Luminous Landscape Forum
Raw & Post Processing, Printing => Colour Management => Topic started by: vjbelle on September 01, 2009, 01:31:30 pm
-
Has anyone tried to profile papers on a 9900/7900 using Snow Leopard? All of my profiles, so far, have been generated with Leopard (10.5.X) using the work flow of having an untagged target and assigning Generic RGB as the profile and telling the driver to let the printer manage color with the printer driver set to 'No color management'. Doing this under SL generates a print that is obviously different then the print generated under Leopard.
However there is another Web article that has different driver settings which have produced identical results to Leopard. The targets still remain untagged RGB but Photoshop manages color. The profile used is 'Generic RGB' and the printer color management is turned off. The printed targets now match what was generated under Leopard.
Any thoughts?
-
Has anyone tried to profile papers on a 9900/7900 using Snow Leopard? All of my profiles, so far, have been generated with Leopard (10.5.X) using the work flow of having an untagged target and assigning Generic RGB as the profile and telling the driver to let the printer manage color with the printer driver set to 'No color management'. Doing this under SL generates a print that is obviously different then the print generated under Leopard.
There are two workflows you could try assuming Photoshop CS4, MacOSX 10.6 and the latest Epson version 6.55 driver:
Workflow 1.
Use your existing workflow but instead of assigning Generic RGB, assign sRGB.
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.
I have tried both methods and they seem to be identical to many previous tests I have tried and documented below.
http://luminous-landscape.com/forum/index....showtopic=31410 (http://luminous-landscape.com/forum/index.php?showtopic=31410)
-
Sounds like you may have the dreaded ColorSync bug that has been talked about quite a bit. Check out Wayne's article online:
http://bit.ly/i3LP0 (http://bit.ly/i3LP0)
Andy
-
All of my profiles, so far, have been generated with Leopard (10.5.X) using the work flow of having an untagged target and assigning Generic RGB as the profile and telling the driver to let the printer manage color with the printer driver set to 'No color management'. Doing this under SL generates a print that is obviously different then the print generated under Leopard.
While I haven't tested building profiles yet, I suspect you should try leaving the target untagged and not assigning Generic RGB and see what happens. I'm also unsure why you're letting Printer Manage Color instead of Application Manage Color.
The print path and display TRC gamma have been updated in 10.6.
-
Sounds like you may have the dreaded ColorSync bug that has been talked about quite a bit. Check out Wayne's article online:
http://bit.ly/i3LP0 (http://bit.ly/i3LP0)
Wayne's article draws on my findings (in the linked thread above) which were in turn inspired by the the ABW discussions on the Adobe Photoshop CS4 forum and the Colorsync Utility workaround discussions on the Adobe Lightroom forum that go back to December 2007. However, Wayne's article is now slightly out of date as the recent Epson driver update to version 6.55 changes some of the behaviour in the Colorsync Utility. Workflow 2 above updates the CU workaround for Epson driver 6.55.
( http://luminous-landscape.com/forum/index....showtopic=31410 (http://luminous-landscape.com/forum/index.php?showtopic=31410) )
While I haven't tested building profiles yet, I suspect you should try leaving the target untagged and not assigning Generic RGB and see what happens.
Unfortunately this is still not the case. I have tested the traditional method for printing charts with Snow Leopard + Photoshop CS4 + Epson driver 6.55 and as with Leopard + Photoshop CS4 + driver 6.39 the results are still not exactly as would be expected unless a Colorsync Utility workaround is first applied.
BTW, assigning sRGB to the target is exactly how Lightroom automatically handles untagged images. If an untagged target is opened and printed with Lightroom set to Printer Manages Colours and the Print Driver Colour Management set to off then the resulting target is printed correctly under Snow Leopard.
My preferred method for printing targets is to continue with Photoshop CS1. It still prints the targets correctly under Snow Leopard and Epson driver version 6.55. (Sadly, my installation of CS1 cannot directly see any printer profiles under Snow Leopard whether soft proofing or printing but it can open and understand CS4 saved custom soft proof psf files which seem to have the relevant printer profile embedded.)
-
There are two workflows you could try assuming Photoshop CS4, MacOSX 10.6 and the latest Epson version 6.55 driver:
Workflow 1.
Use your existing workflow but instead of assigning Generic RGB, assign sRGB.
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.
I have tried both methods and they seem to be identical to many previous tests I have tried and documented below.
http://luminous-landscape.com/forum/index....showtopic=31410 (http://luminous-landscape.com/forum/index.php?showtopic=31410)
I am assuming that your workflow 1. has 'Printer manages Color' in the Photoshop dialogue and 'Color settings off' in the Printer dialogue.
I would try your workflow 2 but I can't seem to find or point Colorsync to the correct paper profile. Epson's native profiles are nowhere to be found on my Mac.
-
Wayne's article draws on my findings (in the linked thread above) which were in turn inspired by the the ABW discussions on the Adobe Photoshop CS4 forum and the Colorsync Utility workaround discussions on the Adobe Lightroom forum that go back to December 2007. However, Wayne's article is now slightly out of date as the recent Epson driver update to version 6.55 changes some of the behaviour in the Colorsync Utility. Workflow 2 above updates the CU workaround for Epson driver 6.55.
I'm not sure which update has changed the behavior mentioned with ColorSync, but you are right. My test machine with Snow Leopard happens to be the one that drives my 7900. It seems likely this behavior change is due to the new Epson Driver, but I can't rule out Snow Leopard affecting it as well.
Anyway, Epson Premium Luster is no longer the "default" ColorSync profile for the 7900/9900 printer. In my ColorSync Utility sRGB is now a profile listed with the Epson 7900 printer .. not genericRGB but sRGB ... and it is now the Default ColorSync Profile. Whether resetting this default to point to a different profile (as described in the article) will have the same results as it did when Epson Premium Luster was the default is uncertain. I also have found a couple of other anomalies in my Library folder, including a new folder called Inkjet 2. This may be due to a problem with the installation of the new driver.
As I mentioned I haven't had time to make sense of it yet, sounds like Ryan is further along in trying to figure it out. My 7900 printer operated just fine after installing Snow Leopard. I didn't run into the problems mentioned by Mark Dubovoy, probably because I had already installed Rosetta because of a non photo related application.
It is encouraging that someone was able to get get Epson and Apple to listen, and because of this future updates for the driver will be available through Apple and the OS (thankyou Mark) ... now if they would both just figure out how to resolve these problems with disabling color management so we wouldn't have to jump through these hoops to use ABW or print unmanaged targets.
-
In my ColorSync Utility sRGB is now a profile listed with the Epson 7900 printer .. not genericRGB but sRGB ... and it is now the Default ColorSync Profile.
Considering Generic RGB has a 1.8 TRC Gamma while sRGB has a 2.2 which now matches the OS, it makes sense to me that the driver would associate sRGB in the CS utility.
-
The changes I tested between 10.5 and 10.6 were associated with Printer Manages color, not Application Manages color.
I proved that the new behavior in 10.6 is the correct behavior! By using an sRGB source file, setting Photoshop to "No Color Management" and the Epson Driver to "Standard (sRGB)" I found that 10.5.7 and 10.6 produced the same result after measuring output and comparing the deltaE in ColorThink. As we know using this setting, the RGB data is passed directly to the driver unchanged we were able to use this as a reference. Comparing this to the data gathered using "Printer Manages Color" in 10.5.7 and 10.6 it proved that the OS was not interpreting the sRGB hand-off from PS correctly. In 10.6 the result does match when we use this setting. So it looks like Apple fixed a bug here.
-
The changes I tested between 10.5 and 10.6 were associated with Printer Manages color, not Application Manages color.
I proved that the new behavior in 10.6 is the correct behavior! By using an sRGB source file, setting Photoshop to "No Color Management" and the Epson Driver to "Standard (sRGB)" I found that 10.5.7 and 10.6 produced the same result after measuring output and comparing the deltaE in ColorThink. As we know using this setting, the RGB data is passed directly to the driver unchanged we were able to use this as a reference. Comparing this to the data gathered using "Printer Manages Color" in 10.5.7 and 10.6 it proved that the OS was not interpreting the sRGB hand-off from PS correctly. In 10.6 the result does match when we use this setting. So it looks like Apple fixed a bug here.
So what happens if you want neither to manage color? Is this being handled correctly in Snow Leopard/CS4 now?
-
I am assuming that your workflow 1. has 'Printer manages Color' in the Photoshop dialogue and 'Color settings off' in the Printer dialogue.
That is correct.
I would try your workflow 2 but I can't seem to find or point Colorsync to the correct paper profile. Epson's native profiles are nowhere to be found on my Mac.
They can be found using the following steps:
1. Locate the the file below:
/Library/Printers/EPSON/InkjetPrinter/ICCProfiles/Pro9900_7900.profiles
2. "Control-click" on the file and then select "Show Package Contents"
3. In the new window, navigate to the location below:
Contents/Resources/
4. Here you should find the profiles for your Epson 7900/9900.
5. To make life a bit easier, I made a copy of my profiles to the location below:
/Library/ColorSync/Profiles
-
That is correct.
They can be found using the following steps:
1. Locate the the file below:
/Library/Printers/EPSON/InkjetPrinter/ICCProfiles/Pro9900_7900.profiles
2. "Control-click" on the file and then select "Show Package Contents"
3. In the new window, navigate to the location below:
Contents/Resources/
4. Here you should find the profiles for your Epson 7900/9900.
5. To make life a bit easier, I made a copy of my profiles to the location below:
/Library/ColorSync/Profiles
Thanks an awful lot for your help!
-
If printing from Photoshop, why would you ever want to have the printer manage colors?
-
If printing from Photoshop, why would you ever want to have the printer manage colors?
Usually this is a workaround for ABW printing, or to print unmanaged targets for profiling. this has been a challenge for those printing from Leopard and CS4 for some time.
-
So what you are saying is that the problem isn't with the normal operation of just sending an RGB color image to the printer, but rather just the printing of targets for profiling?
If this is the case, I need to go back and check my methodology, as my problem could only be the resulting profiles, but not my printing workflow. In other words, I could print my targets from OSX Tiger, create the profile, and then use the profiles as normal from CS4 / Leopard or Snow Leopard.
-
So what you are saying is that the problem isn't with the normal operation of just sending an RGB color image to the printer, but rather just the printing of targets for profiling?
That seems to be the case. Last week I printed a tagged image in the normal way with Snow Leopard, Photoshop CS4 and Epson 7900 driver version 6.55. The results were fine. I have only had problems printing untagged test charts with the combination of Leopard or Snow Leopard and Photoshop CS4.
If this is the case, I need to go back and check my methodology, as my problem could only be the resulting profiles, but not my printing workflow. In other words, I could print my targets from OSX Tiger, create the profile, and then use the profiles as normal from CS4 / Leopard or Snow Leopard.
I haven't fully tested Tiger and Photoshop CS4 so I don't know if this combination is ok.
However, the first page of the thread below reveals the combinations that have been ok for me and I have been using these as bench marks.
http://luminous-landscape.com/forum/index....showtopic=31410 (http://luminous-landscape.com/forum/index.php?showtopic=31410)
-
That seems to be the case. Last week I printed a tagged image in the normal way with Snow Leopard, Photoshop CS4 and Epson 7900 driver version 6.55. The results were fine. I have only had problems printing untagged test charts with the combination of Leopard or Snow Leopard and Photoshop CS4.
Its certainly not the case printing untagged targets to an Epson 2880 using CS4. Printed a TC918 target for an iSis in both 10.5.8 and 10.6, the report from ColorThink indicates the difference here are tiny and attributed to measurement variance for such a Spectrophotometer:
Overall - (924 colors)
--------------------------------------------------
Average dE: 0.16
Max dE: 0.92
Min dE: 0.01
StdDev dE: 0.13
Best 90% - (831 colors)
--------------------------------------------------
Average dE: 0.13
Max dE: 0.30
Min dE: 0.01
StdDev dE: 0.07
Worst 10% - (93 colors)
--------------------------------------------------
Average dE: 0.47
Max dE: 0.92
Min dE: 0.30
StdDev dE: 0.15
--------------------------------------------------
-
Its certainly not the case printing untagged targets to an Epson 2880 using CS4. Printed a TC918 target for an iSis in both 10.5.8 and 10.6, the report from ColorThink indicates the difference here are tiny and attributed to measurement variance for such a Spectrophotometer:
What driver and printer settings were used?
-
That seems to be the case. Last week I printed a tagged image in the normal way with Snow Leopard, Photoshop CS4 and Epson 7900 driver version 6.55. The results were fine. I have only had problems printing untagged test charts with the combination of Leopard or Snow Leopard and Photoshop CS4.
To clarify: You assigned Generic RGB for Leopard and sRGB for SL and in both instances the settings were "Printer manages color" in the Photoshop dialogue and "No color management" in the Printer dialogue and the results were identical?
Is this correct?
-
What driver and printer settings were used?
6.20. No Color Adjustment for output. IF for whatever odd reason you guys are using Color Controls, well it makes sense there'd be a difference.
Just did the same test with a 4800 printer, using the same target and Spectrophotometer with the same settings above:
Overall - (924 colors)
--------------------------------------------------
Average dE: 0.35
Max dE: 1.22
Min dE: 0.01
StdDev dE: 0.23
Best 90% - (831 colors)
--------------------------------------------------
Average dE: 0.29
Max dE: 0.66
Min dE: 0.01
StdDev dE: 0.15
Worst 10% - (93 colors)
--------------------------------------------------
Average dE: 0.87
Max dE: 1.22
Min dE: 0.66
StdDev dE: 0.15
--------------------------------------------------
The numbers for max deltaE (a few blue patches) above will be lower if I gave the target, which just came out of the printer an hour+ to dry down. But bottom line is, with a 2880 (80 series) and 4800 (00 series), both OSs treat untagged data the same way.
-
The numbers for max deltaE (a few blue patches) above will be lower if I gave the target, which just came out of the printer an hour+ to dry down. But bottom line is, with a 2880 (80 series) and 4800 (00 series), both OSs treat untagged data the same way.
What paper and media setting did you use for this test?
-
To clarify: You assigned Generic RGB for Leopard and sRGB for SL and in both instances the settings were "Printer manages color" in the Photoshop dialogue and "No color management" in the Printer dialogue and the results were identical?
I have never used the Generic RGB and Leopard method and I won't be using the sRGB method with Snow Leopard either. I tried the sRGB and Snow Leopard method just out of curiosity and it visually matched my previous bench marks.
-
What paper and media setting did you use for this test?
Premium Luster and the associated luster media setting.
-
Premium Luster and the associated luster media setting.
Could you now 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, with the original Epson ESP7900 driver version 6.38 and Leopard, the default 'Mode' for my ESP7900 was locked to 'Premium Luster Profile Paper 260' and this was set to the factory default profile (in other words, both mode and profile were set to 'Premium Luster Profile Paper 260' profile).
When printing test charts under Leopard with Color Management disabled in Photoshop CS4 and disabled in the printer dialogue box, both Wayne and I have noted that glossy/semi gloss papers did not have a noticeable variation as you have also now reported above. But when Wayne and I tried a matte paper the variation was so dramatic it could be observed by eye when compared to an earlier version of photoshop or even Photoshop CS4 on Windows.
So could you possibly re-run your tests with a matte paper please? For my tests, I have been using Epson Enhanced Matte paper with both the VFA and WPRW media settings. (I think Wayne tried the Enhanced Matte media setting). Please print the test chart in the usual traditional method, colorsync off in Photoshop CS4, and off in the print dialogue box. With the Colorsync Utility most probably locked to its default Premium Luster profile, I would be very interested to know if you then start to see a variation when compared to a test chart produced with say Photoshop CS1 under Leopard and the same print driver. It would be great if you could then run the test a second time but with the ColorSync Utility default "Mode" for your device changed so that it matches the media setting you used for your matter paper. When Wayne and I made this Colorsync Utility change for our matte paper test chart, the result then matched previous versions of Photoshop.
Cheers, Ryan
-
Could you now 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.
For the 2880, the blue dot is set to Radiant White paper and for the 4800, USFAP_MK.icc. Both different of course and both having nothing to do with Luster. And I've never messed with or needed to mess with this area of the innards of the utility.
-
For the 2880, the blue dot is set to Radiant White paper and for the 4800, USFAP_MK.icc. Both different of course and both having nothing to do with Luster. And I've never messed with or needed to mess with this area of the innards of the utility.
With my 7900, when printing untagged charts with CM off and the earlier driver, the CU default was always locked to 'Premium Luster Profile Paper 260' and with the later driver version 6.55, the CU default is always locked to sRGB.
So perhaps the particular untagged chart behaviour that Wayne, I have been having is unique to the 7900/9900.
-
With my 7900, when printing untagged charts with CM off and the earlier driver, the CU default was always locked to 'Premium Luster Profile Paper 260' and with the later driver version 6.55, the CU default is always locked to sRGB.
So perhaps the particular untagged chart behaviour that Wayne, I have been having is unique to the 7900/9900.
I don't think it matters.
Best approach I think is to print out a chart in both OSs and measure em, then use something like MeasureTool or ColorThink to plot the deltaE. They should be identical. I don't know of anything Apple did in SN to affect the print path other than fix a bug using Printer manages color and alter the default TRC gamma assumption from the silly old 1.8 to 2.2.
-
BTW, assigning sRGB to the target is exactly how Lightroom automatically handles untagged images. If an untagged target is opened and printed with Lightroom set to Printer Manages Colours and the Print Driver Colour Management set to off then the resulting target is printed correctly under Snow Leopard.
Just noticed this and wanted to point out that while that is correct, you can't use LR to print such targets for profiling as there's no No Color Management option in that app. The RGB numbers need to go directly to the driver which is what NCM should be doing in Photoshop. You could assign sRGB or any working space to the target and at least in Photoshop, with No Color Management, it should print the same as if you used an untagged document. Now Printer Manages Color is a different story. It would be affected by the tag (at least when PMC actually works correctly which is hit or miss depending on the OS, driver and application).
-
Best approach I think is to print out a chart in both OSs and measure em, then use something like MeasureTool or ColorThink to plot the deltaE. They should be identical. I don't know of anything Apple did in SN to affect the print path other than fix a bug using Printer manages color and alter the default TRC gamma assumption from the silly old 1.8 to 2.2.
I don't need to measure the charts because they are so obviously different - but only with matte papers. The differences occur when printing Epson 7900 test charts with colour management off using either Leopard or Snow Leopard with Photoshop CS4 and the CU 7900 Device Mode at its sRGB default.
I have also just tried the USA version 6.55 7900 driver but unfortunately it gives the same incorrect result as the UK 6.55 driver. The 6.55 drivers are supposed to fix 7900 ABW problems but the driver hasn't fixed the untagged chart problem on the 7900.
Leopard or Snow Leopard with Photoshop CS1 is ok on the 7900. Windows XP and Photoshop CS2 or Photoshop CS4 is ok on the 7900.
-
Just noticed this and wanted to point out that while that is correct, you can't use LR to print such targets for profiling as there's no No Color Management option in that app. The RGB numbers need to go directly to the driver which is what NCM should be doing in Photoshop. You could assign sRGB or any working space to the target and at least in Photoshop, with No Color Management, it should print the same as if you used an untagged document. Now Printer Manages Color is a different story. It would be affected by the tag (at least when PMC actually works correctly which is hit or miss depending on the OS, driver and application).
Yes, I wouldn't have thought it was a good idea to do this but nonetheless my 7900 test chart prints correctly under Snow Leopard + Lightroom + the Epson 6.55 driver. That is with Lightroom set to Printer Manages Colours and the Print Driver Colour Management set to off.
-
Yes, I wouldn't have thought it was a good idea to do this but nonetheless my 7900 test chart prints correctly under Snow Leopard + Lightroom + the Epson 6.55 driver. That is with Lightroom set to Printer Manages Colours and the Print Driver Colour Management set to off.
The problem with this is you end up bypassing ACE, which does good things like black point compensation and choices of rendering intents.
-
The problem with this is you end up bypassing ACE, which does good things like black point compensation and choices of rendering intents.
Yes, I certainly wouldn't rely on or recommend this method for printing test charts. However I am rather curious as to what can be interpreted from the sRGB behaviour of Lightroom and Leopard or Snow Leopard when printing untagged test charts. Indeed, the 7900 version 6.55 Mac print driver creates a new default sRGB Mode under the Colorsync Utility which wasn't the case with the previous 7900 driver...
-
Yes, I wouldn't have thought it was a good idea to do this but nonetheless my 7900 test chart prints correctly under Snow Leopard + Lightroom + the Epson 6.55 driver. That is with Lightroom set to Printer Manages Colours and the Print Driver Colour Management set to off.
But I don't think it is correct because there is a conversion happening to the data sent to the printer using Printer Manages Color. Charts have to be output using a No Color Management path which isn't the case any time in Lightroom or with Printer Manages Color. What software product are you using for the profiles? Something like EyeOne Match will print a TC918 target within its app correctly after which you need to attempt to produce a match using the same driver in Photoshop.
The problem with this is you end up bypassing ACE, which does good things like black point compensation and choices of rendering intents.
ACE shouldn't really be entering the picture here either. There is a pretty convoluted path once the application passes the RGB data to the driver, but I can't see why ACE would be used with a no color management setting sending that RGB data. The app's that send the data directly like EyeOne Match or ColorMunki don't use ACE AFAIK.
-
Andrew, I was just thinking of normal print operations for a final print, not for the printing of targets.
-
I have never used the Generic RGB and Leopard method and I won't be using the sRGB method with Snow Leopard either. I tried the sRGB and Snow Leopard method just out of curiosity and it visually matched my previous bench marks.
What PS working space are you using when you print your profiles? Even if I don't color manage an untagged chart I find that the working space controls the output (unless I assign a profile).
-
What PS working space are you using when you print your profiles? Even if I don't color manage an untagged chart I find that the working space controls the output (unless I assign a profile).
Before I upgraded to Photoshop CS4 on the Mac I had been happily producing profiles by printing the untagged charts with Color Management in both Photoshop and the printer driver turned off. However, I have found that Leopard or Snow Leopard with Photoshop CS4 and the Epson 7900 don't produce the expected test chart result on matte paper so I don't know what is going on in this case...
-
But I don't think it is correct because there is a conversion happening to the data sent to the printer using Printer Manages Color. Charts have to be output using a No Color Management path which isn't the case any time in Lightroom or with Printer Manages Color. What software product are you using for the profiles? Something like EyeOne Match will print a TC918 target within its app correctly after which you need to attempt to produce a match using the same driver in Photoshop.
I use an Eye One Pro and i1Match to create my profiles using the TC918 target. Back in January I tried printing a matte TC918 target test chart to my 7900 using i1Match and under Leopard with UK driver version 6.39. The chart was correctly printed. Under the same conditions (Leopard, and 7900 driver 6.39) Photoshop CS1 also gave the correct result. However under the same conditions, Photoshop CS4 did not give the correct result.
I might try the i1Match test again using Snow Leopard and 7900 driver 6.55...
-
Andrew, I was just thinking of normal print operations for a final print, not for the printing of targets.
I have had no problems at all with normal print operations using either Leopard or Snow Leopard and using Photoshop CS4 with the Epson SP 7900.
I have only had problems when using Leopard or Snow Leopard and Photoshop CS4 to print untagged test charts to the 7900.
-
I have had no problems at all with normal print operations using either Leopard or Snow Leopard and using Photoshop CS4 with the Epson SP 7900.
I have only had problems when using Leopard or Snow Leopard and Photoshop CS4 to print untagged test charts to the 7900.
Thanks for the clarification, Ryan. I am going to do some testing later on this week to try and nail down where things are going south in my environment. Since I have many printers, hopefully this will be helpful.
-
Before I upgraded to Photoshop CS4 on the Mac I had been happily producing profiles by printing the untagged charts with Color Management in both Photoshop and the printer driver turned off. However, I have found that Leopard or Snow Leopard with Photoshop CS4 and the Epson 7900 don't produce the expected test chart result on matte paper so I don't know what is going on in this case...
I ran two test prints of charts generated in PM 5. These are the same untagged charts I have been using for a couple of years for my printers.
I changed the default paper profile in CU to premium luster photo paper 260 which is the same paper I am printing on and use for media settings in the printer driver (6.55). Driver is set to color management off and printer is set to no color management.
I then changed the CU back to the default paper profile of sRGB. Assigned sRGB profile to the image and set the dialogue to 'Printer manages color'. Set the printer dialogue of 'No color management'.
The differences between these two prints are remarkable with the first print being very saturated and in fact mimicking very closely my monitor.
The second print (with the assigned sRGB profile) is very flat (as I would normally expect) and also mimics my monitor. In both of those instances my working space was set to ProPhoto.
Then I printed a third test print with my PS working space set to sRGB and no color management anywhere in the driver. That print matched my second print (assigned sRGB Profile) and also mimicked my profiled monitor.
So I have determined that, for sure, if I set PS to no color management and the printer to no color management the working space determines the final printed output. That then leads me to believe that a profile must be assigned to an untagged image and that profile must/should match the default profile in CU or that the PS working space must/should match the CU default profile. Is any of this making sense?
-
I use an Eye One Pro and i1Match to create my profiles using the TC918 target. Back in January I tried printing a matte TC918 target test chart to my 7900 using i1Match and under Leopard with UK driver version 6.39. The chart was correctly printed. Under the same conditions (Leopard, and 7900 driver 6.39) Photoshop CS1 also gave the correct result. However under the same conditions, Photoshop CS4 did not give the correct result.
I have just tried printing the TC918 chart directly from i1Match again but this time using Snow Leopard and 7900 driver version 6.55. The print driver Color Matching option was left on its default setting of Epson Color Controls.
The printed result was identical to Windows XP with Photoshop CS2, WIndows XP with Photoshop CS4, Leopard with Photoshop CS1 and Snow Leopard with Photoshop CS1.
However the result is different to the test chart print from Leopard or Snow Leoapard + Photoshop CS4 + 7900 driver 6.39/6.55.
-
Then I printed a third test print with my PS working space set to sRGB and no color management anywhere in the driver. That print matched my second print (assigned sRGB Profile) and also mimicked my profiled monitor.
So I have determined that, for sure, if I set PS to no color management and the printer to no color management the working space determines the final printed output. That then leads me to believe that a profile must be assigned to an untagged image and that profile must/should match the default profile in CU or that the PS working space must/should match the CU default profile. Is any of this making sense?
As this is something I haven't tried, I have just run two tests with some changes to my default working space (under Edit, Color Settings...)
1. Edit, Color Settings.... Color Management changed to Off.
Opened TC9.18 test chart and printed on matte paper with Photoshop CS4 CM off and print driver version 6.55 CM off. The result was still noticeably wrong and no different to the print that I had produced with my default working space set to ProPhoto.
2. Edit, Color Settings.... default RGB working space set to sRGB.
Opened TC9.18 test chart and printed on matte paper with Photoshop CS4 CM off and print driver version 6.55 CM off. The result was still noticeably wrong and no different to the print that I had produced with my default working space set to ProPhoto.
Note that under Photoshop Edit, Color Settings... I always set Color Management Policies to Preserve Embedded Profiles for RGB, CMYK and Grey. I also select Profile Mismatches: to yes for "Ask When Opening" and "Ask When Pasting"; and I also select Missing Profiles: to yes for "Ask When Opening".
It might be helpful if a common methodology is used to make the tests. Indeed it was only when Wayne precisely followed my test methods back in January that he finally saw what I have been seeing.
Leopard / Snow Leopard
Mac Photoshop CS4
A different version of Photoshop on either Windows or Mac or i1Match for comparison.
Epson 7900/9900
Test chart TC9.18 (a small section of it is enough to see the error)
Matte paper (Wayne and I have been using Epson Enhanced Matte)
Open target with Mac Photoshop CS4 with CM off.
Print with Photoshop CS4 CM off and print driver CM off.
Repeat with a different version of Photoshop or print target with i1Match.
Compare...
Myself, Wayne and others concluded that a version other than Photoshop CS4 is required to correctly print targets under Leopard and 7900 driver 6.39. I am close to reaching the same conclusion with Snow Leopard and 7900 driver 6.55.
-
As this is something I haven't tried, I have just run two tests with some changes to my default working space (under Edit, Color Settings...)
Not surprising considering the settings here have zero role on output (well other than CMM which is always fixed to ACE in Print).
Print with Photoshop CS4 CM and driver CM off.
But the printing should be with CS4 using NCM (no color management).
-
Not surprising considering the settings here have zero role on output (well other than CMM which is always fixed to ACE in Print).
But the printing should be with CS4 using NCM (no color management).
Indeed, what I meant to say was "Print with Photoshop CS4 CM off and driver CM off.
I have corrected my mis-phrasing in the offending post.
-
As this is something I haven't tried, I have just run two tests with some changes to my default working space (under Edit, Color Settings...)
1. Edit, Color Settings.... Color Management changed to Off.
Opened TC9.18 test chart and printed on matte paper with Photoshop CS4 CM off and print driver version 6.55 CM off. The result was still noticeably wrong and no different to the print that I had produced with my default working space set to ProPhoto.
2. Edit, Color Settings.... default RGB working space set to sRGB.
Opened TC9.18 test chart and printed on matte paper with Photoshop CS4 CM off and print driver version 6.55 CM off. The result was still noticeably wrong and no different to the print that I had produced with my default working space set to ProPhoto.
Note that under Photoshop Edit, Color Settings... I always set Color Management Policies to Preserve Embedded Profiles for RGB, CMYK and Grey. I also select Profile Mismatches: to yes for "Ask When Opening" and "Ask When Pasting"; and I also select Missing Profiles: to yes for "Ask When Opening".
It might be helpful if a common methodology is used to make the tests. Indeed it was only when Wayne precisely followed my test methods back in January that he finally saw what I have been seeing.
Assigning a profile and/or changing the work space to the same profile are one in the same. My screen chart changes instantly and my color patches measure the same when using Apple Digital Color Meter.
If an untagged image is brought into PS/CS4 under SL the working space dictates color policies unless a profile is assigned to that image. So if I want to use CS4 and SL which profile should I assign? There should be a method in PM 5 for measuring the patches which I should be able to match in PS. I will find out.
-
Assigning a profile and/or changing the work space to the same profile are one in the same. My screen chart changes instantly and my color patches measure the same when using Apple Digital Color Meter.
If an untagged image is brought into PS/CS4 under SL the working space dictates color policies unless a profile is assigned to that image. So if I want to use CS4 and SL which profile should I assign? There should be a method in PM 5 for measuring the patches which I should be able to match in PS. I will find out.
Altering the assignment doesn't change the numbers a lick. They just have a different definition so Photoshop now provides a new preview based on this. Again, the document can be tagged or not, it doesn't matter. The proper way to be doing all this chart printing is to simply send the existing reference values (in RGB but also known in Lab which is the important set of values here), to the device. The numbers should not be altered in any way, hence Printer Color Management, printing in Lightroom or any app that doesn't have a No Color Management option is out. This is WHY you can then use the profile in Photoshop or any ICC aware application like LR. The key is having a no color management path in the app to send the data to the driver. With the exception of the applications that ship with the targets (like EyeOne Match), the only applciation I know of that can do this is Photoshop using No Color Management.
-
Again, the document can be tagged or not, it doesn't matter. With the exception of the applications that ship with the targets (like EyeOne Match), the only applciation I know of that can do this is Photoshop using No Color Management.
I don't pretend to know anywhere near as much as you regarding all of this..... just trying to make some sense of this. If I have a tagged image in PS then PS handles that image according to the tag. If I have an untagged image in PS then it is assigned either the working space profile or a profile of my choosing with edit color profiles. If I print a tagged image with color settings off in the driver then what is passed to the driver is the tagged image profile. If I print an untagged image with color settings off in the driver then the driver is sent the image/chart with my working space profile dictating color. As I see it there is no way to completely turn off color management. I am either in a working space which dictates color or I have assigned a profile which dictates color. The resulting prints are identical as long as the profile (assigned or working space) are the same. How do I get to a completely vanilla color space?
-
I don't pretend to know anywhere near as much as you regarding all of this..... just trying to make some sense of this. If I have a tagged image in PS then PS handles that image according to the tag.
As it should when you use color management. The idea here is not to use any color management with the NO Color Management option on Print to send the RGB values to the driver to output a chart to build profiles.
The settings in the driver happen after you click Print. That too should be set to No Color Management (No Color Adjustment) in the Epson driver but by the time that happens, Photoshop has left the scene of the crime and it must be sending the Raw RGB values to the driver from the No Color Management setting. You have to have both set.
I am either in a working space which dictates color or I have assigned a profile which dictates color. The resulting prints are identical as long as the profile (assigned or working space) are the same. How do I get to a completely vanilla color space?
The document is just a pile of RGB values, the tag only gives them a scale and is used when color management is invoked to define a source color space for a conversion. We want no such conversions when printing a target for building a profile. So it doesn't matter what the tag is or even if there is a tag because No Color Management means just that. There's zero question in my mind that when you set Photoshop to No Color Management, you get just that. Its no different in 10.5 than 10.6, I verified that. Doesn't matter if there's matt paper in the printer or not.
-
Its no different in 10.5 than 10.6, I verified that. Doesn't matter if there's matt paper in the printer or not.
Yes I agree that there is no difference between 10.5 and 10.6. Neither works correctly with Photoshop CS4 and the Epson 7900 when printing untagged profile charts onto matte paper.
Photoshop CS1 on 10.5/10.6; Photoshop CS2 on XP; and even Photoshop CS4 on XP produce a different result to Photoshop CS4 on 10.5/10.6 on matte paper when printing with CM off in Photoshop and CM off in the driver to an Epson 7900 regardless of driver version.
-
The document is just a pile of RGB values, the tag only gives them a scale and is used when color management is invoked to define a source color space for a conversion. We want no such conversions when printing a target for building a profile. So it doesn't matter what the tag is or even if there is a tag because No Color Management means just that. There's zero question in my mind that when you set Photoshop to No Color Management, you get just that. Its no different in 10.5 than 10.6, I verified that. Doesn't matter if there's matt paper in the printer or not.
I am in complete agreement! There are two dialogs that must be addressed. No color management in the PS dialog and No color management in the printer dialog. However, my working space has now taken over. My image is untagged. No color management in PS and printer driver. My working space can be set to Prophoto and that will produce a print that is completely different than if my working space is sRGB or something else. In fact they will all vary from each other. Why don't you try it? All color management turned off (you can't turn if off completely in the working space) use an untagged image, Prophoto for the working space and then sRGB for your working space. If your system is like mine (CS4/MAC_SL/9900/Latest driver) the prints will be very different. What's correct? That's all I want to know. What am I missing?
Edit: Added SL for correct operating system.
-
When printing test charts under Leopard with Color Management disabled in Photoshop CS4 and disabled in the printer dialogue box, both Wayne and I have noted that glossy/semi gloss papers did not have a noticeable variation as you have also now reported above. But when Wayne and I tried a matte paper the variation was so dramatic it could be observed by eye when compared to an earlier version of photoshop or even Photoshop CS4 on Windows.
Correct. When printing targets to Epson Premium Luster on both a 3800 and 11880 with Leopard and CS4, targets visually looked very much like targets printed from a different Mac running CS3 and Tiger(10.4). However if printing targets using MK ink to matt papers, there were obvious visual differences in several patches. Most colors looked very close, but in some the difference was dramatic ... no need to measure. Printing from CS2 and Tiger, as well Windows XP all produced visually identical targets to the CS3/10.4 targets. All targets were printed with CS4 set to no color management, and color management turned off in the printer driver.
Note that while there were obvious differences, the targets visually "appeared" fine ... you wouldn't even realize anything was wrong unless you had a target from a different combination to compare to side by side.
My 7900 exhibited the same behavior when creating a profile for Breathing Color Canvas, which uses matt ink. I forgot about this issue and the resulting profile was very odd looking. After reprinting the targets changing the default paper type in the CS utility, the results looked much more normal. I didn't have a machine with CS3/10.4 to try it on, so I'm still not completely certain even these targets were reliable. This was still using the previous Epson driver (6.11 I beileve) and Leopard.
I have since switched to Eric Chan's recommendation in this thread (http://luminous-landscape.com/forum/index.php?showtopic=34806&hl=target) for printing targets
I have not tried printing targets since Snow Leopard was released to see how the behavior has changed.
-
I am in complete agreement! There are two dialogs that must be addressed. No color management in the PS dialog and No color management in the printer dialog. However, my working space has now taken over. My image is untagged. No color management in PS and printer driver. My working space can be set to Prophoto and that will produce a print that is completely different than if my working space is sRGB or something else. In fact they will all vary from each other. Why don't you try it? All color management turned off (you can't turn if off completely in the working space) use an untagged image, Prophoto for the working space and then sRGB for your working space. If your system is like mine (CS4/MAC/9900/Latest driver) the prints will be very different. What's correct? That's all I want to know. What am I missing?
Embedded image profiles and working spaces are pretty much the same thing. However, Photoshop policies determine what photoshop does to a file when you open a file that is either untagged or has a different profile to your default working space as specified in Color Settings. You can even turn off color management completely under Color Settings... as I stated my previous post.
So why not check your color management policies as follows?
Under Photoshop Edit, Color Settings... set Color Management Policies to Preserve Embedded Profiles for RGB, CMYK and Grey. Also select Profile Mismatches: to yes for "Ask When Opening" and yes "Ask When Pasting"; and also select Missing Profiles: to yes for "Ask When Opening". If you set your policies this way then whenever you open a file, you will always be in charge of whether to leave the file as it is or to assign a different working space.
The following page may also be useful:
http://www.earthboundlight.com/phototips/p...king-space.html (http://www.earthboundlight.com/phototips/photoshop-working-space.html)
-
I have since switched to Eric Chan's recommendation in this thread (http://luminous-landscape.com/forum/index.php?showtopic=34806&hl=target) for printing targets
I have not tried printing targets since Snow Leopard was released to see how the behavior has changed.
Ah yes, thanks Wayne. Not for the fist time, I had forgotten about that thread but I knew I had seen the method before somewhere! Indeed, this is the basis of Method 1 detailed at the beginning of this thread except Snow Leopard with the Epson 7900 driver 6.55 require the assignment of sRGB instead of Generic RGB.
More recently, Eric made a related and interesting comment on the Apple Colorsync forum:
"Brief summary: The issue is that with OS X, CS4 is using the newer Leopard print APIs (the only print APIs supported under Cocoa), whereas CS3 was using the older print APIs (will not be supported by the OS going forward). The consequence of using the newer print APIs is that the printer drivers must also be compliant with the new print APIs; being compliant with the old print APIs is no longer sufficient.
Moving forward, OS X is taking steps to prevent accidental printing without color management. This makes it unlikely that you can use the standard OS X apps like Preview to print profile targets reliably. Tagging it with Generic RGB will work today, but not tomorrow when Snow Leopard launches (and switches to a default profile of sRGB, not Generic RGB). And tagging the target with sRGB tomorrow may not work for the future (should future versions of OS X switch to another default profile).
BTW, you should avoid printing profile targets from LR. LR does not provide an option to disable color management, which is what you want for printing the targets. (Yes, this is by design.) Yes, there are workarounds, but they will only work in some cases, but not others. Save LR for printing images -- with color management, of course -- once you have your (custom) profile ready."
http://lists.apple.com/archives/colorsync-...g/msg00118.html (http://lists.apple.com/archives/colorsync-users/2009/Aug/msg00118.html)
-
My working space can be set to Prophoto and that will produce a print that is completely different than if my working space is sRGB or something else.
Well under Snow Leopard or Leopard, I can't duplicate that. I've also tagged a target with both sRGB and ProPhoto RGB, sent them both though the Print command with NO Color Management to a 2880, they are both identical (as they should be). The tag plays no role because No Color Management in Photoshop’s Print dialog is so set.
-
I have since switched to Eric Chan's recommendation in this thread (http://luminous-landscape.com/forum/index.php?showtopic=34806&hl=target) for printing targets
I have not tried printing targets since Snow Leopard was released to see how the behavior has changed.
Wayne,
I have also been using Eric's recommendations to print profiles, but, now that I have SL am trying to figure out what is correct. If you assign Generic RGB and then switch to sRGB you will see a visual change on your monitor. You will also see a similar change in the final print. If you don't use Eric's recommendations and completely turn off color management you are then going to be printing profiles that are linked to the profile used in your working space. Try it! Eric assigns a profile and at that point it doesn't matter what the working space is - the assigned profile dictates the final print.
-
Wayne,
I have also been using Eric's recommendations to print profiles, but, now that I have SL am trying to figure out what is correct. If you assign Generic RGB and then switch to sRGB you will see a visual change on your monitor. You will also see a similar change in the final print. If you don't use Eric's recommendations and completely turn off color management you are then going to be printing profiles that are linked to the profile used in your working space. Try it! Eric assigns a profile and at that point it doesn't matter what the working space is - the assigned profile dictates the final print.
Don't assign Generic RGB and then switch to sRGB.
Instead, open the untagged chart and assign sRGB directly.
Otherwise, follow Eric's method (that is method 1 on the second post of this thread).
-
Well under Snow Leopard or Leopard, I can't duplicate that. I've also tagged a target with both sRGB and ProPhoto RGB, sent them both though the Print command with NO Color Management to a 2880, they are both identical (as they should be). The tag plays no role because No Color Management in Photoshop’s Print dialog is so set.
Maybe this is a 9900/7900 issue. I take it that you also must operate under Rosetta for your printer driver. SL was a completely new install for my system. All applications were added one at a time as well as drivers. I try to be very careful but who knows? What I see on my screen (NEC 2690WUXI2/Spectraview calibrated) is what comes out of my printer regardless of whether or not I have color management on or off. Very unusual.....
-
Maybe this is a 9900/7900 issue. I take it that you also must operate under Rosetta for your printer driver. SL was a completely new install for my system. All applications were added one at a time as well as drivers. I try to be very careful but who knows? What I see on my screen (NEC 2690WUXI2/Spectraview calibrated) is what comes out of my printer regardless of whether or not I have color management on or off. Very unusual.....
Maybe. It would be odd that a newer printer like this would suffer the issue. The 2880 and 4800 don't and they represent a pretty decent difference in age in terms of driver releases. I'm still waiting on a 3880 driver for Snow Leopard but I suspect it will behave properly too.
-
Maybe this is a 9900/7900 issue. I take it that you also must operate under Rosetta for your printer driver. SL was a completely new install for my system. All applications were added one at a time as well as drivers. I try to be very careful but who knows? What I see on my screen (NEC 2690WUXI2/Spectraview calibrated) is what comes out of my printer regardless of whether or not I have color management on or off. Very unusual.....
Yes I think it seems reasonable to conclude that this particular issue relates to the 7900/9900.
The Epson 7900 version 6.55 driver is apparently Snow Leopard compatible and only needs Rosetta because the current Vise installer is not Snow Leopard compatible. This will be corrected at some point.
-
Maybe this is a 9900/7900 issue. I take it that you also must operate under Rosetta for your printer driver. SL was a completely new install for my system. All applications were added one at a time as well as drivers. I try to be very careful but who knows? What I see on my screen (NEC 2690WUXI2/Spectraview calibrated) is what comes out of my printer regardless of whether or not I have color management on or off. Very unusual.....
Sort of correct. It turns out it is only the installation application that needs Rosetta, not the actual driver itself....
-
Maybe. It would be odd that a newer printer like this would suffer the issue. The 2880 and 4800 don't and they represent a pretty decent difference in age in terms of driver releases. I'm still waiting on a 3880 driver for Snow Leopard but I suspect it will behave properly too.
Would you be so kind as to confirm......... when you are looking at an untagged chart in PS and your working space is Prophoto and then you switch your working space to sRGB - or vice versa - does you screen image change? Mine does..... dramatically! If so then the working space dictates how an untagged image is displayed in PS. In my environment it also dictates what gets sent to the printer...... my driver (even though it is set to off) passes through the profile used in my working space.
-
Under Photoshop Edit, Color Settings... set Color Management Policies to Preserve Embedded Profiles for RGB, CMYK and Grey. Also select Profile Mismatches: to yes for "Ask When Opening" and yes "Ask When Pasting"; and also select Missing Profiles: to yes for "Ask When Opening". If you set your policies this way then whenever you open a file, you will always be in charge of whether to leave the file as it is or to assign a different working space.
This is the exact Color management policies that I use.
-
Don't assign Generic RGB and then switch to sRGB.
Instead, open the untagged chart and assign sRGB directly.
Otherwise, follow Eric's method (that is method 1 on the second post of this thread).
Yes...... I think this is the correct - assign sRGB to the untagged image and turn off all color management - at least it is how I'm going to generate my profiles.
-
This is the exact Color management policies that I use.
Ok, so if you use method 1, you open the untagged chart and when prompted you assign sRGB. In this case you are choosing not to use the default working space and instead you are using sRGB as your working space. Then follow Eric's method as normal.
In method 2, you open the untagged chart and when prompted you select Leave as is (don't color manage). The default working space is irrelevant and has no effect with the rest of this method for printing charts (as I confirmed with some tests earlier today). Don't make any other changes to the chart or profiles/working space and just print it. In the Photoshop print dialogue, select CM off. In the printer driver, select CM off. The only extra thing you need to do before you print the chart is to make sure that the CU Utility 7900 sRGB Device Mode is changed from sRGB to the appropriate profile so as to match the media setting you will be using when printing the test chart. This method has worked every time for me whether Leopard or Snow Leopard or 7900 driver 6.39 or 6.55.
-
Would you be so kind as to confirm......... when you are looking at an untagged chart in PS and your working space is Prophoto and then you switch your working space to sRGB - or vice versa - does you screen image change? Mine does..... dramatically! If so then the working space dictates how an untagged image is displayed in PS. In my environment it also dictates what gets sent to the printer...... my driver (even though it is set to off) passes through the profile used in my working space.
Yes, I said so and why above. The NUMBERS do not change, only the definition of the numbers, hence the preview changes. R0/G255/B0 is a different appearing color in sRGB versus ProPhoto RGB. When you tell Photoshop what the scale of the numbers belong to, the numbers remain the same but the color changes as this defines a different color in the two color spaces. Its just like the difference between 500 miles and 500 kilometers is different even though the numbers are the same. So Assign profile (or as you're doing, updating the working space in color settings which assigns that profile to untagged documents) changes the color appearance not the color values. Sending the same values to a driver, without color management is, well sending the same values to the driver.
-
Yes...... I think this is the correct - assign sRGB to the untagged image and turn off all color management - at least it is how I'm going to generate my profiles.
Ok, it would be great if you can somehow visually compare the color chart results with a version of Photoshop other than Mac CS4 or print directly from i1Match (or even print a TC918 chart directly from PM5 if that is possible?) I found it very handy to have a known working benchmark to compare results.
-
Ok, it would be great if you can somehow visually compare the color chart results with a version of Photoshop other than Mac CS4 or print directly from i1Match (or even print a TC918 chart directly from PM5 if that is possible?) I found it very handy to have a known working benchmark to compare results.
Observations:
1. Untagged chart image. CU default sRGB for 9900 printer. Working space Prophoto. Image assigned sRGB. No color management anywhere.
2. Untagged chart image. CU default sRGB for 9900 printer. Working space sRGB. Image left untagged. No color management anywhere.
3. Untagged chart image. CU default sRGB for 9900 printer. Working space Prophoto. Image assigned sRGB. Printer manages color. No color management in the printer driver.
4. Untagged chart image. CU default PLPP 260 (my same media setting). Working space Prophoto. Image left untagged. No color management anywhere.
Charts 1 - 3 looked visually the same. Chart 4 was obviously different with much more saturation (this also matched my monitor). Delta readings from my i one pro into PM 5 showed a maximum delta of 2 in the Blues - some readings were as close as delta .1 for the first three instances. Delta readings for number 4 as compared to any of the first three instances were as high as 13 and averaged around 11.
Just observations…… trying to find out what is right.
So far this is 40 inches of 24 inch roll paper but whats paper and ink for anyway?
-
Ok, it would be great if you can somehow visually compare the color chart results with a version of Photoshop other than Mac CS4 or print directly from i1Match (or even print a TC918 chart directly from PM5 if that is possible?) I found it very handy to have a known working benchmark to compare results.
All of my charts are generated from PM5 if this helps. I'm going to assume that anything generated in PM5 is fairly accurate. I can always send you or anyone else one of my generated charts or a standard chart from PM5. I don't think that really matters....... this is all a SL and PS and Epson problem that can be worked around.
-
Observations:
1. Untagged chart image. CU default sRGB for 9900 printer. Working space Prophoto. Image assigned sRGB. No color management anywhere.
2. Untagged chart image. CU default sRGB for 9900 printer. Working space sRGB. Image left untagged. No color management anywhere.
3. Untagged chart image. CU default sRGB for 9900 printer. Working space Prophoto. Image assigned sRGB. Printer manages color. No color management in the printer driver.
4. Untagged chart image. CU default PLPP 260 (my same media setting). Working space Prophoto. Image left untagged. No color management anywhere.
Charts 1 - 3 looked visually the same. Chart 4 was obviously different with much more saturation (this also matched my monitor). Delta readings from my i one pro into PM 5 showed a maximum delta of 2 in the Blues - some readings were as close as delta .1 for the first three instances. Delta readings for number 4 as compared to any of the first three instances were as high as 13 and averaged around 11.
Based on my tests, I would have expected chart 3 (method 1) and 4 (method 2) to be the same irrespective of how close they may or may not be to your monitor. As they are different, it is hard to conclude which is right. I still think you need a known correct chart printed from a different version of Photoshop all else being the same. This was the only way I was able to determine which chart worked correctly under Snow Leopard. In any case I have decided to stick with using Photoshop CS1 for printing charts under Snow Leopard as it still gives the same result as my known correct benchmarks using the traditional settings. If you have access to Windows XP then you could download the 30 day trial of the Windows version of Photoshop CS4. I tested this combination back in January and the untagged chart printed correctly. This could be one way for you to get a bench mark result. However, if you only have access to a Mac with Leopard/Snow Leopard and Photoshop CS4 then you could spend even more time, ink and paper and still not be totally sure if you have the correct result for your 7900/9900.
By the way, to keep my time and costs down I use just one strip of patches from an i1Match TC918 chart (B7 to B17) rather than the whole chart. This way I have been able to get about ten tests on one A4 sheet of paper and this has been sufficient to visually make a comparison. It doesn't matter which patches are used for the strip as long as there is a broad selection of colors, the strip is saved as an untagged tif and the same strip is used every time for testing. If one particular combination of settings produces a test strip that looks exactly right when compared with a benchmark strip, then a full test chart can be printed, visually compared and even measured so as to be doubly sure.
-
Based on my tests, I would have expected chart 3 (method 1) and 4 (method 2) to be the same irrespective of how close they may or may not be to your monitor. As they are different, it is hard to conclude which is right. I still think you need a known correct chart printed from a different version of Photoshop all else being the same. This was the only way I was able to determine which chart worked correctly under Snow Leopard. In any case I have decided to stick with using Photoshop CS1 for printing charts under Snow Leopard as it still gives the same result as my known correct benchmarks using the traditional settings. If you have access to Windows XP then you could download the 30 day trial of the Windows version of Photoshop CS4. I tested this combination back in January and the untagged chart printed correctly. This could be one way for you to get a bench mark result. However, if you only have access to a Mac with Leopard/Snow Leopard and Photoshop CS4 then you could spend even more time, ink and paper and still not be totally sure if you have the correct result for your 7900/9900.
By the way, to keep my time and costs down I use just one strip of patches from an i1Match TC918 chart (B7 to B17) rather than the whole chart. This way I have been able to get about ten tests on one A4 sheet of paper and this has been sufficient to visually make a comparison. It doesn't matter which patches are used for the strip as long as there is a broad selection of colors, the strip is saved as an untagged tif and the same strip is used every time for testing. If one particular combination of settings produces a test strip that looks exactly right when compared with a benchmark strip, then a full test chart can be printed, visually compared and even measured so as to be doubly sure.
There is a minimum length that a 9900 wants to see for correct/any printing. Its near 10 inches so I use that as my minimum. This chart was generated by me in PM5 and has 2380 patches consisting of 5 separate charts and to print it in its entirety takes 33 inches of a 24 inch roll. After all of this trial and error/paper and ink I have started to use the profiles supplied by Epson. I am very surprised to find them to be visually extremely accurate. More so than anything I have generated. This also holds true for the profiles supplied by Hahnemuhle for fine art Baryta. Maybe the days of this mandatory profiling are over.
-
There is a minimum length that a 9900 wants to see for correct/any printing. Its near 10 inches so I use that as my minimum. This chart was generated by me in PM5 and has 2380 patches consisting of 5 separate charts and to print it in its entirety takes 33 inches of a 24 inch roll. After all of this trial and error/paper and ink I have started to use the profiles supplied by Epson. I am very surprised to find them to be visually extremely accurate. More so than anything I have generated. This also holds true for the profiles supplied by Hahnemuhle for fine art Baryta. Maybe the days of this mandatory profiling are over.
Indeed, I quickly gave up trying to produce custom profiles for my Z3200 as I have found that the canned profiles from HP and third parties are so good for this printer.
Some time ago I made a custom 7900 profile for Hahnemuehle PhotoRag Ultrasmooth paper using a Bill Atkinson 5202 chart and i1Match. I have just downloaded the Hahnemuehle profile for this paper from their web site and their profile appeared almost identical to my Logo Colorful profile when I made a casual comparison with the Colorsync Utility. (Perhaps they are using i1Match or PM5). In fact the Hahnemuehle profile has a slightly larger volume in CU than my profile. I may use Photoshop soft proofing to make further comparions with one or two images and if their profile seems noticeably better then I may switch. On the other hand, their profile gives me no choice for Perceptual rendering. I have produced three i1Match 7900 profiles for Hahnemuehle PhotoRag Ultrasmooth paper, Logo Colorful, Logo Chroma and Logo Classic. I normally use Relative Colorimetric rendering on the 7900 so in this case it doesn't make any difference which of my profiles I use. However it is nice to have some choice just in case I decide to use Perceptual rendering for a particular image.
As almost all of my printing on the 7900 is with Hahnemuehle Photorag Ultrasmooth and I am happily using the Z3200 with canned profiles for any any other printing my i1 Pro tends to be used just for monitor calibration these days.