Digital BW, The Print

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "tynmansystems" <tynmansystems@y...> wrote:
> 
> I agree. I have never been overly convinced that a monochrome 6 
> megapixil sensor would be noticably sharper than an RGB-bayer-pattern 
> 6 megapixil sensor. The interpolation that occurs for RGB sensors is 
> performed to maintain color accuracy, and shouldn't have a 
> significant effect on sharpness. I.e., the effect might be 
> measurable, but I'll bet not visible.

You'd lose that bet, for several reasons...

A monochrome camera can pretty much do without an anti-aliasing filter. Aliasing is nearly invisible in most monochrome captures. If you do use an AA filter, it can be at the horizontal or vertical pixel pitch, or even a little less.

In a color camera, the AA filter has to be at a minimum of the diagonal pixel pitch of the green pixels (40% greater than the horizontal or vertical) and can be up to twice the horizontal or vertical pitch (matching the red or blue vertical pitch to prevent color aliasing). If the camera is severely filtered (2x the H/V pitch) the image is going to be very soft compared to a 1x H/V (or less) monochrome camera.

If the camera is mildly filtered (1.4x the H/V pitch or less) unless the scene is very desaturated, so that all details read strongly in all channels, you get a great deal of color aliasing (moire). Once converted to monochrome, this manifests as strong luminance bands.

I've done quite a bit of experimentation with monochrome conversion algorithms (such as walking color regions), with IR filters (which cause a monochrome response in most CCDs), with actual monochrome sensors, and with attempts to bleech out the color filters on a color CCD. Monochrome sensors have a major sharpness advantage over patterned color sensors.