Pages: [1]   Go Down

Author Topic: Printer Profile Question  (Read 3026 times)

nirpat89

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 661
    • Photography by Niranjan Patel
Printer Profile Question
« on: September 11, 2017, 10:05:36 am »

While playing with my new Epson P400 benchmarking different papers under different printing conditions, I came across a question that I am not able to find answer to in usual places (probably it is too basic.) 

Case#1.

Let's say I want to print an AdobeRGB file, TestAdobeRGB.tiff on an Epson Photo Paper Glossy paper.  From Photoshop I go to File > Print.  Pick Photoshop Manages Color and select from the dropdown profile named SC-P400 Series Photo Paper Glossy provided by Epson.  Select Rel Cal and Black Pt Comp as recommended by the profile maker.  Click Print Settings, go to Epson driver.  Select Photo Paper Glossy in Media Settings.  Select Off (No Color Adjustment) as Mode and select appropriate paper size.  Leave everything else in the default mode.  Come out and hit Print.  That is the standard way.

Case #2.

Now if I open the same file TestAdobeRGB.tiff and go Edit > Convert to Profile.  Select RGB with same profile SC-P400 Series Photo Paper Glossy with Rel Cal and Black Pt Comp checked.  OK and save file as TestAdobeRGB_ConvertedtoPrinterProfile.  Take this file and open it in ACPU and print from there under identical printer settings as above.

The question is:  Will the output from Case #1 and Case #2 be identical?  I am trying to understand what exactly does Photoshop do to the file before sending it out to the printer.

Thanks.

:Niranjan.
« Last Edit: September 11, 2017, 12:54:01 pm by nirpat89 »
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Printer Profile Question
« Reply #1 on: September 11, 2017, 10:11:05 am »

They *should* be identical all settings being equal except you converted in PS but used ACPU to print that converted file. Now you cannot pick CMM in ACPU, it's always using ACE so you'd need to keep that set for this to match.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

nirpat89

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 661
    • Photography by Niranjan Patel
Re: Printer Profile Question
« Reply #2 on: September 11, 2017, 10:41:33 am »

They *should* be identical all settings being equal except you converted in PS but used ACPU to print that converted file. Now you cannot pick CMM in ACPU, it's always using ACE so you'd need to keep that set for this to match.

Thanks, Andrew for the quick reply.  I make sure to use ACE.  Learned that from one of your videos, which I have been gorging on lately.  Need to understand more of this color management stuff beyond the rudimentary. 

:Niranjan.
« Last Edit: September 11, 2017, 10:48:53 am by nirpat89 »
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Printer Profile Question
« Reply #3 on: September 11, 2017, 05:42:39 pm »

Thanks, Andrew for the quick reply.  I make sure to use ACE.  Learned that from one of your videos, which I have been gorging on lately.  Need to understand more of this color management stuff beyond the rudimentary. 

:Niranjan.

I can vouch for ACE's accuracy. Microsofts ICM, an alternative, is much less accurate.*  "Accurate" is a term often bandied about without clear definition and is especially problematic when talking about color perception. I use the term here in reference to the ICC (color.org) which defines rather specifically the conversion processes used in matrix based profiles. Microsoft's ICM does a sloppy job of following them. LUT printer profiles use interpolation and  the ICC does not specify the algorithms CMEs should use for those.

Also, while I'm sure Andrew has gone over this multiple times in his various video tutorials, indeed it should produce the same results. For those reading that haven't seen Andrew's stuff (which I highly recommend) here's what happens:

When you convert in Photoshop to the device's RGB space the values are those that the profile, as interpreted by the CME (Color management engine) maps to RGB values the printer uses to create the requested color, pixel by pixel. These RGB values are sent to the printer driver by ACPU without further color management.

When you use Photoshop to manage color, the exact same conversion occurs and the values are sent to the printer driver, also without further color management.

There is one difference with Windows. For unknown reasons ACPU prints a slightly smaller image. When you are printing to the printer's native resolution, 600 PPI for Canon, 720 or possibly 600 for Epson, ACPU will resample upward and duplicate every 20th pixel or so. It uses a nearest neighbor approach which can create tiny jaggies on near vertical or horizontal pictures. And, of course, there is the issue of printing the wrong size. I highly recommend using Photoshop manages color to avoid these issues. ACPU really should just be used for printing profiling targets.
« Last Edit: September 11, 2017, 05:56:30 pm by Doug Gray »
Logged

nirpat89

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 661
    • Photography by Niranjan Patel
Re: Printer Profile Question
« Reply #4 on: September 12, 2017, 08:07:26 am »

There is one difference with Windows. For unknown reasons ACPU prints a slightly smaller image. When you are printing to the printer's native resolution, 600 PPI for Canon, 720 or possibly 600 for Epson, ACPU will resample upward and duplicate every 20th pixel or so. It uses a nearest neighbor approach which can create tiny jaggies on near vertical or horizontal pictures. And, of course, there is the issue of printing the wrong size. I highly recommend using Photoshop manages color to avoid these issues. ACPU really should just be used for printing profiling targets.

Also ACPU has no ability to place the image in the page layout.  It seems it only prints starting at (0,0) on the left top corner. So if you want to center the image, you might have to create appropriate margins by extending the canvas size of the image itself.  It is kind of cumbersome. 

The reason I went there in the first place was to keep a record of the first prints from my new printer on a couple of papers without color management that I can compare against in the future.  For that I used a home-made target that includes a set of black and white lines designed to inspect the sharpness.  When I printed the same target in Photoshop with the appropriate canned profile, I saw a particular (and peculiar) difference in the how the sharpness target was printed.  So I went back to ACPU with the same profile pre-converted to make sure that the difference was not related to the profile.   The lines printed same as before, i.e. the discrepancy remained, but with more neutral tone as would be expected form the application of the profile. 

I think I will start a new thread to discuss what I am seeing (after I have done some more tests to make sure I am not doing something silly, which I probably am) in the Printing forum where it might be more suited.

Addendum:  Since there is no way to send a file without color management in Photoshop (like it used to before,) I was wondering if there was a way to create a dummy icc profile that does not do any number crunching but just sends the input file as is to the printer, kind of like an icc profile of a theoretically perfect printer.
« Last Edit: September 12, 2017, 08:45:17 am by nirpat89 »
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Printer Profile Question
« Reply #5 on: September 12, 2017, 06:22:38 pm »

Addendum:  Since there is no way to send a file without color management in Photoshop (like it used to before,) I was wondering if there was a way to create a dummy icc profile that does not do any number crunching but just sends the input file as is to the printer, kind of like an icc profile of a theoretically perfect printer.

Ever since Adobe removed printing w/o color management, I've used something called the "null transform technique" on my Windows machines (XP through Win10)  The procedure is to assign an arbitrary printer profile to the image then print using Photoshop manages color using the exact same printer profile. It makes no difference at all what printer profile you choose. It also makes no difference what the settings are for Intent or BPC*. They are ignored. Photoshop pops up a warning with a "get it" link to ACPU. I just cancel it and proceed. This works with my Canon 9500 and Epson 9800 as well as previous Epsons I no longer have. However, because of the stern warning, I carefully check it by printing out targets using ACPU and Photoshop and scanning/comparing them to make sure they match within the normal print to print variation. I do this every time Photoshop CC gets upgraded.

It's possible that it may not work with all printer device drivers since they receive metadata indicating image colorspace which they could potentially use to do some sort of additional conversions depending on the design of the driver. It's entirely possible that some printer drivers are not compatible with the null transform technique. I have not heard of any though. It appears most people are simply scared enough of the Adobe warning that they just go to the ACPU or use the print utility in X-Rite's I1Profiler. I now use either the X-Rite print or the null transform since the pixel spacing on Isis charts is quite critical. Using the default 6mm sizes with ACPU requires messing with the patch size to accommodate ACPU shrinkage.

*Note:  It may seem odd that the Colorimetric Intent and BPC settings make no difference. It's simply that they are only used in converting to and from some other profile. They were already applied when the image was converted and have no additional use post conversion. It's much the same as assigning a profile where you are not given the option of choosing anything other than the device profile to assign.
« Last Edit: September 12, 2017, 06:29:23 pm by Doug Gray »
Logged

nirpat89

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 661
    • Photography by Niranjan Patel
Re: Printer Profile Question
« Reply #6 on: September 13, 2017, 08:40:20 am »

Thanks Doug for sharing this alternative technique to circumvent the absence of no color management.  Looks like a case where double profiling means no profiling... :)
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Printer Profile Question
« Reply #7 on: September 13, 2017, 10:16:43 am »

You can build a NULL profile in PS (on Mac) too, the two profiles which are in reality just one, need differing internal names so Photoshop doesn't know you're trying to trick it. It examines the internal profile name to ensure you don't pick the same profile after which it will pops that 'get ACPU' dialog. In the end, it's just easier for most people, certainly on Mac, to skip this rabbit hole and just use ACPU for targets or, the product that built the targets that itself should have no issues printing that data without color management! But if you really, really want to go this route in PS, just take one working space profile, duplicate it, rename one differently (internal profile name via ColorSync Utlity), then you can produce a NULL profile there.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Printer Profile Question
« Reply #8 on: September 13, 2017, 10:18:11 am »

Thanks Doug for sharing this alternative technique to circumvent the absence of no color management.  Looks like a case where double profiling means no profiling... :)
It isn't double profiling. It's a NULL profile. And kind of a hack.
Something built a target for printing without color management to create a profile. That something should be able to print the target for you.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

nirpat89

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 661
    • Photography by Niranjan Patel
Re: Printer Profile Question
« Reply #9 on: September 13, 2017, 01:07:05 pm »

It isn't double profiling. It's a NULL profile. And kind of a hack.
Something built a target for printing without color management to create a profile. That something should be able to print the target for you.

Was just using a little CM humor....:D
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Printer Profile Question
« Reply #10 on: September 13, 2017, 01:49:13 pm »

It isn't double profiling. It's a NULL profile. And kind of a hack.
Something built a target for printing without color management to create a profile. That something should be able to print the target for you.

Yep, a hack. One that works and produces the correctly dimensioned target prints on my Win x64 with my printers. I would prefer a non-hack but it's the only technique I know of, other than using a very old Photoshop which I no longer have. The I1 Isis is quite critical about dimension accuracy. Especially using the default 6 mm patch sizes, which works out to exactly 24x24 pixels at the peculiar (101.6 PPI) resolution of its targets. Like all hacks, its a good idea to verify it works on your system.
Logged
Pages: [1]   Go Up