T3D Re: T3D Re: T3D Re: T3D Re: digital camera resolution

[michaelk@xxxxxxxxxxxxxxxxxx (Michael Kersenbrock)]

> |
> |And if one reads that article in EDN that I referenced a while ago,
> |one sees that the digital camera's "pixels" often are black-magic 
> |computation based ones.  Example: One of the cost savings used in some of 
> |them is the use of VCR CCD's instead of camera CCD's.  The VCR CCD's are
> |produced in volume and are cheaper.  Difference is that 
> |the VCR ones have long aspect ratio "pixels", so they manipulate the data to
> |get three picture pixels out of one "CCD pixel" in its long-direction.  

First of all, I'm checking the article and I mis-remembered.  The aspect
ratio of the CCD is about 2:1 so they get three pixels out of *two* "CCD pixels".
Not quite as bad as I recalled.  :-)

> |Another example:  How does the color information get captured?  In the
> |best case, they have filter stripes and the number of pixels is essentially
> |one third of those on the CCD chip because it takes three to capture each
> |of three primary colors.  But some use two colors and "fake" the third in
> |software.  Some use black-magic formulae to conjure up the colors from other

Also checking, my recollection of this is a little misleading (not the
black magic software part though) . They have
several different color filter schemes, some of which include having
very assymetrical quantities of sensors for each color, so for any
given picture-pixel, some color components are somewhat faked in the
sense that the raw information may not have all three color sensors
nearby.  They can have very different resolutions depending
upon the color component (Green seeming to be the highest resolution).

Eye emulation?


> |partial acquisitions, biasing the colors in what they perceive to be that
> |which is desirable by the users (well, I guess the film makers do this too).
> 
> The Oly unit is stated as a "progressive scan". I can't really envision
> how this works - can anyone enlighten?

"Progressive scan" is mentioned but not in a way that's useful -- only
that it's ideal for still photos of subjects that may move and that
video applications of the CCD sensors have the movement aspect to hide
"low image resolution, inaccurate color balance, limited
dynamic range and other shortcomings".  So in video, the data would 
be output in an interlaced fashion rather than progressively.  This
doesn't seem to make sense in terms of performance unless the array
isn't capturing all of the image pixels simultaneously, either by
having a mechanical shutter or by having adjacent "storage cells"
next to each sensor cell which "remove the need for a costly 
mechanical shutter".

Mike K.



P.S. - It mentions that the Olympus' D-500L and D-600L digital cameras
       share the same SLR architecture as the company's 35mm and
       APS cameras.


------------------------------