Ad
Ad
Ad
Pages: « 1 [2] 3 4 5 »   Bottom of Page
Print
Author Topic: ABW or Photoshop  (Read 9610 times)
Eric Myrvaagnes
Sr. Member
****
Offline Offline

Posts: 7789



WWW
« Reply #20 on: May 28, 2013, 09:04:18 AM »
ReplyReply

IF you use another product to convert the color to B&W, viewing your image as you do so and producing a look you desire, sending it through ABW will re-convert it, you'll lose what you did. So you'll need to print that data without ABW using a good ICC profile.
Hi Andrew,

Can you explain? If you have already converted a color image to grayscale, how can ABW do anything significant to change the image (other than adding overall color via the toning sliders)? As far as I can see, the "re-conversion" won't do a thing.

Eric M.
Logged

-Eric Myrvaagnes

http://myrvaagnes.com  Visit my website. New images each season.
Sal Baker
Full Member
***
Offline Offline

Posts: 140


« Reply #21 on: May 28, 2013, 10:16:24 AM »
ReplyReply

Hi Andrew,

Can you explain? If you have already converted a color image to grayscale, how can ABW do anything significant to change the image (other than adding overall color via the toning sliders)? As far as I can see, the "re-conversion" won't do a thing.

Eric M.

ABW apparently messes with color space, how the driver lays down the ink, and gama, but it won't change your image editing at all.  I convert to BW in Silver FX Pro and do substantial editing.  I then save to Photoshop and make a few more image edits.  I print through PS using ABW and the result is a near perfect match to the edited image on my monitor (I don't use the color tinting options in ABW.)

The only adjustment I make before printing is applying a subtle curve adjustment layer to slightly boost the near-black range which eliminates the shadow blocking I would get without it.  But ABW has no way of knowing how to delete the edits you've made to the final image.

Sal
Logged
cybis
Full Member
***
Offline Offline

Posts: 163



WWW
« Reply #22 on: May 28, 2013, 11:26:26 AM »
ReplyReply

What sold me to ABW is printing circular test gradients like the one from Graham Preston found at Northern-Images. These patterns will show up unpleasant color tone irregularities like nothing else.

At least on the 7900, as far as color tone linearization goes, ABW beats the color pipeline by a long shot.
QTR has the potential for perfect tone and color linearization, but I find the dithering pattern inferior to Epsonís (QTR doesnít use variable drop size yet on the x900).

Now, ABW is a black box with no way to adjust color tone linearization. It was optimized for Epson papers which all contain OBAs. While ABW still beats the color pipeline on non-OBA papers, I wish there was a way to tweak the color tone in the highlights.

Logged

Schewe
Sr. Member
****
Offline Offline

Posts: 5411


WWW
« Reply #23 on: May 28, 2013, 11:28:34 AM »
ReplyReply

QTR has the potential for perfect tone and color linearization, but I find the dithering pattern inferior to Epsonís (QTR doesnít use variable drop size yet on the x900).

But if you set the output rez to 2880, I'm pretty sure it always only uses the smallest droplet size (at least that
s true of the Epson driver).
Logged
cybis
Full Member
***
Offline Offline

Posts: 163



WWW
« Reply #24 on: May 28, 2013, 11:47:42 AM »
ReplyReply

But if you set the output rez to 2880, I'm pretty sure it always only uses the smallest droplet size (at least that
s true of the Epson driver).

QTR uses the Gutenprint driver which I don't believe uses small drop size in its QTR implementation for the x900. The x900 were only recently added to Gutenprint and QTR, as opposed to other Epson printers, so it could be unique to the x900s. Not sure.
Logged

Ernst Dinkla
Sr. Member
****
Offline Offline

Posts: 2801


« Reply #25 on: May 28, 2013, 01:29:28 PM »
ReplyReply

QTR uses the Gutenprint driver which I don't believe uses small drop size in its QTR implementation for the x900. The x900 were only recently added to Gutenprint and QTR, as opposed to other Epson printers, so it could be unique to the x900s. Not sure.

Seems odd to me, the minimum droplet size hasn't changed on the Epson wide format models since the 9800; 3.5 picoliter where the 9600 had 4 picoliter. The next droplet size on the 990 will be something like 10 picoliter, a size that should be quite visible. It is more likely that the 9900 dithering/weaving did improve beyond the Gutenprint driver.

--
Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
December 2012, 500+ inkjet media white spectral plots.

Logged
Ernst Dinkla
Sr. Member
****
Offline Offline

Posts: 2801


« Reply #26 on: May 28, 2013, 01:36:01 PM »
ReplyReply

On the HP Z models the B&W mode should be used with driver color management. HP recommendation. That CM expects either sRGB or AdoberGB as the assigned profile in the image file and is based on internal LUTs for the conversion. Like ABW it will accept Greyscale and RGB images. If the Greyscale image was kept in Gamma 2.2 during editing etc then I do not see what difference there could be between a "neutral" RGB image with AdobeRGB assigned or a Greyscale image with Gamma 2.2 assigned. Most likely the last is converted again to neutral RGB and it is assumed to be either AdobeRGB or sRGB assigned if the driver CM is not that sophisticated. The slight difference in the gamma curve of sRGB probably would not make a huge difference either. I guess the Epson ABW acts similar.

That still makes it possible to create a B&W QTR "ICC" profile for that ABW workflow, even when Greyscale images are used, at least in Windows. Third party media and individual printer characteristics can be compensated that way. The Dmax can be higher in a driver's B&W mode than in color mode and less color is mixed in on the B&W driver mode. Along the tone range the Lab a and b is not CM controlled so much depends on the choice of paper white and how warm the grey inks are. With a full CM controlled color workflow that aspect should be better but up in the shadows/black a warm carbon black ink on a cool paper still delivers a split with whatever CM profiling. If not you can bet the lower Dmax limits the image quality.

To say that an ABW workflow is meant to be used with RGB images and not Greyscale is something I can not agree with. I would not advocate either that ABW should not be used with "profiles".

--
Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
December 2012, 500+ inkjet media white spectral plots.
Logged
cybis
Full Member
***
Offline Offline

Posts: 163



WWW
« Reply #27 on: May 28, 2013, 01:49:44 PM »
ReplyReply

The next droplet size on the 990 will be something like 10 picoliter, a size that should be quite visible. It is more likely that the 9900 dithering/weaving did improve beyond the Gutenprint driver.

Looking through a microscope at a x900 QTR/Gutenprint print, I only saw one size drop. And those drops were visible with the naked eye in the highlights.
Logged

Ernst Dinkla
Sr. Member
****
Offline Offline

Posts: 2801


« Reply #28 on: May 28, 2013, 01:59:24 PM »
ReplyReply

Andrew,

Several questions arise from this:
(1) I'd be curious to know whether any objective testing has been done to confirm the view that the ABW process is more archival, or is this a logical deduction based on views of the longevity of the individual inks involved in both processes?

As far as I know all the Epson pigment ink B&W tests Wilhelm Research did, started with the introduction of ABW (2400?) in the driver and used that driver and they all had a better fade resistance. I can not recall another driver being mentioned. The black generation in the different generations of ABW may have changed over time but it is still a B&W mode that has some color components in the mix. In some of the B&W tests the inks were not the limit in years but the paper white shift caused it. For example the 7900 B&W test with EEM that was set at 110 years because the EEM paper white in the archival test (another testing method) could not get beyond that number. A more or less arbitrary introduction of paper white shifting in the results. Aardenburg Imaging at least tests both paper white shift and the color shifting in the same test method.

--
Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
December 2012, 500+ inkjet media white spectral plots.
Logged
Ernst Dinkla
Sr. Member
****
Offline Offline

Posts: 2801


« Reply #29 on: May 28, 2013, 02:07:29 PM »
ReplyReply

Looking through a microscope at a x900 QTR/Gutenprint print, I only saw one size drop. And those drops were visible with the naked eye in the highlights.

That does not tell me whether it does not use the minimum droplet size. You could have seen the minimum droplet size. At the highest resolutions in the Epson driver the larger droplet sizes are not used. It is a 3.5 droplet size, with enough light and young eyes you can see dots like that. Ask Roy what the x900 lays down with QTR.

--
Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
December 2012, 500+ inkjet media white spectral plots.
Logged
cybis
Full Member
***
Offline Offline

Posts: 163



WWW
« Reply #30 on: May 28, 2013, 02:31:12 PM »
ReplyReply

That does not tell me whether it does not use the minimum droplet size. You could have seen the minimum droplet size. At the highest resolutions in the Epson driver the larger droplet sizes are not used. It is a 3.5 droplet size, with enough light and young eyes you can see dots like that. Ask Roy what the x900 lays down with QTR.

Those drops were much larger than the smallest droplet size delivered by the Epson driver. I just emailed Roy. And I'll do some more testing this w-e.
Logged

datro
Jr. Member
**
Offline Offline

Posts: 53


« Reply #31 on: May 28, 2013, 10:23:09 PM »
ReplyReply

To say that an ABW workflow is meant to be used with RGB images and not Greyscale is something I can not agree with. I would not advocate either that ABW should not be used with "profiles".

I've been following with interest all of these threads lately regarding ABW printing using QTR-built profiles for linearization (which as Alan points out, can only be done reliably on the Windows platform these days).  I've been printing with ABW now for quite some time on my Epson 7900 (without linearization profiles) and have the process fairly-well tuned in for my workflow which starts with drum-scanned 4x5 B&W negatives.  I agree with Ernst that ABW is not meant just for RGB files.  In fact, as Eric, Luc and others have implied, once you zero in on whether to use "Neutral", "Dark" or "Darker", and assuming you start with files which are Gamma 2.2 (very important for the ABW workflow), you will find that ABW works extremely well and predictably when starting with grayscale images.

Regarding using linearization profiles:  I've just recently started testing this to find out if there is truly an advantage over the normal ABW workflow.  I've carefully built my profiles using the QTR 51-step target, the QTR-Create-ICC-RGB utility (creates a standard ICC file with only Luminance information in it), and then printed my own custom B&W test target using the profile and ABW.  The results?  Well, I am quite surprised to see that shadow detail is actually slightly worse when using the profile.  So far I cannot explain this, even after multiple attempts, re-checking my steps, and using different papers.  Others using this procedure seem to find that shadow detail is improved with the profile; for me it is the other way around.  One thing I have discovered:  my 7900 in ABW mode is already EXTREMELY linear across the luminosity values in the 51-step target.

I plan to do some further investigation, including using M2 mode (UV-cut) instead of M0 on my i1Pro 2, but I have to say that so far I haven't found (in my case) that using profiles with ABW provides any benefit other than the opportunity to softproof.  For some this could be important; for me it is less so because my workflow (using only a few papers) is already predictable enough.

Dave
Logged
royvharrington
Newbie
*
Offline Offline

Posts: 7


« Reply #32 on: May 28, 2013, 10:59:27 PM »
ReplyReply

As far as I know and can figure out, 3.5pl drop size is what all the drivers are using.
That's the minimum stated in the Epson docs and its been the same for several generations
of Pro printers.  At 2880x1440dpi only the one smallest dropsize is used. At 1440x720dpi
three dropsizes are used -- the larger ones seem to have varied a little from model to model.
Up through the x880 printers both QTR and Epson drivers used all the same API to the printer
so it was easy to compare the exact commands used.  With the x900 and x890 printers Epson
has added a new API which is used by the Epson driver but not anyone else.  So it might
seem possible that this works differently but they still quote 3.5pl as the minimum size.
Another thing to note is that although QTR and Gutenprint have a common root from way
back 10 years ago, all the tuning and handling of newer printers has been completely independent.

The other issue I want to mention is Mac color management with B&W printing -- ABW or QTR.
This used to work pretty well back in the old days -- CS3 and OSX 10.4 or 10.5 I think.
Eric Chan did ICC profiles for ABW on 3800 and 3880 I believe; QTR has had the capability
of creating grayscale ICCs for a very long time.  Note that these do not handle "colors" --
just match the tonal gradation of B&W printing.  But there's been a steady degradation of this
over all the versions since.  With current versions (CS6 and Mt Lion) you can't even select
ABW with an ICC -- the Epson driver is also in the act of preventing this.  With QTR you can
select profiles but you always get double profiling so its not very useful.

I got frustrated enough with this that I wrote a completely new print layout app that handles
all the color management the way is used to work -- B&W and color.  Sorry about the commercial
plug but the program is called QTR-Print-Tool at www.quadtonerip.com -- free 30 day trial.

Roy
Logged
Schewe
Sr. Member
****
Offline Offline

Posts: 5411


WWW
« Reply #33 on: May 29, 2013, 12:13:44 AM »
ReplyReply

Roy,

Welcome to LuLa...thanks for responding. I tested out your Quad Tone Rip recently for my upcoming The Digital Print book and really liked the capabilities. Somewhat frustrating that only USB seems to be supported (at least via Mac) but I got really nice results using 2 or 3 curve spit toning. I haven't gone down the rabbit hole of making curves or profiles but prolly will (but not for the book because of deadlines).

I still really like a full split-tone capability of using an ICC approach (for the right images) but prefer the QTR for more subtle toning not doable in the Epson nor Canon monochromatic options.

So, given the QTR default curves, do you have any ink usage data regarding which inks are used for the various curves?
Logged
Ernst Dinkla
Sr. Member
****
Offline Offline

Posts: 2801


« Reply #34 on: May 29, 2013, 04:20:26 AM »
ReplyReply


Regarding using linearization profiles:  I've just recently started testing this to find out if there is truly an advantage over the normal ABW workflow.  I've carefully built my profiles using the QTR 51-step target, the QTR-Create-ICC-RGB utility (creates a standard ICC file with only Luminance information in it), and then printed my own custom B&W test target using the profile and ABW.  The results?  Well, I am quite surprised to see that shadow detail is actually slightly worse when using the profile.  So far I cannot explain this, even after multiple attempts, re-checking my steps, and using different papers.  Others using this procedure seem to find that shadow detail is improved with the profile; for me it is the other way around.  One thing I have discovered:  my 7900 in ABW mode is already EXTREMELY linear across the luminosity values in the 51-step target.

I plan to do some further investigation, including using M2 mode (UV-cut) instead of M0 on my i1Pro 2, but I have to say that so far I haven't found (in my case) that using profiles with ABW provides any benefit other than the opportunity to softproof.  For some this could be important; for me it is less so because my workflow (using only a few papers) is already predictable enough.

Dave

What you experienced is quite possible. There have been discussions on the dynamic ranges possible on different paper qualities and whether the L values are enough to get the best picture on every paper. Part of the problem could be that ABW is still a black box compared to QTR, the partitioning of the different inks over the tone range can not be controlled as much as it can be done in QTR. One can not edit the Epson media presets in ABW to that degree.

--
Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
December 2012, 500+ inkjet media white spectral plots.

Logged
Alan Goldhammer
Sr. Member
****
Online Online

Posts: 1604


WWW
« Reply #35 on: May 29, 2013, 07:03:14 AM »
ReplyReply

I take a day off from LuLa and see that there are a lot of posts on this topic now!  Thanks to Roy Harrington for chiming in on QTR.  Some points of observation.  Andrew Rodney should have clarified that the ABW printer uses less color ink than the normal driver.  With respect to 'permanency' of the prints made from the ABW driver, there is test data for some papers up on the Aardenburg website.  I have submitted a number of tests (including a color patch set for comparison) for Hahnemuhle Photo Rag Ultra Smooth so those might give some indication about the relative stability of ABW versus the normal printing pathway.  The Epson driver settings of normal, dark, darker, etc. are just gamma adjustments to the print and this has been shown HERE for the Epson 3800 (you need to go down and click on the Appendix link to get the data; and this reference has a lot of other good stuff in it); another thorough discussion is HERE.

Regarding the use of ABW profiles and whether they are more useful than a normal ICC printing path, this may be one of user preference.  I would certainly advocate testing both on an actual print (the Northlight Images testprint is particularly useful as Mark Segal has already noted) and through measurements of a 21 step B/W patch set to insure that one's ABW profile is properly constructed.  In using the ABW driver it's important to note that ABW profiles are strictly dependent on the 'tone' setting in the Epson driver.  Of course this is all academic for MacOS users (I haven't looked at Roy's new approach that he outlined in his post as I'm on WinOS and QTR profiles still work for now).
Logged

Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6818


WWW
« Reply #36 on: May 29, 2013, 08:10:15 AM »
ReplyReply

Alan, reverting to Reply 17, I see that QTR IS available in a Mac version. I doubt I'll have time to work with it over the near future, but I shall be complementing the material in reply 12 with additional measurements of L*a*b* deltas compared between ABW-Normal and Photoshop Manages Color, using one of Bill Atkinson's grayscale test ramps. I'll be reporting those results in this thread I hope some time today. This work of course is limited to my Epson 4900 and driver version 8.64 on Mac OSX.

Regarding the ABW tone settings, "just gamma adjustments to the print" are of course kind of important when you can actually see differences that impact on the character of the outcomes.

BTW - very dangerous indeed to take a day off from "LuLa".  :-)
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
NeroMetalliko
Jr. Member
**
Offline Offline

Posts: 64



« Reply #37 on: May 29, 2013, 09:04:06 AM »
ReplyReply

Hello,
regarding the question of ABW vs ICC print in B&W I can share my humble opinion.

First of all, I think that maybe not all the printer model have the same behavior, and obviously there are differences depending on the paper used.

With a given paper, using my little R3000 (which is not a pro printer), adopting Adobe RGB (gray gamma 2.2) as working colorspace, there are some differences, as per L* linearity parameter, between printing a B&W image in ABW (neutral dark) and printing the same image with an ICC profile (provided by paper manufacturer for example).
That said, the first advantage of ABW is that, if for example in neutral-dark I get a similar non linear behavior as the ICC one (typical dark print shape) I can try a neutral-normal setting to counterbalance it, and this is not possible in native ICC without editing the image.

This gives you an option the have a better starting point for the print, but in my experience is far from perfect too

As already mentioned, other ABW advantages are the slight deeper black I have usually got (but not so much in some cases) and the probably higher archival grade due to some minimization of colored inks during the print. Again, here results may vary a lot more with a different paper choice.

What ABW it does not allow easily is the use of a linearization ICC.
A way to workaround this is the already mentioned use of QTR rgb ICC creation and a double ICC action (assign/convert) in RGB before sending the image to ABW driver.

Finally, what ABW cannot in any case allow is a gray tone neutralization (if needed for some papers, because normally in my test the neutral gray of ABW is better that the one of the ICC print on the same paper) and/or the desire to have a split-toned image (sepia shadows or whatever) without messing with the L* linearity.

Another issue could be the over-inking, but I will leave this argument out, for the moment.

In my opinion, the use of an ICC print in B&W makes sense if you need to overcome the gray neutralization and/or the spit-tone lack in ABW.
So, usually, my first choice for B&W print is ABW with a paper I have L* linearized and for which the gray tone is good enough.

If I need to print on a less neutral paper and/or want a custom split/tone I could consider to switch to an ICC B&W print, obviously using a L* linearization which includes the desired degree of neutral gray compensation (or split/tone).

There is a post regarding my approach to the L* linearization with/without tone compensation using DeviceLink profiles here:
http://www.luminous-landscape.com/forum/index.php?topic=78142.0
with some details regarding the current state of my experimentation/development and including some examples of real world results.
Unlucky at the moment is not something really usable or that I can share/provide: it is a bunch of scripts and batch files used under different applications, so it should be considered a experimental work in progress with unknown timing...

Nonetheless, I hope it will be useful.
Any opinion in regard is always welcome.

Ciao Smiley

Andrea
Logged
Paul2660
Sr. Member
****
Offline Offline

Posts: 1633


WWW
« Reply #38 on: May 29, 2013, 12:41:27 PM »
ReplyReply

Very good discussion.

When I was working with the 9880, I spent some time getting to know QTR and was able to get some good profiles (or curves as QTR calls them) for B&W printing.  However after viewing the LuLa video series that covered the ABW I wanted to try it.  I found that I was able to get very good results from it also on the 9880.  When my 9880 finally gave up the ghost and I purchased the 9900, i stayed with the ABW (at that time the 9900 was not supported as it is now).  Being able to soft proof in CS was a nice feature of QTR.  (I am a windows user).  Output from the ABW driver is excellent with deep rich blacks and very little if and color shifting when viewed in indoor lighting.   

My current workflow for a B&W is to either convert in LR4 following steps from the Lula video series or Martin Evening's LR4 book or I will use Silver efex pro.  Both are going to leave the image in RGB not greyscale.  I then will use the Epson ABW driver and make the print, leaving the image in RGB.  It's my understanding that the ABW driver needs to have the image sent in RGB not greyscale.  I tend to print from the Neutral settings and just add a bit of contrast.  The ABW driver gives very little feedback and I had hoped that with the 9900 Epson might improve on this as far viewing the settings changes. 

I am going to revisit QTR, and if Roy is still checking this post, it would be great to see a new user's guide written.  I realize this is time consuming and I would gladly being willing to pay for it.  The problem is understanding the process of creating the curves for QTR, if you want to print from QTR.  I never had any issue converting the curve to a profile for CS, I am assuming that process is the same as a year ago?


Paul Caldwell

Logged

Paul Caldwell
Little Rock, Arkansas U.S.
Photography > http://photosofarkansas.com
Blog> http://paulcaldwellphotography.com
Schewe
Sr. Member
****
Offline Offline

Posts: 5411


WWW
« Reply #39 on: May 29, 2013, 12:59:46 PM »
ReplyReply

It's my understanding that the ABW driver needs to have the image sent in RGB not greyscale.

Nope...if printing from Photoshop (vs Lightroom) you can print RGB or grayscale...it doesn't matter. Lightroom doesn't really have a "grayscale" mode since everything is processed in it's internal colorspace.
Logged
Pages: « 1 [2] 3 4 5 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad