Printing and Gamut Check

[Richard Liu]

Hi,

I'm wondering how much effort I should invest in fixing the out-of-gamut areas that the Soft proof adjustment layer displays in grey when I check the Gamut Check option in the control.  When I add additional adjustment layers to do that (e.g., Curves and HSL adjustment layers), the resulting photo lacks the punch of the original.  On the other hand, when I just uncheck the Soft proof adjustment layer and print, the results when viewed under good lighting bear a reasonable resemblance to what I see on the display.

[I just tried three times to upload JPEG exports, only to be met by error code -200.]

How are other users of Affinity Photo handling soft proofing and out-of-gamut areas? 

The display is a BenQ SW271 calibrated to 5800K, 100 cd/m², gamma 2.2, relative black point.  The printer is a Canon Pro-1000 printing on Canon Photo Paper Plus Glossy II using the "Canon PRO-1000/500 Photo Paper Plus Glossy II" profile supplied by Canon.

[walt.farrell]

The purpose of Soft Proofing is to let you see how the image will probably look when printed, and compare that with how it looks on the screen. They can never match completely, of course, as one uses transmitted light and one uses reflected light, but the idea is that by soft proofing and adding additional adjustments you should get the printed version closer to what you see on the screen.

If you think the prints are coming out acceptably without soft proofing and additional adjustments, I'd say don't worry, be happy

[Richard Liu]

20 hours ago, walt.farrell said:

If you think the prints are coming out acceptably without soft proofing and additional adjustments, I'd say don't worry, be happy

Well, that's not exactly what I said.  In fact, I am using soft proofing.  It almost always shows that the print will be lighter and "duller" than the displayed document, and in most case it does find significantly large swatches of the document to be out-of-gamut.  A tutorial for AFP 1.6 gives some examples, how one might insert adjustment layers before the Soft proof Adjustment Layer to avoid that happening, i. e., using a Curves adjustment and/or an HSL adjustment; however, it doesn't say when it might be preferable to not undertake any correction at all.  If there are no rules of thumb or general guidelines, i. e., if the only way to decide is to try both, then that "solution" has a direct cost.  Furthermore. it would seem that, no matter how often one tried both methods and selected the best, the next time one begins back on square one. 

[walt.farrell]

I suspect that over time one becomes more familiar with the behavior of one's printer and the typical paper that one uses and how it reacts to certain colors. And that might reduce the need for soft proofing and/or for printing trial copies.

I interpreted your original message as meaning that the addition of the adjustments was making the print look duller. It will, I think, be true that if you print the unadjusted print it should end up duller if you have out-of-gamut areas. But the adjustments (again, I think) should make it print closer to what you see on the screen. Still duller, if you're using colors that the inks/printer/paper can't handle. I doubt there's anything you can do about that except use a different printer or inks or printing medium. And it will probably never completely match what you can see on the screen.

[Ron P.]

On 6/20/2019 at 9:07 PM, walt.farrell said:

If you think the prints are coming out acceptably without soft proofing and additional adjustments, I'd say don't worry, be happy

 

 

[Richard Liu]

6 hours ago, walt.farrell said:

I interpreted your original message as meaning that the addition of the adjustments was making the print look duller. It will, I think, be true that if you print the unadjusted print it should end up duller if you have out-of-gamut areas. But the adjustments (again, I think) should make it print closer to what you see on the screen. Still duller, if you're using colors that the inks/printer/paper can't handle. I doubt there's anything you can do about that except use a different printer or inks or printing medium. And it will probably never completely match what you can see on the screen.

@walt.farrell,

Since the Soft proof adjustment layer is using the same profile to display as the printer does to print, I guess the question boils down to this:  What am I seeing on the display when Check Gamut is off?  Have the out-of-gamut colors been mapped according to Rendering intent to colors that the printer profile can handle?  If so, then I would think that that is a much finer grained correction than a global Curve or HSL adjustment, which affects colors that aren't out-of-gamut.

By the way, I seem to have run into an interesting problem:  How does one set the Rendering intent of a print in macOS?  This thread

seems to indicate that it's not possible for me to specify the Rendering intent when actually printing.  I have no idea whether it's hard-coded in the printer profile, but I imagine a mismatch between the Rendering intent of the Soft proof and that of the actual print might be a further headache to contend with.

[walt.farrell]

4 hours ago, Richard Liu said:

Have the out-of-gamut colors been mapped according to Rendering intent to colors that the printer profile can handle?  If so, then I would think that that is a much finer grained correction than a global Curve or HSL adjustment, which affects colors that aren't out-of-gamut.

That's a good question. I'm not completely sure, but it must be showing you the colors as they would print, and that should mean they've been remapped.

What the gamut check should do then, I think, is highlight those colors that have changed due to being out of gamut. But that's a guess on my part.

Note that you don't have to apply a curve or HSL adjustment across the entire image. You could mask them to include only the out of gamut areas, so they would not affect the other colors. I have not tried doing that. Thanks for the question, as that would be an interesting experiment

I don't use a Mac, and can't answer your question about rendering intent.

[Murfee]

 

Hi @Murfee,

Could you please repost your valuable observations and recommendations in the thread that I linked to and erase your detailed post here? 

Edit: Reply moved from Richard's question in the Bug's forum...to avoid confusion 

 

   4 hours ago,  Richard Liu said: 

I posed the question, should I fix areas of a photo that the Soft Proof adjustment layer shows to be out of gamut before printing

Hi Richard, I would fix them and then switch off the soft proof layer, the image on screen might look a bit different but the print should come out as you want it to. (the biggest complaint a lot of users have is "My Prints are too dark") 

I find it far easier to get good prints by developing my raw files in 32bit mode and assigning either the ProPhoto or ROMM profiles, I make all of my adjustments then when I am happy I do a merge visible of all the layers, switch off everything below that pixel layer (this just speeds up your final steps of proofing). This is just an insurance policy for myself that all of my adjustments will be picked up in the printing process, I am aware that some users have experienced issues with some filters in the past, can't remember which ones and not sure if the problems were fixed. I suspect that some (but not all) of the issue is to do with the amount of data being sent to the printer and the drivers struggle a bit with very large & complex files...but that is just a wild guess. 

When printing at home I add a softproof adjustment layer using the relevant profile, with 32bit mode and the large colour profile that I use I find that the only areas that are out of gamut are usually the deepest shadows & brightest highlights, but that will depend on the image you are printing. I make very slight adjustments below the softproof layer often a very gentle tweak with the brightness/contrast adjustment layer will do the trick, if not then a bit of curves adjustment. If the out of gamut areas are really small and not in a critical part of the photo then I don't fuss . When I am satisfied (and that takes a lot ) I switch off the soft proof layer. If it is a large image I might also just copy the adjustments & merged pixel layer to a new file and crop to a critical part if the image and run a small print off, if I am not happy then I make further adjustments in the original file. It sound like a lot of faffing but if you want the best possible prints then it is worth the hassle, when you have done this a few times you will instinctively know what you need to adjust and what is fine to leave alone. 

If I am printing from a jpg then trying to soft proof can be a bit more of a challenge   In those rare cases I tend to just select the paper type in the printer driver and let the printer take care of the colours...so far the results have been very good. 

Edited July 6, 2019 by Murfee
Added comment

[Richard Liu]

Hi @Murfee,

Thanks for moving your posting here.

As a matter of fact, when I first began using Affinity Photo, having previously used nothing, not even the ubiquitous products from Adobe, I did everything in 32 bit mode with the ROMM profile, i.e., development and editing.  That seemed to be the cause of a curious effect that I can only describe as looking at my work in Affinity Photo and exported JPEGs as if through a veil, or dirty window.  The colors were muddy, not clear.  For a time I thought I just needed to add a tone curve, and I was satisfied until I began using DxO Photo Lab 2.  The colors seemed deeper, clearer.  I put the question to one of these forums, whether it is possible to produce something comparable to the output from PL2 with AFP, and one person took a crack at it and produced a .afphoto that compared favorably.  I noticed that the output of Develop was 16 bit, and I questioned him about that.  Then I reconfigured AFP to produce 16 bit from Develop and never turned back ... until now.

I created a 32 bit workflow from a RAW file that I had previously developed, edited and printed in 16 bit using the ROMM profile.  The veil phenomenon, which I had never experienced in 16 bit mode, was back, and everything was darker on the display than with the 16 bit file.  When I printed the 32 bit workflow, including a curve adjustment to clear up out--of-gamma areas, but excluding the soft proof layer, I saw on the thumbnail in the print dialog that a recolor adjustment that I was using for bringing texture in rock and wood (cf. James Ritson's tutorial https://player.vimeo.com/video/150884324/ from the beginning until about 1:40 min.) was going to be printed much to orangey, even though the orange color was hardly visible on the display.  So I had to reduce its saturation the opacity until the thumbnail displayed properly!  The resulting printout was OK, but it was much brighter than on the display.  In contrast, when I work with 16 bits the printout agrees tolerably well with the display.

I've uploaded the files to Google Drive (I think!).  Here's a link, I think it allows downloading:  https://drive.google.com/drive/folders/1zkX5qqe1shqc2igAalTTEVzNNefsDwZc?usp=sharing .  .NEF is the Nikon RAW file.  "36" in the filename denotes the 32 bit workflow, resp. the JPEG export.

What display and printer are you using?  I have a BenQ SW271 calibrated to Adobe RGB, gamma 2.2, 100 cd/m² and 5800K, and am printing on a Canon Pro-1000 using Canon Photo Paper Plus Glossy II and the printer profile for that paper supplied by Canon.

 

[Chris26]

Just want to add my three-penneth to all this out of gamut stuff;  today for example I had out of gamut on a swathe of beautiful Blues, I switched my soft proof from Relative colormetric to perceptual and half of that out of gamut dissappeared. I usually find that switching from rel colorrmetric to perceptual reduces out of gamut by up to 50% on average in my images. 

And by the way, on the mac, if you go to System Preferences - Displays - Colour (they spelt it wrong) then select your display profile (mine is a custom one from my profile software for example) then click on 'Open Profile" you will find the rendering intent listed. 

[Richard Liu]

Your thruppence is much appreciated.  I have opened some profiles and see variables with descriptions like "Intent-i, 16-bit, device to PCS converter," and "Intent-i, 16-bit, PCS to device converter," where i is an integer from 0 to 2.  I was expecting something like "Relative colorimetric," "Absolute colorimetric," "Perceptual," etc.  The problem is specifying the intent for printing, since the print dialog evidently does not support it.  I assume that the printer driver then uses a default one, but which one is that?

[Murfee]

Hi @Richard Liu  I will download & have a look at your files when I get a bit of time possibly tomorrow. 

I have not experienced the veil mode you describe, however when I first open a raw in 32 bit mode it is darker than in 16 bit with very little contrast. For me this allows far more manipulation ... but you do need to process with various adjustment layers. I sometimes duplicate the image and go into the tone mapping persona, I don’t go mad but a little bit of tone compression can reveal a lot of hidden details. 

This is one of the new tutorials by James Ritson  HDR from a single image it shows just how much information is in a 32 bit file.

To answer the question about monitor & printer, up until 3 years ago I used an all singing all dancing Windows machine, I had an Eizo CR something or other (cant remember the numbers but I got it in new about 2006) monitor, my Printer was a huge Epson 7880 bought slightly later, all were working well...until the flipping dog opened the door to my office and along with his partner in crime managed to knock the screen flying into the printer.  I was going to replace immediately but other things were going on that sort of forced a short rest from photography. I needed a laptop so decided to get a MBP ... my chance to ditch Microsoft & Adobe all in one go... So the actual answer is I use the screen of my MBP, uncalibrated, in varying lighting conditions. My serious print work goes to an outside printer that I have been using for years, I rely on him for decisions on which of his machines & paper that will work best for each image. I leave  soft proofing for these to him. The funny thing is the print colours match my uncalibrated screen almost perfectly. I needed a printer in a hurry at home for some very basic documents, so popped out and came back with an Epson  XP -7100 for the bargain price of £50, its an all in one thing, but for a very cheap little machine it can turn out very good quality photographs, I was really surprised how good they are. Again, they match my screen output. I use my home printer for smaller prints of local scenes that I get mounted & framed, these are then passed to a friend for sale, all proceeds then go to local charities.

[Chris26]

1 hour ago, Richard Liu said:

Your thruppence is much appreciated.  I have opened some profiles and see variables with descriptions like "Intent-i, 16-bit, device to PCS converter," and "Intent-i, 16-bit, PCS to device converter," where i is an integer from 0 to 2.  I was expecting something like "Relative colorimetric," "Absolute colorimetric," "Perceptual," etc.  The problem is specifying the intent for printing, since the print dialog evidently does not support it.  I assume that the printer driver then uses a default one, but which one is that?

Did you see something like this, you highlight the word 'Header' in the right box? At the bottom of the right box you see the rendering intent

 

1 hour ago, Richard Liu said:

 The problem is specifying the intent for printing, since the print dialog evidently does not support it

 

Which print dialogue are you referring to?

 

 

[Richard Liu]

@Chris26,

Thanks, I didn't see the Header record.  Yes, actually one can use the ColorSync utility to examine any of the profiles, so not only the ones for the displays, but also those for the printers, scanners, etc.  Seems that they all say "Perceptual" in the Header, but I'm wondering what those entries with descriptions "Intent-i, 16-bit, device to PCS converter," and "Intent-i, 16-bit, PCS to device converter" do?  I would imagine it's the printer driver that takes the colors that an application sends it and maps it to the color specified by the printer profile, and it would have to handle any rendering intent specified as well.

The printer dialog is what pops up when you tell Affinity Photo to print.  There, you specify the printer and the profile.  It needn't be the same printer and profile that you specified for soft proofing, of course, since nothing even forces you to soft proof.  So, taking the profile for my Canon Pro-1000 and the Canon Plus Glossy II paper as an example, I see that the header in the profile specifies rendering intent Perceptual.  So if I give that profile to the Soft Proof adjustment layer and specify Absolute Colorimetric, how can it possibly honor that profile unless there's something in it that either supports that intent directly, or tells the application or print driver how to convert from Perceptual to it?  In other words, why doesn't AFP give me an error message saying this printer and profile only support Perceptual?  Now things get more interesting when I print, because, on a Mac there's nowhere in the print dialog, including all its submenus, where one can specify rendering intent.  I understand that specifying rendering intent for printing is supported on Windows.

[Murfee]

Hi @Richard Liu I downloaded your _DSC1411_AFP36 afp file, the NEF and your _DSC1411_AFP_X jpg

The veiling seems to be coming from something done in the develop module, the only way I can get something similar while developing is by reducing the contrast. Was this originally created in the 1.6 version?

I turned off all of your adjustments in your afp file and dragged your backup background out, I could then see the veiling. 

The screenshot is showing what you are starting with in the Photo Persona, compared to what I am seeing as a starting point. The only adjustments I made in the develop persona was to increase the exposure a bit (yes 32bit do open a bit dark) and a touch of brightening to lift the mids, I lowered the highlights just a tiny bit to stop them blowing out, set the colour profile to ROMM and hit develop. My develop assistant is set to the Serif engine and to take no action on the tone curve.

[Chris26]

Hi richard, at this juncture I have to bow out.  I am not that deep into colour management theory as you are.  All that works for my photography is a profiled monitor screen, consistent colour management workflow and a knowledge of the terminology used.  If my print comes out the way I see on screen then it works.

Since you are also dealing with AfPhoto's printer dialogue I can not help since I do not have this siftware yet.

As far as Printer drivers and Rendering intents go, I only assume (from photo printing successes) that if the software is in charge of all print output then whatever the dynamics or physics are at work under the bonnet, these are working fine.  

11 hours ago, Richard Liu said:

So if I give that profile to the Soft Proof adjustment layer and specify Absolute Colorimetric, how can it possibly honor that profile unless there's something in it that either supports that intent directly, or tells the application or print driver how to convert from Perceptual to it?

How it works I have no clue.  All I know is for example, that Photoshop is an independent colour managed environment and any generalised computer colour management settings involved in the general operation of your computer are ignored.  Photoshop has its own colour managed workflow that you set up yourself in a seperate dialogue box, and with Bridge if you have that.

[Richard Liu]

@Murfee,

I really appreciate your time.  After reading your account, I redeveloped the .NEF file.  I've made a screen movie of my developing steps.  It's also on Google Drive:  https://drive.google.com/open?id=1zkX5qqe1shqc2igAalTTEVzNNefsDwZc.

00:00   Load the .NEF into AFP 1.7.1

00:07   The Develop Assistant confirms that it has applied the +2/3 stops exposure correction that I dialed in on the camera.  Clipped highlights, shadows and tones are displayed.

00:16   I begin to play with the Blackpoint.  Usually, I push it to about -5% in order to create what @James Ritson in one of his 1.6 tutorials calls RAW latitude.  You can see the veil appear as the slider moves below zero.

00:31   Pull the highlights down until they aren't clipped anymore.

00:41   Increase brightness.

01:00   Add a bit of sharpening.

01:17   The sharpening has evidently clipped some shadows and tones.

01:35   Just demonstrating which profile we're working in.

01:38   Opening the shadows up a bit to remove the clipping caused by sharpening.

01:51   Oops!  Forgot to activate sharpening.  Do it, then try to get a good setting for Shadows.

02:30   Check increasing exposure.  Nope, that just blows out some highlights.

03:16   One last check before hitting Develop.  Some tones are still being clipped, but I think that's OK.

03:26   I just noticed that I deactivated Details before hitting Develop!  Dumb!  But you get the idea, I think.

So the veil is being caused by reducing the blackpoint.

Now I've done everything all over again, beginning with a more careful development as per the above, except for the oversight at 3:26 (_DSC1411_AFP36_2.afphoto and _DSC1411_AFP36_2.jpg)  In Photo I first dehazed a copy of Background, then ran that through Tone Mapping, adjusting basically tone compression and micro-contrast, then I recombined the original and the tone mapped, dehazed Backgrounds.  I didn't see much point in adding more sharpening, but I did use the High Pass live filter to sharpen small details, e. g. the pebbles in the swing area and the leaves.  Defringe was supplied selectively to the folded umbrellas and the pole to the left of the bear, lest even the little red bulbs be defringed and lose their color.

After that things are pretty straight forward:  Levels, Brightness/Contrast and Curves deepen the shadows and colors, then Vibrance and White Balance to adjust the colors, Soft Proof and Curves to address out-of-gamut colors.  Basically, if the gamut check is deactivated, the soft proof is just somewhat lighter than the original.  In fact, if I turn off soft proof and uncheck the curves adjustment required to address the out-of-gammut tones, and print, the printout is much lighter.

One possible bug:  The fringing that is selectively corrected with Defringe is very visible in the print and in the exported JPEG, but not when viewed in AFP at 100%.  If you have time, can you confirm that this happens on your setup, too?  Thanks.

 

 

 

[Richard Liu]

@Chris26,

OK, thanks for your help.

I think, like you, most of us live by the maxim, if it ain't broke, don't fix it, a corollary of which is, if it works, don't worry how it works.  It's when things don't work, at least, not the way we expect them to, that our strategies diverge.  Mine is to try to understand what components are involved, what each is supposed to do, and where the cause(s) of the problem might lie.  I do not know Photoshop, have never worked with it.

 

[Murfee]

1 hour ago, Richard Liu said:

00:16   I begin to play with the Blackpoint.  Usually, I push it to about -5% in order to create what @James Ritson in one of his 1.6 tutorials calls RAW latitude.  You can see the veil appear as the slider moves below zero.

 

Do you mean the legacy video maximising raw latitude, if so I have just watched and do not see the black point slider being touched, the tone curve is set to take no action, the highlights are reduced, the brightness is increased, the result is a tonally compressed image but there is no veil. It looks a bit flat but that is the intention. 

Maximising Raw Latitude

I can’t download your video at the moment, it will be a few days before I will get the chance, (painfully slow internet here) 

Changes were made in 1.7 that cause 32bit files to open a bit darker than in 1.6, try again with the new file, adjust the exposure slider to the right a bit just enough to bring it inline with what you think it should be, drop your highlights, push the brightness up. Do not adjust the blackpoint. At this stage I would leave the shadows alone for this image. Develop the file. You can adjust the shadows in the Photo Persona by using the live filter shadows & highlights, this filter does a far better job and gives you control over the shadow range, you can adjust it so that you are only adjusting the deepest shadows if that gives a more pleasing result for you.) I believe that the shadows & highlights adjustment layer will be retired at some point.

I don’t use the defringing filter, for some reason I can never get a good result. I prefer the selective colour adjustment, depending on the fringe colour, I choose magenta’s or blues, I adjust the sliders below to reduce the magenta or blue, I also tweak the amount of black/white. When I am happy with the result I invert the adjust,ent layer and paint in where I need to apply the effect.

I wouldn’t worry and over analyse soft proofing.  Prints & screens will be different due to the way they will be viewed, it is not always necessary for the two to match. The important thing is how the print looks where it is intended to be viewed, if you are happy with that then you don’t need to try and get an exact match. The important thing is to get to know your printer, I would run a few small jpg files as tests. Forget any soft proofing, just select the paper type & let the printer control the rest. If you like the prints then aim to try and match what the jpgs looked like on screen. 

For my little printer at home, purple things are the problem, they come out a bit too blue, so knowing this will happen, I push the purple towards red a bit on the screen, the print then comes out as I want. 

[Richard Liu]

@Murfee,

For the better part of a week now I've been reconsidering my Develop-Photo workflow on the basis of yours.  I've set RAW output format in Develop Assistant to RGB (32 bit HDR) and leave Blackpoint alone in the Develop persona¹.  I content myself simply with addressing clipped hightlights and shadows and adding some details to the image, then develop it.  In the Photo Persona, I apply Haze Removal to a duplicate of Background.  Then, depending on the particular photo, I will usually work on a duplicate of either Background or the dehazed duplicate² in the Tone Mapping persona, limiting myself to either a reasonable adjusting compression and local contrast, or setting compression to 0% and local contrast to 100%³.  Back in the Photo persona I then combine all the layers, choosing their opacities and blend modes to produce a pleasing, realistic result while avoiding the "burn your eyeballs out" effects typical of bad HDR. From there on, my work in the Photo persona is not much different than when I work in 16-bit RGB.  I do notice that results have more "pop" than before, and the agreement between what is displayed on the BenQ SW271 calibrated to RGB, 100 cd/m², 5800K and gamma 2.2 and what is printed on the Canon Pro-1000 with the paper-specific profile provided by Canon is reasonably good.  The Soft Proof adjustment layer still has its limitations, of course, and I do not find a clear statement from Serif about how it is handling out-of-gamut colors.  If Check gamut is activated, pixels of out-of-gamut color are displayed in black.  But if that option is not selected, are those same pixels displayed in the color calculated by Affinity Photo from the specified rendering intent, or does Affinity Photo simply pass the rendering intent to the operating system, in effect leaving the handling of out-of-gamut colors to it and the display?

So thanks for your help.

__________

¹   That was a misunderstanding on my part.  In a tutorial for AFP 1.6 entitled Salvaging Underexposed Images (http://player.vimeo.com/video/202715178/) James does adjust among other things the blackpoint of the RAW file and summarizes these endeavors collectively at 1:16 as "creating a very flat image" in an attempt to capture "as much highlight and shadow detail as possible."

²   If haze removal has dramatically improved sky detail, I usually pass the Tone Mapping persona a duplicate of the original Background layer.  If the effects of dehazing are more subtle, I apply Tone Mapping to a duplicate of the dehazed layer.  I don't want overly dramatic skies in the tone mapped layer to dictate its opacity in such a way that the extra detail that I intended it to provide must be barely visible, lest the sky be too dramatic.

³   James' new tutorial, 

and his older one on bringing out water detail (http://player.vimeo.com/video/202899215/) are reliable guides to which tactic to use, depending on which of these two purposes the tone mapped layer should fulfill.

[Murfee]

2 hours ago, Richard Liu said:

So thanks for your help

Glad that you are happier with your results, it’s a matter of experimenting to find what works for you.