From: ken Watson (watsok@frii.com)
Date: 09/13/00-08:59:52 AM Z
Sil,
I read the article and it is referring to video CMOS devices vs still camera
devices. In addition it has a lot theoretical positions taken because the
company admits it switched to CMOS from CCD's. Rockwell is not a player in
the market for sensors for digital still cameras. The players are Sony,
Kodak and a few others. I believe their market remains in the military and
weapons arena. They even , sorta admitted this when talking about infra red
sensors and Astronomy sensors, all requiring to placed in a Dewar to
function properly. You might need gloves to hand hold a camera with a sensor
at these temperatures.
The circuit topology the are discussing requires/specifies a lot of support
circuitry for each pixel. So if you have 4 million pixels/photodiodes you
have 4 million copies of this circuit as well. This makes the sensor larger
and a lot more expensive. In addition this circuitry takes up space that
could be gathering light thus lowering the signal to noise level. Also by
having a larger die area and more active circuitry create a larger
opportunity for a defect lowering yield. Another clue is they talk glowingly
of the cmos process allowing for a camera on a chip but where is it? There
are none today and I can say this discussion has been going on for over
three years.
If you looked at the noise graph ( which only measured one type of noise out
of about four sources) it only discussed shift noise which is the weak point
of high speed VIDEO CCD sensors, not much of an issue in still cameras. The
speed in today's still cameras is limited by how fast the electronics can
compress the image for storage and still have enough battery power to take a
reasonable number of shots.
From reading this it looks like Rockwell has taken their spy satellite type
of sensor and is trying to get into the commercial arena. Sensors only
optimized for performance are available in both CMOS and CCD. They fail in
the marketplace because the price of the equipment interests only a few.
Also a comment on infra red. All silicon sensors are very sensitive to red
and near infra red. So much so camera makers usually add a IR cut filter. If
you want a near infra red camera and can find this filter and remove it you
now can take IR shots. The problem may be that there is a IR cut coating on
one of the lens though.
..-----Original Message-----
..From: Sil Horwitz [mailto:silh@earthlink.net]
..Sent: Tuesday, September 12, 2000 9:34 PM
..To: alt-photo-process-l@skyway.usask.ca
..Subject: RE: Advances in CCD to go with Canon
..
..
..At 2000/09/12 08:59 PM -0500, Jamie wrote:
.. >Phillips is making a great CCD chip right now which is used in the leaf
..and megavision camera >backs,and probably others.
..
..They may be, but CCD chips are history - just wait about 6
..months. For full
..details, see:
..http://www.rsc.rockwell.com/html/cmos.html (If you do visit that
..page, be
..prepared for some really heavy technical reading! Very
..interesting, even if
..one doesn't understand it all.)
..
..Interestingly, and opposite to what I had thought and others have noted,
..CMOS technology is less noisy than CCD, and much, much less expensive.
..There already are CMOS InfraRed chips, so how long will it be for UV? I'm
..still looking forward to some UV transistorized device that you can lay
..right on an alternative photographic material, connected to your
..computer,
..which will expose the material directly without an intermediate film or
..paper to get in the way. How about that? It is very possible.
..
..
..Sil Horwitz, FPSA
..Technical Editor, PSA Journal
..teched@psa-photo.org
..silh@earthlink.net
..Visit http://www.psa-photo.org/
..Personal page: http://home.earthlink.net/~silh/
..
This archive was generated by hypermail 2b29 : 10/01/00-12:08:59 PM Z CDT