Luminous Landscape Forum
Raw & Post Processing, Printing => Colour Management => Topic started by: Ryan Grayley on January 23, 2009, 08:26:44 am
-
[**Note that this bug appears to be fixed by the Apple MacOSX 10.6.3 update released on the 29th March 2010**]
In December 2008, I started to encounter problems creating profiles for my Epson Stylus Pro 7900. After some experimentation I thought I had a resolution. However it seems that others may not be having the same negative experiences so I am now having some doubts and hence this post.
When the ESP 7900 arrived I also upgraded to Mac PhotoShop CS4 and this is when my profiling problems started. I wondered if the PhotoShop CS4 ABW problems reported on the Adobe forums and elsewhere might be related in some way. This lead me to devise a test methodology as follows.
- To save paper and ink costs and also time I used just one strip of patches from the i1Match 918 chart and used Enhanced Matte paper for all the comparisons. My reasoning was that I only wanted to compare different software and hardware combinations and printing full charts would be wasteful and unnecessary.
- The test strip consisted of patches A7 to A17 from the i1Match 918 test chart.
- I printed the strip using various versions of PhotoShop on Windows and Mac.
- Colour management was disabled both in PhotoShop and the print driver.
- Each strip printed twice, one with Velvet Fine Art media setting and the other with Water-colour Paper Radiant White media setting.
- Matte black ink was used on the ESP 7900.
- The default printer was set to the ESP 7900.
The following 'Table A' results were different from the 'Table B' results. In particular patches A17, A15 and A11 were visibly and subtly different.
Table A:
(All results correct and identical in this table)
MacOSX 10.6.8 + CS4 + Epson driver v8.64 (driver Color Matching locked to ColorSync). Intel Mac. (2010-08-09)
MacOSX 10.6.8 + CS4 + Epson driver v6.55 (driver Color Matching locked to ColorSync). Intel Mac. (2010-08-07)
MacOSX 10.6.4 + CS4 + Epson driver v6.55 (driver Color Matching locked to ColorSync). Intel Mac.
MacOSX 10.6.3 + CS4 + Epson driver v6.55 (driver Color Matching locked to ColorSync). Intel Mac. [Bug appears fixed by the MacOSX 10.6.3 update.]
MacOSX 10.6.4 + CS1 + Epson driver v6.55 (driver Color Matching set to Epson Color Controls). Intel Mac.
MacOSX 10.6.3 + CS1 + Epson driver v6.55 (driver Color Matching set to Epson Color Controls). Intel Mac.
MacOSX 10.5.8 + CS1 + Epson Europe driver v6.55 (driver Color Matching set to Epson Color Controls). Intel Mac.
MacOSX 10.5.6 + CS1 + Epson Europe driver v6.39 (driver Color Matching set to Epson Color Controls). Intel Mac.
MacOSX 10.5.6 + CS1 + Epson Europe driver v6.39 (driver Color Matching set to Epson Color Controls). PowerMac.
MacOSX 10.5.6 + CS1 + Epson USA driver v6.35 (driver Color Matching set to Epson Color Controls). Intel Mac.
MacOSX 10.6 + i1Match + Epson driver v6.55. (driver Color Matching set to Epson Color Controls). Intel Mac.
MacOSX 10.5.6 + i1Match + Epson Europe driver v6.39. PowerMac.
Windows XP + CS2.
Windows XP + CS4.
Table B:
(All results incorrect in this table)
MacOSX 10.6.0 + CS4 + Epson driver v6.55 (driver Color Matching locked to ColorSync) Intel Mac.
MacOSX 10.5.6 + CS4 + Epson Europe driver v6.39 (driver Color Matching locked to ColorSync). Intel Mac.
MacOSX 10.5.6 + CS4 + Epson Europe driver v6.39 (driver Color Matching locked to ColorSync). PowerMac.
MacOSX 10.5.6 + CS4 + Epson USA driver v6.35 (driver Color Matching locked to ColorSync). Intel Mac.
MacOSX 10.5.6 + CS4 + Epson USA driver v6.35 (driver Color Matching locked to ColorSync). PowerMac.
Colorwiki suggests manually changing the default profile to Generic RGB in the devices section of the Apple ColorSync Utility but this didn't help.
( http://www.colorwiki.com/wiki/Tech_Support_Grab_Bag (http://www.colorwiki.com/wiki/Tech_Support_Grab_Bag) )
I then tried manually changing the ColorSync Utility default profile to match the media setting and in both cases the results matched the charts printed from the other versions of OS and PhotoShop.
Note that this problem has only occurred when disabling colour management in both PhotoShop and the print driver in order to print the test charts. Normal printing with colour management enabled in PhotoShop CS4 seems to be ok.
So is this test methodology sound?
And what are other peoples experiences?
Thanks,
Ryan
Edit 1: Added test result for Epson USA driver on Intel Mac and titied up the result tables.
Edit 2: Added two more test results including i1Match. (The 918 charts were modified in the i1Match chart folder so as to just print the single strip in the same way as with PhotoShop.). Fixed some typos.
Edit 3: Added print driver Color Matching settings on MacOSX.
Edit 4: Added more details of the test strip.
Edit 5: Added results for 7900 Mac driver 6.55
Edit 6: Added results for the Apple MacOSX 10.6.3 update released on the 29th March 2010. This bug appears to be finally fixed a whole 15 months after the bug was first discovered!
Edit 7: Added results for the Apple MacOSX 10.6.4
Edit 8: Added results for the Apple MacOSX 10.6.8 with Epson driver 6.55 (2010-08-07)
Edit 9: Added results for the Apple MacOSX 10.6.8 with Epson driver 8.64 (2010-08-09)
-
In December I started to encounter problems creating profiles for my Epson Stylus Pro 7900. After some experimentation I thought I had a resolution. However it seems that others may not be having the same negative experiences so I am now having some doubts and hence this post.
When the ESP 7900 arrived I also upgraded to Mac PhotoShop CS4 and then my profiling problems started. I wondered if the PhotoShop CS4 ABW problems reported on the Adobe forums and elsewhere might be related in some way. This lead me to devise a test methodology as follows.
..........
- The combinations of OS and Photoshop tried were:
Windows XP + CS2
Windows XP + CS4
MacoOSX 10.5.6 + CS1 + Epson Europe driver v6.39 on Intel.
MacoOSX 10.5.6 + CS4 + Epson Europe driver v6.39 on Intel.
MacoOSX 10.5.6 + CS1 + Epson Europe driver v6.39 on PowerPC.
MacoOSX 10.5.6 + CS4 + Epson Europe driver v6.39 on PowerPC.
MacoOSX 10.5.6 + CS4 + Epson USA driver v6.35 on PowerPC.
In every case the results were identical except MacOSX 10.5.6 and PhotoShop CS4.
Colorwiki suggests manually changing the default profile to Generic RGB in the devices section of the Apple ColorSync Utility but this didn't help.
( http://www.colorwiki.com/wiki/Tech_Support_Grab_Bag (http://www.colorwiki.com/wiki/Tech_Support_Grab_Bag) )
I then tried manually changing the ColorSync Utility default profile to match the media setting and in both cases the results matched the charts printed from the other versions of OS and PhotoShop.
Note that this problem has only occurred when disabling colour management in both PhotoShop and the print driver in order to print the test charts. Normal printing with colour management enabled in PhotoShop CS4 seems to be ok.
So is this test methodology sound?
And what are other peoples experiences?
Thanks,
Ryan
I had the exact same experience, after careful testing. CS4/XP matched CS3/OSX 10.5 as well as all previous setups, but not CS4/OSX 10.5. Drove me crazy.
(Again, not an issue for normal printing with profile, only for printing test charts.) I gave up and kept using my XP machine just for printing targets.
Could you please explain in more detail the Colorsying settings workaround, and any helpful links? Also, is there any indication that Adobe or Apple is fixing this?
Thanks.
-
Two questions:
1. Were you printing using the RGB color mode for the Epson driver (a.k.a. AccuPhoto HD) or the ABW driver?
2. Did you try the latest Mac Intel driver from the Epson USA support site?
-
I had the exact same experience, after careful testing. CS4/XP matched CS3/OSX 10.5 as well as all previous setups, but not CS4/OSX 10.5. Drove me crazy.
(Again, not an issue for normal printing with profile, only for printing test charts.) I gave up and kept using my XP machine just for printing targets.
Could you please explain in more detail the Colorsying settings workaround, and any helpful links? Also, is there any indication that Adobe or Apple is fixing this?
Thanks.
The ColorSync Utility workaround as follows.
- Open ColorSync Utility and click on the Devices tab.
- Scroll down to the relevant printer section for your printer. (For example the Epson SP 7900)
- The default Mode will most probably be set to Premium Luster Photo Paper (and indicated by an Aqua blue ball) in the left hand pane.
- Click on this default entry.
- In the right hand pane will be two entries associated with the selection.
- The 'Factory Profile' entry is for information and indicates the default factory profile for the selected mode.
- The 'Current Profile' entry allows the default factory profile to be changed to a different profile.
- Click on the 'Current Profile' drop down menu and select 'Other...'
- A foreground window will then open and here a different profile can be selected.
- Select the canned Epson profile for the media setting that you will be using to print the test chart.
- When the test chart has been printed, restore the default profile to the factory default.
I can't guarantee this will work for you but it seemed to work for me. Please post back here with your findings.
I found various links to threads on this forum and on the Adobe forums that are mainly concerned with problems when using ABW. Eric Chan from Adobe has mentioned that Adobe and Epson are working on a fix for the ABW problems but I don't know if this will also solve the test chart problems.
Here are some links which may be useful.
http://luminous-landscape.com/forum/index....showtopic=29496 (http://luminous-landscape.com/forum/index.php?showtopic=29496)
Cheers,
Ryan
Edit1: Fixed grammar.
Edit2: Removed broken links.
-
Two questions:
1. Were you printing using the RGB color mode for the Epson driver (a.k.a. AccuPhoto HD) or the ABW driver?
2. Did you try the latest Mac Intel driver from the Epson USA support site?
1. My difficulties only relate to the printing of untagged RGB colour test charts and so I have been using the RGB color mode.
2. So far I have tried the European driver v6.39 on Intel Mac and PowerMac; and the USA driver v6.35 on PowerMac.
Should I try the USA driver on Intel too?
Thanks
Ryan
Edit1: Fixed typo.
-
1. My difficulties only relate to the printing of untagged RGB colour test charts and so I have been using the RGB color mode.
2. So far I have tried the European driver v6.39 on Intel Mac and PowerMac; and the USA driver v6.35 on PowerMac.
Should I try the USA driver on Intel too?
Edit1: Fixed typo.
I have now tried the USA ESP 7900 driver on an Intel Mac and this also gives an 'incorrect' result.
Ryan
-
I have now tried the USA ESP 7900 driver on an Intel Mac and this also gives an 'incorrect' result.
Ryan
OKay thanks for this thread. I have been going crazy about this. Here is my setup
MAcbook pro intel mac
have the esp 7900 with usa driver.
Now my problem is when i print a test chart out of any program they all come out different. i have cs4,cs3, aperture 2 and lightroom2 also eye one match 3.2.
they are so different that i can tell from my naked eye. some colors are a little off and some are a different color completely. i also have vm fusion installed and qimage on the xp side of things. when i print from eye one and qimage under xp the charts come out the same.
i was wondering if you get the same results if you print from eye one and cs1. will the charts look the same on yours?
thanks
chris
-
OKay thanks for this thread. I have been going crazy about this. Here is my setup
MAcbook pro intel mac
have the esp 7900 with usa driver.
Now my problem is when i print a test chart out of any program they all come out different. i have cs4,cs3, aperture 2 and lightroom2 also eye one match 3.2.
they are so different that i can tell from my naked eye. some colors are a little off and some are a different color completely. i also have vm fusion installed and qimage on the xp side of things. when i print from eye one and qimage under xp the charts come out the same.
i was wondering if you get the same results if you print from eye one and cs1. will the charts look the same on yours?
thanks
chris
Yes I have also tried i1Match and the prints are the same as PhotoShop CS1.
I have updated the table on the first post to include i1Match.
Ryan
-
Interesting findings. Unfortunately I do not have a copy of CS1 around that I can find, still looking. I'd like to try this as well.
So a profile made using CS4 should work fine as long as the image is printed from CS4.
What I'm not sure of is which is is "correct". It very well could be CS1 is not compatible with current operating systems.
Edit: and I've never trusted output directly from EyeOne Match.
-
i am still new to this so my question is this. if i print a chart from qimage in xp on fusion can i use that profile in cs4 and mac osx? i think that will be my best way of a correct profile. i don't know why i can't match my charts in eye1 and cs3.
thanks
-
i am still new to this so my question is this. if i print a chart from qimage in xp on fusion can i use that profile in cs4 and mac osx? i think that will be my best way of a correct profile. i don't know why i can't match my charts in eye1 and cs3.
thanks
I haven't used Qimage or PhotoShop CS3 and (like Wayne) I don't normally use i1Match to print targets. I prefer to use PhotoShop to print targets because it normally gives me more control over colour management. For the above tests, I downloaded the 30 day trial of the Windows version of PhotoShop CS4 and used this to print some test charts. The results were identical to the Mac version of PhotoShop CS1. If you achieved good results with the Windows trial version in this way you would then need to decide what to do when the trial runs out. One option would be to ask Adobe to offer a platform swap of your Mac version to the Windows version. I did this once and it was quite straight forward but I think I had to dig out my original proof of purchase. Otherwise I don't really know what else to suggest.
-
Interesting findings. Unfortunately I do not have a copy of CS1 around that I can find, still looking. I'd like to try this as well.
So a profile made using CS4 should work fine as long as the image is printed from CS4.
What I'm not sure of is which is is "correct". It very well could be CS1 is not compatible with current operating systems.
Edit: and I've never trusted output directly from EyeOne Match.
Do you have access to Windows XP? If so, you could try the 30 day trial of PhotoShop CS4 for Windows and use it to print some test strips. This would be really interesting.
It would be great if my tests were reproduced...any more volunteers?
Cheers,
Ryan
-
Okay so today will download a trial of cs4 under xp. then i will print strips from xp cs4, xp qimage xp eye one, and then from mac osx eye1, cs3, cs4. and then compare to see which ones come out the same. i keep thinking i am doing something wrong but when i use xp i get the same results. i will report back when done
-
I have found a copy of CS1 which I installed. I cannot find a way under 10.5.6 to completely bypass all color management using CS1. It's been so long since I used it (I'm not even sure I was making my own profiles when CS1 was out). I opened page of the the EyeOne Match 918 RGB target for some tests.
Anyway, I see only 2 choices for color management under CS1, same as source or Printer color management.
No matter which one I choose, the problem is color management is actually not disabled in the print driver for CS1 when using 10.5.6. If you disable color management and look under the Color Matching tab, you see the driver still has ColorSync Enabled, and you can choose instead Epson . ColorSync actually also enables a popup where you select an output profile ... which defaults to the printers default profiles as set in the ColorSync utility. When printing from CS4, both of these are grayed out. So I cannot be sure that when I print this target with CS1 and 10.5.6 that I have actually sent the unmodified data to the printer. Somewhere along the line it appears as though some CMS may be modifying the data.
If you print using Same As Source, and instead of using ColorSync (which asks you to select an actual output profile, which defaults to the printers ColorSync default) you set the printer to use Epson Color Controls, for the most part the chart actually matches the CS4 chart. In fact, visually everything matches perfectly except column 1.
The CS1 print does not look like a target that you would normally measure ... there are no really dark patches except the two end columns. I don't think a proflie made with this target would perform very well ... seems like the gamut volume would be seriously hampered. Output would look OK as long as I printed under the same setup, but the profile would be limiting the printers gamut. (Similar to a problem I had when profiling an ipf6100, where I didn't disable color matching. The profiles worked fine as long as I used the same settings, but the profile gamut volume was substantially less than others were getting with their 6100's).
I believe this may be a problem that some Canon users are having as well under 10.5 ... probably with the ipf5000 since Canon did not update those drivers for Leopard.
As to why the CS1 matches windows output, I'm not sure, and am curious. I guess my quandary is so far I cannot conclude which of these prints is actually correct, and just because one matches the windows output ... well not sure, because I know nothing about windows printing and windows driver.
I'll keep doing some testing. I do know that targets printed from CS4 and profiled yield very good results for me. I have sent some of these to others and have had no complaints about them not working. I'm still suspecting the problem is 10.5.6 with older versions of CS ...
Edit ... I just printed the target from CS4 using Photoshop manages color and the printer profile... it matches my first 2 CS1 targets exactly. The first target was printed using Printer manages color, color management turned off in the printer settings tab, and the color match tab default (which is ColorSync and the default colorsync printer profile). The second was same as source and again the default under the Colormatch tab (Colorsync, default profile). To me it appears it is impossible to send completely unmodified information to the printer using CS1/10.5.6.
Edit 2 ... I just confirmed that EyeOne Match 2.6.2 has the same problem as CS1 under 10.5.6 ... disabling color management under the Printer Settings menu does not disable Color Matching. Thus you can match prints from CS1 but you can't match prints from CS4 ... unless someone can find a way to actually completely disable the affects of the Color Matching tabs.
-
Thought I would add that problems printing from CS1 and CS2 in leopard are not new, and in fact this is similar to problems printing from LR 1.2 in Leopard as well.
Bottom line for me ... CS4 and 10.5.6 targets are being printed accurately. If Windows is delivering prints that match CS1 and Leopard, I would suspect somehow the printer profile is getting applied in Windows Photoshop somewhere as well, since printing from CS4/10.5.6 and applying the printer profile delivers a target that matches CS1/Leopard. I'll stay opened minded here, but I think the problem is testing against software that even Adobe says is not compatible with Leopard.
Of course, it should be possible to print from Windows CS4 and Mac CS4 and achieve the same results. I may have a friend of mine who runs full window systems and makes his own profiles print out a few targets and see if they match mine. I don't trust myself and windows.
-
OKay so i did some test. by the way i am new to this so take it with a grain of salt. first off i installed cs4 under vm fusion. i ran prints of the eye 1 rgb chart out of windows cs4, eye 1 in windows and qimage in windows. all three charts are the same. then i ran prints from mac osx from eye 1, lightroom 2, aperture 2, and cs4. all these prints varied from one another. as compared to the windows ones mac cs4 matched all the ones out of windows, lightroom 2 came really close only a couple of squares i can see a little difference, eye 1 and aperture were off. Now a question for you gurus in here. if i make a profile from one of these can i send it to you so you could see if it looks all right gamut wise? i think i will print my charts out of mac and cs4.
thanks
-
Today I decided to try one more test (call me curious).
I have just finished setting up a mac reverting back to system 10.4.11 to try and troubleshoot a problem with an i1i0 table. This is a fresh install, with no software other than Apple's and Profilemaker.
I installed CS1 on this computer. As I mentioned yesterday it seems impossible to print with color management disabled using CS1 and 10.5.x. Printing problems with Leopard were widespread when it was released.
From this I printed Bill Atkinson's 918 patch i1i0 target, with color management disabled, to my 3800. I printed the same target from CS4 and system 10.5.6 to the same printer (2 different macs).
the targets visually are identical ... to my eyes a perfect match. I cannot detect a single square that appears different visually. I resized the target down to fit an 8.5 x 11 paper, so I can't verify this by reading it, but at this point that seems unnecessary.
CS1 and 10.5 are not compatible. There might be a work around, but that seems pointless. I also believe this applies to CS2. My memory is weak, but I thought CS3 had an update after Leopard was released so I believe CS3 with that update and CS4 are compatible with Leopard.
My conclusion is targets printed from CS4 using OS 10.5, if settings are correct, are accurate. The difference seen with CS1 and leopard are due to CS1 being incompatible, not a problem with CS4.
The only remaining question I have (for Ryan) would be what problems you were encountering in December with your 7900 that led you to do this testing? Don't know if any of us can be of any help.
-
My conclusion is targets printed from CS4 using OS 10.5, if settings are correct, are accurate. The difference seen with CS1 and leopard are due to CS1 being incompatible, not a problem with CS4.
The only remaining question I have (for Ryan) would be what problems you were encountering in December with your 7900 that led you to do this testing? Don't know if any of us can be of any help.
Thanks Wayne for carrying out your tests...
May I ask which paper and black ink type you used?
Also, could you possibly have a quick look at the Devices section in ColorSync utility please?
I am interested to know the default mode/profile that is selected for the printers that you have tested.
In my case, the default 'Mode' for my ESP7900 is locked to 'Premium Luster Profile Paper 260' and this is set to the factory default profile (in other words, both mode and profile are set to 'Premium Luster Profile Paper 260' profile).
I started investigating this issue because one of the profiles I created before Christmas resulted in a bad print.
I have now re-printed the test chart with Mac CS1 (BA 5202) and re-profiled and the results are now fine.
I may follow your lead and try earlier versions of MacOSX on my G5. I will start with MacOSX 10.3 and then I may try MacOSX 10.4. This may take a little while longer as my print business is starting to get busy again.
By the way if CS1 isn't compatible with MacOSX 10.5.6 can anyone suggest why my installation of CS1 produces test charts that are exactly the same as test charts produced by PhotoShop CS2 and CS4 on Windows XP?
Cheers,
Ryan
-
Thanks Wayne for carrying out your tests...
May I ask which paper and black ink type you used?
Also, could you possibly have a quick look at the Devices section in ColorSync utility please?
I am interested to know the default mode/profile that is selected for the printers that you have tested.
In my case, the default 'Mode' for my ESP7900 is locked to 'Premium Luster Profile Paper 260' and this is set to the factory default profile (in other words, both mode and profile are set to 'Premium Luster Profile Paper 260' profile).
I started investigating this issue because one of the profiles I created before Christmas resulted in a bad print.
I have now re-printed the test chart with Mac CS1 (BA 5202) and re-profiled and the results are now fine.
I may follow your lead and try earlier versions of MacOSX on my G5. I will start with MacOSX 10.3 and then I may try MacOSX 10.4. This may take a little while longer as my print business is starting to get busy again.
By the way if CS1 isn't compatible with MacOSX 10.5.6 can anyone suggest why my installation of CS1 produces test charts that are exactly the same as test charts produced by PhotoShop CS2 and CS4 on Windows XP?
Cheers,
Ryan
Ryan,
I was printing to an Epson 3800, which is my default printer, and the ColorSync default profile is PLPP. I was printing on Epson Premium Luster paper with PK ink. However, I don't think any of that matters.
After some testing, I think you are headed in the wrong direction if you think you should be using CS1 to print targets. I believe something about your extensive testing is yielding results that are causing you to draw the wrong conclusions.
I have now been able to print from CS4/10.5.6, CS1/10.4.11, C1/10.5.6, and Windows XP running under Parallels and all 4 targets are visually perfect matches - and look like what I expect targets to look like. My first target under windows was very different, but I believe that was because I clicked a wrong button so I didn't actually get the Windows color settings disabled. It took a couple of tests under CS1/10.5.6, but I did find one combination that seemed to work around the problem.
I can't explain why you are having the difficulties you are but because your Windows target matches your CS1/10.5.6 target you may be reaching an incorrect conclusion. Once you completely disable any involvement of Photoshop and the operating system (windows or Mac) managing color, the printer will deliver identical results. If it doesn't, then some setting isn't right, the program and OS are incapable of not interfering with the data, or some other problem is occurring with the printer (a few clogged nozzles for example).( I assume you are insuring that when opening the targets, Photoshop is not assigning or converting a profile, and that in the printer dialog box it shows the files as untagged RGB. I prefer doing this by enabling the Missing Profile -Ask When Opening checkbox under the Edit/Color Settings menu, which forces a dialog box when opening an untagged file. If I don't get that dialog I know I have a problem. )
The challenge is you can't actually disable Color Management in the operating system with CS1/10.5. Leopard required new drivers for all major high end printers (Epson beta drivers appeared very shortly after and took them quite a while to get into release). Leopard also required Photoshop to communicate differently to disable color management, a change I believe came with an update to CS3. CS1 is not compatible with 10.5, Adobe has so stated, and this is one of the incompatibilities. It also continues to cause problems with various programs, such as EyeOne Match. (Note that these changes did enable programs running in Leopard to actually access some color management when printing. Until Leopard getting programs like iPhoto to access and use different paper types and profiles required jumping through hoops similar to the ColorSync/default profile problems - in fact that's how you did it was well. iPhoto could only print to the printers default output profile. Now using iPhoto you can select different profiles under the Color Matching/ColorSync options.)
As mentioned, I did get CS1/10.5.6 to print a target that matches. (a work around). When selecting print with preview in CS1, the document space must be untagged RGB. The Print Space Profile should be Same as Source. If both of these are set, then Photoshop is sending the unmodified data to the printer. When the printer dialog box opens, under Printer Settings Color Mode should be set to off. Finally, under Color Matching settings, select Epson Color Controls. (If you leave it to ColorSync, ColorSync will actually apply the profile shown in the popup - which is set to the ColorSync default unless you pick a different one). I assume that the default for Epson Color Controls is no change to the color .. this is why it works. But to insure accuracy Color Matching should actually be disabled when Photoshop is controlling the color. Note this is what happens under CS4 ... The color matching options are grayed out.
At this point, I again believe that CS4/10.5.6 targets are accurate as long as you are using the correct procedures. I don't believe your problem in December is related to a problem with CS4 and OS X 10.5.
-
Ryan,
I was printing to an Epson 3800, which is my default printer, and the ColorSync default profile is PLPP. I was printing on Epson Premium Luster paper with PK ink. However, I don't think any of that matters.
Thanks for checking your ColorSync Utility. As you are printing on Luster you won't notice any differences with Mac CS4 as the default ColorSync Utility profile is Luster and therefore correct for the paper you are testing. However, if you were to exactly recreate my test methodology then that would great. In order to do this you will need to use your Epson SP 9900, matte black ink, Enhanced Matte paper and use the same strip from the i1Match 918 test chart. Your 3800 may be a valid alternative for the tests but I think it would be better to eliminate as many variables as possible. In case you don't have any Enhanced Matte, I will run a few test strips on ordinary copy paper to see if the differences can still be observed.
My theory is that regardless of the chosen media type, PLPP is always being used to determine the ink densities when disabling colour management in Mac CS4. If I am correct then this could imply that the Epson canned profiles contain information on ink density and therefore must be set correctly even when printing test charts with Colour Management disabled. This would explain correct results if I manually change the default profile in ColorSync Utility to the chosen media setting.
Could a few more colour experts offer some comment please? Eric?
Thanks again Wayne,
Ryan
-
I was printing to an Epson 3800, which is my default printer, and the ColorSync default profile is PLPP. I was printing on Epson Premium Luster paper with PK ink.
Thanks for checking your ColorSync Utility. As you are printing on Luster you won't notice any differences with Mac CS4 as the default ColorSync Utility profile is Luster and therefore correct for the paper you are testing. However, if you were to exactly recreate my test methodology then you might see what I am seeing. In order to do this you will need to use your Epson SP 9900, matte black ink, Enhanced Matte paper and use the same strip from the i1Match 918 test chart. Your 3800 may be a valid alternative for the tests but I think it would be better to eliminate as many variables as possible. In case you don't have any Enhanced Matte, I will run a few test strips on ordinary copy paper to see if the differences can still be observed.
My theory is that regardless of the chosen media type, PLPP is always being used to determine the ink densities when disabling colour management in Mac CS4. If I am correct then this could imply that the Epson canned profiles contain information on ink density and therefore must be set correctly even when printing test charts with Colour Management disabled. This would explain correct results if I manually change the default profile in ColorSync Utility to the chosen media setting.
Could a few more colour experts offer some comment please? Eric?
Thanks again Wayne,
Ryan
-
Two questions:
1. Were you printing using the RGB color mode for the Epson driver (a.k.a. AccuPhoto HD) or the ABW driver?
2. Did you try the latest Mac Intel driver from the Epson USA support site?
Sorry for late reply.
1. RGB mode.
2. Yes, it is version 6.11.
FWIW, I was using Epson Premium Semigloss in a 7800. Test charts were inspected not only visually, but also compared statistically in Profiler Pro Measure Tool.
-
The ColorSync Utility workaround as follows.
- Open ColorSync Utility and click on the Devices tab.
.......snip......
- Click on the 'Current Profile' drop down menu and select 'Other...'
- A foreground window will then open and here a different profile can be selected.
Sorry for the delay in replying.
My Current Profile drop down arrow was grayed out and didn't work, so I couldn't go any further at the time. I just now tried logging in as Admin and now it works; Apple didn't seem to provide any authentication dialog for this utility! I will try your workaround when I get a chance. Your theory about the Premium Luster default seems plausible. As I mentioned in reply to Eric, I am using Premium Semigloss in a 7800, and comparing the test chart statistics in Profiler Pro.
-
Thanks for checking your ColorSync Utility. As you are printing on Luster you won't notice any differences with Mac CS4 as the default ColorSync Utility profile is Luster and therefore correct for the paper you are testing. However, if you were to exactly recreate my test methodology then you might see what I am seeing. In order to do this you will need to use your Epson SP 9900, matte black ink, Enhanced Matte paper and use the same strip from the i1Match 918 test chart. Your 3800 may be a valid alternative for the tests but I think it would be better to eliminate as many variables as possible. In case you don't have any Enhanced Matte, I will run a few test strips on ordinary copy paper to see if the differences can still be observed.
My theory is that regardless of the chosen media type, PLPP is always being used to determine the ink densities when disabling colour management in Mac CS4. If I am correct then this could imply that the Epson canned profiles contain information on ink density and therefore must be set correctly even when printing test charts with Colour Management disabled. This would explain correct results if I manually change the default profile in ColorSync Utility to the chosen media setting.
Could a few more colour experts offer some comment please? Eric?
Thanks again Wayne,
Ryan
OK ... I think I see where you are coming from. What you are saying is that unless the default colorsync profile is set to a proflie built with the paper type you are choosing in the print settings dialog, the ColorSync default profile will override some setting when printing the target. The only way around that is to reset the ColorSync default Profile to one that was built using the same media type you are going to choose in the Print Settings .. in this case choosing Enhanced matte MK as both the paper type and having ColorSync default set to Enhanced Matte_MK.
In theory, a color profile shouldn't be involved in the ink load. It should be set and controlled by the media type in the print settings dialog, and possibly modified using the Advance media control. In fact, typically with 3rd party papers you have to print test prints using various Epson Media choices to find one that will deliver an appropriate ink load. Perhaps there is something in a profile which modifies this however.
Anyway, I did have a roll of Enhanced Matt laying around. I loaded up Enhanced Matte on my 7900. I printed the EyeOne Match target using CS4/10.5.6 with all settings like I would normally do. The target appeared pretty normal, but admittedly a little "flat" ... to the point that if this were a 3rd party paper I was profiling I would probably have punched up the ink load a little. I then loaded the target into CS1, and using the method I described yesterday printed the target, and sure enough it didn't match. It appeared, as you described, to have more "ink". Most of the darker colors were a little darker, most of the lighter colors were about the same. But visually it was obvious they weren't identical.
I then changed the default ColorSync profile to Epson Enhanced Matte _MK, and printed the target from CS4. It did not match the first CS4 target, as you stated, and it did match the CS1 target.
So at this point I agree there appears to be a problem ... at least with the 79/9900 driver. I may test this with my 11880. Despite all of the correct settings in CS4 and the print driver, if you are trying to print a target with no color management involved at all, it appears to only work if the media type you are choosing in print settings is the same as the media type used for the profile that is the Default ColorSync profile for that printer as set in ColorSync utility. Of course this is assuming the the work around for CS1/10.5 is actually accurate as well.
This doesn't appear to affect actual output ... if allowing Photoshop to manage colors and printing using the correct printing profile, it did not matter which profile was set as the default in ColorSync. It only appears to be a problem when trying to print a target (or anything that is trying to eliminate all color management).
Current work arounds for me appear to be 4 ...
1. Reset the colorSync default profile to one that was built using the media type you are using for the paper you are building the profile for. The question is what if you want to change the inkloads ... does the default in ColorSync over ride this as well? May have to test that.
2. Print from CS1 (and perhaps 2?), making sure the ColorMatching dialog is set to Epson Color Controls (this may not work if these are not set to a default. I don't know enough about them to know where and when Epson Color controls may actually affect the data.)
3. Print from 10.4.11 - it seems to work with CS1, I'm guessing it will also work with CS4 (will test that tonight). It appears all of this is related to Apples change in the printer mechanics introduced with Leopard.
4. Print from XP.
Again, this appears only necessary if trying to print with no color management involved.
Perhaps we should move this over to the Adobe Forums?
-
I would say this seems to make sense. When i printed targets on glossy paper they were fine and matched out of cs4 and eye 1. but when i tried with matte paper known of them matched. i will try and change the colr sync profile if i can figure it out.
thanks for all the work.
chris
-
OK, its seems we are closing in on the problem.
Right now, I am getting correct image printouts on any paper, as long as I select my custom profile under Photoshop Manages Colors, even though the Colorsync Utility has remained set to the default Premium Luster setting. That is, everything works fine as long as I'm not trying to print an unmanaged target.
The thing I am not clear on in the suggested workaround is this:
Suppose I want to print an unmanaged target. I go ahead and set the Colorsync Utility default profile to the same paper that I plan to use, and then print the target and make a profile. The consensus is that this works.
But it is also true that Photoshop prints images correctly via Photoshop Manages Colors using my custom profiles for *any* paper even though the utility is left set to its Premium Luster default.
My question is, does the Colorsync Utility profile setting have any effect on normal printing through a profile, or only on unmanaged target printing? Put another way, is there any harm in just setting the Colorsync Utility default profile to the paper in use, and leaving there even when not printing unmanaged targets? Or does it need to be restored to its Premium Luster setting to print normally through Photoshop Manages Colors?
Probably would have been easier to try it than type all this.
-
Don't touch the ColorSync utility, that's my first suggestion!
-
Don't touch the ColorSync utility, that's my first suggestion!
If you don't touch the colorsync utility and need to print a target with all color management off to a matt paper, such as enhanced matt, the target is not printed correctly.
I don't understand why if you turn management off in both Photoshop and the Epson Print Settings dialog, how the default profile in ColorSync would have any affect at all ... but it does.
What do you suggest?
-
My question is, does the Colorsync Utility profile setting have any effect on normal printing through a profile, or only on unmanaged target printing? Put another way, is there any harm in just setting the Colorsync Utility default profile to the paper in use, and leaving there even when not printing unmanaged targets? Or does it need to be restored to its Premium Luster setting to print normally through Photoshop Manages Colors?
Probably would have been easier to try it than type all this.
No. I've tested this and as long as Photoshop is managing colors, the default profile in ColorSync doesn't matter ... prints are identical. This is only a problem if you are trying to print a target with no color management to a paper other than the paper used to build the profile that is the default in colorSync.
In the test Ryan did and I confirmed, if you try and print a target in the normal fashion on Epson Enhanced Matt with MK inks, that target will be incorrect unless you also change the default profile in ColorSync utility to Epson Enhance Matt_MK. When installing the driver it appears the default profile is always set to Luster.
-
If you print using Same As Source, and instead of using ColorSync (which asks you to select an actual output profile, which defaults to the printers ColorSync default) you set the printer to use Epson Color Controls, for the most part the chart actually matches the CS4 chart. In fact, visually everything matches perfectly except column 1.
If memory serves me, you'd want Same as Source (untagged RGB) and in the Epson driver, No Color Adjustment.
-
If memory serves me, you'd want Same as Source (untagged RGB) and in the Epson driver, No Color Adjustment.
I tried that. IIRC it didn't give me a correct profiling target print on OSX 10.5.6 with CS4.
Though I think it used to on all previous OSX and XP with all previous PS for the last 6 years. Something is different with either CS4 or 10.5 or both.
-
I tried that. IIRC it didn't give me a correct profiling target print on OSX 10.5.6 with CS4.
Though I think it used to on all previous OSX and XP with all previous PS for the last 6 years. Something is different with either CS4 or 10.5 or both.
Indeed, Eric Chan has mentioned such changes with 10.5 and CS4 several times on the Adobe Forums in the context of problems with ABW.
I regard the behavior of CS4 discussed in this thread as another bug. In case this issue hasn't been reported to Adobe, I have submitted a bug report at the link below.
http://www.adobe.com/cfusion/mmform/index.cfm?name=wishform (http://www.adobe.com/cfusion/mmform/index.cfm?name=wishform)
In the meantime I am going to use CS1 to print my test charts.
Ryan
-
Indeed, Eric Chan has mentioned such changes with 10.5 and CS4 several times on the Adobe Forums in the context of problems with ABW.
I regard the behavior of CS4 discussed in this thread as another bug. In case this issue hasn't been reported to Adobe, I have submitted a bug report at the link below.
http://www.adobe.com/cfusion/mmform/index.cfm?name=wishform (http://www.adobe.com/cfusion/mmform/index.cfm?name=wishform)
In the meantime I am going to use CS1 to print my test charts.
Ryan
Unclear if this is actually an Adobe issue. Could be Epson or Apple as well ... could be the 3 are just not interacting correctly.
Thought I would update for those interested. It also appears you cannot get an accurate target for Enhanced Matt Papers using an 11880 and system 10.5.6 as well. Not only is my target inaccurate, but I have not found a work around yet ... changing the default ColorSync profile doesn't affect the target.
To try and make sure I wasn't doing something wrong I printed Bill Atkinsons 918 patch target to my 11880 from one Mac using system 10.5.6 and CS4, another Mac using system 10.4.11 and CS1. (Looking for my installer disks for CS2 and CS3). The two are visibly different, but not as obvious as the 7900. The CS4 target looks "milky". At first I thought perhaps somehow I messed up and used PK ink instead .. that's kind of what it look like. So I added a large black patch, and printed the target twice more, once with MK and once with PK just to be sure there was a diffirence. As expected it was easy to verify from the black patch that the MK target did indeed use MK ink.
So I repeated the printing with the black patch from CS1 using both MK and PK, just to make sure I had what I wanted. Again, the C1 target looked better ... cleaner colors, more saturation.
I then went ahead and read and proflied the two targets, and I compared the results in ColorThink against the 11880 default Enhanced Matt _MK profile From this it was obvious the target printed with CS4 had a major problem ... gamut volume was only 347,000 as compared to Epson's profile of 472,000 and the CS1 profile I made of 494,000. Attached is a color think graph comparing the CS1 profile (wireframe) against the CS4 profile.
At this point I cannot find a work around ... I cannot find a way to print an accurate target on matt paper to an 11880 using CS4 and 10.5.6. It may be possible you can only print an accurate target for an 11880 from CS4 if you are using Epson Premium Luster(260) as your media type.
just to make sure there is no user error here I may duplicate this test tomorrow.
[attachment=11233:CS1_CS4.jpg]
-
I then went ahead and read and proflied the two targets, and I compared the results in ColorThink against the 11880 default Enhanced Matt _MK profile From this it was obvious the target printed with CS4 had a major problem ... gamut volume was only 347,000 as compared to Epson's profile of 472,000 and the CS1 profile I made of 494,000. Attached is a color think graph comparing the CS1 profile (wireframe) against the CS4 profile.
This is interesting as I also noticed that the 3d gamuts in the ColorSync Utility were unusually small when I initially used CS4. Now that I have re-profiled the same (matte) papers using CS1 the gamuts are now a similar size to the nearest equivalent canned profiles such as VFA.
Cheers,
Ryan
-
Unclear if this is actually an Adobe issue. Could be Epson or Apple as well ... could be the 3 are just not interacting correctly....
Wayne, this is an ongoing issue, and I appreciate you reporting your experiences here and the work you are putting into it. I brought this up on the colorsync list some weeks back, and someone from Adobe, after some confusing posts from others, simply said they changed things to comply with Apple's requirements to work with changes in the OS, and the ball is now in Epson's court.
The bottom line is that the "color managed" data path has been changed in ways few of us seem to be able to get a handle on, and the "expert" community seems remarkably silent about it, while the rest of us have yet to feel confident of a way to print an unmangled chart, and use the resulting profile with confidence as well.
I wish I had more to offer, but am watching other relevant forums for as much input as possible...
But for now I wonder, how are profiles being made at the highly visible color management shops like Chromix, to name just one?
I don't personally have to deal with this, using a RIP on a PC avoids any OS level, application level, or OEM driver level problems, but I do need to get a handle on it to help others and provide profiles to them when necessary...
The whole thing feels like a mess right now.
Tyler
-
Wayne, this is an ongoing issue, and I appreciate you reporting your experiences here and the work you are putting into it. I brought this up on the colorsync list some weeks back, and someone from Adobe, after some confusing posts from others, simply said they changed things to comply with Apple's requirements to work with changes in the OS, and the ball is now in Epson's court.
The bottom line is that the "color managed" data path has been changed in ways few of us seem to be able to get a handle on, and the "expert" community seems remarkably silent about it, while the rest of us have yet to feel confident of a way to print an unmangled chart, and use the resulting profile with confidence as well.
I wish I had more to offer, but am watching other relevant forums for as much input as possible...
But for now I wonder, how are profiles being made at the highly visible color management shops like Chromix, to name just one?
I don't personally have to deal with this, using a RIP on a PC avoids any OS level, application level, or OEM driver level problems, but I do need to get a handle on it to help others and provide profiles to them when necessary...
The whole thing feels like a mess right now.
Tyler
Thanks for the info. I'm glad Ryan started this thread and was persistent until I understood exactly where the problem was, because I was pretty skeptical. I was having a heck of a time trying to build a profile for Museum Etching for my 7900. Pretty sure this is the reason .I'm curious how many out there are making profiles that aren't even aware of this problem.
I've decided to have one hard drive on the machine where I print and read targets using a clean install system 10.4.11. My i1i0 table is less flakey under this system, and using CS1 (or CS3, testing that now), I can print targets without having to use one of my 2 CS4 copies, saving them for my Laptop and workstation.
BTW, I just tested CS3 and 10.5.5, which prints a target virtually identical to CS1/10.4.11 according the MeasureTools compare mode. function.
-
Wayne, let us know if the resulting profile, from the "correctly" printed chart from previous versions, will then work well when using it from CS4... if you get to that point.
Obviously this could be problematic as well since it was differently handled for the chart. What a can of worms.
I suspect, also, that printing right out of EyeOne Match, or similar, would result in chart measurements like those you just made from Adobe pre V4. I was also informed, but have not tried it, that tagging the chart with GenericRGB may allow it to pass through unaltered. But then we may get Epson driver updates soon that change everything!
T
-
Wayne, let us know if the resulting profile, from the "correctly" printed chart from previous versions, will then work well when using it from CS4... if you get to that point.
As long as photoshop is managing colors, there seems to be no problems with output. The profiles work very well ... output as expected. The real challenge is finding a method that sends your target data to the printer without any color management interfering ... once you find that the profiles from those targets work fine from CS4/10.5.x.
-
Regarding the problem with Epson ABW and CS4, I've been advised to expect driver updates in April/May to address this issue.
-
Regarding the problem with Epson ABW and CS4, I've been advised to expect driver updates in April/May to address this issue.
Thanks for the info on this. Hopefully the update will also fix the profiling issues discussed in this thread.
Cheers,
Ryan
-
Thanks for the info on this. Hopefully the update will also fix the profiling issues discussed in this thread.
I expect it will, but obviously it will need to be tested.
-
I have now tried printing my test chart with the UK version 6.55 ESP 7900 driver on an Intel Mac with MacOSX 10.5 Leopard and this also gives an 'incorrect' result.
-
I have now also tried printing my test chart with the UK version 6.55 ESP 7900 driver with MacOSX 10.6 Snow Leopard and CS4 and this also gives an 'incorrect' result.
Two Snow Leopard work-arounds seem to work as follows.
Workflow 1.
- Assign sRGB to the target.
- Select printer manages color in the Photoshop dialogue.
- Select 'No color management' in the print driver dialogue.
Workflow 2.
- Open the Colorsync Utility, click on the Devices tab and expand the entry for your 7900/9900.
- The default "Mode" profile should be sRGB if you are using the version 6.55 Epson driver.
- Change the "Current Profile" from sRGB to the profile for the media type you will be using to print the chart.
- Open the chart with CS4 but do not assign a profile.
- Choose "No Color Management" in the Photoshop dialogue.
- Choose "Color Settings: Off" in the print driver dialogue.
- Make sure that the correct media type is chosen so as to match the changes made in CU.
Photoshop CS1 can't see any printer profiles under Snow Leopard whether soft profiling or when printing.
However CS1 still gives a correct target print using the traditional method and Epson driver 6.55.
This is the method I will continue to use for now.
-
This is a very informative thread, to say the least. I have many printers in my studio, from Epson, HP and Canon. And the only printer that I can get to correctly send an untagged, no color managed target for profiling is my Canon IPF8100, which uses a Photoshop plugin (bypasses ColorSync I assume). I have spent more hours than I care to know, and in the end I threw my arms in the air and bought a new Macbook Pro 13", and am now using my old Macbook Pro 15" with OSX 10.4 Tiger just to avoid all of the hassles.
Printers that don't work with CS4 / 10.5.7:
Epson 3800
Canon iPF5100 with regular driver
Canon iPF8100 with regular driver
HP Z3200
Canon Pro 9000II
Canon Pro 9500II
The issue is larger than just Epson, it seems, and even though I needed a new computer anyway, most people are not going to be in the same position as I. I have no confidence in any output from OSX Leopard on any of my machines, whether it is a target for the creation of a profile, a 'real' color print or Advanced Black and White prints.
The problem seems to be hit or miss, based on conversations with friends that are all on OSX 10.5 Leopard. All are using Photoshop CS4. Different printers from different manufacturers.
I am going to upgrade my production desktop machine to Snow Leopard this week, and I will do some more tests. If early results don't yield anything different than before, I am going to keep printing from an OSX 10.4 Tiger machine.
-
At last, this bug appears to be fixed by the Apple MacOSX 10.6.3 update released on the 29th March 2010.
That is 15 months from first discovery. Nice one Apple.
However I strongly advise against taking my results as gospel.
I believe everyone should thoroughly test their own setup including Photoshop CS5 when it becomes available.