Digital BW, The Print

Todd,

We have discussed all of your points at one time or another and I believe we
are pretty much in complete agreement. Posterization can have many sources.
The most common is probably over manipulating 8-bit files and I think the
second maybe scanning.

(By the way for those new to the subject, we use the word posterization to
mean the loss of smooth transition from one tone to the next in the print.
Areas that print out very flat or smooth areas that abruptly change tone.
Posterization actually means a reversal of tone where instead of getting
lighter or darker the tone actually goes in the opposite direction. You can
get this extreme but it is rare.)

My remarks about the RGB separation curve method and posterization come
simply from looking at the histogram of a file before you apply the
separation curve and after you apply it. From what I have seen regardless of
the original condition of the image data there is less of it after the
curves are applied. You can easily see this in the histograms or even better
since you have a Mac there is a utility in the Files section of the Hi-End
Scanner Group that gives you a count of the number of colors in an image
file. Tyler check some counts before and after curve application and the
count does go down after the curves are used.

As we have noted before, when you do something to an 8-bit file that removes
some of the 256 shades of gray it may or may not be noticeable in the print
or on the screen. So most of the time we are okay. I think that we get into
problems when the lost data falls in an image critical area or there is so
much loss any sense of continuous tone is gone. So if you are walking the
fence with an image and then apply an RGB separation curve you are more
likely to fall into posterization.

For those not familiar with the RGB separation curve and Epson driver
method, a grayscale file is converted to RGB and a set of curves are applied
to partition the ink so that the dark ink is used for the dark tones and the
light inks are used for the highlights. With the MIS-VM inks this also
controls the hue of the print. In order to fool the Epson driver into
partitioning the inks correctly the RGB curves are often very severe and
cause data loss. We would seem to be really torturing the whole system here
since we take a grayscale file and send it to the driver as modified RGB
data and the driver converts this into CMYK info to control the printer.

I can see what is happening to the image data with this workflow. I have no
idea what is happening to the data with the Piezo driver or a RIP. However,
I assume that the Piezo driver is a CMYK driver but I don't know. The RIP
drivers are CMYK drivers and theoretically it would seem likely that the
data would have to be manipulated less than in an RGB/Epson driver workflow.
This is purely assumption on my part since I have not experience yet. (I did
pick up a copy of PressReady but have not had time to play with it.)

My own system has been to use MIS-VM in a 1280 primarily with medium warm
and warm curves. I scan in 16-bit and make initial adjustments there and
then drop to 8-bit for final editing. I work in RGB using a soft proof
profile. If I get posterization I try different curves, either earlier
versions or different hues or switch between Paul's and Tyler's curves, to
either eliminate the problem or shift it to a less objectionable tonal
range. If that fails I will go back and apply the adjustments I have used on
the 8-bit file to a 16-bit version of the file. If that doesn't work then I
try a re-scan.

I think bottom line some images don't work. They don't work in silver
either. Sometimes a neg is just too flat in some areas to be useable.
Sometime a digital approach will pull it out but not always. The key may be
to have a variety of drivers and ink sets available just as you had a
variety of papers, developers and toners in the darkroom. I have a second
1280 that I have been using to try different ink sets. I would like to leave
that set up with PiezoTone and keep VM in the other (I am considering
switching to the Sepia-VM though). I need to set up a third 1280 then for on
going trials but I am out of room. Adding a RIP for the VM and PiezoTones
would further increase options.

I think the multi-printer, ink set, driver approach is the way to go rather
than trying to find the one perfect system. Different images are going to do
better on one than the other. I generally like the Epson driver for sky and
clouds, and the Piezo driver for images with lots of fine detail for
instance. I think your remarks on the different ink sets have different
transition points is valid and supports this idea.

I can't match prints from MIS-VM and Piezo. It is just like trying to match
silver prints on different brands of paper. You can get close but.... I
certainly can't get a file that prints well with the Piezo driver and then
print it out using the RGB/Epson driver without adjusting levels, etc. They
just aren't that close. Since I have managed lots of posterization in prints
using Piezo I can't recommend it as a posterization cure. The RIP's may be
another story but I suspect you can mess up there too.

Regarding the "pure" 8-bit print, I can only say that if it printed out with
posterization without any manipulation being done I would suspect that
either the scan was not good enough or the image itself was pretty flat,
perhaps too flat. Also keep in mind that a posterized image can have a
perfect histogram. If I scanned a posterized print for example, I would
expect that the histogram would look okay.

As you have pointed out before and I confirm, you can have files with
terrible looking histograms that print out just fine. I think the reverse is
also true. You can have files with really great looking histograms that may
still print out with flat or dead areas due to a lack of info in the scan or
the image.

Well that ramble should have put most of you to sleep!

Martin


----- Original Message -----
From: "Todd Flashner" <tflash@...>
To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
Sent: Tuesday, June 25, 2002 10:09 PM
Subject: Re: [Digital BW] Re: pshop 6->7 VM (converts file differently?)


> on 6/25/02 10:31 PM, Martin Wesley wrote:
>
> > This has always been a tricky area and is the weakness of the RGB
separation
> > and Epson driver workflow. The application of the curves themselves in
8-bit
> > mode can totally trash your data. How do the PS6 and PS7 histograms
compare
> > prior to applying a separation curve? Is the problem occurring when the
mode
> > change is made or when the curves are applied?
>
> Hi Martin,
>
> Well, I'd like to see this explored a bit. As I don't have the Piezo
driver
> I can't compare, but you may remember a few weeks ago I posted regarding a
> print I received from a friend which had the same problem (right down to
> dark skin - or at least light skin in shadow) but, A) the print was from a
> file that had no manipulation done in 8-bit mode, i.e., had a "pure"
> histogram, and B) was printed through the Piezo driver. Still,
posterization
> (or something close enough to it to be called that) was present.
>
> Now, please don't misinterpret my interest. I don't know that I'm trying
to
> prove anything per se - if the Piezo driver/workflow were to consistently
> prove itself better I'd just go and buy it, and I may just do that in
time,
> (but that's not what I've seen so far). Anyway, I am interested in
> understanding this. Sooo, when you say it's a weakness of the RGB
workflow,
> do you in fact mean: as opposed to the Piezo workflow; or do you mean it's
> the weakness of any partitioned workflow. Have you (anyone) tried the same
> image side by side through the two different workflows and compared?
>
> What I'd consider a side by side test would be to work the file fully in
> grayscale, then dupe it. Send the original to the Piezo driver to print on
> say EAM with that profile, convert the dupe to RGB and apply a good EAM
> curve to it and print through the Epson driver.
>
> Similarly, I take your point, which was that the posterization may be
caused
> by working a file in 8-bit mode such that it wont appear to be damaged
UNTIL
> you apply the separation curves. This too needs be side by side tested by
> also working the same file in the same way in 16-bit mode and printing
from
> that, and comparing. Consider also that the Epson driver includes
Microweave
> and/or Error Diffusion, which may (throw salt over my shoulder) help with
> broken files.
>
> Anyway, sorry to be so intense, but this topic has bedeviled me. While
> struggling with my own occasional posterization I believed all the
> conventional wisdoms about the sources: bad curves, overworked files,
> misapplied driver settings...all very properly the first things to rule
out.
> But then what? When we've ruled out the first suspects, when a person uses
a
> good scanner, does no 8-bit editing, uses the Piezo driver with the right
> profile, etc, and they still get posterization, what does it tell us?
>
> And it always seems to be in the 3/4 tones, no?
>
> I just wonder what it is in this process. I'd love for someone who has the
> piezo driver to make the comparison sometime. Take a file that posterizes
> through one workflow - any workflow - and then do the same file through
the
> alternative workflow and see if it helps. If it does I'd say stick with
the
> workflow that helps, until you get posterization with it. Then try the
> alternate workflow and see if it helps. My suspicion is that no one
workflow
> will out work the other in all cases. I'd bet it's good to have two
> workflows available precisely in case you hit the wall with one, you'll
have
> somewhere else to go.
>
> Someone mentioned it's probably the way tones in some images fall relative
> to the transition between ink densities. That sounds very plausible to me,
> and it would explain why it is common to Piezo and Epson prints, and 8-bit
> and 16-bit files, and some of our prints get it while others don't. The
> irony is (while it's certainly a blessing in other ways), all the carbon
> pigment inksets seem to be increasingly built around the Piezo densities,
> which means switching inks/workflows becomes less likely to change
anything.
>
> Todd
>
>
>
>
> Please visit the Group Homepage to check the Files, Bookmarks, Polls and
other resources as they are often being updated. The page is at:
>
> http://groups.yahoo.com/group/DigitalBlackandWhiteThePrint
>
> Please follow these basic guidelines:
> - Include your full name with your message.
> - Include the address of your website, if you have one.
> - As threads develop, trim off excess portions of earlier messages to keep
them short.
> - As the topic of a thread changes remember to change the subject header.
> - Good manners are required at all time. No personal attacks or
"flames."
> - Complete your Yahoo profile.
> - Before posting a question, search the message archives and the various
resources on the homepage.
>
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>
>