This content has been marked as final. Show 22 replies
I have not plotted PhotoGamut RGB directly against the K3 inkset per se, but I have recently plotted it against lots of devices and profiles. In general, Photogamut RGB exceeds the output of nearly all printed output devices (press, digital press, inkjet, lightjet, etc) with only a few minor exceptions. In the cases where a device exceeds the color gamut of Photogamut, it is just barely, and in my opinion, inconsequential. I doubt it would ever be noticed since it is on the very fringes.
I prefer this space to an ultra large editing space like ProPhoto, though others do not. I find there are easier transitions and fewer surprises when converting a job from PhotoGamut to the final output space. PhotoGamut does clip unprintable highlights, which is good if you are concerned primarily with printed output. If web display is a huge consideration, sRGB will preserve these unprintable highlights and pastels and might be preferable. Adobe RGB is in the middle as far as printable gamut...it preserves the bright highlights and pastels but has a much smaller gamut that PhotoGamut overall.
So, it depends on the image and your requirements. I still find Photogamut a great choice for brightly colored originals that are destined for print.
PhotoGamut? Is this new in CS3? I have not seen this RGB profile.
It has nothing to do with CS3 in particular. PhotoGamut RGB was developed in Europe by a bunch of color geeks in an attempt to create a more compact working space that encompasses the entire print world (press, digital press, laser, inkjet, lightjet, dye sub, etc). Unlike most RGB working/editing spaces, it is table based, but is still perceptually uniform, so it supports all rendering intents. It is wider than Adobe RGB in printable colors, but does not include most of the unprintable highlights, light pastels, etc, that are out of gamut on most printing devices. This makes it a very good working space if you are focused primarily on printed output. Since it is shaped like a typical output space, final conversions are gentler and result in less shift, especially when using perceptual rendering intent. In my opinion, it offers the advantages of ProPhoto without the baggage, at least for printed output.
Since it is a print based editing space, it does clip some of the brightest highlights and light pastels that can be displayed on many monitors, but this is appropriate if you are concerned mainly with printed output, where these colors are generally not achievable.
Do a Google search on "photogamut rgb" and you will find a free download of the profile.
When you say highlights, what percentages are out of gamut? Is it under 5%? I would be very interested in this profile because I've never been a big fan of AdobeRGB because of its inherent limited gamut or density range. It sounds as though PhotoGamut would be a good photographic profile for large format applications ( i.e., inkjet ). In such cases, I prefer Perceptual rendering intent via Photoshop. When you say ProPhoto's baggage, are you refering to its 16bit application?
Thanks for the info, I'm going to check it out.
I cannot answer your question numerically, since I have not measured percentages or looked at all those individual points.
But I think the following should address your question. If you look at nearly any inkjet, press, lightjet or digital press profile in ColorThink or a similar 3D gamut plotting program, and compare it to PhotoGamut RGB, you will see that PhotoGamut RGB encompasses the entire gamut, including printable highlights, with VERY few exceptions. In those rare cases where the output device has a larger gamut, it is right at the fringes, so I'd be extremely surprised if anyone would notice it, even under close scrutiny. Even a bright white paper with a wide gamut, such as Epson Premium Glossy, is fully contained.
The highlights and light pastels that DO get clipped by PhotoGamut RGB are generally unprintable on just about any device/paper combination. sRGB, Adobe RGB and other matrix based editing spaces allow for additional bright highlights and light pastels that can be viewed on an RGB monitor, but they are well out of gamut on most printers. So, if monitor display is the most important consideration, (ie, web work, emails, digital picture frames, etc), they will provide a little extra snap.
But if you are focused on Print Only, or if printing and soft proofing are your primary consideration, PhotoGamut RGB is a great working space. You will normally see a lot less shift when soft proofing and smoother, easier transitions during conversion, since it is shaped like a typical printer color space, only larger. You'll have fewer data points that are way outside your printer gamut. It also has the benefit have having a wider gamut than Adobe RGB. To exceed the PhotoGamut RGB working space with a matrix based editing space, you'd need to move into the ultra wide working spaces, such as ProPhoto, Wide Gamut RGB, etc.
Hope that gives you the information you need.
I've also been using Photogamut for printed output for some time now, and one thing I've been bothered about is the fact that Photogamut's neutral values are not perfectly aligned with the L channel in L*a*b.
Granted, 99% of my images aren't affected by this, but recently I started printing some high-key dance portraits, and noticed a big color shift on the white background when I made even the smallest density adjustment on-screen. This color shift doesn't appear on the prints, only on-screen. And since our perception of color is altered by the surrounding colors, this may lead an editor to mistakenly correct for color shifting that really isn't happening at the printer.
Of course some of this banding is from my monitor (I'm using a LaCie 324) not being factory calibrated at every neutral RGB value (like an Eizo CG would be), coupled with the fact that monitor profiling only goes so far toward linearizing.
What's your take on this? Do you experience this issue too?
Thanks for bringing this up. I hadn't noticed neutral shifting in my work. I just opened up an 11 step grayscale which was created in Lab color space in Photoshop. It has 11 blocks of gray from Lab 0/0/0 to 100/0/0, each L* value being 10 points apart. In its native Lab space, of course, it reads dead neutral. If you convert to sRGB, Adobe RGB or any of the standard matrix based working spaces, each of the gray steps reads dead neutral (Lab's a* and b* coordinates remain 0/0, and in RGB, R=G=B).
I converted the same Lab file to PhotoGamut RGB and got the same exact result. In RGB, each gray value show R=G=B, and the Lab a* and b* coordinates remain 0. If I convert to an RGB printer profile, by contrast, most of the Lab readouts show a* and b* coordinates that are 0, but a few are off, and the RGB numbers are all over the board in an attempt to correct the data so it will "look" neutral when printed.
Above is theoretical and based on numbers in the profiles.
I also converted my Lab grayscale to PhotoGamut RGB and saved it as a small TIFF file. I brought the converted TIFF into ColorThink and graphed it. It wasn't a "perfectly" straight line, but it was pretty darn close to it. The largest deviations from perfect gray rendering were in the 3/4 tones and darker (about 70%K and darker). When it gets that dark, it becomes visually much more difficult to see differences in tone. It would be of much greater concern to me if gray rendering were askew in the midtones, 1/4 tones and highlights. The eye is very sensitive to slight color differences in lighter tones.
I converted the same Lab grayscale to both ColorMatch and Adobe RGB and plotted them in ColorThink too. ColorMatch, being a D50 space laid directly on the neutral axis. Adobe RGB, being a D65 workig space, was also a straight line, but displayed bluer (since D50 is usually used as a reference).
So, this is probably what you are seeing in your work. You have to give the edge to the matrix profiles for perfectly gray neutrals. Then again, they don't have some of PhotoGamut's strengths.
I guess I'd probably use a small matrix editing space such as sRGB for dedicated grayscale or near-neutral work if I needed an RGB file for web display or printing to an RGB device.
I use mostly sRGB and Adobe RGB depending on how colorful the original image is. I usually reserve PhotoGamut for unusually bright, colorful images with printable colors that get clipped using Adobe RGB.
Interesting study. Thanks again for bringing it up.
Lou Dina wrote:
>PhotoGamut RGB [...] is wider than Adobe RGB in printable colors
In my opinion, that is not correct, Lou:
1) When mentioning "printable colors", one must specify which CMYK destination device he means. Assuming a destination CMYK space of U.S. Web Coated (SWOP) v2, both PhotoGamutRGB and Adobe RGB *entirely* encompass it, with *no* portions whatsoever of that destination gamut left out.
2) PhotoGamutRGB extends to some areas of device-independent color that Adobe cannot reach; that is true. But Adobe RGB reaches into a far more substantial portion of colors that PhotoGamutRGB falls very short of. And, overall, Adobe RGB has a gamut volume that is 22% larger than that of PhotoGamutRGB.
One final observation: a wider-gamut source RGB working space is no "harsher" on an RGB-to-CMYK conversion than a smaller-gamut working space. Either source space may or may not force one to deal with issues of gamut compression, depending on image content. The difference is that the wider-gamut source RGB space will allow for the presence in the image of brighter, more brilliant and saturated colors that could conceivably be reproduced by better printing processes or it is possible to preserve in other end uses. Throw out those colors by using a smaller-gamut RGB working space, and you will no longer have that option available.
Also, we should all get over the *fact* that saturated colors get compressed in certain areas of the visible spectrum when converting to any of a number of CMYK device spaces. We are only trying to assuage our disappointment by starting from a less-saturated set of colors, but we are gaining *nothing* in the end.
I am going to have to disagree with you on many of your statements.
First, in reference to "printable colors" you picked a very narrow CMYK space indeed, ie US Web Coated SWOP. If that's your destination, then Adobe RGB covers it, and even sRGB is a good choice. We're not talking about such a narrow space. We are talking about finding a space that doesn't clip a wider gamut output space, without having to resort to an ultrawide space like ProPhoto. Look at any inkjet gamut for a wide gamut paper, such glossy, luster, the new fiber based papers, etc, and you will see that they extend well beyond Adobe RGB in those areas I am discussing. Yes ARGB has a wider gamut in certain areas, but most of those are bright highlights or shadows that are all impossible to print on ANY device. So, for the purposes of printing, they are useless. If you destination is the web, monitor, or email, then they are usable and worthwhile, as mentioned in my previous posts.
All you need to do is compare ARGB, PhotoGamut and a wide gamut high Dmax paper at the same time in ColorThink and you will see what I mean.
I also disagree with your comment that a wide gamut matrix space will be just as gentle when converting to a print destination profile. In PhotoGamut, the extreme endpoints found in Adobe RGB or ProPhoto, ie, unprintable bright highlights, pastels and other points, are simply unavailable, so during editing you don't end up with extraneous points way outside the printer gamut. If converting to your printer profile, especially using Perceptual intent, you will have smaller moves using PhotoGamut than you will with a large matrix space if you have a colorful image. You'd have to, since the table based PhotoGamut is a similar shape only larger, more akin to concentric circles that circles and Triangles.
With Relative Colorimetric, unprintable colors just get clipped and details along those fringes get clobbered. Editing in PhotoGamut already acknowledges the fact that paper and ink can not produce super bright highlights, bright pastels, etc, on any current print device. If you edit in ARGB or ProPhoto, you can and usually do retain these bright unprintable highlights, so you MUST have a shift when converting. Not necessarily so in PhotoGamut. Those out of gamut colors MUST be brought into gamut somehow.
BTW, I am not promoting PhotoGamut over ARGB, sRGB or most other "normal" working spaces. I use sRGB and ARGB most of the time. But, PhotoGamut does have some definate advantages over matrix spaces, especially for wide gamut originals destined for a wide gamut printer/paper combination. It also has a few potential disadvantages. Even Gretag recommends it as a working space in ProfileMaker 5 Help files for bright colored originals. It is more widely used in Europe than in the USA. Lack of common usage is one of its potential pitfalls, but only when dealing with people who don't understand color management.
So, depending on one's needs, sRGB, Adobe RGB and PhotoGamut all have their place, as do other spaces. I choose not to work with ProPhoto or wgRGB, except for extremely rare images or situations. My main use for ProPhoto is during RAW conversion with a bright, saturated original, when I know I will immediately convert to PhotoGamut and don't want to clip my colors by going directly from RAW to ARGB. If I have a typical low or medium gamut original, like you, I usually work in the smallest editing space possible.
All of my statements were accurate in my previous post if taken in context. When referring to "printable colors" I was speaking of the wide world of print, which includes Lightjet, Inkjet, Press, Digital Press, wide gamut papers and inksets, etc. US Web Coated SWOP is a very small subset of that universe.
I hope that clarifies my previous statements.
Lou Dina wrote:
> I am going to have to disagree with you [Marco] on many of your statements.
> First, in reference to "printable colors" you picked a very narrow CMYK space indeed, ie US Web Coated SWOP.
Since I saw no precise definition on your part of what "printable colors" you had in mind, I picked as an example the "narrow space" that typifies the most commonly-used 4-color-process printing scenario in North America.
Rather than criticizing my choice of an example, you could tell me exactly which printing scenario
had in mind, and we can then try to have an intelligent conversation about
> If that's your destination, then Adobe RGB covers it, and even sRGB is a good choice. We're not talking about such a narrow space. We are talking about finding a space that doesn't clip a wider gamut output space, without having to resort to an ultrawide space like ProPhoto. Look at any inkjet gamut for a wide gamut paper, such glossy, luster, the new fiber based papers, etc, and you will see that they extend well beyond Adobe RGB in those areas I am discussing.
Yes. So what? We just have to deal with that -- meaning, we must be the ones to apply intelligent compression to the image data. Soft-proofing is a perfectly suitable tool for that purpose.
Still, it all depends on the specific image content: any given image will rarely, if ever, cover the whole range of colors encompassed by the working space profile, whether this is wide- or narrow-gamut. Some images have no significantly out-of-gamut colors to begin with, and come far short of the furthest reaches of the embedded (or assigned) working space.
> Yes ARGB has a wider gamut in certain areas, but most of those are bright highlights or shadows that are all impossible to print on ANY device.
> All you need to do is compare ARGB, PhotoGamut and a wide gamut high Dmax paper at the same time in ColorThink and you will see what I mean.
This whole argument is misguided. Use a wide-gamut color space, carefully apply a few adjustment layers while soft-proofing (and simple masks, if necessary) to tailor the compression to the destination, and you will lose
by using the larger container of the wide-gamut working space, because you already have
in the file that you need for
destination RGB device space. It's just a matter of
b tailoring the image for the destination
by using appropriate combinations of adjustment layers and standard Photoshop editing techniques. Nothing too farfetched in any of this -- all of it is within the reach of any imaging professional.
There's no Perceptual RI available in RGB-to-RGB conversions? Then, have soft proofing + adjustment layers + standard editing be the substitute for the missing perceptual RI.
> I also disagree with your comment that a wide gamut matrix space will be just as gentle when converting to a print destination profile.
"Gentle" or "harsh" in this context is beside the point. Again, having your master image in a wide-gamut space
b doesn't cause any loss per se
in either dynamic range or saturated-detail areas. It may cause clipping only in a "dumb" RGB-to-RGB conversion, one that doesn't "tailor for the destination". But you
tailor the compression for your destination if you want to obtain good results, using soft-proofing, etc., as your tools.
> With Relative Colorimetric, they just get clipped.
Not if you control the results by "tailoring for the destination". The "clipped" detail comes back very quickly, and most of the time with very gentle and subtle moves that you must make
you convert to the smaller RGB space.
> BTW, I am not pushing PhotoGamut over ARGB, sRGB or most other "normal" working spaces.
But of course you are. What else does your argument mean, otherwise? Please give yourself some credit.
> [PhotoGamutRGB's] lack of common usage is one of its potential pitfalls, but only when dealing with people who don't understand color management.
Does that mean that disagreeing with your argument regarding the use of PhotoGamutRGB gets one branded as "not understanding color management"?
> So, depending on one's needs, sRGB, Adobe RGB and PhotoGamut all have their place. I choose not to work with ProPhoto or wgRGB, except for extremely rare images. My main use for ProPhoto is during RAW conversion when I know I will immediately convert to PhotoGamut and don't want to limit my color by going directly from RAW to ARGB.
is unwise: convert from ProPhoto RGB to PhotoGamut RGB without adjusting the image content first, and you are doing the very thing you advise against -- that is, you are possibly clipping detail in the image at the very outset, saturated and/or not.
Also, what about RGB device profiles whose gamut surpasses today, or will surpass tomorrow, that of PhotoGamutRGB?
> All of my statements were accurate in my previous post. When referring to "printable colors" I was speaking of the world of print, which includes Lightjet, Inkjet, Press, Digital Press, etc. US Web Coated SWOP is a very small subset of that universe.
The "world of print" so loosely defined is quite an undifferentiated mass. It's impossible to debate what one cannot identify.
Still, the point remains that -- no matter what the end use is, and as long as that destination RGB device space has a profile that describes it -- you are losing
by using wide-gamut working spaces in combination with soft-proofing, adjustment layers and careful standard editing moves.
The original poster listed the intended output as K3 inks, which, of course, means Epson inkjets. That's the only output that I ever inferred from the beginning of this discussion and that's the context I in which I read Lou's comments.
OK. Sometimes one can lose sight of exactly how forum threads got started. My apologies for that.
K3 inks alone were not what Lou was referring to, though, by his own admission.
Hi Marco. I'd like to ramp it down a notch. You have a lot of good information and insights to share and I appreciate your contributions.
My emphasis with PhotoGamut is wide gamut originals destined for output on devices that have a gamut wider than Adobe RGB. I thought that was clear in the previous posts, and if not, it could have been lack of clarity on my part.
My opinion, which differs from some people, is that that PhotoGamut RGB is generally preferable to working in an ultra wide matrix space, such as ProPhoto or wgRGB, when dealing with bright images that exceed the Adobe RGB gamut and are destind for print on a fairly wide gamut device. It was designed specifically for print.
I'm not sure what needs to be "ramped down." If you think that I have used an uncivil tone, please say so.
On the other hand, I think I understand your opinions. Yet I see no evidence that my own argument is getting across.
Marco, my friend,
I think that PhotoGamutRGB is a fine working space for print, for which it was rationally designed, for the reasons that Lou enumerated. I've said so for a long time. The fact that it's LUT-based means that you can take ProPhoto output from your RAW converter or archived files and gracefully move it into a better working space for print, without clipping and smashing detail as would happen if going to sRGB or aRGB. That can be useful, and I think it's preferable to trying to "reconstruct" detail by using soft proofing and adjustment layers after it's been clipped. sRGB was not designed as a working space for print, and it certainly has disadvantages. The same is true for AdobeRGB, and yes, even ProPhoto.
The gamuts of sRGB and aRGB are clearly too small for some images destined for print, and while ProPhoto certainly covers every CMYK and inkjet printer space currently in existence, I'm not crazy about having to use Perceptual conversions to bring colors from way outside the printer gamut into gamut. It makes sense to move the OoG colors into a working space that is closer to the destination space(s) for editing. I don't think this needs to be done with every image, but it's a useful tool that I've also used.
By the way, for those not aware of this, simply changing the output space in ACR from ProPhoto to sRGB (or aRGB) is the same as outputting the image in ProPhoto and then making the (RelCol) conversion to sRGB (or aRGB) - OoG colors will get clipped, and you can see this in the ACR histogram. If you want the "best" sRGB (or aRGB) image, without clipping, it is necessary to edit the image in ACR for sRGB (or aRGB) while watching the histogram to avoid clipping. Of course, you can't do this in Lightroom, since the histogram is based on Melissa and not the output color space.
I neither doubt nor question that PhotoGamutRGB is appropriate for converting either to a relatively narrow output space such as US Web Coated (SWOP) v2 or to the wider ones of a large number of
RGB inkjet devices.
But one of the issues here is whether or not there is a compelling reason to use PhotoGamutRGB rather than ProPhoto RGB as source working space in conversions to RGB output for inkjet printers -- also keeping in mind that tomorrow's inkjets will easily surpass PhotoGamutRGB.
As I have stated, there need be
i no downside
to using a wide-gamut RGB working space such as ProPhoto RGB for output to these devices,
i as long as certain precautions are being taken.
That is, as long as the move from ProPhoto RGB to the RGB inkjet profile is performed with the aid of adjustment layers, masks and other standard editing techniques performed under soft-proofing before the conversion is actually finalized.
It would certainly make
b no sense whatsoever
to "reconstruct detail by using soft proofing and adjustment layers
it's been clipped". As a friend, you know that I would never advise anyone to clip detail and then attempt a doomed "recovery" of something that is no longer there. Therein would lie madness -- and it's a misreading of my message, if that is what you think I meant.
One thing ought to be clear: no detail has yet been clipped (a) as long as the source file is still in ProPhoto RGB under soft-proofing
i before conversion,
and (b) while it's being "tailored for the RGB output". Clipping can only happen
conversion, and clipping of detail in the file is exactly what the adjustment layers, etc., are there to prevent before we commit to an actual conversion to the destination RGB output space.
I'll be glad to explain this workflow to you in greater detail if what I'm proposing still remains unclear.
I think I follow what you''re saying, but to me it sounds like a kludge.
"As I have stated, there need be no downside to using a wide-gamut RGB working space such as ProPhoto RGB for output to these devices, as long as certain precautions are being taken. That is, as long as the move from ProPhoto RGB to the RGB inkjet profile is performed with the aid of adjustment layers, masks and other standard editing techniques performed under soft-proofing before the conversion is actually finalized. "
I've mostly used PhotoGamutRGB in preparing CMYK images, and not inkjet prints, but my thoughts are still the same - wouldn't it be more elegant to convert to PhotoGamutRGB to pull the OoG colors in, rather than to use "adjustment layers, masks and other standard editing techniques?" I just don't find that kind of editing to be enjoyable.
No "kludge", Richard. It works, though it may not be enjoyable. <g><br /><br />> wouldn't it be more elegant to convert to PhotoGamutRGB to pull the OoG colors in, rather than to use "adjustment layers, masks and other standard editing techniques?" I just don't find that kind of editing to be enjoyable.<br /><br />If you use the perceptual intent, you may be sacrificing saturation. (You also must defer to the particular perceptual "flavor" used by PhotoGamutRGB's B2A0 tables, which may have its own limitations or idiosyncrasies. That could work like a "kludge" too, in a way, and not necessarily be the most elegant solution.) Instead, if you use relative colorimetric to preserve saturation, then you face issues similar to those in matrix-to-matrix conversions anyway.<br /><br />This whole subject needs to be looked at more extensively, in my opinion.<br /><br />Best.<br /><br />Marco
I am in synch with your thinking and approach on this subject, which you have expressed clearly and accurately.
Marco, I worked in ProPhoto fairly extensively for quite awhile and chose to abandon it, for reasons I stated earlier and restated by Richard. I agree that with care and awareness of the land mines, it is possible to get excellent results with ProPhoto. For the average user, it can lead to some unexpected surprises and even inferior results. In the hands of a knowledgeable professional, it is fine. I simply prefer PhotoGamut RGB for wide gamut originals destined for print on a device that exceeds the Adobe RGB gamut. It has fewer outliers and, in my opinion, is a better editing space for print.
When I prepare an image destined for press, I always make the final edits to the image in the final CMYK destination space, since that space defines the boundaries, limits and numbers for printable colors (for that paper/process/ink, etc). The destination profile, by definition, prohibits the creation of colors outside its gamut, so it is a great place to make final observations, and if necessary, adjustments. In other words, if the profile is a map that accurately describes the territory, it is a perfect fit.
In the same way, PhotoGamut is a "closer fit" to printed output than is ProPhoto. I have to worry a lot less about retaining highlight detail, since unprintable highlights are generally out of PhotoGamut's gamut.
Neither editing space is perfect or ideal. We all know that. You prefer ProPhoto and I prefer PhotoGamut.
I believe this thread has gone way off topic. The original question was whether PhotoGamut continues to hold up well in light of advances in output technology. To me, the answer is a resounding yes. Is it Perfect? No.
This will be my final post on this thread.
Lou Dina wrote:
> Marco, I worked in ProPhoto fairly extensively for quite awhile and chose to abandon it, for reasons I stated earlier and restated by Richard. I agree that with care and awareness of the land mines, it is possible to get excellent results with ProPhoto. For the average user, it can lead to some unexpected surprises and even inferior results.
True. But you are not an average user. I mean that as an acknowledgment.
At any rate, yes, the average user (meaning the non-professional) will not necessarily be aware of the pitfalls. I would not advise this type of user to adopt ProPhoto RGB without making it abundantly clear that there are definite and severe precautions to be taken.
> In the hands of a knowledgeable professional, it is fine.
Agreed. That was my point too.
> I simply prefer PhotoGamut RGB for wide gamut originals destined for print on a device that exceeds the Adobe RGB gamut. It has fewer outliers and, in my opinion, is a better editing space for print.
It seems adequate for the great majority of today's inkjets, and for aiming to print on presses as well. That is not in question.
> Neither editing space is perfect or ideal. We all know that. You prefer ProPhoto and I prefer PhotoGamut.
Given the choices, I see no necessary and inevitable downsides in ProPhoto RGB -- as long as one takes the precautions that I illustrated earlier.
> I believe this thread has gone way off topic. The original question was whether PhotoGamut continues to hold up well in light of advances in output technology. To me, the answer is a resounding yes. Is it Perfect? No.
It holds well today, but technology evolves rapidly, and tomorrow's needs will very likely call for a revision of this assessment. That is what I am inviting you and others to do.
> This will be my final post on this thread.
Yes, time to wrap it up.