Printing Neon Blue?

[Chris Kern]

Can anyone offer me any tips on how to process an image of a scene bathed in blue Neon light (attached) so as to be able to create a print that plausibly captures the color?

Admittedly, I'm new to making my own prints (with a recently-acquired Epson P800), but this is the first time I've been unable to produce an acceptable soft-proof in Lightroom.  I've tried the profiles for both Epson Premium Luster and Hahnemühle FineArt Bayrta, and both perceptual and relative rendering intents, and have failed to discover any combination of sliders that reproduces the electric vibrance of the scene that was captured by the camera and which my NEC monitor is able to display.  (These two papers have served me well with other images containing bright, saturated colors.)

I realize there's probably no way to get an exact match—obviously a high dynamic range transmissive medium like a monitor is going to be better at representing this sort of color than any photo paper—but I'm wondering if anybody out there with more experience than I have knows some magic incantation that so far has eluded me.

[keithcooper]

What printer?

I'm just finishing testing of an Epson P400 for a review, and faced something similar

I printed an outdoor christmas photo when some of the lights were blue LED based.

With the P400 it's tricky not to get a purple look - just to see, I also tried it on my iPF8300, where having blue ink made it a whole lot easier...

A combination of selective hue changes in the image and tweaks of profiling, finally got a reasonable look to the image with the P400, but not something I'd do other than as an experiment ;-)

One trick was to look at a big printed profiling target, and see what patch looked close to what I want. If the printed target had nothing close, then that wasn't a good sign... if the colour was there, then it should be able to be reproduced again, although it might look a bit off on the screen.

[Chris Kern]

What printer?

Epson P800.

[keithcooper]

Epson P800.

Thanks - I missed that  - Doh!

One for experimentation and a few small test prints? - the P800 has a good wide range of colours.  It's prints like this where I'm much less inclined to rely on soft proofing

[Chris Kern]

Thanks - I missed that  - Doh!

Actually, you didn't.  I added a parenthetical with the name of the printer after you pointed out that I had forgotten to include it in my original post.  (I looked for a way to annotate the Last Edit tagline to explain the update, but couldn't find one.)

One for experimentation and a few small test prints? - the P800 has a good wide range of colours.  It's prints like this where I'm much less inclined to rely on soft proofing

I guess that's the next step.  But soft-proofing has worked so well for me until now that I thought I should ask around first to see if I was missing something.

Maybe I should even try color management "by printer" ...

[Richowens]

Chris,

Just a thought....metallic paper?

Rich

[keithcooper]

No need to ditch the use of a profile, but it's images like this that push the limits of using profiles and soft proofing - don't forget to try different Rendering Intents too.

I find I rely on soft proofing less and less as I get more experience with any particular paper - more to flag up areas of concern and overall look, than as a final check that they will look OK. It's a useful tool, but with more limits than often assumed.

This is an area I'm often asked about when doing talks about printing (and CM) at camera clubs, where people may have heard of the benefits, but the various caveats were missed...

Printing images like this is a good reminder that the best looking image on screen does not always equate with the best print, and what very different media they are.

[Chris Kern]

Just a thought....metallic paper?

That seemed worth a try (at least a virtual try), so I installed the ICC profiles for Epson's two metallic paper products and tried to soft proof the image with them.  However, once again, the electric blue light cast by the neon fixtures is way out-of-gamut for what the printer and paper can reproduce, and trying the best slider values I could devise for the other papers—a combination of brightness, tone curve and color adjustments in both rendering intents—produced the same blah results.

Admittedly, I haven't actually printed this image yet.  I plan to take Keith's advice and make some small test prints.  But my experience so far with this monitor and printer, along with all the papers I have tried, is that the Lightroom soft proof is a very good approximation of what I will see in the print.  This is the first time I've experienced any colors that were so far beyond the range of the printer's capabilities that I couldn't come up with a set of adjustments that produced a reasonable approximation of what I was seeing on the monitor—modulo the difference in dynamic range, of course.  In fact, more often than not I haven't needed to make any color adjustments, just add a bit of "clarity" and tweak the tone curve to punch up the printed version of the image.

[Paul2660]

What profile is the image in?  Prophoto or Adobe RGB (1998)?  Neon blue is going to be a problem for sure.  The pervious post of Metallic paper would be where I would go. But look at the Moab Slickrock line also. Odds are you will still be out of gamut but it should be better.

Out of gamut issue with blues, pretty common for me until I started putting all my work back in Adobe RGB (1998).   Even with this you can easily take a heavily saturated blue out of gamut for an inkjet.  9900 9880 and 7800.

Personally I stopped using the Prophoto color space as no inkjet printer from Epson I know of can approach the ability to print that full color space. It caused me too many out of gamut issues.

Paul C

[John Caldwell]

Agree with the metallic paper, for certain. Be sure to turn off "simulate paper white" when soft proofing with metallics. You'll never get this range of neon blues in gamut with a paper print. To make the blue seem even more vibrant (than it truly is), desaturate the non-blue tones in your image. Above all, be sure to light the print properly once it's displayed. This will have as great an impact as any other step in your process.

John Caldwell

[JRSmit]

What profile is the image in?  Prophoto or Adobe RGB (1998)?  Neon blue is going to be a problem for sure.  The pervious post of Metallic paper would be where I would go. But look at the Moab Slickrock line also. Odds are you will still be out of gamut but it should be better.

Out of gamut issue with blues, pretty common for me until I started putting all my work back in Adobe RGB (1998).   Even with this you can easily take a heavily saturated blue out of gamut for an inkjet.  9900 9880 and 7800.

Personally I stopped using the Prophoto color space as no inkjet printer from Epson I know of can approach the ability to print that full color space. It caused me too many out of gamut issues.

Paul C

Little to do with the prophotorgb. The camera sensor/electronics cannot deliver the saturation possible in prophotorgb (check the CaptureOne camera profiles), but can exceed adobeRGB at certain parts of the gamut. Nor is there a monitor that can show the full scope of prophotorgb. Nor is the digital image technology capable of capturing blue very well (see research done by american museums w.r.t. digital preservation of art).
Going from prophotorgb to AdobeRGB means the max possible color saturation is reduced to adobeRGB. Then the question is, how is it reduced?

[Paul2660]

Little to do with the prophotorgb. The camera sensor/electronics cannot deliver the saturation possible in prophotorgb (check the CaptureOne camera profiles), but can exceed adobeRGB at certain parts of the gamut. Nor is there a monitor that can show the full scope of prophotorgb. Nor is the digital image technology capable of capturing blue very well (see research done by american museums w.r.t. digital preservation of art).
Going from prophotorgb to AdobeRGB means the max possible color saturation is reduced to adobeRGB. Then the question is, how is it reduced?

What I found was prophoto with blue skies, (with saturation) tended to always have a part of the sky that was way out of gamut, this was very clear during the printing process and in the soft proof.  I prefer saturation in my work, and the skies tended to be a problematic area. However it can also be seen in yellow and reds.  As I understand it, the Prophoto color space in  16 bit has a massive amount of color that my NEC 30" monitor can't display anyway.  The same monitor claims to have 99% of the Adobe RGB color space. 


I just decided to use a color space that was more complimentary to my monitor, PA302W with spectra vision.  This also has helped with my prints which involve bright blue skies being more in gamut. 

Paul C

[Chris Kern]

As far as I am aware there is no way to modify Lightroom's working color space.  My understanding is that internally, LR always uses ProPhoto RGB—which includes many colors my monitor, a NEC PA272W, is incapable of displaying (and probably some my aging eyes and brain are incapable of recognizing).  The PA272W covers almost all of the Adobe RGB color space, and then some.  The attachment I included in my original post was exported as sRGB.

As an experiment, I exported copies of the photo in all three color spaces and compared them on my calibrated NEC monitor.  They look identical.  So I conclude that the color I am unable to reproduce satisfactorily when soft-proofing in LR falls within the sRGB color space.

There definitely seems to be something unusual about the character of the blue emitted by the neon lighting in the sculpture.  I've attached another photo with a highly-saturated blue that was easy to soft proof and print, so I don't think it's simply that the neon blue in the sculpture image is highly saturated.  But that particular hue so far has defeated my ability to find a combination of Lightroom adjustments that will reasonably reproduce it on photo paper.

[Paul2660]

As far as I am aware there is no way to modify Lightroom's working color space.  My understanding is that internally, LR always uses ProPhoto RGB—which includes many colors my monitor, a NEC PA272W, is incapable of displaying (and probably some my aging eyes and brain are incapable of recognizing).  The PA272W covers almost all of the Adobe RGB color space, and then some.  The attachment I included in my original post was exported as sRGB.

As an experiment, I exported copies of the photo in all three color spaces and compared them on my calibrated NEC monitor.  They look identical.  So I conclude that the color I am unable to reproduce satisfactorily when soft-proofing in LR falls within the sRGB color space.

There definitely seems to be something unusual about the character of the blue emitted by the neon lighting in the sculpture.  I've attached another photo with a highly-saturated blue that was easy to soft proof and print, so I don't think it's simply that the neon blue in the sculpture image is highly saturated.  But that particular hue so far has defeated my ability to find a combination of Lightroom adjustments that will reasonably reproduce it on photo paper.

Good point. I have never been able change the working color space in LR either. If someone knows a way I would love to know how.

I always export from LR as a tiff in the Adobe RGB color space. Soft proof in CC where I have Adobe RGB as my default color space. Then import back to LR. As I understand it LR respects the asigned color space for the tiff. Then I print from LR.

Convoluted I guess but works.  The key to me is printing an image in the Adobe RGB color space as I don't believe that the Epson inkjet line can even fully cover that color space and for sure I don't believe it can approach the prophoto even in 8 bit.   

Paul C

[digitaldog]

Good point. I have never been able change the working color space in LR either. If someone knows a way I would love to know how.

You cannot change the processing color space in LR or ACR, it's ProPhoto RGB primaries with a 1.0 TRC.
You CAN encode that into any RGB working space or output space you wish and have a profile for.
Getting OT for the OP, upload here a small (say 300x300 pixel) image in the RGB working space you're going to use for output, I'll plot that gamut against P600 profile(s). It's quite possible it's out of gamut.

[Wayne Fox]

I always see this argument that ppRGB is too big. Maybe there is a misunderstanding of the process of moving from a working space to an output space in thinking that the entire working space is mapped into the output space so using a space as big as prophotoRGB is problematic. At least that’s the only possible “theory” I can come up with as to why it would be a problem.

But ppRGB is simply a container that is large enough to contain pretty much any color. This means moving from ppRGB to an output space will move whatever colors in the image into the output space, using one of two logics ... move any colors necessary to try and maintain the relationship between colors (perceptual) or try and leave most of the colors alone and clip out of gamut to nearest color (relative).  It doesn’t matter if the working space is way larger than necessary, the empty space outside the image gamut doesn’t come into play. I’m not sure why clipping into adobeRGB would be advantageous instead of moving to the output space.

In the case of the OP, the idea of first clipping to AdobeRGB, then moving to the final output space seems counter intuitive, because blues are where the Epson printer has an advantage over AdobeRGB. 

I haven’t used AdobeRGB for a long time -several years ago when my 9900 printer arrived found many landscape images with blue sky suffered if I used AdobeRGB as my working space  I have found issues where I didn’t take the time to soft proof, even if it was for nothing more than deciding which rendering intent is better suited for the image, and yes there is the occasional image which requires some tweaking to get good results, but clipping to adobeRGB first, while perhaps solving the problem, has never yielded a better result.

I have seen issues with some very specific shades of blue ( in this case based on the preview the blues appear to have easily become synthetic in post processing, thus contain colors outside of the cameras original colors, adobeRGB and the printer gamut).  I wouldn’t be surprised to see this file take some effort.

And I would agree, some extreme synthetic blues a canon printer can pull off easier than an epson.

[digitaldog]

I always see this argument that ppRGB is too big.

Or Adobe RGB (1998) and similar is too small. For ProPhoto RGB, it's size is the price we pay for a working space that is simple in shape, based on an emissive display (theoretical), a gamut used to process raw data in the ACR engine and a size we need to encode colors we can capture and print.
FWIW, the image in ProPhoto RGB falls way outside the gamut of a P600 (on Exhibition fiber):

[Chris Kern]

FWIW, the image in ProPhoto RGB falls way outside the gamut of a P600 (on Exhibition fiber):

Thanks for doing this.  You've demonstrated to my satisfaction that trying to reproduce this blue on my P800 is a lost cause.  I'm going to keep plugging away for a while in an attempt to create a print that pleases me, but clearly I'll need to compromise the color of the light that attracted me in the first place to shoot this sculpture (untitled, by Dan Flavin, "to Helga and Carlo, with respect and affection," 1974, Hirshhorn Museum, Washington, D.C.).

I have seen issues with some very specific shades of blue ( in this case based on the preview the blues appear to have easily become synthetic in post processing, thus contain colors outside of the cameras original colors, adobeRGB and the printer gamut).  I wouldn’t be surprised to see this file take some effort.

Synthetic, yes.  Although I don't think from post-processing.  That blue is as close as I could get to what I saw at the museum and captured with my D800E.  Other than applying Lightroom's Camera Standard color profile, the only color adjustment I made was to reduce the blue luminance value slightly.  I have no clue what wavelength(s) the neon bulbs were emitting, or whether something in the wall paint was causing it to fluoresce, but that light was really way out in the wild blue yonder—as Andrew's gamut graphic demonstrates.

[Stephen Ray]

That blue is as close as I could get to what I saw at the museum and captured with my D800E.

Seems the museum's photographer captured the same scene but recorded it differently using her Canon Mark III. (Apparently in Adobe98 space.)

Because a Canon printer and metallic media was mentioned earlier, I happen to have such a profile on my system and am able to get a good soft proof from her more tame file, although her file appears the same as yours as far as HSL. I am also able to easily coerce the neon blues using HSL to other profiles on my system to get good soft proofs.

I often find myself using internet images for color reference.

[JRSmit]

Or Adobe RGB (1998) and similar is too small. For ProPhoto RGB, it's size is the price we pay for a working space that is simple in shape, based on an emissive display (theoretical), a gamut used to process raw data in the ACR engine and a size we need to encode colors we can capture and print.
FWIW, the image in ProPhoto RGB falls way outside the gamut of a P600 (on Exhibition fiber):

Looking at the cloud of colors it seems these blues are almost literally against the envelope of prophotorgb. There is to my knowledge nothing that can reproduce that, no disply, no print. Nor is there a camera that can capture that. This is a synthetically created result.