LUT Display Profiles: Lightroom Doesn't Like 'Em

<p>I recently got a wide-gamut Dell U2711 monitor, which forced me to upgrade from my i1 Display 2 to a ColorMunki for monitor profiling.</p>

<p>ColorMunki display profiling software is childish in its simplicity, so I installed Argyll & use dispcalGUI for my profiling needs.</p>

<p>I created 2 profiles for my monitor using dispcalGUI + ColorMunki, a matrix & a LUT profile, both high quality using the maximum number of color patches available for each profiling method (64 & 2386, respectively). I used the adaptive mode for the Munki, which increases integration time for dark patches (I believe), leading to a more accurate profile for darker colors (else the Munki's profile severely brightens shadows, but that's for another discussion).</p>

<p>My problem is, when the matrix profile is loaded, Lightroom (LR) & Photoshop (PS) both render images exactly the same. When the LUT profile is loaded, LR really messes up smooth color gradients (banding everywhere). PS renders the image just fine.</p>

<p>To demonstrate my point, I took screen shots of the same color gradient (the calibration target file from MPix) at 100% in LR (Develop module) & in PS, then converted both images to sRGB, & compiled this comparison:</p>

<p><img src="http://rishisanyalphotography.com/ForumPostFiles/photo.net/LUT%20Profiles%20-%20PS%20vs%20LR.jpg" alt="" width="800" /><br>

Here's a <a href="http://rishisanyalphotography.com/ForumPostFiles/photo.net/LUT%20Profiles%20-%20PS%20vs%20LR.jpg">link</a> to the full-resolution file. Please view at 100%.</p>

<p>Some of the banding was lost probably due to the color space conversion I had to do, but you should still be able to see the banding in the gradient on the right. It's actually much worse than what you see here.</p>

<p>Does LR not like LUT display profiles? I asked this before back in 2007 on Adobe's forums but never got a straight answer: <a href="http://forums.adobe.com/thread/349454">http://forums.adobe.com/thread/349454</a></p>

<p>Here's a <a href="http://rishisanyalphotography.com/ForumPostFiles/photo.net/Dell%20U2711%20B14C60_R98G91B92_D65G2.2_120cdm2%20HQ-XYZLUT_2386Patches_Munki.icc">link</a> to the LUT ICC profile itself for my display. I believe it's V2, not V4.</p>

<p>Any thoughts? Thanks in advance,<br>

Rishi</p>

 

<p>Rishi,</p>

<p>Whoops! Had to edit. Just saw your link to the original LUT profile. I'll give it a look in Colorsync Utility.</p>

<p>Tim -- good to see you again! I believe we had quite an interesting debate over profiling a MacBook Pro LCD & getting poor blues... I'll have to check if this still happens with my new MBP (late 2010) & with my Munki as well as with my old i1 Display 2 with a custom correction matrix (which I can make with my Munki!)</p>

<p>Anyway, since you asked for it & I already did it, here are the individual screenshot files for LR & PS, with the LUT ICC profile (it's the same profile for both, since I was tryina point out that it was strange that LR & PS render things differently using the <em>same</em> display profile!):</p>

<p>Lightroom Screenshot: <a href="http://rishisanyalphotography.com/ForumPostFiles/photo.net/LR_LUTProfile.jpg">Click here</a><br /> Photoshop Screenshot: <a href="http://rishisanyalphotography.com/ForumPostFiles/photo.net/PS_LUTProfile.jpg">Click here</a></p>

<p>Perhaps it'll be more obvious now without the conversion to sRGB.</p>

<p>Again, this is weird, b/c if LR & PS are using the same color engine for conversion to display, how can there be a difference?</p>

<p>Cheers,<br />Rishi</p>

<p>LR most certainly has issues with V4 profiles, check that the Munki which used to build V4 by default isn’t. V4 profiles=bad!</p>

<p>Hi Andrew,</p>

<p>I know that the Munki software was making V4 profiles, but Argyll/dispcalGUI shouldn't be. Especially b/c iccdump (an Argyll binary) which can ONLY read V2 profiles <strong>can</strong> read my dispcalGUI profiles but <strong>can't</strong> read my Munki software-generated profiles (complaining that they're V4 files).</p>

<p>So I'm 99% sure these are V2 profiles. Here's a <a href="http://rishisanyalphotography.com/ForumPostFiles/photo.net/Dell%20U2711%20B14C60_R98G91B92_D65G2.2_120cdm2%20HQ-XYZLUT_2386Patches_Munki.icc">link</a> to the profile if you wanna check for yourself.</p>

<p>Why should LR & PS render the images different if they're both using ACE? Actually, I tried using Apple's color engine in PS & even then I didn't get the banding. So I dunno what the heck LR is doing.</p>

<p>Aren't LUT profiles recommended for wide-gamut displays? Seems like a glaring omission for LR not to support them properly...</p>

<p>Cheers,<br>

Rishi</p>

<blockquote>

<p>Aren't LUT profiles recommended for wide-gamut displays? Seems like a glaring omission for LR not to support them properly...</p>

</blockquote>

<p>Not necessarily, especially with displays that conduct the calibration internally in high bit. That argyle profile is V2 according to the tags. What about the Munki profile? </p>

<p>As far as I could tell, the Munki software didn't even allow me to make anything but the simplest display profile. Which is why I spent about 5 minutes on it then ditched it for more intricate software...</p>

<p>So is the final answer simply: <em>don't use LUT display profiles</em>?</p>

<p>On my Dell U2711, I adjust the backlight & the R, G, B gains to reach desired luminance & white point (usually with a dE of ~0.5, though that varies up to 2.5 when I sample different areas of the screen...), so at least that much is being done in the internal 12-bit LUT.</p>

<p>The rest of the calibration is done by the vcgt portion of the calibration/profile Argyll builds.</p>

<p>What's the advantage of a LUT profile for a display? As I understand it LUT profiles are better for nonlinear devices, but does a display profile benefit from being LUT as opposed to matrix?</p>

<p>Just for fun, here's the Dell U2711 (wireframe) gamut vs. Adobe RGB:<br>

<img src="http://rishisanyalphotography.com/ForumPostFiles/photo.net/DellU2711vsAdobeRGB.jpg" alt="" /></p>

<p>And here's the Dell U2711 (wireframe) gamut vs. Epson R2400 Premium Luster:<br>

<img src="http://rishisanyalphotography.com/ForumPostFiles/photo.net/DellU2711vsEpsonR2400Luster.jpg" alt="" /></p>

<p>Cheers,<br>

Rishi</p>

<p>Again, for some display systems, like my NEC SpectraView’s, you don’t even have an option (nor do you need) a LUT based profile. You are correct, on a very non linear system, the LUTs would be preferable. If there’s some issue in an app with LUT profiles, well make a Matrix profile and if said display is well behaved, you’re even better off. </p>

<p>I’ve not heard of specific issues with LR and Matrix profiles but do know there are issues with V4 profiles. And whenever you evaluate a preview in LR, it should be in the Develop module at 1:1. The other modules use a different (faster, lower quality) preview mechanism. </p>

<p>Well, I made a matrix profile, and it doesn't have these banding issues. </p>

<p>But I really don't know how to tell if the display is 'well behaved' with this matrix profile... you mean by 'verifying' (validating) the profile?</p>

<p>My log file from the profiling process says that the max detected dE was ~1.4 & the average was like 0.03 or something really low.</p>

<p>Thanks,<br />Rishi </p>

<blockquote>

<p>But I really don't know how to tell if the display is 'well behaved' with this matrix profile... you mean by 'verifying' (validating) the profile?</p>

</blockquote>

<p>No, that’s kind of a useless process. Well behaved would be a unit that produces the calibration inside the panel in high bit, instead of in the vcgt or graphic system. </p>

<p>So then for my 'not-well-behaved' Dell U2711, would I see any benefit from a LUT profile? Or is that just hard to say?</p>

<p>Well it sounds like you are saying there’s an issue with the LUT profile you build and LR so the answer seems to be no. </p>

<p>Ok, sorry, let me rephrase the question: if LR didn't have any issues with LUT profiles, like PS doesn't, might I see any benefit from using LUT display profiles with wide-gamut (or any) displays?</p>

<p>Also, do you have any experience using the so-called 'adaptive' mode with spectroradiometers for display profiling? The process takes a lot longer (takes me 3 hours per profile) but I'm guessing that the increased integration time (i.e. exposure) for the darker colors yields more accurate results since the spectroradiometer has a hard time 'seeing' dark colors (such low S/N) as you pointed out in a previous thread.</p>

<p>Cheers,<br>

Rishi</p>

<p>Go with the matrix profile if it gets rid of the banding in LR. What other color managed display preview advantage do you gain by staying with a LUT profile?</p>

<p>BTW, I found a negative number in the RGB formulation tag within that LUT profile that shows red in Colorsync Utility. See the screenshot below. Don't know if that is causing the banding but you might check to see if it is in the matrix version.</p><div></div>

<blockquote>

<p>What other color managed display preview advantage do you gain by staying with a LUT profile?</p>

</blockquote>

<p>I don't know! That's what I was asking Andrew :)</p>

<p>Interesting observation about the negative number. You're right, it doesn't appear in the matrix profile, which you can download <a href="http://rishisanyalphotography.com/ForumPostFiles/photo.net/Dell%20U2711%20B14C60_R98G91B91_D65G2.2_120cdm2%20HQ-Curves+MTX_Munki.icc">here</a>.</p>

<p>The banding also happened for my friend when he profiled his Dell U2711 with my Munki.</p>

<p>What to do? Go in & manually change it to see if it's the source of the problem? Not even sure how to do that...</p>

<p>Thanks,<br />Rishi </p>