Digital BW, The Print

Koloshor,

> > > What's a color densitometer? As far as I know, density has
> nothing to do
> > > with color.
> >
> > Color is determined by "sensing" density (just like in a
> scanner or digital
> > camera does), through RGB or other color filters, so in that regard,
> > determining what color something is, is based on "reading" the density.
>
> Hi, Austin.
>
> You're just on the edge of understanding. Or maybe I'm on the
> edge of understanding what you're understanding.

Let's see...

> In any way,
> let's give it a little push, and someone will go over some sort
> of edge, somewhere.
>
> A scanner is just like a color densitometer.

...in the regard you state in this next part:

> It has three
> narrowband filters, so it can measure the reflectance of narrow
> red, green, and blue bands, and "see" the density of cyan,
> magenta, and yellow inks.

Agreed, and there is also transmissive, which is what film scanners do, and
what a densitometer for film will do.

> This lets a scanner get the maximum
> color separation when scanning cyan, magenta, yellow prints, such
> as color photographs, or publications printed on a CMYK printing press.

I'm not sure that this is pertinent.

> A camera is just like a different piece of equipment, a
> colorimeter. The colorimeter has broader filters. It's red
> response extends all teh way from deep red to green (the peak is
> in the red), the green filter extends pretty much across the
> entire spectrum (but the peak is in the green) and the red and
> green responses overlap, so a pure spectra orange (for example)
> produces a reading in both the red and green channels. Blue,
> similary, covers a large portion of the spectrum, and overlaps
> the green (and even the red, a bit).

OK...

> The colorimeter responses
> mimic the spectral responses of the three different color
> receptors in the human eye, so a colorimeter can tell you what
> "human visible" color you're measuring.

I'm not sure how true this is, as I have never used a "colorimeter", but
certainly, by design, one can change the bands of the color filters to
achieve what ever results you want...

> > To answer the original question, for B&W, using an RGB analyzer, each
> > channel will give the same value...that is what grayscale is.
>
> Unfortunatly, that's not quite how it works.

I believe it does work that way, and there is certainly enough evidence (and
my own experience) to support my contention.  Have you ever scanned a
grayscale wedge and looked at the resultant RGB values?  How about as an
experiment, you open PhotoShop, and create a step wedge of all equal RGB
values from 0-255 and tell me what you see...

> The color
> densitometer is designed for maximum color separation if the
> colors used are produced from cyan, magenta, and yellow dyes. In
> any other situation (including B&W photographic prints) it will
> typically not read equal CMY when shown a perfectly neutral gray.

I said RGB, not CMY.  RGB behaves exactly as I said it does, gives equal
values for RGB when analyzing a grayscale target.  Granted, they may be some
%1 (or less for true grayscale), but any difference is in the electronics,
not in the concept, or the color being analyzed is not truly grayscale (like
the tinted base of a film).

Also, my calibration cards for my densitometer support my contention.  The
four calibration points of the transmissive card, and the three of the
reflective card all have RGB values within +-0.02 density value (on a scale
of 0-3.6 or so, which would make it less than %1).

I'm not quite sure what it is you were trying to get at here, but please see
if you can elaborate...though, I'm not an expert in colorimeters or
densitometers (if that is what your interest is), but I do know scanner and
digital imaging system design quite well, and can talk knowledgeably about
those in a detailed sense.

Austin