While I always send 16bit data to the printer, I have also become more of a
skeptic in my old age. My take on it is if you can't see the difference in
real world viewing, then the difference is not relevant.
While 16bit data may improve the rendition of gradients and subtle colour
changes, in my experience only a small subset of images benefit from the
extra depth at the printer end.
Cheers,
Paul Langmead
On Fri, 12 Oct 2018 at 6:25 PM, Ben Goren <ben@xxxxxxxxxxxxxxxx> wrote:
On Oct 12, 2018, at 8:43 AM, Yves Gauvreau <gauvreau-yves@xxxxxxxxxxx>Paul Langmead
wrote:
Sounds like a good idea but I'm not sure I understand what to look for in
those 3 printed images.
I’m not sure, either.
The question may be impossible to answer.
You can always trade spatial resolution for bit depth (and vice-versa).
Traditional halftone printing is, literally, one bit: you either put ink on
the page or you don’t. And, until modern times, newspapers and most other
printing presses only had one ink to put on the page. Nevertheless, they
still printed photographs by dithering the image using halftone screens.
The finer the screen, the less “grainy” the photograph.
(And the “vice-versa”? Anti-aliasing, standard on all modern displays and
used by many printer drivers, uses grays or other in-between colors to fill
in gaps and smooth out text and other figures, making the result appear to
have much more linear resolution than a simple one-bit equivalent would; a
modern “retina” display might be in the region of 300 PPI yet far surpasses
the crispness of a 1200 PPI Linotronic — never mind a 300 PPI laser
printer. And that’s just for black-and-white text!)
...but even that’s just a function of viewing distance. A massive
billboard that’s viewed from across the block can have amazing detail (and
smoothness, etc.) yet have an halftone screen with dots the size of your
fingerprint. The pointillists figured that out a century ago; look at a
wall-sized Seurat from across the room and it’s a photorealistic (if
stylized) rendering, yet the pea-sized dots are strikingly obvious when
viewed at arm’s length. And, before them, the Impressionists...a Monet
viewed from across the room is, honestly, a technically superior rendering
than almost any photograph you’ll encounter — even if it’s all an
“impressionistic” blur at arm’s length. That man had a comprehension (in
whatever form) of color and visual perception that still challenges the
best color science we have today.
Almost all modern color printers are inkjets, and they’re still laying
down dots of single colors of ink. They’re able to do what they do because
those dots are micron-sized and laid down with comparable precision. But
it’s still binary (per ink): either there’s a bit of ink on the paper, or
there isn’t. (Slightly complicating matters is that printers are using
variable-sized dots, and that there are still some dye sublimation printers
made, and other caveats — but those are mostly irrelevant.)
A computer file is, conceptually, the same as if you had a palette of ink
with as many different colors as there’re discreet values in the bit depth.
For eight-bit three-channel color, that’s (2**8)**3 = ~16.7 million
different colors. Sixteen bit is (2**16)**3 = ~281 trillion different
colors. Obviously, no printer has hundreds of trillions of inks, so it’s up
to the printer driver to figure out how best to lay down what inks it has
to best approximate the data being sent to it. (And, again, I’m glossing
lots of other details near and dear to the hearts of color scientists —
such as, what, exactly, _are_ those hundreds of trillions of colors, and
how many of them are discernible by the human eye?)
All that’s to say...well, for example, print your gradients from 8- and
16-bit files and you’ll almost certainly be able to tell the difference
with a microscope. Print them on letter-sized paper and tape them to the
far wall and nobody will have a clue which is which.
So the only meaningful answer is for you to make realistic prints of the
same image both ways, look at them the way they’re actually going to be
looked at, and make your decision that way.
Because, if you yourself can’t see a difference, then, as far as you’re
concerned, there’s no difference....
Cheers,
b&
--
