Adobe Color Printer Utility

[Ernst Dinkla]

It is not an act of smearing. I have been following these threads for more than a year now. I have the Adobe CS4 suite, Windows version. The CS5 version is crippled in my view because there's a problem between Adobe/Apple color management policies. There was no need to adjust it for Windows. A free application was promised many months ago and now a BETA version of it shows flaws, not just on CM but even with drivers. A company that has more than 7K employees and not getting this right within 6 months should get some criticism. I hardly have written in the threads on this subject, at the beginning I mentioned that a CM-Off button should do exactly what it says. There are OSses + their applications that work like that. I repeat that message here after a year.

I honestly do not understand that there is no anger about this issue. I see the same explorations of what colorspace has to be used to make it work safe as the Chan detour threads have shown for Snow Leopard + Photoshop CS5. It should be a solution not another puzzle. It is like getting a package of band aid and you have to buy glue to let them stick. And yes, part of this problem is caused by that Apple desire to hide anything that may be too complicated for its customers. Fine for an iPhone, different for a print environment. Maybe they should not listen to marketing if it is about CM.

If it is too costly let them write a manual for printing targets from Photoshop, InDesign, Lightroom, etc, with precise instructions for the printer models, what color space loops will work, etc. For one reason or another that is left to the gurus here to explore and to write down and not Adobe or Apple staff. Must be too embarrassing that it has to be done that way. And do not tell that it is all clear to everyone right now. I see enough capable men here in this list that are still confused.


met vriendelijke groeten, Ernst Dinkla

New: Spectral plots of +175 inkjet papers:
http://www.pigment-print.com/spectralplots/spectrumviz_1.htm

[Mark D Segal]

It is not an act of smearing. I have been following these threads for more than a year now. I have the Adobe CS4 suite, Windows version. The CS5 version is crippled in my view because there's a problem between Adobe/Apple color management policies. There was no need to adjust it for Windows. A free application was promised many months ago and now a BETA version of it shows flaws, not just on CM but even with drivers. A company that has more than 7K employees and not getting this right within 6 months should get some criticism. I hardly have written in the threads on this subject, at the beginning I mentioned that a CM-Off button should do exactly what it says. There are OSses + their applications that work like that. I repeat that message here after a year.

I honestly do not understand that there is no anger about this issue. I see the same explorations of what colorspace has to be used to make it work safe as the Chan detour threads have shown for Snow Leopard + Photoshop CS5. It should be a solution not another puzzle. It is like getting a package of band aid and you have to buy glue to let them stick. And yes, part of this problem is caused by that Apple desire to hide anything that may be too complicated for its customers. Fine for an iPhone, different for a print environment. Maybe they should not listen to marketing if it is about CM.

If it is too costly let them write a manual for printing targets from Photoshop, InDesign, Lightroom, etc, with precise instructions for the printer models, what color space loops will work, etc. For one reason or another that is left to the gurus here to explore and to write down and not Adobe or Apple staff. Must be too embarrassing that it has to be done that way. And do not tell that it is all clear to everyone right now. I see enough capable men here in this list that are still confused.


met vriendelijke groeten, Ernst Dinkla

New: Spectral plots of +175 inkjet papers:
http://www.pigment-print.com/spectralplots/spectrumviz_1.htm

Ernst, the underlying substance of what you are getting at is a real issue. I migrate between frustration and bewilderment over this stuff. My simple message is not to cloud the issue by making off-base comments about people who use Apple computers - not that I'm sensitive to such comments; it's just that they aren't helpful. More difficult to understand is that the version of this application we are discussing is NOT beta software. And yet more difficult to understand is the apparent inconsistency of the issues between users with different software and hardware configurations, which of course is a software engineer's worst nightmare. For example, at least the first step of the profile creation process using this app started very well for me, and I don't pretend to have a magic touch that other people experiencing problems with this don't have - it seems to be luck of the draw. I am now about to haul out the spectro and the backboard and try to create the profile and test it. Will report back in later.

[Doyle Yoder]

Here is my report printing the TC9.18 target from IDCS5 (Emulate Adobe InDesign 2.0 CMS Off), PSCS4, and the Adobe Color Printing Utility to my Canon iPF9000.

IDCS5 Vs. ACPU

--------------------------------------------------

dE Report

Number of Samples: 918

Delta-E Formula dE94

Overall - (918 colors)
--------------------------------------------------
Average dE:   0.55
    Max dE:   1.53
    Min dE:   0.08
 StdDev dE:   0.23

Best 90% - (825 colors)
--------------------------------------------------
Average dE:   0.50
    Max dE:   0.88
    Min dE:   0.08
 StdDev dE:   0.17

Worst 10% - (93 colors)
--------------------------------------------------
Average dE:   1.03
    Max dE:   1.53
    Min dE:   0.88
 StdDev dE:   0.14

--------------------------------------------------
--------------------------------------------------


IDCS5 Vs. PSCS4
--------------------------------------------------

dE Report

Number of Samples: 918

Delta-E Formula dE94

Overall - (918 colors)
--------------------------------------------------
Average dE:   0.31
    Max dE:   1.38
    Min dE:   0.01
 StdDev dE:   0.20

Best 90% - (825 colors)
--------------------------------------------------
Average dE:   0.26
    Max dE:   0.57
    Min dE:   0.01
 StdDev dE:   0.12

Worst 10% - (93 colors)
--------------------------------------------------
Average dE:   0.78
    Max dE:   1.38
    Min dE:   0.57
 StdDev dE:   0.17

--------------------------------------------------
--------------------------------------------------


PSCS4 Vs. ACPU
--------------------------------------------------

dE Report

Number of Samples: 918

Delta-E Formula dE94

Overall - (918 colors)
--------------------------------------------------
Average dE:   0.35
    Max dE:   1.54
    Min dE:   0.01
 StdDev dE:   0.14

Best 90% - (825 colors)
--------------------------------------------------
Average dE:   0.32
    Max dE:   0.51
    Min dE:   0.01
 StdDev dE:   0.10

Worst 10% - (93 colors)
--------------------------------------------------
Average dE:   0.61
    Max dE:   1.54
    Min dE:   0.51
 StdDev dE:   0.16

--------------------------------------------------
--------------------------------------------------

[digitaldog]

I am assuming you updated the AppColorMatchingInfo.xml file to add the ACPU so that it defaults to No Color Correction?

Why do I have to do this (and how?). I can still set it manually right (that’s what I did in the driver)?

[Mark D Segal]

Further to my post #34, I am pleased to report that this whole process worked properly for me. I am more than pleased about this, because I am using a profile making kit (XRite Pulse Color Elite) which dates from 2005, and therefore ante-dates Snow Leopard by some 5 years. The software and spectro both installed and operated seamlessly back in September when I purchased the MacPro. My operating environment is MacPro with OS 10.6.4., Epson 3800 printer and Ilford Gold Fibre Silk paper. The steps I followed were:

1. Install the new Adobe target printing application, save-out the 729 patch target from the Pulse Elite as two TIFF files (it is a two-page target), print each page individually from the Adobe application and let it dry over night.
2. Calibrate the spectro and measure the patches. This all worked fine. No misses, no repeats, which means that the patches printed with correct scale, as I reported in post 34.
3. Create the profile as a V2 for D50 illuminant using the Pulse Elite software.
4. Open Bill Atkinson's printer test target (the one with the strawberries, foliage, gradients, etc) and print it with Photoshop Manages Color using the newly created profile. Printer settings are 2880, finest detail, high speed OFF, Printer Color Management OFF (my usual). The profile for this paper is keyed with Epson Premium Luster paper in the Epson driver.
5. Compare the printed outcome with the soft-proof for the same condition, with and without "Simulate Paper White".

Visually the results are fine. I found the screen-to-print match closer without Simulate Paper White, which is not an unexpected outcome using this printer and paper combination. The Atkinson printer test target is usually a very reliable indicator of what to expect from prints of one's own photographs. Hence based on what I see as of now, I should be able to print my own images with a high degree of predictive confidence. I shall report back if I experience disappointment in this respect. As of now I have the impression that I am satisfactorily colour-managed. BTW, my display is an NEC PA271W calibrated and profiled with BasicColor Display 4, but operating on DVI-to-DVI 8-bit output (rather than the promised 10-bit) due to the unresolved mini-DisplayPort technical issue between Apple Computer and NEC Solutions.

[Doyle Yoder]

Why do I have to do this (and how?). I can still set it manually right (that’s what I did in the driver)?

Are you telling me that with the Canon iPF6300 drivers that when choosing application manages color when printing with apps that use Apple's new printing path that under Color Matching in the driver you can choose "Vendor Matching", and it is not grayed out and defaults to "ColorSync". If you can choose "Vendor Matching" then under Main/Color Mode you should be able to choose "No Color Correction" or "Color". Obviously with the old printing path printing from ID this is the behavior, but I have never ever heard of this being possible with the new printing path when printing with application manages color or as in the case of PSCS4 with No Color Management selected or with ACPU.

If that is the case then Canon has taken a radical change with their iPFx300 drivers as the iPFx000 and x100 drivers need to the have the applications that use the new print path listed in the AppColorMatchingInfo.xml file for the driver to turn off CM when printing with application manages color. I don't believe this to be that case because those with the iPFx000 and x100 printer they can install the iPFx300 drivers which will update the AppColorMatchingInfo.xml file. The AppColorMatchingInfo.xml file from the iPFx300 drivers includes PSCS5 and LR3 which is not in the AppColorMatchingInfo.xml file from the iPFx000 and x100 drivers. But the AppColorMatchingInfo.xml file from the iPFx300 drivers does not include LR3.3RC or ACPU which will cause the driver to default to ColorSync and in the Main/Color Mode to Color which will result in double profiling.

[Photo Op]

Mark- thanks for your post #44. I am a serious enthusiast, with nowhere near the experience you possess. But, I have been following these threads regarding "dark prints" and unsatisfactory print results for a few years. Similar to you, I have a MacPro with OSX 10.6.5, Epson 3880 printer and Red River Arctic Polar Satin (one of Eric's favorites). Your post #44 essentially sums up my results of the past two days working with the released (Beta?) Adobe Color Printer Utility. I do not have reams of technical printouts to quantify my results. I can however offer these comments. Having used for years the methods and printer test targets (of Atkinson and the DigitalDog) and the sage advice of Mr. Schewe and Michael in their various videos, I can say my journey (or search) is over. For me FINALLY, WYSISYG from screen to print. As I said, prints (the various test targets) are NOT dark. The blues are blue, without a purple tint. And the strawberries look good enough to eat. I've saved the results (printouts) of the targets from years of testing v10.4, 10.5, and 10.6, on Epson R2400, R2880 and now 3880. In the last 24 hours, I've compared the previous prints to what I printed today with LR v3.3RC and the differences are stunning.

Now, I don't know whether all this is the result of JUST using the Adobe Color Print Utility. Let me suggest that prior to running all the calibration I also updated "all" the software involved. I also have the most up to date print driver installed. So for me. the journey (or search for print nirvana) is over. That is until Apple does something STUPID, again.

YMMV!

[Alan Goldhammer]

So here I am in the dark ages running a Win 7 machine, using ColorMunki to profile my printer (my monitor is a NEC that I profile with their puck and SpectraView).  And you know what?  My prints match my monitor and the standard print that I use to validate the profile (the Jack Flesher one from the Outback website that uses Atkinson images) shows the colors to be dead on.  I guess I must be doing something right.

[digitaldog]

Here is my report printing the TC9.18 target from IDCS5 (Emulate Adobe InDesign 2.0 CMS Off), PSCS4, and the Adobe Color Printing Utility to my Canon iPF9000.
Delta-E Formula dE94

Can you set ColorThink for dE2000? Better for this kind of work.

Got the iPF6300 to print from CS4 after reinstalling print drivers. Odd, this morning unlike last night, it would not even “print” (click print, nothing happened at all, not the no ink on page result). So CS4 will have to dry down to compare to the matrix.

Here’s what I have from iPF6300, this time Prem Glossy paper.

IDCS5 vs. ACPU
Number of Samples: 930

Delta-E Formula dE2000

Overall - (930 colors)
--------------------------------------------------
Average dE:   0.23
    Max dE:   0.82
    Min dE:   0.02
 StdDev dE:   0.13

Best 90% - (836 colors)
--------------------------------------------------
Average dE:   0.20
    Max dE:   0.41
    Min dE:   0.02
 StdDev dE:   0.09

Worst 10% - (94 colors)
--------------------------------------------------
Average dE:   0.51
    Max dE:   0.82
    Min dE:   0.41
 StdDev dE:   0.09

--------------------------------------------------

CS5 vs. ACPU:
--------------------------------------------------
Average dE:   0.16
    Max dE:   0.79
    Min dE:   0.01
 StdDev dE:   0.10

Best 90% - (836 colors)
--------------------------------------------------
Average dE:   0.14
    Max dE:   0.29
    Min dE:   0.01
 StdDev dE:   0.07

Worst 10% - (94 colors)
--------------------------------------------------
Average dE:   0.37
    Max dE:   0.79
    Min dE:   0.29
 StdDev dE:   0.09

--------------------------------------------------
Very similar to the results on the Epson with good correlation of the two apps, low max dE. For reference, on the 3880, the dE was: Average dE:   0.19 Max dE:   0.72

CS5 vs. ID5:
--------------------------------------------------
Average dE:   0.25
    Max dE:   0.94
    Min dE:   0.02
 StdDev dE:   0.14

Best 90% - (836 colors)
--------------------------------------------------
Average dE:   0.22
    Max dE:   0.42
    Min dE:   0.02
 StdDev dE:   0.10

Worst 10% - (94 colors)
--------------------------------------------------
Average dE:   0.54
    Max dE:   0.94
    Min dE:   0.42
 StdDev dE:   0.11

--------------------------------------------------
Not as bad as I saw with the Epson, in this case, the worst RGB values were very dark 0/0/22.

So the Canon with its print driver appears to be more consistent in the three app’s while the Epson is farther off than I think it should be. Another Apple-Adobe-Epson combo issue?

[digitaldog]

I just remeasured the targets from CS4 vs. CS5 I printed on the 3880 yesterday. No dry issues at this point! I’m still not happy with the max dE and if possible, and those who are willing, I’d like to suggest some concentrate on these two app’s with Epson drivers (Canon at this point seems a bit more consistent). With today’s report, I’m still getting a max dE of 1.38 between the two versions of Photoshop, again on greens! 7 greens all above dE of 1 just seems wrong to me.

It appears that ACPU and CS5 correlate well, even with this Epson printer. But CS4 and CS5, not so much.

[Mark D Segal]

So here I am in the dark ages running a Win 7 machine, using ColorMunki to profile my printer (my monitor is a NEC that I profile with their puck and SpectraView).  And you know what?  My prints match my monitor and the standard print that I use to validate the profile (the Jack Flesher one from the Outback website that uses Atkinson images) shows the colors to be dead on.  I guess I must be doing something right.

Alan: no you are NOT in the dark ages. And you know it :-)

Are you using Photoshop CS4 or CS5 or Lightroom for printing? Have you made new profiles from Photoshop CS5 before this new Adobe utility was announced the other day?

i.e. it would be good to learn a bit more about what you are doing right!

[digitaldog]

Another useful find (I think <g>). Fellow color geek Chris Murphy and I have been doing work behind the scenes. He believes that using CS5 with the hack whereby you Assign a profile and then pick that profile for printing (to get a Null conversion) doesn’t work with sRGB as it does with Adobe RGB (1998). I can confirm that on this end printing to the 3880. Be useful if anyone else can confirm or deny this. If you recall, CS5 and the new ACPU appear to produce the same results when CS5 is set using the Adobe RGB (1998) profiles for assignment and printing.

[madmanchan]

I did my own tests yesterday on a 3880, Mac 10.6.4, driver 6.60. I tested CS4 and CS5 (with the workaround, with Adobe RGB), and ACPU. I let all the prints dry for 4 hours, then measured them all with an Eye-One UV in patch mode. Very slow, but reliable.

The results were all the same (within measurement error).

Namely, comparing pairwise (any 2 pairs), average DE2000 is 0.12. Max DE2000 is ~1.4.

For reference, when I print the target twice with the same app and compare those results, the numbers are the same.

Also, all of the targets printed at exactly the same size (7.5" x 10"). The ACPU printouts were not resized.

[digitaldog]

The results were all the same (within measurement error).
Namely, comparing pairwise (any 2 pairs), average DE2000 is 0.12. Max DE2000 is ~1.4.

Well I can’t say I necessarily agree here that a max dE of 1.4 is within measurement error or acceptable but I’m willing to listen otherwise.

I assume you sort the list in something like ColorThink by max to min dE. Why then are the worst few patches up over 1 dE while others are a fraction of a delta? If I measure the same chart twice, I don’t get the results you report so I have to assume the differences here are how the item is printed, not measured.

For fun, lets both print the same target identically two times in a row and measure and look at the report. So say ACPU printing the target twice. You really think you’ll get a 1 dE on any patch? That would be a real concern for a device like a modern Epson or Canon.

[Doyle Yoder]

Another useful find (I think <g>). Fellow color geek Chris Murphy and I have been doing work behind the scenes. He believes that using CS5 with the hack whereby you Assign a profile and then pick that profile for printing (to get a Null conversion) doesn’t work with sRGB as it does with Adobe RGB (1998). I can confirm that on this end printing to the 3880. Be useful if anyone else can confirm or deny this. If you recall, CS5 and the new ACPU appear to produce the same results when CS5 is set using the Adobe RGB (1998) profiles for assignment and printing.

You may find variation if the a not properly written printer driver/OS does not turn off CM. In the case 10.5 seems to use Generic RGB and 10.6 sRGB.

[digitaldog]

When I measure the TC918 from ACPU (two printed in a row and wet) I get a max dE of 0.70. That seems like a far cry from the max dE I’m getting from the same instrument (iSis XL) between the various app’s (worst case is 1.18). And I see far closer correlation between some app’s and the measurement data in terms of the max dE (CS5 vs. ACPU being 0.72). Should not all the app’s produce roughly the same max’s at least under 1.0?

[digitaldog]

Also, all of the targets printed at exactly the same size (7.5" x 10"). The ACPU printouts were not resized.

What printer and driver? Its absolutely resizing on this end going to an Epson 3880, v6.60 on 10.6.5. Others have reported the sizing differences too but that doesn’t seem to negatively affect the ability to read the targets.

[Rhossydd]

I did my own tests yesterday..........Also, all of the targets printed at exactly the same size (7.5" x 10"). The ACPU printouts were not resized.

Hasn't this been tested on any Windows systems ? If so, is the resizing seen by the Windows users here regarded as acceptable ?

Paul

[digitaldog]

Hasn't this been tested on any Windows systems ? If so, is the resizing seen by the Windows users here regarded as acceptable ?

IF its sizing the same on Windows and Mac, its acceptable to the degree that one can measure the targets. I mean, with an EyeOne, the patches can be this far off, no problem. On most auto Spectrophotometers like the iSis, if you are off more than say 5% either way, it will barf. I haven’t tested this on my other auto Spectrophotometer, a DTP-41UV-T but suspect that it would also measure without error. Should this “bug” have been detected and fixed prior to release (assuming this utility is wrong and the others are printing correctly which my rulers say they are)? Yes. Is it a deal breaker, no. And so far, the output from ACPU and CS5 seem to produce the same results which is good. Why ID5 and CS5 are more different <g> may be as much an issue with the OS as anything else. That’s why having some Windows users do the same testing from the same app’s would be real useful.

[Alan Goldhammer]

Alan: no you are NOT in the dark ages. And you know it :-)

Are you using Photoshop CS4 or CS5 or Lightroom for printing? Have you made new profiles from Photoshop CS5 before this new Adobe utility was announced the other day?

i.e. it would be good to learn a bit more about what you are doing right!

I use LR for all printing as I agree with Jeff Schewe that once you get the printing presets established everything is pretty much fool proof (I did have a friend print out one of my color images on his 9900 using Qimage and it was really difficult seeing any difference in the color shadings versus my 3880 with a CM profile).  I have CS4 (haven't updated to CS5 as I still don't see that there is anything new that will help me out) for pixel edits as necessary.  ColorMunki uses it's own utility to create profiles so you don't have to print targets through any Adobe product.  The only thing to be aware of with ColorMunki is the target drydown work around that has been mentioned on another thread.  For B&W printing I rely on Eric Chan ABW profiles.  As I noted, the test prints look excellent with the ColorMunki profiles.  If there is a particular print that might warrant a profile improvement, ColorMunki can look at the image and refine the profile.

Alan