RE: filmscanners: Re: Any real 4000x6000 camera back???

["Austin Franklin" <darkroom@ix.netcom.com>]

> This discussion of whether the Kodak chip is true 16Mp or only 4Mp
> seems to be based on either/or assumptions. The reality is that the
> luminance (B&W) information is true 16Mp, ie, distinct and
> non-interpolated allowing each sensor element to contribute discrete
> information, while the color information is 4Mp, with interpolation
> distributing the values to the individual 16M pixels. Our eyes pick
> up detail based much more on luminance than chromanance, so the
> detail perceived is based on the 16Mp, and the interpolated 4Mp
> chromanance is no problem. This amount of information will enable
> production of a very high quality full bleed magazine page,
> considering its printing constraints.

Edge detail is, of course, much more important (visually to us that is) than
color information...but that's not really the point.  No one said these
backs didn't produce good images, if not stunning...but all that doesn't
make their claim any less of a lie.  To claim the sensor is 16M color pixels
is just not true, and to me, that IS a problem.

The luminance is not necessarily "accurate" to 16M...simply because it
depends on what the color components are.  There isn't, technically, a
luminance and chrominance value from the sensor...it's RGB, and you get
four/three discrete color values...and the value of each is dependant on
what the color components of that sampled area are.  Luminance and
chrominance are really a completely different color encoding methodology.