Digital BW, The Print

----- Original Message -----
From: "Anthony Atkielski" <atkielski@...>
To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
Sent: Thursday, August 01, 2002 3:07 PM
Subject: Re: [Digital BW] Storage of digital images


> Bob writes:
>
> > When you refer to a 160 ppi or 200 ppi printer,
> > what exactly are you referring to, and how does
> > this relate to dpi resolution of the printer?
>
> Inkjet printers, like offset presses, use essentially opaque inks in four
or
> more colors, usually cyan, yellow, magenta, and black.  Since these inks
are
> opaque, their darknesss on paper cannot be varied, nor can one color be
> printed on top of another to blend the two.

Anthony,

The inks are not opaque, even the pigment inks. If you do some swabs with
the inks the color of the underlying substrate does show through. In general
the lighter positions in the gray scale inks are dilutions of the black ink
with a clear base material resulting in semi-translucent inks. So drops of
ink can be placed on top of each other to create intermediate tones. This is
definitely true for dye based color inks. I believe that the inks used in
offset printing are quite different from those used in inkjet printers.

> So, in order to create
> variations in the darkness of the printing, the inks must be divided into
> dots of variable size.  When the dots are large and closely spaced, the
> printing appears dark (from a suitable distance).  When the dots are small
> and widely spaced, the printing appears light.
>
> Inkjet printers, like the imagesetters used to create plates for offset
> printing, do not know about variable sizes for dots; they always print
dots
> of a fixed and very small size.  When dots of variable size are required
(as
> for printing halftone images, as described above), the tiny dots are
> combined to form larger dots.  The range of densities on a print (from
> darkest to lightest) depends on the number of different sizes of halftone
> dot that can be printed, and this in turn depends on the number of machine
> dots (the actual dots printed on the paper) that can be combined to form
> each halftone dot.
>
> Printing a single pixel on paper requires a halftone dot of each color.
>
> As a result of all this, a 2880-dpi inkjet printer cannot print anywhere
> near 2880 pixels per inch.  It prints far less, and there is an inverse
> relationship between the number of intermediate tones it can print and the
> resolution (because more tones requires more variability in dot size, but
> that means that individual halftone dots must be made up of more machine
> dots, which further reduces resolution).
>
> So ... if the inkjet printer manages 2880 dpi, and you want 256 levels of
> black or green or whatever, you'll need 256 machine dots per halftone dot.
> That means roughly a square of eight machine dots on a side, and 2880/8 =
> 360, so 360 is the maximum number of pixels per inch that the printer can
> manage with 2880 dpi, if 256 tone levels are desired.  If you want more
> tones, resolution goes down; if you want less, resolution goes up.
>
> Actual implementations are much more complicated, and shortcuts exist to
> improve resolution a bit, but the above explanation illustrates why you
> don't get anywhere near 2880 pixels per inch from a 2880-dpi printer.

Your logic is partially correct for something like an image setter or
standard laser printer but the fact that the different inks can be overlaid
to create different tones or hues, and the use of dither patterns suggests
that you can simulate much more than the math implies. If what you say was
correct, I doubt that we would be very pleased with out digital prints,
color or B&W. I believe it would be more correct to say that a modern inkjet
printer interprets pixel input rather than to say it prints pixels.

Martin