Black & White Digital Cameras?

[gary_ferguson1]

I believe Kodak used to offer a dedicated black and white digital

camera. Does anyone know what was different about this camera that it

made it better for black and white than simply converting from colour?

[andy_piper2]

At the time (ah, the good old days!), it was sharper, because it didn't have to allot alternating pixel elements for R, G, and B. In the same space where a color camera could capture 1 full color pixel (from 3 RGB elements), it could capture 3 discreet B&W pixels, thus showing finer detail.

 

Squeezing out max. resolution was a big deal then (c. 1992), since the chips available were only about 1 to 1.5 megapixel equivalents at most. By going B&W-only they could deliver resolution good enough for newspaper reproduction - as sharp as a (at the time non-existent) 4Mp color camera.

 

I remember that everyone was impressed that you could pull an 11x14 B&W dye-sub print from one of those cameras, while the limit for non-pixellated color prints was still about a 5x7 image.

 

In theory, a monochrome digicamera should still be sharper than a full-color camera (of equivalent megapixel capacity), but the algorithms for rebuilding an image from RAW color data have improved, so the final resolution difference would no longer be as large. And the market demands color, so no one bothers.

[mike sisk]

Kodak did make a 760M but it wasn't heavily promoted. You

can read more here: <a href="http://www.robgalbraith.com/ubbthreads/

showflat.php?Cat=&Board=UBB2&Number=28321&Forum=UBB2&Words=760m&Matc

h=Entire%20Phrase&Searchpage=0&Limit=25&Old=3months&Main=28321&Search=t

rue#Post28321">http://www.robgalbraith.com/ubbthreads/

showflat.php?Cat=&Board=UBB2&Number=28321&Forum=UBB2&Words=760m&Matc

h=Entire%20Phrase&Searchpage=0&Limit=25&Old=3months&Main=28321&Search=t

rue#Post28321</a>

[andy_piper2]

The original (1992-93) Kodak monochrome camera was a DCS200m - the 760m seems to have been an attempt to revive monochrome that died on the vine. Ahh, well...

[nehril]

I've often wondered if you couldn't get an advantage in sharpness by shooting RAW images

and then not interpolating for color: i.e. if you knew the effect of each RGB filter on your

camera's bayer grid, could you get the original light intensity on each sensor? you could

then get b&w images at a higher effective megapixel value.

 

you'd lose the ability to apply "filters" after the fact in photoshop, but it would be an

interesting exercise. perhaps you could even preserve some ability to apply color filtration

in software by comparing an color-interpolated and adjusted image back to the raw

values.

 

If I understood Dave Coffin's excellent dcraw.c coding a bit better, I'd code it myself just to

see what would come out.

[bobatkins]

The trouble with that is that in your image you'd have individual pixles that were exposed effectively using red, green and blue filters over them. This wouldn't make for much in the way of "smooth tonality" in your image and uniform areas would be full of what you might call "filter noise" (or B&W chroma noise if that makes sense). So if you had a red object, all the red pixels would be white, but the adjacent blue pixles would be black and the adjacent two green pixels would be some shade of grey!

 

If you smoothed it all out you might just have well have used Bayer interpolation in the first place!

[gary_ferguson1]

Interesting points. So would a Foveon chip, pixel for pixel, offer black and white advantages over alternative sensors?