Lambda print color space

<p>Hi,<br />recently I decided to do some Lambda prints on Fuji Crystal DP II through WhiteWall.<br />They ask to submit files in sRGB and I'm a little puzzled:<br />since the gamut of this machine+paper is bigger than sRGB, am I loosing part of the colours ?<br />I wrote them an email asking why wouldn't it be better if I send them a file converted to the paper profile<br />(embedding the profile too) and got this reply:<br /><br />"We can accept any RGB or CMYK color space, but if no ICC profile is embedded we will asume it is sRGB.<br />We do not advice you to use our ICC profiles as the colour space, the conversion to the destination output should be done by us to maximize image quality. The ICC profiles we provide are suitable for softproofing, but your file should not be converted to it's colour space. if you prefer a large Gamut working space Adobe RGB is a better option, but you must make sure the ICC profile is embedded so it is treated as Adobe RGB, not sRGB."<br /><br />Why converting to a narrower gamut in the first place ? (AdobeRGB is slightly narrower in places compared to the paper profile of the Fuji Crystal DP II that they provide for soft proof).<br /><br />Wouldn't it be better sending a file in its working space (ProPhoto in my case) and let them do the conversion ?<br /><br />Your thoughts ?<br />Thanks,<br />Giovanni.<br /><br /></p>

<p>If I read the part you quote, that is exactly what they ask you to do: send the image with a proper image colour profile (sRGB or AdobeRGB are the usual choices, but I see no reason why ProPhotoRGB wouldn't work). They just explicitely ask to *not* assign a printer colour profile to the image, which in my view sounds correct.<br /> But maybe I am reading it wrong, could be. You could just ask them if they are OK with a file with ProPhotoRGB as colour profile to make sure.</p>

<p>Hi Giovanni<br>

They prefer JPEG or TIFF in 8 bit and yes they can do any color space if you have it embedded.<br>

A couple things to consider however.<br>

Printer spaces are somewhat a moving target. No doubt over time they update their own ICC profiles for ideal printing. Now that should match the ICC profile for soft-proofing yet that is only the case if they update those files on their website. They may consider the ICC profiles on their site "close enough" and may not update those as often.<br>

sRGB, aRGB, and ProPhoto RGB are fixed and are not changing by comparison.<br>

Note that if you use ProPhoto RGB (and some degree aRGB) and are submitting a JPEG or 8 bit TIFF (instead of 16 bit), you run the risk of having banding added to the image as the 256 colors per channel must be stretched out over a wider gamut. So that is a tradeoff as well.<br>

Finally, it only matters if you image has colors that are outside of the sRGB color space. If not, submitting is sRGB is just fine. You could verify this with soft proofing.<br>

Hope the incremental information is helpful.</p>

<blockquote>

<p>since the gamut of this machine+paper is bigger than sRGB, am I loosing part of the colours ?</p>

</blockquote>

<p>Potentially yes, you are clipping saturated colors that may (and often do) exist that lie outside sRGB. <br>

This lab‘s writing is suspect and makes little sense. They say:</p>

<blockquote>

<p>We can accept any RGB or CMYK color space, but if no ICC profile is embedded we will asume it is sRGB.</p>

</blockquote>

<p>OK, so they accept <strong>any</strong> RGB color space which would include the output color space produced by using an ICC profile of their process. Assuming sRGB for untagged data makes sense. But...</p>

<blockquote>

<p>We do not advice you to use our ICC profiles as the colour space, the conversion to the destination output should be done by us to maximize image quality.</p>

</blockquote>

<p>That invalidates the first sentence! And the bit about them doing it maximizes image quality is nonsense. <br>

Do they even supply an output color space profile for the printer? If so, you’d want to use that to control the rendering intent, for soft proofing, for the use of Black Point Compensation etc. </p>

<blockquote>

<p>The ICC profiles we provide are suitable for softproofing, but your file should not be converted to it's colour space.<br /></p>

</blockquote>

<p>I’ll decipher: <em>We supply a profile we hope you believe can be used for soft proofing but since it’s not used for output, soft proofing is hit or miss. </em><br>

IOW, they don’t have a clue about Color Management but want their customers to believe they do by providing a profile that probably doesn’t describe their actual output conditions. So don’t use it. Nor can you tell them how you setup the soft proof in terms of rendering intent such they will use it as well for the output. Find another Lab. </p>

<blockquote>

<p>Why converting to a narrower gamut in the first place ? (AdobeRGB is slightly narrower in places compared to the paper profile of the Fuji Crystal DP II that they provide for soft proof).<br /></p>

</blockquote>

<p>Exactly, no justification. They should either fully support a real, ICC, Color Management workflow (they don’t) or just tell you that sRGB is the only color space they will accept. Forget soft proofing. But they want their cake and eat it too, making up a ridiculous half baked Color Management workflow. Find another Lab, tell them you’re not accepting this CMS BS. </p>

<blockquote>

<p>Printer spaces are somewhat a moving target.</p>

</blockquote>

<p>They <em>can be</em> (a modern Epson for example isn't, even when you look at thousands of color measurements from a dozen of the same models). <br>

Printers that are a moving target <strong>need to be controlled</strong>, simple as that. The RGB or CMYK values you send today and in a year should produce the same visual output IF the people running the equipment are handling their process control properly. Just a Pro Lab’s running say E6 did for years and years. Imagine taking your film for processing today, asking for a snip test. Then running the balance the next day and the color was different because such processes ARE moving targets but the lab didn’t replenish and produce consistent behavior. It would be a disaster. Good labs control the process such that the moving targets don’t move when they are actually running the jobs. </p>

<p>There are too many assumptions with tech here. I have worked for printing houses and their machines are often quite old, although well maintained. The software (at that place) was typically run on Windows 2000. The manufactures will not maintain anything that was not in the "when delivered" configuration, which puts the machines locked in time. With that in mind, what they were asking you to do was not that much of a stretch. I was working as a programmer and was writing software to bridge the gap. The place you are dealing with may not see any gap that needs to be bridged.</p>

<p>Hi Giovanni - a couple more thoughts.<br>

In your comment:</p>

<blockquote>

<p>Why converting to a narrower gamut in the first place ? (AdobeRGB is slightly narrower in places compared to the paper profile of the Fuji Crystal DP II that they provide for soft proof).</p>

</blockquote>

<p>I pulled up the ICC profile in ColorThink Pro and Adobe RGB really covers virtually all of the ICC profile for the Fuji Crystal DP II.<br>

So unless you need to know for academic reasons why they don't want you to convert to their ICC profile, just use Adobe RGB. 8 bit should be pretty good (RGB or TIFF) and if you have concerns about banding give them a 16 bit TIFF.<br>

Adobe RGB is considerably wider gamut than the ICC profile for the printer of interest. I would used soft proofing in your workflow. Not all images have colors that would need to be remapped to the smaller printer gamut yet some do and always good to have a heads up on what colors would needed to be shifted to be "in gamut." If that step is needed, you have more control if you bring it in gamut in LR or PS rather than depending on the Labs conversion process.<br>

All IMHO.</p>

<p>I agree with John that Adobe RGB (1998) is a more appropriate space for this output device in which to send the data to the lab and hopefully the original was in something larger like ProPhoto RGB (because there are plenty of output devices that do exceed Adobe RGB gamut). <br>

I see zero reason to soft proof with <strong>anything</strong> but the output profile used to convert the data for the actual output device. </p>

<p>Andrew,<br>

yes, the original was edited in ProPhoto. I assumed that the profiles they provide were suitable to do a decent soft proof....</p>

<p>@Andre Rodney:</p>

<blockquote>

<p>Exactly, no justification. They should either fully support a real, ICC, Color Management workflow (they don’t) or just tell you that sRGB is the only color space they will accept.</p>

</blockquote>

<p>My thinking is they assume 90% of people has a regular, small gamut display and they don't want to disappoint customers and play it safe telling them staying in sRGB so they print match their screen.<br>

What about the other 10% who actually owns a large gamut display and want to exploit the full gamut of the paper they choose and are able to actually see those colors ? Even if it is buying me that extra few colors (at no extra cost indeed), why shrinking the whole thing down to sRGB ?<br>

I'm totally up with richseiling (West coast imaging printing) when he says in another post:</p>

<blockquote>

<p>What I am relating is first hand experience working extensively (hundreds of prints pushed to the limit for my fine art clients) in AdobeRGB, and I can tell you that visually that "tiny bit of yellow and magenta" equals a whole lot of colors that are very meaningful to the eye that AdobeRGB just can't reproduce. But in reality it is more than just those colors...</p>

</blockquote>

 

<blockquote>

<p>My thinking is they assume 90% of people has a regular, small gamut display and they don't want to disappoint customers and play it safe telling them staying in sRGB so they print match their screen.</p>

</blockquote>

<p>But that’s not going to happen! The device isn’t producing sRGB, there is no such printer on this planet. They convert <strong>from</strong> sRGB <strong>to</strong> some output color space for this device. They don’t allow you to do this. Makes no sense. The only way to soft proof to this or <strong>any</strong> output device is to have the ICC profile that defines the output and it’s not sRGB or Adobe RGB or ProPhoto RGB. It’s “<em>Lambda RGB</em>” based on the profile used to convert that data and using a <strong>specific</strong> rendering intent.</p>

<blockquote>

<p>Even if it is buying me that extra few colors (at no extra cost indeed), why shrinking the whole thing down to sRGB ?<br /></p>

</blockquote>

<p>Because they are using a fast and dirty conversion process where they assume all data is in sRGB so that’s all they allow. Color Management isn’t on their radar. <br>

This isn’t about an RGB Working Space per se, it’s about an output color space, defined by an output profile you use to convert the data and soft proof that you are not allowed to use. </p>

<p>Are we reading too much between the lines of a communication from the company about their ICC profiles being only for soft proof.<br /> <br />For the price performance point that Whitewall offers they give pretty decent information in their FAQ about how to work with them: https://us.whitewall.com/service/Faq-image#AaiB-5<br>

<br /> The link includes 19 ICC profiles for the various machine/paper combinations, they accept data from any RGB color space (and CMYK as well), they walk you through the settings for soft proofing including which rendering intent to use and black point compensation (I assume to get the best match with what they print).<br /> For this price/performance point of lab, of all those that I know using Whitewall, I have not heard a problem with quality or color matching issues.<br>

<br /> In the search for practical/actionable information, are there better labs at these price/performance points that do better including online info to make working with them easy (aka at least as much info as Whitewall)? I certainly know a lot of labs at a similar price point that do a lot worse. I am sure there are higher price point Labs do better (and no doubt some higher price labs that do no better).<br>

<br /> Why not follow their FAQ instructions with an image in Adobe RGB, send them a print, and evaluate it when you get it back. Though it may not speak to repeatability of results, it certainly would add some actual results to the discussion.<br>

<br />All IMHO of course</p>

<blockquote>

<p>Are we reading too much between the lines of a communication from the company about their ICC profiles being only for soft proof.</p>

</blockquote>

<p>I don’t think so. IF one can soft proof, one should be able to convert. And pick a Rendering Intent which I would hope you’d agree is important and plays a role in what you end up seeing AND getting. That’s not possible with a ‘<em>use the profile only for soft proof</em>’ workflow. <br>

Great, they allow one to send CMYK, the totally wrong color space for the output. <br>

As I said, for the price, I have no issue with a lab forcing an sRGB workflow on the files they accept. None at all. What I object to is the <em>idea</em> of providing profiles that don’t do what they were designed to do, and the impression that by using ICC Profiles, these lab’s are implementing a sound color managed workflow. They are not. It’s like printing a book from Aperture or Lightroom, you send the data ‘<em>as is</em>’ and you get what you get. They are not suggesting a color managed process in terms of soft proofing and converting and send the data in a fixed RGB Working Space as these other lab’s <em>could</em> and <strong>should</strong> simply do as well. </p>

<blockquote>

<p>Why not follow their FAQ instructions with an image in Adobe RGB, send them a print, and evaluate it when you get it back.<br /></p>

</blockquote>

<p>Further, do that twice, a month apart and let us know if they match. They should of course match, the same RGB numbers, unlike cheese stored in the sun, change over time. <strong>Repeatability is critical and should be expected from customers. </strong><br>

</p>

<blockquote>

<p>IF one can soft proof, one should be able to convert. And pick a Rendering Intent which I would hope you’d agree is important and plays a role in what you end up seeing AND getting. That’s not possible with a ‘<em>use the profile only for soft proof</em>’ workflow. </p>

</blockquote>

<p>I agree. Whitewall does allow that in their workflow. You can convert to their ICC profile with whatever rendering intent you desire. It was only mentioned in an email not to do that that raises eyebrows. I am just suggesting to try it out with a real print as one would to test out any print lab.</p>

<blockquote>

<p>Great, they allow one to send CMYK, the totally wrong color space for the output.</p>

</blockquote>

<p>I agree, I almost winced including the CMYK part of the comment as copying what they had in their FAQ</p>

<blockquote>

<p><br /> As I said, for the price, I have no issue with a lab forcing an sRGB workflow on the files they accept. None at all.</p>

</blockquote>

<p>They are not forcing an sRGB workflow from anything I read. Just that they are assuming an sRGB profile when one is not embedded.</p>

<blockquote>

<p>What I object to is the <em>idea</em> of providing profiles that don’t do what they were designed to do, and the impression that by using ICC Profiles, these lab’s are implementing a sound color managed workflow. They are not. It’s like printing a book from Aperture or Lightroom, you send the data ‘<em>as is</em>’ and you get what you get. They are not suggesting a color managed process in terms of soft proofing and converting and send the data in a fixed RGB Working Space as these other lab’s <em>could</em> and <strong>should</strong> simply do as well.</p>

</blockquote>

<p>Again, I think we might be reading too much into a single email exchange from Whitewall and am suggesting testing the print lab via prints which can include printing in the supplied ICC profile for the printer/paper combination with ones choice of rendering intent. Then see if one likes the results. (multiple prints over time is a nice addition for repeatability).<br>

<br />From following you for many years Andrew, I do have a pretty good feel for what print lab characteristics to avoid. However, though it may have been buried in my previous post, I am looking for the Andrew Rodney Good Printing Seal of Approval on a list of labs that I hope you could provide in different price/performance classes. Now that would be really valuable for a lot of folks to know what specific labs "to go to", not just "what characteristics to stay away from."<br>

A consumer reports rating chart of print labs from you would even be better. IMHO that would be an amazing practical output from your expertise that would be used very widely (and many might pay money for). If you already have that, a link would be appreciated.</p>

 

<blockquote>

<p>A consumer reports rating chart of print labs from you would even be better.</p>

</blockquote>

<p>IF someone wants to pay for output on a regular basis from any said lab, I’d be happy to setup a target and a process to produce a metric of how well they hold their calibration and where in color space they deviate from the profile’s expectation (per rendering intent). <br>

OR find a lab that is using Chromix’s Maxwell system to do this:<br>

http://www.chromix.com/maxwell/index3.cxsa<br>

Maxwell is doing what I’d do manually using an iSis with a target of my design and ColorThink Pro to produce reports. </p>