Digital BW, The Print

Austin I've combined two of your posts into one.

> But I DO see it all the time...most any time I try to do moves to 8 bit
> files, and then print them using Piezo, I end up with posterization.  If I
> need to do moves, I re-scan.

I'm tempted to pull an Austin and say, "then you must be doing something
wrong, or something is wrong with your system", but without knowing the
extent of the moves you are talking about, or the types of images, that
wouldn't be fair. ;-)

Dan's whole point was that there are a lot of people who watch their
histograms rather than their images and the proof is in the image not the
representation of the image. IOW, trust your eyes over the math or the
histogram.

Whether I agree with all Dan's conclusions or not (and if you had Maris send
you the color theory thread you know where I stand. BTW, the unmentionable
gets mentioned...), I whole heartedly agree with him that the only way to
really address the issue is to do it with images, and do it scientifically.
This way we learn more about where a process might break down rather than
debating whether it breaks down. As you say, it's a simple experiment you
can try for yourself. ;-)

One must take the same image, one 16-bit variant and one 8-bit, apply
identical moves to each, then compare. Then one must show the starting image
and the final of each image and explicitly describe the exact moves applied.
Finally whatever conclusion one draws from the experiment should be based
upon looking at the image itself, not the math or the histogram.

Anybody who can scan raw can conduct such tests and if anyone would like me
to explain how I will. I think I'll refrain from anymore conversations (at
least for the moment ;-) about what one should expect the results to be.
Let's do the work first and discuss what our results are. I've done this on
the color theory list already, along with Dan, so I wont bog this list down
with it. If you want the thread from the color theory list Maris gave the
link, or I could send you a text file Maris complied which covers a good bit
of it. My quick answer is it is certainly possible to break and image, but
it takes a lot more than one might think, and some images are more prone to
problems than others. Some images can take a LOT of abuse.

Sorry if all that sounds edgy, I don't mean to be insulting or
argumentative. It's me not you ;-).

I've just been around a lot of these conversations (as you well know ;-)),
but the only way to really advance the issue is to do it with real visual
examples, so we can really suss out what's going on.

 
> Yes, you are absolutely correct.  Because you have "holes" (missing
> codes/values) in the histogram, doesn't mean you are going to get
> posterization, but I guarantee you, if you have posterization (due to image
> "problems", not ink/printer problem) you WILL see holes in the histogram.

Not taking issue with the possibility of what you say, just the absoluteness
of it. My experience is suggesting otherwise, and I think a lot of us make
the assumption that some of these failures are due to insufficient bit
depth, when in fact it's something else. I know I did.

What I'm getting is rough "blocky" tonalities in 3/4 tone skies that I get
in both 8-bit mode and 16-bit mode.

I'm disappointed, cause when I first got them in 8-bit mode I thought "aha
tonal dropout, better do it in 16-bits" but it shows there too. It's
something that's in the scan and it gets revealed and exacerbated by my
tonal moves (darkening and adding contrast) but is not caused by tonal
dropouts. And it's evident on the monitor so it's not a printer/ink problem.

I need to take it up with one of the scanner lists. Just have to figure out
how to set up a damn website and make a little demonstration....

Todd