PS Print Dialog Box: Intent settings

[chriso84972911]

Hi Everyone,

Just wanted to ask a question in relation to the Intent options in PS Print Dialog Box.

If you have saved a file to print at a later date, which has been converted to a specific paper's colour profile space (using a specific printer and ink set), you've obviously chosen a specific intent to make that conversion.

So, when you do decide to print, will the choice of Perceptual or Relative Intent that's in PS Print Dialog box have any affect on the print that's produced? Or since the conversion has already taken place, does it not effect the final outcome?

[NB, colourmanagement]

Hi Chris

Thanks for the question.

I am not sure why you'd want to pre-convert your files to the "output / printer" profile ahead of time?

You'd really be best to leave that conversion to be done in the background at the time of printing each time you print.

[why? you might get a new printer, or change paper, in which case those pre converted files would likely be worthless.]

Best approach is to archive an RGB full gamut file for future use.  maybe you're doing that anyway, if so, great.

Back to your original question:

So you've pre - converted a Photoshop file to the ICC profile of (probably an inkjet) printer.

During that original conversion is when the rendering intent (and the black point compensation option) controls the conversion parameters such as compression etc.

Now you come to print, on the exact same printer and paper, I presume? In this case (as you realise) no profile conversion is needed.

Might I please ask, though - how do you set the "Photoshop's print settings" dialog?

[in the old days with a pre-converted file, you'd have selected "no color adjustment" but that option was removed some time ago]

In "color handling" do you select "Photoshop manages color"?

I guess so, but in that case what is your selected printer profile?

IF you are choosing the printer's ICC profile here (the exact same one as used in the original conversion - as you should) then what you'll get is a "null" ICC conversion. Even so, I feel it's worth setting the rendering intent (and black point compensation option) the same as they were in the original conversion just in case. Theory is fine but practice is the way is to reproduce your settings or make a test.

In cases like this my approach is to do a test. It's the only way to be sure.

For testing, I recommend you use a "standard" test image - ideally with a composite of images so it's a good test of output.

[you can download mine here downloads | colourmanagement.net  you want the Adobe RGB version]

now 1:

print just as I described above:

"color handling"  = "Photoshop manages color"

"printer profile", the ICC profile is set to the exact same one that you used in the original conversion

"rendering intent" AND "black point compensation" set the same as they were in the original conversion

make a print

2:

print as I described above,

"color handling"  = "Photoshop manages color"

"printer profile", the ICC profile is set to the exact same one that you used in the original conversion

"rendering intent" AND "black point compensation" set differently to the way they were set in the original conversion

make a print

compare the prints

are they the same?

if so try "black point compensation" the other way and see what happens?

I hope this helps

if so, please do mark my reply as "helpful" and if you're OK now, as "correct answer" so others who have similar issues can see the solution

thanks

neil barstow, colourmanagement.net

[chriso84972911]

Thanks for that detailed response Neil.

The question actually came from a student i was teaching, who had converted and saved in PS after soft-proofing with the appropriate profile and rendering intent before bringing it into college. Once here, he was using an Epson 3880 via PS and raised the question.

My response was to select 'Absolute', since this is basically 'print as-is'...but after consulting a couple of people, i think even if it did try to re-convert with 'Perceptual', it shouldn't matter as all colours are in-gamut (the printers gamut) so nothing would be done.

But when i get a chance, i will try a test just to confirm my suspicions.

Best

Chris

[NB, colourmanagement · Accepted solution]

Hi Chris

I am pleased that my reply helped you, any chance you could mark the question as "correct answer"? Perhaps after doing your test should you prefer that.

When questions on this forum appear as "answered" that helps others who may have similar issues see the thread as worthwhile, should they want to know (in this case) more about Photoshop's Print Settings dialog and ICC profile to ICC profile conversions.

I think he was right to softproof his image before bringing it for printing, and even to convert to do that thoroughly - as, sometimes, post conversion the file can look a little different to the softproofed version. But I definitely recommend first making a temporary "duplicate" of the original, then converting that to the relevant output colourspace (using the relevant profile) and perhaps then viewing this converted file alongside the RGB original, next delete the duplicate and print the orignal file via Photoshop as normal.

It would be great to hear the result of your testing the process I described.

Thorough tests are always good to know about.

have a good week/end Chris

If this helps please do mark my reply as "correct answer" so others who have similar issues can see the solution

thanks

neil barstow, colourmanagement.net

[chriso84972911]

Hi Neil,

It will be a while before i get round to doing the coIour chart test, but can tell you from recent experiment using a B&W image (still in RGB of course) that when you have a file that's already been converted and you put it through the printer using either Relative or Perceptual, there is no 'obvious' difference. Neither was there any 'obvious' difference when i turned off the Black Point Compensation....in fact, all 4 prints are practically interchangeable. However, it would be good to get an Adobe guru to give the technical thumbs up to qualify these findings.

[Test Screen Name]

Some thoughts

1. Black point compensation and rendering intent apply when a profile-to-profile conversion is done.

2. Most software won't bother to convert if the profile already matches.

3. You wrote "i think even if it did try to re-convert with 'Perceptual', it shouldn't matter as all colours are in-gamut (the printers gamut) so nothing would be done."  A profile contains three tables: absolute, perceptual and saturation, used to convert in every case. So, while there is an entirely reasonable expectation that "perceptual" will be the same as absolute if YOUR colours are already in gamut, so far as I see it, your colours are not analysed. So the same squeezing of colours will happen as if there were out-of-gamut colours.

[NB, colourmanagement]

Hi

Thanks for the contribution.

1: I would agree with that summary

2: whilst that should be the case it's unfortunately not guaranteed, for example there's been a long discussion on the colorsync-users users forum about an issue within Canon's printer driver software where even with the same profile on both sides of the equation (unwanted) changes were being made to files during processing for printing.

I hope it's safe to say that Photoshop would not make changes if the profiles either side of the equation match, though.

But I'd still want to test it.

ICC profiles that seem the same may not be the same, as an example: many users have tripped over this with the multiple versions of sRGB (sometimes with very slightly differing names) which, on the surface of it, seem the same, but are not.

Also profiles can get corrupted, so that behaviour is unexpected. Happens more often than one might think.

3: You're right, I would agree that whether file content is within output gamut or not is irrelevant to whether the conversion process is active or not.

It's quite true that the ICC conversion process does not "look inside an image file to check the gamut" before deciding whether a conversion is needed.

In any case, a profile to profile conversion makes changes to far more than just overall gamut. Tonal range, colours and and grey balance are often altered  to suit the output Colour space as well.

It's also true that perceptual profile to profile conversion does compress colour gamut, whether it's necessary or not. That's why Thomas Knoll wanted RelCol (Relative Colorimetric) as the default in Photoshop and, thus, developed BPC (black point compensation) to allow for successful use of RelCol* in image work, he actually told me he was hot happy with the traditional description not what Perceptual does.

*without BPC, a RelCol conversion will not scale black (like perceptual does), so that any image data with values below the max black value of the output space is mapped to black. This means that  image greys can become black unintentionally.

It gets a bit hard to explain without writing an essay, but my understanding  is that the above described compression happens because the amount of (perceptual) compression is set as a default action within the profiling software that made the profile (i.e. embedded within the tables of the destination profile). This means that a pre-determined (gamut) volume (irrespective of image saturation) will be compressed into the output colour space's gamut.

Another, often unconsidered, aspect of this behaviour is that perceptual mapping CAN actually clip image data should it fall outside the area that the profile conversion "expects". 

Summary:

1: Profile to profile conversions are complex and the only way to earn properly what's happening is to make tests.

2: Perceptual is nowhere near as all-forgiving and "safe" as many users think it is.

Don't get me started on conversions where the destination is a working colour space - perceptual seems to be an option as it is offered, but working space type profiles cannot contain a perceptual table. So you better watch out for clipping there guys.

Hope athat all makes sense

Neil Barstow

More on Rendering Intents here: www.colormanagement.net

[chriso84972911]

Hi Neil,

Yes that does make sense...from now on i will advise students to always note what Intent they chose to use with a particular print profile (if they converted and saved the image prior to printing) and to select the same at the print stage in PS. I would assume a similar workflow would be required if they chose to use LR to print in college having first saved a converted JPG on their home PC.

Regarding the choice of 'Perceptual' when making a conversion, i did always wonder why so many people seemed to advocate using it (one particular Lab in London actually recommending that images sent to them should always be converted into their print profiles using it), when in fact my own experience revealed that soft-roofing with RelCol seemed to produce far more subtle changes (and consequently, less work to correct) than Perceptual...with the exception being images with very highly saturated colour's or lots of subtle shadow detail.

[TheDigitalDog]

Regarding the choice of 'Perceptual' when making a conversion, i did always wonder why so many people seemed to advocate using it

Simple, they haven't a clue! The reason there's so much ignorance on the subject of color management, is that those who have it are so eager to share it! - The Digital Dog

Fact: there are no rules in how a profile vendor creates their Perceptual Mapping. Like transparency film, no rules, up to the manufacturer to decide what they think their customers will prefer visually.

Profiles and rendering intents know nothing of color in context. They only treat individual pixels, one at a time.

The best RI is the one you prefer visually as you soft proof the options. Even Saturation is game.

Not all ICC profiles are created equally!

Not all ICC profiles are created equally

In this 23 minute video, I'll cover:

The basic anatomy of ICC Profiles

Why there are differences in profile quality and color rendering

How to evaluate an ICC output profile

Examples of good and not so good canned profiles and custom profiles on actual printed output.

High resolution: http://digitaldog.net/files/Not_All_Profiles_are_created_equally.mp4

Low resolution (YouTube): https://www.youtube.com/watch?v=TNdR_tIFMME&feature=youtu.be