewhat easier to attain, you also start running into irreducible
diffraction problems around that pitch, since your sensor is only
about 4-6 wavelengths across and hence a measurable fraction of
your light will be coming from photons aimed at adjacent pixels.
Your MTF, politely speaking, goes to hell. (That, I think is why
the reported sizes of 3-Mp chips are substantially larger than would
be needed just to scale the 2Mp ones)

If sensor density keeps increasing, you could eventually wind up with
a completely different imaging architecture, however, one much more like
that of film. Imagine sensors scattered as finely as halid crystals, and
the data for each output pixel coming from a large, adaptively-chosen
set of input pixels...

paul   in about 10 years
Paul Wallich                                            pw@xxxxxxxxx

< 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 >


<Prev in Thread] Current Thread [Next in Thread>
Sponsored by Tako
Impressum | Datenschutz