Advanced info about b&w printing with Epson 9880 and 9890

[digitaldog]

Try to keep up Andrew. I am using ABW for the reasons you cited. This uses a grayscale profile rather than an RGB one. And I can soft-proof. But I do so under CS3 and 10.5 because later versions produce incorrect output. Clear now?

Nope its not, in that ABW doesn’t use an ICC profile nor was it intended to. If you use ABW as it was intended, it works fine under 10.6 in CS5.

[Scott Martin]

There are some advantages and disadvantages to using either a true ICC workflow or ABW.

I think this should read "either a color or ABW workflow..." since we can make ICC profiles for both workflows.

...you don’t use a true ICC workflow with ABW, its a black box.

You've used QTR-Create-ICC to make ICC profiles for ABW printing right? It's fantastic. If you print with the profile you've get even better tonality without the subtle contrast and density adjustments you get without one. Plus you can soft proof complete with the toning of the process. It's really fantastic.

[gromit]

And a much better option, and has always been so, is to use a well linearized curve out of QTR. The sooner Epson opens up the 9900 to QTR the better off they will be.

The only real world advantage that QTR has over ABW is split-toning.

[digitaldog]

No, I do NOT build ICC profiles using QTR for ABW. I either tone my work as I wish and use my own ICC color profiles OR I use ABW as it was designed. I’m happy with both processes depending on the needs at hand. When I use ABW is when I want a very neutral (non color toned) output and I want the driver to do the heavy lifting (convert the color data on the fly). IF I use my own conversions to B&W, I use my own profiles. Again, when ABW is used AS DESIGNED there’s nothing broken in 10.6/CS5. You add some outside process to the works, anything is possible.

[Scott Martin]

I agree - ABW works fine under CS4 and CS5 under 10.6 as it's intended. Add QTR profiles on top of it if you like to geek out.

[Alan Goldhammer]

I agree - ABW works fine under CS4 and CS5 under 10.6 as it's intended. Add QTR profiles on top of it if you like to geek out.

Yes, and the QTR profiles are relatively modest corrections (at least in my testing on my 3880).  The one advantage is you can soft proof using them which is useful.  Remember, ABW gives you about a 5-10% greater Dmax depending on the paper and this will obviously expand the gray scale which is useful for shadow detail.

[gromit]

Nope its not, in that ABW doesn’t use an ICC profile nor was it intended to.

Though it was never documented, if you sent ABW an image with an L* (not 2.2 as commonly reported) gamma and used No Color Management you got *perfectly* linear results on Epson's papers with default driver settings. Guess what, L* is used in the PCS so adding a profile to remap to L* is a perfectly viable approach. These profiles can be used to linearize third-party papers (or changes to the default driver settings). And because they incorporate colour readings of the toning at each step, they can also be used for soft-proofing. This isn't rocket science.

[digitaldog]

This isn't rocket science.

Neither is posting a comment on a forum that is NOT unclear and implies something that several people have specifically corrected. Look at your first post here, you wrote something that isn’t correct:

I can't speak for Windows, but ABW is broken in Photoshop CS5 and/or Mac OS X 10.6 (and presumably Lion). I've kept a copy of Photoshop CS3 and 10.5 solely for this reason and do all my B&W printing through this.

If you want to take the topic OT and talk about RIPs, and L*, start a post and clearly explain your POV.

Two of us have said (and a third person agreed), in terms of your post that there is nothing broken in ABW in 10.6 in CS5. Period. Clear? This isn’t rocket science either. If you want to talk about rocket science, or color science, do so without taking a current topic way OT, please.

[gromit]

Two of us have said (and a third person agreed), in terms of your post that there is nothing broken in ABW in 10.6 in CS5. Period. Clear?

Attached are the results I just got on my 9900. The first with Photoshop CS3, the second with Photoshop CS5. Both on 10.5.8. Exact same profile and media settings. Note the difference at the L*=50 mark (51.5 and 48.3). This difference is visible and repeatable. I suggest anybody using a comparable workflow would be well advised in doing the same measurements.

[Alan Goldhammer]

Attached are the results I just got on my 9900. The first with Photoshop CS3, the second with Photoshop CS5. Both on 10.5.8. Exact same profile and media settings. Note the difference at the L*=50 mark (51.5 and 48.3). This difference is visible and repeatable. I suggest anybody using a comparable workflow would be well advised in doing the same measurements.

I just plopped the data into Excel to see what the plot looks like (I don't particularly care for the QTR plots) and no question about the change and it is most prominent between steps 16 & 40.  It looks as though the linearity is completely lost (based on how my before and after curves look) and that for some reason CS5 may not be recognizing the profile at all.  If you have LR, do you see comparable results?  This might help isolate the problem.  I print mainly from LR but when I recently prepared some QTR profiles, I did double check them against a print of an unprofiled 21 step B/W wedge using CS5 on a Win7 machine and the profiles looked fine (e.g., they did provide the desired linearization).  You might want to post these results on the QTR tech group blog to see if any other Mac/CS5 users have seen the same thing.

One other question (though this might open up another can of worms), if this happens for QTR profiles, can one be confident of color profiles made prior to these software updates?  Maybe it's just an artifact of QTR (one would really hope this would be the case otherwise there would be a lot of angry people out there).

[digitaldog]

Just sent a 918 patch target though CS5 and CS4 identically (don’t have CS3 installed, I’ll do it if you insist). Used ABW, measured on the iSis, Avg dE is 0.68! Let the target dry down, that will go a lot lower (these prints are fresh out of the 3880).

So No, again, terms of your post, there is nothing broken in ABW in 10.6 in CS5. But there could be something broken in your beloved QTR profiles or how you built em, used em etc.

Use ABW correctly as designed, and works as it should. In CS4, and in CS5. No reason to believe CS3 is any different and the original post clearly put the blame on CS5.

[gromit]

One other question (though this might open up another can of worms), if this happens for QTR profiles, can one be confident of color profiles made prior to these software updates?  Maybe it's just an artifact of QTR (one would really hope this would be the case otherwise there would be a lot of angry people out there).

For completeness I've attached the results for 10.6 (same paper, profile, settings etc). This is yet darker still (45.9). I don't think this is the fault of QTR's profiles as I've seen the same general movements with grayscale profiles created with basICColor print. Frankly, I've wasted too much time with all this and it's easier to just go back to CS3 and 10.5 to print B&W. In my testing, colour output is fine and now use CS5 (running on 10.5) for the majority of my work.

If you're using Eric Chan's profiles for the 3800/3880 note that these don't go nearly far enough. You really need to tune the placement of the a/b curves with ABW's Horizontal & Vertical controls (this will generally require a re-linearization). The Color Density driver setting can also be useful for some papers. If you're serious about B&W it would be worth your while investing in an i1Pro and doing all this yourself.

[Alan Goldhammer]

For completeness I've attached the results for 10.6 (same paper, profile, settings etc). This is yet darker still (45.9). I don't think this is the fault of QTR's profiles as I've seen the same general movements with grayscale profiles created with basICColor print. Frankly, I've wasted too much time with all this and it's easier to just go back to CS3 and 10.5 to print B&W. In my testing, colour output is fine and now use CS5 (running on 10.5) for the majority of my work.

If you're using Eric Chan's profiles for the 3800/3880 note that these don't go nearly far enough. You really need to tune the placement of the a/b curves with ABW's Horizontal & Vertical controls (this will generally require a re-linearization). The Color Density driver setting can also be useful for some papers. If you're serious about B&W it would be worth your while investing in an i1Pro and doing all this yourself.

Strange results on the Mac/CS5 combination indeed.  Thanks for the tip on refining the profiles; I'll do some more work on this.  I generally find that the 'dark' setting works better for me than the 'darker' one as Eric has advocated for some papers.  I think this is clearly an issue of having a lot of variables to tweak.

[digitaldog]

One other question (though this might open up another can of worms), if this happens for QTR profiles, can one be confident of color profiles made prior to these software updates?  Maybe it's just an artifact of QTR (one would really hope this would be the case otherwise there would be a lot of angry people out there).

Its absolutely not doing this with profiles I’ve built, its not doing it with ABW using Adobe RGB (1998) as the profile selection (otherwise, how would I use CS5 which has no No Color Management option and requires the use of Photoshop Manages Profile)?

Is it QTR? Could be, don’t care, don’t use it. You guys futz with it. Using my ProfileMaker Pro, PROFILER or i1P profiles, Epson profiles, ABW AS designed, CS4 and CS5 under 10.6 is all good.

We all got dragged down this rabbit hole due to a post that was both unclear in terms of actual usage and is basically incorrect in terms of ABW outside QTR!

[Alan Goldhammer]

Something strange must be going on as I saw exactly the same problem with Mac OS/CS5 and the Epson ABW driver over at Photo.net.

[digitaldog]

Well I think Patrick summed it up: what do you find weird? that i use the ABW correctly? ; )

[Alan Goldhammer]

Well I think Patrick summed it up: what do you find weird? that i use the ABW correctly? ; )

NO, only that another user was having the same problems on the Mac OS and CS5.  I was not casting any aspersion on your use only noting that there are now two reports of the ABW driver not working properly.  The user on photo.net had all the correct settings and was using Eric Chan's profile and still getting a poor print.  That implies that in HIS system something is not working.

[digitaldog]

NO, only that another user was having the same problems on the Mac OS and CS5.  I was not casting any aspersion on your use only noting that there are now two reports of the ABW driver not working properly.  The user on photo.net had all the correct settings and was using Eric Chan's profile and still getting a poor print.  That implies that in HIS system something is not working.

Maybe you guys need to dive deeper into the exact workflow with profiles such as Eric’s, or QTR, driver versions etc. I’ve done the testing here, I’m confident that ABW works properly and identically on CS4 and CS5. Dano said the same. Onsight said the same. More than a few users of ABW are saying the same by not complaining its broken. Based on the history of print drivers and Apple OS’s, its quite possible something is up but its sounding like a very, very rare condition and in the two I’ve heard of thus far, they are deviating from the recommended process of using ABW. 

[Alan Goldhammer]

Maybe you guys need to dive deeper into the exact workflow with profiles such as Eric’s, or QTR, driver versions etc. I’ve done the testing here, I’m confident that ABW works properly and identically on CS4 and CS5. Dano said the same. Onsight said the same. More than a few users of ABW are saying the same by not complaining its broken. Based on the history of print drivers and Apple OS’s, its quite possible something is up but its sounding like a very, very rare condition and in the two I’ve heard of thus far, they are deviating from the recommended process of using ABW. 

It's not up to me since I have a Win7 computer and it works fine with or without profile, either printing via CS5 or LR.  I'm bowing out of this discussion since I see no further point to raising any issues and will leave it up to Mac users who are having problems to work it out on their own.

[eleanorbrown]

I don't fine ABW problematic to use on my Mac.  I print through Lightroom print module and tone my images to 25/50 and a few other settings in the toning options in ABW. Neutral prints are great too tho most of the time I tone slightly.  I have a 9880 and 7900 printers. Used to use Imageprint but find I prefer ABW better. Eleanor