Image file conversions from 5000K to 6500K

<p>Hi,<br>

I would like to automatically (through a software...?) convert my image files from 5300K to 6500K, in order to use them for the Web and other desktop publishing. As I primarily work for print on inkjet, I prepare these images with my screen set to 5300K (measured my paper white under JUST 5000K), gamma 2.2 and 250cd/m2 luminosity to match the paper.<br>

Thanks a lot in advance for your thoughts on this.</p>

<p>To answer your question, ACR or Lightroom would do this (the later quickly). WHY you’d want to do this is questionable. Also be aware that Kelvin values are a range of colors, not a single, specific color. So each app will treat the images differently and you’ll probably get differing image appearances than you desire. And images have <strong>nothing</strong> to do with the kelvin value you calibrate your display to. The display is totally independent of the images, the output devices etc. Just what are you hoping to do here?</p>

<p>Andrew, thank you very much for you prompt reply. Please excuse me for being imprecise. Here is what I meant to ask:<br>

Once I finished my work on the file, which means I printed the image, I would like to continue to use it in (for me) new ways, which is the web and other on screen presentations.<br>

My setup, as mentioned, is 5000K, gamma 2.2 and 250cd/m2. Because I have an other screen, which is calibrated to 6500K, I observed that the image does not look the same at all, it is way to blue. I can take any image software and kind of 'soft-proof' from screen to screen, until they match, but that's fastidious additional work.<br>

So my question is, isn't there a quicker, more rational way of doing that, meaning, transporting my visual choices for hue, saturation, contrast etc. from a 5000K monitor to a 6500K monitor so that the two images should look identical (or at least close)?<br>

Thanks again</p>

<p>250 cd/m^2, really? And that matches your paper? That's unusually bright.</p>

<p>If you had a color profile identical to sRGB except for having a 5300K white point, you could assign that profile to an image and then convert to sRGB using absolute colorimetric rendering intent. Really old versions of Photoshop could create such a profile, but I don't know of a good way to do that anymore. Perhaps I can create and post one.</p>

<p>Unfortunately most color management seems to default to perceptual intent, which uses the output device’s white point. The white point of these converted images would therefore look too red for almost anybody with a monitor white point of less than 6500K (including your own).</p>

<p>If not for both the perceptual intent default and near total lack of color management at all, it would be better to just embed the modified sRGB 5300K profile set to absolute colorimetric intent. While this is the correct way to do it, it will almost never work in practice.</p>

<p>Mark, yes 250 sounds bright indeed, but it matches (measured and confirmed through eyeballing) my viewing booth dimmed almost 50%. I never paid attention to that, or better forgot about that, but you are right it is bright, and I don't know why it is so bright. It is not to the point of the question, though it might make things more difficult when I try to convert my image files for the Web, as a classic sRGB profile assumes far lower luminosity.</p>

<p>Joe, I hear your advice with the embedded sRGB profile modified to 5300K with absolute color rendering intent. Just why wouldn't that work?</p>

<p>Mark, I forgot to say, that 250cd/m2 produces for me the closest match to paper. It's not perfect. I just posted a lengthy question about that. Please do look it up: <a href="../digital-darkroom-forum/00WWVn">Matching paper white not satisfying even after hard work... </a> (<a href="../digital-darkroom-forum/00WWVn">http://www.photo.net/digital-darkroom-forum/00WWVn</a>).</p>

<blockquote>

<p>Here is what I meant to ask:<br />Once I finished my work on the file, which means I printed the image, I would like to continue to use it in (for me) new ways, which is the web and other on screen presentations.</p>

</blockquote>

<p>This again has nothing to do with the kelvin values defined in a working space. In an ICC aware app, with whatever working space you have, it either looks good or doesn’t. If it does, fine. Then if you need to repropose it for another use or output, you convert to that output color space (ie Adobe RGB to myprinter RGB). If you want to convert for the web, you take whatever RGB working space you currently have, with no regard to the color temp (white point) and convert to sRGB. </p>

 

<blockquote>

<p>My setup, as mentioned, is 5000K, gamma 2.2 and 250cd/m2.</p>

</blockquote>

<p>That’s the targets you calibrate your display to, based hopefully on the way you view your prints. It has zero to with the documents themselves. ICC aware applications divorce the display from the images you edit. That’s why we have RGB working spaces which are solely independent from the display calibration and profile. </p>

 

<blockquote>

<p>Because I have an other screen, which is calibrated to 6500K, I observed that the image does not look the same at all, it is way to blue.</p>

</blockquote>

<p>Then change the targets for the display calibration. Naturally two displays, calibrated and profiled to differing white points will look different. But you don’t alter the documents to produce a match. You alter the target calibration of the display. </p>

<blockquote>

<p>Joe, I hear your advice with the embedded sRGB profile modified to 5300K with absolute color rendering intent. Just why wouldn't that work?</p>

</blockquote>

<p>It would work if the file were opened in a color managed application that doesn't override the absolute colorimetric rendering intent in the embedded profile. Very few applications will meet both of those criteria.</p>

<p>Andrew, thank you for the detailed observations.<br>

I hopefully, follow you now. That means that (I am repeating and paraphrasing, to be sure to understand)<br>

- I don't need the second display calibrated to whatever target settings I assume for my output (in this case the Web), I just continue to have all my targets set in a way that they look good to me (and match the paper)<br>

- the files that I selected for the web simply need to be to be transformed to the output color space, meaning sRGB.<br>

Now, just one follow-up question on this, and this is was led me into the mistake (I assume): Why do I than read that for Photo editing, with the main intent being on screen presentation, should have the editing display targets set to 6500K with a Gamma 2.2 ? Is it not because this colortemperature and gamma settings are (somehow?) closer to the sRGB colorspace, and that every conversion, be it from one color space to an other, deteriorates the image (posterising, clipping, etc.)?</p>

<blockquote>

<p>I don't need the second display calibrated to whatever target settings I assume for my output (in this case the Web), I just continue to have all my targets set in a way that they look good to me (and match the paper)</p>

</blockquote>

<p>You can try. But unless both displays are very similar, don’t expect identical color appearance (especially true if one is a standard sRGB like display, the other an Adobe RGB like display). On Windows, depending on the flavor and app, dual display support can be iffy. Most use the 2nd display for palettes. For that usage, calibrating and profiling it seems more work than necessary. </p>

<p>The web is not color managed unless you use one of two browsers. If you do, they will match on each display that has a profile, what you saw in Photoshop. Other browsers will not. </p>

 

<blockquote>

<p> the files that I selected for the web simply need to be to be transformed to the output color space, meaning sRGB.</p>

</blockquote>

<p>Yes. Users that work with calibrated and profiled displays and ICC aware web browsers will see what you saw. Those who don’t will not. </p>

 

<blockquote>

<p>Why do I than read that for Photo editing, with the main intent being on screen presentation, should have the editing display targets set to 6500K with a Gamma 2.2 ?</p>

</blockquote>

<p>Because some have to give you a guideline to start, even if your mileage may vary. In terms of TRC Gamma, 2.2 is a good setting as nearly all displays behave closely to this target. For white point, it all depends on the viewing conditions of the print. 6500K (better, D65 which isn’t the same) is a good starting point. But since we have no idea what the illuminant you are using to view your print is, how warm or cool it may be, that D65 setting may or may not match. Too cool? Use a lower setting (ie D60, D50). Adjust to taste. The only true D50 (or D65) Lightsource is 95 million miles from your display. </p>

 

<blockquote>

<p>Is it not because this colortemperature and gamma settings are (somehow?) closer to the sRGB colorspace, and that every conversion, be it from one color space to an other, deteriorates the image (posterising, clipping, etc.)?</p>

 

</blockquote>

<p>No, because again, the RGB working space is a Quasi-Device Independent color space you edit your images in and is totally separate from how we calibrate our displays or our output devices (see:<a href="http://www.adobe.com/digitalimag/pdfs/phscs2ip_colspace.pdf">The Role of working spaces in Adobe applicaitons</a>)</p>

<blockquote>

<p>Joe, I hear your advice with the embedded sRGB profile modified to 5300K with absolute color rendering intent. Just why wouldn't that work?</p>

</blockquote>

<p>Actually this does work in Safari, but not Firefox or any other web browser I know of.</p>

<p>I’m starting to like that idea because it is the right way to do what you want, but it still won’t work very often.</p>

<p>This image is pure white with an embedded profile that is 5300K white point and absolute colorimetric rendering intent but otherwise like sRGB. It will probably look native white in everything but Safari. Programs like Photoshop may ask how to handle the embedded profile.</p>

<p>Clicked too fast past the option to attach an image.</p><div></div>

<p>Andrew,<br>

I have my JUST 5000K viewing booth and ambient light. But let's say somebody wants to set up a darkroom from scratch. I have read quite some times, that when editing photographic images, people should set their screens to 6500K and their viewing conditions also. Why should that be better than setting the target to 5000K and using a 5000K viewing booth?</p>

<p>Joe, thank you!<br>

That's very interesting. Unfortunately, as you say, it does not work.</p>

<blockquote>

<p>I would like to automatically (through a software...?) convert my image files from 5300K to 6500K, in order to use them for the Web and other desktop publishing. As I primarily work for print on inkjet, I prepare these images with my screen set to 5300K (measured my paper white under JUST 5000K), gamma 2.2 and 250cd/m2 luminosity to match the paper.</p>

 

</blockquote>

<p>sRGB color space is D65 white point.<br>

Image color space has nothing to do with monitor color space.<br>

For example:<br>

1- if your image color space is sRGB or Adobe 1998, the white point is D65<br>

2- if your image color space is ProPhoto, the white point is D50<br>

3- and so on ......</p>

<p>For web publishing a transform to sRGB (if image is not in sRGB) must be performed, this transform compensate for any original white point. This is what color management do.</p>

 

<p>that's a clear answer Jacopo!<br /> As image color space has nothing to do with monitor color space, why is it, that according to what I understood till now, preferable to work in light close to D65 when working for the Web, close to "native" white point of the sRGB and Adobe 1998 color spaces?</p>

<blockquote>

<p>I have my JUST 5000K viewing booth and ambient light. But let's say somebody wants to set up a darkroom from scratch. I have read quite some times, that when editing photographic images, people should set their screens to 6500K and their viewing conditions also. Why should that be better than setting the target to 5000K and using a 5000K viewing booth?</p>

</blockquote>

<p>Got to get past the numbers. They are often meaningless or at least Your Mileage May Vary. In the old days, people figured they had 5000K viewing booths (mostly Fluorescent) so it made sense to calibrate their CRTs to 5000K too. Problem is, they got a dim, yellow appearing display. The numbers didn’t work. 6500K did. But again, YMMV. You might calibrate to D55, D65, even calibrate to a custom set of X/Y chromaticity values which some products let you do. The numbers are only correct when you get a match! Since the viewing conditions are often undefined, and since different products claim to be this or that kelvin value (keeping in mind a kelvin value is a range of colors), no wonder people who expect exact numbers to produce a match are disappointed. </p>

<p><strong>IF</strong> you could heat your display to 5000K, <strong>and</strong> it behaved like a theoretical black body radiator, you’d get a match and a pool of molten plastic on your desktop. </p>

<p>So, to resume, back to my original question, which I would like to try to ask differently now:<br /> When converting image files for the Web, should there be a difference regarding Hue, Contrast, Saturation... between an image file produced in my work flow, entirely close to D50 (Sensor, Strobes, ProPhoto, Monitor, Light booth, ambient light...) to an image file edited in a setup closer to D65 (and assume I would be comfortable with both setups)?</p>

<blockquote>

<p>preferable to work in light close to D65 when working for the Web, close to "native" white point of the sRGB and Adobe 1998 color spaces?</p>

</blockquote>

<p>You have to work with a monitor white point that is "comfortable".<br>

For people "comfortable" means 6000/6500K (near D65).<br>

This is true not only for web.</p>

<p>I do not understand why you assume that</p>

<blockquote>

<p>For people "comfortable" means 6000/6500K (near D65).<br /> This is true not only for web.</p>

</blockquote>

<p>For me I work in a 5000/5500 (near D50) environment, my eyes adapt and white is white (well screen to print is a different story). I think it is relative. Proof is that when I am, which happens to me, in a 6000/6500 (near D65) environment I also find it perfectly acceptable after adaptation. The problem of preferring D65 over D50 or D55, or the other way round, only arises for me when confronted with a white point different from that of the rest of the environment to which my eyes already adapted.</p>

<p>Generally near D50 is peceived as yellowish, as the chromatic adaptation is not complete.<br>

But if you are satisfied, you can stay with your values.<br>

I that case, remember to look at web images using a color managed browser.<br>

This is a good suggestion in any case.</p>

<p>BTW on the Solux Web site you can find their account of the tests they did to find out under which color temperature people really prefer to look at reflective images, in this case art work in a museum. 3500K that is. This preference does evidently say nothing about the fact, whether those people experienced white as white, which is what we discuss here.</p>

<p>Jacopo, could you direct me to more information about that last point you made, about the fact that chromatic adaptation in not complete near D50? You are saying that chromatic adaptation to monitor target value of 6500K or 6000K is better in that respect?</p>