W Shumaker wrote:
>You can never re-create data that is missing.
>
Yup
>All processing will be a tradeoff.
>
Agreed
>Smaller pixels will always be noisier than big pixels,
>
Yes...but. Once noise goes below some threshold, further reduction
doesn't matter for most applications. Once a smaller sensor crosses that
threshold, it replaces the larger one.
>along with more color fringing, lens resolution, etc. issues.
>
>I purchased an E-1 for what the camera is today. The future will
>not produce lower noise higher pixel count with a fixed sensor
>size, unless.... they are able to produce a sensor with more
>pixels without reducing the per pixel area. (Foveon perhaps)
>
I disagree. The history of technology is, if nothing else, a graveyard
of unbreakable limits that were broken. It may be further development of
existing sensor technology or a new technology, but I'm certain that it
will happen. Another thing that I heve read is that noise in current
cameras is often partially a function of affordable amplifier technology
for price sensitive consumer products. And say, how about embedding an
electronic cooling junction behind each pixel sensor?
>.... Additionally there is the new E-1bw to
>color adaptor that takes 3 photos in a row with fast a switching
>color filter for RGB generation, a true pixel count camera.
>
Dosen't Foveon already do something better, all 3 images at once?
>Oh, and did I mention the new Olympus E-1ms? The new sensor
>does multi-sampling like an over-sampled A/D converter to get
>higher dynamic range, sort of like the multi-sample in a scanner.
>The difference is that an anti-shake and anti-subject-movement
>algorithm can be applied to the multi-samples. The only loss is
>around the perimeter of the image. This will be the first very
>innovative anti-subject-movement compensating camera.
>
Some video cameras already do this, although the details are different.
They adjust the position of frames to match repeating elements from the
prior frame, I believe. They do, in any case, compensate for subject
movement at the cost of loss of some of the pixel count of the sensor.
>Already there are sensors that can do very high
>sample rates, in the megahertz range, but when the pixel count
>gets up there, remember you heard it on this list first.
>
Yup, from me, several months ago. I blathered about IS and said that
electronic image stabilization for any image movement on the sensor, no
matter what the cause, is theoretically easy, but will have to wait for
adequate frame/sample rates.
>Another sampling mode is long and short samples, long samples
>pull out the shadow details, while the short samples fill in the
>highlights. Multi-sampling is the ticket to revolutionary new
>digicam design.
>
I can see that. I just wonder if lower noise levels (on which we
disagree) won't increase dynamic range to the point where multi sampling
is unnecessary. The dynamic range of better DSLRs is already better than
much slide film.
> You can't do that with film. Now on to new
>mega-amp-hour battery design... Cameras with built-in venting
>fans, they hum like your laptop fan...
>
>Wayne - who ain't been smokin nottin
>
Thanks,
Moose
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|