Help with matching Epson prints to Monitor

I finally got a Spectraview profiler for my NEC monitor, had struggled with Coloreyes Pro for months without realizing that NEC monitors cannot be profiled that way. Now the image looks really good and perfectly balanced. Using both colorimeter and software from Spectraview.

Problem is, the print looks very slightly cooler, or actually more towards the green side. The luminosity is perfect on B&W prints and greyscale strips, but even there the print has a subtle cooler or toned look. If you do not compare with the monitor it looks perfect but on close inspection the monitor is slightly warmer and actually the skin tones look more natural on it.

I am using an Ott-lite lamp to view the prints. 

The monitor profile is spot on as validated and also by just looking at various targets and general images. So it is the prints that are slightly off. I could live with them except one always strives for perfection isn't it? I could increase the red tones in CS6 etc but that does not seem like a good solution to me.

The printers are Epson 3880 and 9900. Actually the larger printer gives slightly more accurate prints. The profiles are original Epson or from Red River. The papers I've tried (and they give the same results, suggesting that the profiles are perhaps the problem) are the Epson Premium glossy, semiglossy 250, premium luster, Exhibition fiber and  the Red River Satin.

The results are the same whether I print from Aperture, CS6 or QImage (via Fusion).

Any suggestions would be appreciated. Thanks.

Pradeep

Are you taking into account the color of the paper by soft-proofing with the simulate paper option checked and then adjusting before committing to print? The paper you use will definitely give a cast depending on the color level of the paper itself. If that doesn't fix the issue, you might get a custom profile made for your most used paper at first to see if that takes care of it.

There are many reasons why prints are not going to look exactly like the monitor, even when using a good calibrated display like a NEC SpectraView and a color managed printing workflow with decent printer profiles.

First off, the color temperature of the monitor may not match the Ott-Lite. If you've calibrated the NEC with more or less normal settings, you've got it running around 6500K. The Ott-Lite ranges somewhere from 5200K - 5900K according to Ott-Lite specs; you'd have to measure it to know for sure where it is.

Now, if the Ott-Lite was say 5200K, that's a warmer color temperature than the monitor at 6500K. So you might expect the prints to look a bit warmer. But another factor with the Ott-Lite is that it's not really a full spectrum illuminant. Fluourescents generally speaking are not smooth, full-spectrum light emitters matching natural sunlight, even if they're billed that way in marketing materials. Ott-Lites, like most others, have some strong spectral spikes in blue and/or green wavelengths that might lead to a slight cool shifted appearance in a print, even though the Ott-Lite's color temperature on the whole might be slightly warmer than the monitor backlight.

There are other lighting concerns that mix into the equation of course. Your monitor and the Ott-Lite viewing lamp are probably in a room that has some ambient lighting that doesn't have the same color temperature as either the monitor or the Ott-Lite. The mix of color temperatures will add further variables, along with the relative strength of the different lighting sources.

Yet another consideration is the printer profiles. I'll take it as a given that your monitor is probably in tip-top shape, but the printer profiles may not be as good. Generic vendor paper profiles vary in quality, and some of them may introduce color casts. Also some of them may have been created based on calibration criteria that don't match your viewing conditions. If you're seeing a consistent cool tone shift in your prints, it's less likely the fault of one specific profile, but it could be a systemic thing with them. Most printer profiles are created with calibration settings geared for print viewing in 5000K color temperature light. Different reflectance properties of the paper itself (often triggered by the inclusion of optical brightening agents) combined with possible jaggy spectral response curves of the viewing light (especially fluourescents) may contribute to print metamerism in ways that could be corrected by more advanced print calibration procedures than what is typically done for generic downloadable profiles.

So essentially all of this demonstrates one of the key truisms about calibration -- it's not about making one device or medium "look the same as" another, because that's difficult (if not impossible) due to the reality of emitted vs. reflected light, different device color gamuts (e.g. monitor vs. printer, or Epson 3880 vs. 9900), and ambient light sources that don't match each other. What calibration really is about is benchmarking each specific device to a relatively consistent standard in its own right. Devices may not match, but they should differ post-calibration in a somewhat consistent fashion that can be compensated for in a more predictable and less random way.

But you still need to account for the ambient lighting conditions under which each device or print medium is viewed, and that's not typically something that's dealt with inside the calibration workflow itself. Instead you'd either measure or eye-ball the differences between the original working file on-screen, and the final print, both viewed under the appropriate lighting conditions. And then if you saw differences you prefer not to let slide, you'd adjust things accordingly, either by adjusting the calibration of one or more devices to get a closer match, or by specific adjustments to a given print that's going to be view under certain lighting conditions...

EJ and Royce, thank you so much for a detailed reply.

I did simulate the paper and soft-proofed in CS6 and the print is still slightly 'greener'.  I did not adjust and print again, since I normally use QImage to do all my printing and did not want to save all the prints in this 'adjusted' form again.

I calibrated the monitors again this morning just to confirm the profile is accurate. The colors simply look very natural and not excessively warm (looking at multiple standard targets).

I then printed on all kinds of paper with their profiles that I have.

Interestingly, almost all of them look closer to the monitor using 'perceptual' instead of 'relative colorimetric' which is what CS6 defaults to. The black point compensation does not make a difference except the luminosity is slightly lower (very minimal darkening - hard to really notice) with it set to 'on'.

I am I think about 95% there. With Rel Colorimetric set, it drops to 92% or so. Again this is all very subjective and I may simply be obsessive about it.

I looked at the prints in daylight (diffuse, coming in through the window). The color difference is very slight now, with there being more 'orange/red' on the monitor and more 'yellow/green' in the print. I can live with this quite happily since it is only evident on well defined targets and everything else looks pretty close.

Royce, I really thank you for your insight on the 'truism' about calibration, it seems very difficult to make one device match another. In fact, the 3880 and the 9900 are also slightly different on the same paper with the same profile and inks, with the latter being closer to 'the real thing'. Now to make the images going to the former just slightly on the magenta side of the hue slider and see if that does it, although as I said I can live with the prints now.

Thanks once again for an exhaustive explanation. This place really rocks!

Pradeep