Clipping out-of-gamut colours

[RobertArdill]

Hi Robert,

My suggestion (and I don't say this very often!) is to pop across to http://www.luminous-landscape.com, and post your question in the direction of Jeff Schewe (whom I'm sure you'll be well familiar with by now) (or Andrew Rodney); I think the two of you would have very meaningful conversations on this topic.

Hi Colin,

Well, I'll take that in the spirit in which it was no doubt intended ... a constructive suggestion as to where I might find more like-minded companions.

The ironic thing is that it's precisely attention to detail on this very site that encouraged me to join this forum and to post a couple of topics. I was looking for information on soft-proofing and came across the really excellent article on color management: https://www.cambridgeincolour.com/tu...anagement1.htm, FYI. What is so good about this series is that it covers some of the essential points of color management beautifully clearly and concisely. You can't read this and ever again be in any doubt about the differences between Relative intents and Perceptual ones, for example, and you will also understand the benefits and pitfalls of each.
The article on soft-proofing is equally good: https://www.cambridgeincolour.com/tu...t-proofing.htm. Again, it is going into details, but these are details that we need to know and understand.

So, what's your problem? I offer a suggestion ... you don't need to pick up on it; you don't need to try it out; you can agree or disagree; you can offer an alternative and possibly better suggestion. But why knock it because it doesn't fit in with what you think is important?

For a discussion about the dangers of getting bogged down in minutiae (ironically once again), this one pretty well fits the bill, IMO.

Of course I can go talk to Jeff Schewe, but I'm not interested in arguing about something like soft-proofing or out-of-gamut clipping with him, or anyone else, for that matter. If someone makes a suggestion that seems useful to me then I'll be happy to take it on board; if I think the suggestion is a bad one then I'll probably say so; if I think the person's got his head stuck in irrelevant detail then I'll leave him there to find his own way out.

So ... have you tried out the little technique I've offered you guys on one possible way in which out-of-gamut colors can be handled? Or do you think this is an irrelevant subject?

Robert

[RobertArdill]

Sorry; I've been rather busy, so have not had a chance to make comments on what this means technicallly.

The math says you are creating output with colours that are totally "false", but if you like the results, who am I to argue. After all I fully advocate getting a pleasing look over 100% accuracy in most work.

In Perceptual all of the colours are shifted to bring the outliers into gamut. In Relative, the out of gamut outliers are corrected by the non-outliers are left alone. We don't know exactly how this is done because the software and printer manufacturers do not publish their algorithms; but I suspect reality is a bit more complex, especially in how Relative is implemented.

With your technique you are preserving neither rendering intent. You are doing this by taking the value of every pixel and finding the mean based on both rendering intents and outputing what amounts to an "false" colour for each and every pixel. Nicely said both in gamut and out of gamut colours will all be incorrectly rendered using this technique.

You've spent a lot of time writing about the importance of colour accuracy in your work, yet the technique you are advocating goes in the opposite direction. Am I missing something here?

Hi Manfred,

No, you're not missing anything, although you're not entirely correct either. It's true that as soon as you use Perceptual you will get a color shift if any of the colors are out-of-gamut; with Relative Colorimetric only the out-of-gamut colors will be wrong (or clipped), while the others will be OK (subject to the white point being shifted); with a mix of the two (as per Robert's Proposed Technique), only the out-of-gamut colors will be shifted (so it's not true to say, to quote you, that I'm taking an average and that as a result "both in gamut and out of gamut colors will all be incorrectly rendered using this technique").

This is why I generally prefer to use Relative, although I will always try out both intents to see which one gives the more pleasing effect: for me achieving a pleasing effect is everything as I'm not trying to faithfully reproduce the exact colors in a scene.

There is no way in which an out-of-gamut color can be printed. It has to be brought back within gamut. That's a fact. The question is: what kind of change is the least displeasing for a particular image? Sometimes allowing the CMM to do it's own thing with a Perceptual rendering is the best. Sometimes allowing the clipping is just fine. But then again, sometimes neither is really acceptable, so what do you then do?

Well, the usual suggestion is to tweak the hue/saturation/lightness of the out-of-gamut colors to bring them back in to gamut.

I'm suggesting ... as a possible alternative, nothing more ... that sometimes, possibly, maybe, try it for yourself and see kind of thing, the image may end up looking better by using Relative Colorimetric and bringing the out-of-gamut colors back into line by using a Perceptual intent on them. All this is doing is shifting the hue/saturation/lightness of the out-of-gamut colors to bring them back into line.

Because Photoshop has excellent masking tools and layer opacity controls, this adjustment can be done quite delicately and only where you want it done. I haven't used the technique enough to be able to say that I think it's a really good technique (which is one reason I posted the technique here, so that some of you could also try it to see what you think); but I haven't come across it anywhere else (perhaps there's a good reason for that ), it's really easy (once you have an action to create the 'Perceptual' layer), and it does work in that it does bring the out-of-gamut colors back into gamut.

So - why don't you try it and forget the maths for the moment?

Robert

[Daveart]

Hi Dan, a manufacturers ICC is not really that accurate unless you print an image on there paper there ICC profile for that paper and make and model of printer and send the result to the manufacturer the print of there image that you printed and they will then send you a revised ICC for your setup and inks you are currently using, but you must also take into account when you replenish your ink you should get a recalibration, why? because manufacturers inks can vary in intensity which would print slightly or in some cases dramatically change the colours, or you maybe lucky and go a long time before any noticeable changes appear.

Oh and keep to the same paper and paper manufacturer, or get ICC for each paper.

Dave

[Manfred M]

So - why don't you try it and forget the maths for the moment?

Thank you for confirming my suspicions.

You have to understand the math to understand what is happening to your image when you apply your technique. Digital cameras, computers and printers use these to generate your image (including how out of gamut is handled). I'm not suggesting that you need to be a mathematician, but you at least need to have an understanding of the underlying math and how the interactions of the two rendering intents affects your final print.

If I get a chance to do so later on, I'm going to try to produce a few diagrams to show what your technique does and why I am criticizing it.

[Colin Southern]

Well, I'll take that in the spirit in which it was no doubt intended ... a constructive suggestion as to where I might find more like-minded companions.

Correct. Jeff and Andrew are recognised as industry experts in the field of colour management, and whereas I don't always agree with either of them, none-the-less, their areas of expertise are probably better aligned with what you're seeking.

The ironic thing is that it's precisely attention to detail on this very site that encouraged me to join this forum and to post a couple of topics. I was looking for information on soft-proofing and came across the really excellent article on color management: https://www.cambridgeincolour.com/tu...anagement1.htm, FYI. What is so good about this series is that it covers some of the essential points of color management beautifully clearly and concisely. You can't read this and ever again be in any doubt about the differences between Relative intents and Perceptual ones, for example, and you will also understand the benefits and pitfalls of each.
The article on soft-proofing is equally good: https://www.cambridgeincolour.com/tu...t-proofing.htm. Again, it is going into details, but these are details that we need to know and understand.

Thanks. Site-owner Sean McHugh authors all the articles, with occasional input from myself and others.

So, what's your problem?

I'll tell you exactly what my problem is ... I sent off 2 water-damaged HC3sX FBL controllers back to their manufacturer (in separate packages) a month ago today, and neither appears to have arrived - so I'll be out of pocket $1400+ if they don't show up. Great That's the only current problem that I have that I'm aware of.

I offer a suggestion ... you don't need to pick up on it; you don't need to try it out; you can agree or disagree; you can offer an alternative and possibly better suggestion. But why knock it because it doesn't fit in with what you think is important?

I didn't knock anything - my previous post was my first in this thread. If you're referring to the suggestions of others that in essence "they feel you may be at risk of investing a disproportionate amount of human resource in one area at the expense of other areas" then I might be inclined to suggest that perhaps you're also a little guilty of being a little "anti" because you weren't told what you wanted to hear (ie "judging the signpost instead of looking at the direction it's pointing). To be honest, if it's an areas that you want to learn more about then more power to you - totally your call, and it's not my place to comment on that one way or another. I will add though that in general (not directing this at you in any way), we DO see a lot of examples of people "majoring in minor things"; people who will choose a 18MP camera over a 16MP model for the "extra detail" - discard 95% of the information when down-sampling - and present work that doesn't appear to have been sharpened at all. We'll have those who will berate the colour accuracy of ACR and yet not even be aware of the fact that they need to be selecting (or creating) a more appropriate profile. We have those who only work in Prophoto colourspace (so they don't lose information) and then berate a particular brand of printer because the colour is all over the place - after having adjusted it by eye on a non-calibrated and non-profiled sRGB capable monitor.

Reflecting on that, I think it's probably true to say that my influence has possibly rubbed off on a few around here over the years, and that's one where I emphasize what I call "Real World" photography; it's a world where an optimal multi-pass sharpening workflow has far more influence on an image than "lens A" over "lens B" - a world where high-iso "noise" doesn't degrade an image anywhere nearly as much as camera shake / subject motion / or insufficient DoF (so "up the ISO if one needs to") - and to be honest, it's a world where trying a combination of rendering intents probably isn't going to make one heck of a difference with all things considered (and I say that as someone who creates their own camera, monitor, and printer/ink/media profiles and makes a substantial part of their living from large-format printing) (up to around $1000 so far this week to put that in perspective).

For a discussion about the dangers of getting bogged down in minutiae (ironically once again), this one pretty well fits the bill, IMO.

You don't agree with the input offered by others - we get it Robert.

Of course I can go talk to Jeff Schewe, but I'm not interested in arguing about something like soft-proofing or out-of-gamut clipping with him, or anyone else, for that matter. If someone makes a suggestion that seems useful to me then I'll be happy to take it on board; if I think the suggestion is a bad one then I'll probably say so; if I think the person's got his head stuck in irrelevant detail then I'll leave him there to find his own way out.

I'm not suggesting you argue with anyone (infact I suspect that they'll all wish you wouldn't!). What I AM suggesting is that you engage in some appropriate dialog with those who specialise in the areas that you seek to learn and understand more about. Not really understanding why you're appearing so almost "confrontational" about this; to me, it's a no brainer: you wish to discuss finer points of colour management, and on the site referenced you have the opportunity to do just that with two (and more) gentlemen who have - literally - "written the books" on this very topic.

So ... have you tried out the little technique I've offered you guys on one possible way in which out-of-gamut colors can be handled? Or do you think this is an irrelevant subject?

It's not something that interests me as it's not an area that I have any difficulties in that need addressing; a "solution in search of a problem" so-to-speak.

My apologies that you're not getting the kind of responses that you seek here -- it may well turn out to be an unfortunate reality that this is not the best site for your particular needs.

[RobertArdill]

OK, what the hell, may as well be hung for a sheep as a lamb

OK, here's a refinement of Robert's Out-Of-Gamut Technique (ROT). This again would benefit from an action (or a script, which I may put together if I get some encouragement , hmm, maybe not that likely?).

The problem with Photoshop's out-of-gamut warning is that it gives one no idea of how far the pixel is out-of-gamut, so one really has to do a visual check to see where the problem is. The Photoshop OOG warning is a useful pointer, but that's all, so the actual correction has to be visual, otherwise the solution usually ends up worse than the problem.

So, here’s a suggestion to overcome this problem: what we need is an out-of-gamut mask with the most out-of-gamut pixels lighest, fading to black for those that are not out-of-gamut at all. Don’t ask me why Adobe hasn’t produced this (maybe there’s a plugin or filter than can do it??).

Anyway, if you are going to print using Relative Colorimetric, here’s how you can do it.

Take two copies of the image. Convert Copy 1 to Relative Colorimetric for your particular print profile (use BPC if you are going to print with BPC). Convert it back to your working space (Adobe RGB, say). Copy the image (Copy 1) over Copy 2. Change the layer blend mode to Difference. Flatten Copy 2. Copy it and paste it into the layer mask of the ‘Perceptual’ layer in the original image (the one created using ROT).

Now apply a levels correction to the mask, shifting the white point until the Photoshop OOG warning goes away (or, preferably, do it visually until you are happy that the clipped areas are now OK).

OMG, I hear your cries! It’s Sooo complicated! No, it’s not. All we’re doing is taking the difference between an image before and after it has been converted to the Relative intent which will be used to print, and using this difference as a mask.

A word of warning – Relative Colorimetric shifts the white point, so there will be a difference even where the pixels are not out of gamut. So the mask should be blacked out in the areas where Photoshop does not show a gamut warning.

As an example, here is an image with Photoshop's OOG warning:



Here is the mask that I generated using the little technique I've just described:



The areas where there are little dots of white (slope at the left of the image) can safely be masked out because they are very unlikely to be visible, even though they are out of gamut (the Photoshop OOG warning does show them on the full-size image, not on this downsized one). I just haven't bothered to do that.

And here is the image after OOG correction (Photoshop warning still on)

[Black Pearl]

All very clever I'm sure and you're obviously happy with your gamutty fiddlings but what you've ended up with is an image that doesn't look in any way real. The colours are weird, its over saturated and there's too much contrast.

[Colin Southern]

All very clever I'm sure and you're obviously happy with your gamutty fiddlings but what you've ended up with is an image that doesn't look in any way real. The colours are weird, its over saturated and there's too much contrast.

There's too much high-frequency sharpening also.

[RobertArdill]

Hi Colin,

Your points are well taken and I apologize if I am being a bit antsy! It's easy to misread a comment as being negative when the author did not intend it. But it's hard sometimes not to react to a comment like this one: "I'm not suggesting you argue with anyone (infact I suspect that they'll all wish you wouldn't!)" because it does sound as though you're fed up with my arguing , while perhaps all you intended was some humour.

Robert

[RobertArdill]

All very clever I'm sure and you're obviously happy with your gamutty fiddlings but what you've ended up with is an image that doesn't look in any way real. The colours are weird, its over saturated and there's too much contrast.

Hi Robin. I quite agree that the colors and tonal balance of this image are not very good. There is a slight over-saturation (but not much, this was a very colorful scene with strong heathers and bright yellow-green mosses/grasses above the lake. The shadows are much too dark clearly. However, I didn't post this image for its perfection - simply as an example to illustrate the out-of-gamut technique I'm playing around with.

This version of the image is a bit better:



Robert

[RobertArdill]

There's too much high-frequency sharpening also.

Another positive and useful comment Colin? My, I seem to have fallen into the nettles here.

As it happens, there is no over-sharpening at all in the original image. Perhaps you are viewing it too large - or maybe I didn't upload the image correctly to your site.

Robert

[RobertArdill]

OK, guys. It's been interesting talking to you, but I've really had enough. I'm happy to engage in discussion as long as it's constructive, but this is ridiculously negative, irritating and time-wasting (no doubt as much for you as for me) and we are clearly getting nowhere fast.

So 'bye and good luck with your photography

Robert

[Polar01]

Robert as for the sharpening, I find sometimes when I upload an image, that it appears to be over sharpen when in fact there has been none applied at all. When this happens I even have to admit that is looks way over sharpen, it does not happen all the time, but when it does it does.
You method is most interesting, it may take sometime for me to get my head around it, to try. Until then I will the old method, hue/sat adjustment layer, colour picker, inverted mask to show only those colours adjusted. If it takes me a two or three layers so be it,or I just might use a different stock. I use the supplier's ICC profiles and the ones I create on their stocks myself.

Cheers: Allan

[Colin Southern]

Hi Colin,

Your points are well taken and I apologize if I am being a bit antsy! It's easy to misread a comment as being negative when the author did not intend it. But it's hard sometimes not to react to a comment like this one: "I'm not suggesting you argue with anyone (infact I suspect that they'll all wish you wouldn't!)" because it does sound as though you're fed up with my arguing , while perhaps all you intended was some humour.

Robert

No - I wasn't being "funny" - I'm serious.

I was simply reacting to your (somewhat odd) comment where you said "Of course I can go talk to Jeff Schewe, but I'm not interested in arguing about something like soft-proofing or out-of-gamut clipping with him, or anyone else, for that matter." Why would you assume that it would be an argument? Again, you appear to be the one being negative here, not me.

To be honest Robert, all of this comes across as somewhat egotistical; you appear to have come here seeking a "pat on the back" for "discovering" a new way of processing OOG colours - experienced members here have pointed out that that (a) the practice doesn't match the theory, and (b) suggested that you might be better off concentrating on other areas as well - you post an awfully processed image as an example of your new technique - some of it's issues get politely pointed out - and reacting badly to this reality check, with a bruised ego you're now "packing up your toys to go play somewhere else".

All-in-all I'd have to say this is a pretty disappointing attitude; here at CiC we literally get THOUSANDS of people commenting on how helpful people here are - and from what I've seen you've been treated no differently; I apologize that the advice and feedback isn't what you wanted to hear, but at the end of the day, nobody gains anything (apart from false hope) if we just lie to you with false comments of "Well done" and "bravo".

So in summary, perhaps CiC isn't the problem here.

Live long and prosper Robert Ardill - I wish you well with your photography too.