h a l f b a k e r yNot just a think tank. An entire army of think.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Colour digital cameras have a sensor chip which records the
image. Individual pixels on the chip record only one colour
each - this is done by putting a grid of red, green and blue
filters in front of the chip. So, if a given pixel records blue
light, the intensity of this light is then stored
in the
camera's
memory. There's no direct method of finding out how much
red or green light should be recorded for that pixel, so to
get
a value, the camera interpolates the values recorded at
neighbouring 'red' and 'green' pixels. This interpolation takes
time and causes image errors (particularly on high-contrast
edges and highlights on shiny objects). My idea is this: It
should be possible to have a camera on which you can slide
the grid of filters off the sensor chip and have every pixel
record the 'white' light intensity. This will lead to greyscale
images, but with a much higher frame rate and (because no
interpolation will be needed) the image will be recorded at
the camera's true resolution. Sometimes it's acceptable to
trade off colour for speed and resolution.
[Note:
This
is a halfbaked rather than a good idea because within a
couple of years you'll have digital cameras which record
red,
green and blue for every pixel. This can be done by stacking
the sensors on top of each other with 'red' at the bottom
and
'blue' at the top. Red light will penetrate to the bottom
sensor because it has a longer wavelength. Clever, eh?]
shocker! no wonder no one can afford them!
http://www.jessops....search.cfm?node=481 you can buy cars for this amount of money! [sappho, Apr 26 2002, last modified Oct 05 2004]
New Leica
http://www.petapixe...arrive-on-may-10th/ Now Leica are going to make this - only 10 years after I wrote the idea. [hippo, Apr 23 2012]
Leica M Monochrom
http://www.petapixe...igital-rangefinder/ Pretty much as I suggested, and only $7,950 (and expect to pay about the same for a lens for this camera) [hippo, May 11 2012]
Please log in.
If you're not logged in,
you can see what this page
looks like, but you will
not be able to add anything.
Annotation:
|
|
Croissant. I much prefer greyscale pictures, because the eye is not distracted by the spectral information. In general I wish that digital cameras could be developed to allow a little more user intervention in many ways, otherwise they seem no better than snapshot cameras (I don't have one so please correct me). lens filters, choosing shutter speeds and apertures for restricted focal ranges - they're all the fun of SLR photography, why are there no such games for digital cameras? |
|
|
sappho: modern digital SLR cameras can do everything you say: manual exposure, focus, aperture control; exchangeable lenses, filters, etc. In addition some also offer in-camera image processing. |
|
|
Rods: I don't see how measuring the hue would give any necessary benefit, since measuring RGB gives results equivalent to any other three-valued colourspace, owing to the principle that any colour can be represented as the sum of any 3 suitably diverse components. |
|
|
oh, good. Thanks, [pottedstu], I may stop sniffing at them so disdainfully now! But having just got a load of developing/enlarging kit for my trad SLR, I'll stick with it. |
|
|
[Rods] - yup, that's kind of what the end note bit means -
each pixel will record intensity and hue. A pixel will be a
sandwich of three detectors and the 'deepest' detector
in the pixel which is reached by the incoming light is the
one which defines the colour of the incoming light. So
long wavelengths will reach the deepest detector and the
camera will therefore call that red light [sappho] -
unfortunately, no one can afford a digital SLR... |
|
|
You can get digital backs for certain 'analog' 35mm SLRs. |
|
|
Leaf makes digital backs for medium format cameras as well, most notably the Hasselblads. Often used for catalog product shots and tabletop, these backs expose each color channel individually, in sequence, and are tied, by cable, to a computer as they have no storage of their own. The results are very beautiful but use is limited to static objects due the time-exposure nature of the image acquisition. |
|
|
And, wow, are they pricey. |
|
|
But not in the least bit unusual. |
|
|
Was there an implication that they were unusual? |
|
|
Hmm. We should have known we would find Prior Art on the HB. |
|
|
BorgCo are considering the design of a "pure" black and white digital SLR. |
|
|
This is NOT a colour CCD and a "B&W" option in the software; it is a true monochrome ultra-high-resolution CCD. If a current RGB CCD element of 10 megapixel capability is readily available, then a similar device of 30 megapixel monochrome resolution should be equally attainable. |
|
|
The lens stack, coatings, etc. would be designed to mimic as exactly as possible the characteristics of Panchromatic film. |
|
|
There would be little or no facility to manipulate the image within the camera; postprocessing would be done offline. |
|
|
Suitable for serious "art" photographers, rather than the seething proletarian hordes. |
|
|
My brother once put this shortcoming down to supply and demand; it should be as easy to make a superb greyscale digital camera as a good RGB one, but those hordes will seethe... |
|
|
Grayscale digital cameras are common in industrial applications. Conversions are not, admittedly. |
|
|
Shouldn't that be //BorgCo is// ? I assumed that the
Borg were measured and enumerated as a continuum
like, say, sand or urine? |
|
|
Not so much like sand, I think. |
|
|
Avoid the issue by rewriting to use the subjunctive. |
|
|
"If BorgCo were to produce a camera at present, it
would be a pure black and white digital SLR." |
|
|
"Were BorgCo to produce...". No need for "if". |
|
|
//It should be possible to have a camera on which you can slide the grid of filters off the sensor chip..// |
|
|
When they're in place, the filters have to be precisely aligned with the pixels below them. How practical is it to make a physical sliding mechanism that accurate? |
|
|
Presumably similar accuracies are achieved when positioning the masks in chip fabs, but they're hardly mass-produced consumer kit. |
|
| |