Color Filter Array Web Page Grows

Thanks, John. Interesting reading.

I'm a fan of the "random array", which I have not heard discussed anywhere,
other than in Robert Cook's Distributed Resampling paper circa 1984.
--
Mike Russell - www.curvemeister.com

I have a web page on a great many subjects...

and one of its pages,

http://www.quadibloc.com/other/cfaint.htm

talks about the designs of color filter arrays used with the CCD image
detectors in cameras. Of course, nearly all of them use the well-known
Bayer pattern, but a few do use different approaches.

I mention SONY's "emerald" color - sadly, not used in their Alpha 1000
digital SLR (first that model, and now one from Pentax, are becoming
*affordable*!) - and recently, I added information about the unusual
Super CCD designs from Fujifilm. (I don't describe them in as much
detail as elsewhere on the web, however.)

Inspired by them, I added new illustrations showing how the gap
between a *hexagonal* grid, which some might think to be theoretically
superior to a square grid, and the grid of square pixels needed for
the end product might be bridged.

Other stuff that's been around on the page for some time include a
description of how a beam splitter might be used to allow a full 36 x
24 mm effective detector to be built up from cheaper small detectors,
and how a Soleil-Babinet compensator might be used to make a camera
that is a reasonably-priced *imaging spectrometer*.

The RAW files from a camera like *that* would be rather big, and using
that feature of the camera would only work for long exposures... but
it would allow *maximum* color fidelity, since the response of the
human eye could be matched exactly (and any compensation for lighting
could be done)... and the usefulness for things like astrophotography
would be even greater, of course.

John Savard

Quadibloc wrote:
I have a web page on a great many subjects...

and one of its pages,

http://www.quadibloc.com/other/cfaint.htm

Ha!
I like this one: http://www.quadibloc.com/other/images/hh5.gif

talks about the designs of color filter arrays used with the CCD image
detectors in cameras. Of course, nearly all of them use the well-known
Bayer pattern, but a few do use different approaches.

I mention SONY's "emerald" color - sadly, not used in their Alpha 1000
digital SLR (first that model, and now one from Pentax, are becoming
*affordable*!) - and recently, I added information about the unusual
Super CCD designs from Fujifilm. (I don't describe them in as much
detail as elsewhere on the web, however.)

I had not heard of Emerald. It's interesting how many subtleties of
green can be distinguished... if you ever tried matching a green from
standard color chips that becomes obvious.

Inspired by them, I added new illustrations showing how the gap
between a *hexagonal* grid, which some might think to be theoretically
superior to a square grid, and the grid of square pixels needed for
the end product might be bridged.

Other stuff that's been around on the page for some time include a
description of how a beam splitter might be used to allow a full 36 x
24 mm effective detector to be built up from cheaper small detectors,
and how a Soleil-Babinet compensator might be used to make a camera
that is a reasonably-priced *imaging spectrometer*.

The RAW files from a camera like *that* would be rather big, and using
that feature of the camera would only work for long exposures... but
it would allow *maximum* color fidelity, since the response of the
human eye could be matched exactly (and any compensation for lighting
could be done)... and the usefulness for things like astrophotography
would be even greater, of course.

John Savard

On Dec 17, 11:59 pm, Paul Furman wrote:
Ha!
I like this one:http://www.quadibloc.com/other/images/hh5.gif

That was the one I added the most recently. And I've added a bit more
to that page now, to make some facts about that array clearer - one
diagram showing how it would relate to square pixels, and that diagram
itself has been adjusted so that the difference in lightness used to
mark out the hexagons is more visible.

John Savard