Mailinglist Archives:
Infrared
Panorama
Photo-3D
Tech-3D
Sell-3D
MF3D
|
|
Notice |
This mailinglist archive is frozen since May 2001, i.e. it will stay online but will not be updated.
|
|
Re: What is APS?
- From: P3D Gregory J. Wageman <gjw@xxxxxxxxxx>
- Subject: Re: What is APS?
- Date: Wed, 11 Sep 1996 12:50:12 -0700
Mike K. responds:
>Those microcontrollers don't even touch the coat-tails of the throw-away
>packaging of the new/next generation high-end DSP chips *now*.
Yes, but I ask again, what are you going to *process* with all that
horsepower, when the digital imager is the weakest point in the
system? Not the optics, not the processing, the image capture. All
the sophisticated image processing algorithms in the world won't help
you when you can't capture the original image with enough resolution.
I spent six years working for a medical imaging company. I know many
of the digital games you can play to enhance contrast, or reduce
pixellation (at the cost of resolution), but you can't enhance
resolution that isn't there in the first place. At best, you can
compensate for other flaws in the system, such as distortions in the
optics. But you can't add resolution.
>Yes. Digital camera optics/resolution is expected to improve as well
>at the lower price classes. One *can* buy adequate digital cameras now
>at high prices (how's that one that's built inside a Nikon body by Kodak?).
The camera/imager combination you're talking about only has, at best,
about one third the resolution of 35mm film. The imaging chip isn't
quite the size of a 35mm film frame, and number of pixels is "only"
around 6 million (that's about 2500 by 2500 if it were square).
Pretty impressive, but I've read that film captures the equivalent
of anywhere from 4000 to 8000 "pixels" in the same space.
One of the limitations I left out is the fact that to capture a color
image, you either have to take three separate exposures onto the same
imager through color filters (not practical if you're out to emulate
an SLR), or you sacrifice resolution once again by using R/G/B
sensitive triads of cells (like a color TV tube in reverse) to capture
it in one go. So now you've cut your resolution by 3.
>Not that hard. You only have to cool the chip which is very small. You
>could just attach (or integrate on the same chip) an electronic cooler
>(the principle escapes my brain at the moment... they're used in el-cheapo
>Igloo picnic coolers now... but only a very tiny version is needed to keep
>the very small featherweight chip cool ).
It's called a "thermionic heat pump". But these things suck power.
Those coolers typically plug into your car battery through the
cigar lighter socket. You want to put a car battery into your camera?
I admit it might make it feel more like a Realist... Oh, and don't
forget to turn it on an hour before you want to shoot.
>One also can just make the imaging chip bigger instead.
Assuming you mean keeping the same number of pixels, this doesn't
increase resolution, it decreases it. The area of each pixel goes up,
meaning that a larger portion of the image falls on each pixel,
resulting in an 'averaging' effect. To get more resolution, you have
to make the pixels smaller, which decreases their sensitivity
following the inverse square law. Film has the same problem; the
finer you make the grain, the slower the film. It still takes a
certain number of incident photons to record an image. Those pesky
laws of physics again.
Why is medium-format superior to 35mm? It uses the same film, with
the same intrinsic resolution. It just uses more of it. The analogy
with a digital imager is more pixels, not larger ones. But now the
problem becomes one of yield. The larger you make your chip, the harder
it is to get one that is 100% functional. That drives up the cost.
>Kodak has decent digital cameras now -- some of which are used by a
>few professional photographers (based on magazine stories). Cost is too
>high currently -- but the price of the low end has been dropping rapidly
>with quality going up.
People on this list complain about Kodachrome 64 being too grainy for
3D work. The best digital camera on the planet is at least 2-3 times
worse than that, and I'm probably being generous. The technology will
have to improve by a factor of 2-3 times JUST TO CATCH UP TO 35mm
FILM, let alone surpass it. This kind of improvement is not a matter
of successive refinement; some real breakthroughs are needed. The sad
part is that if people are willing to settle for what's available now,
we'll never get there.
-Greg
------------------------------
|