I want to make this as simple as possible, and it’s a subject that invites you to dive in to all manner of techno-vortices of Color Science.  It’s really a simple question, though.  Does it matter which light source I use to illuminate artwork for reproduction?

I’ve argued that the color and quality of the light source is crucial in reproducing the vision of the artist.  I think I’ve demonstrated pretty well how the light quality affects the texture and surface of the canvas, but I’ve been struggling with a good way to demonstrate how the light color affects the perception of the various tones and values in a painting.  I’ve been tempted to go into a deep, technical discussion of it, but then realized it’s really a very simple thing.  Show how the colors are different with three basic light sources – tungsten, strobe and “full spectrum” fluorescent.  The simplest way to show that is with a Colorchecker.

Basically, think of it in terms of throwing something at an object, and getting what bounces back.  You can’t get back what you didn’t toss in the first place.  If you’re not throwing red, for example (in the case of a fluorescent, specifically), you’re not going to get red bouncing back to the camera.

First, the video:

Now, the details.

The “Tungsten” version was shot using halogen gallery-type floods, the camera set to 3700K, and shooting RAW with the Nikon D800.

The “Strobe” version was simply changed to 5000K, and the “Fluorescent” version was 5000K as well.  The fluorescent bulbs are so-called “continuous spectrum” 5000K bulbs, which is really a misnomer, a fact you can see by simply looking at a spectral response graph of any good bulb of this type, including those used on the Cruse Scanner and other similar devices.

The files were processed using Adobe Camera RAW, and a click-gray balance was made on the lighter of the two center gray patches.  The shadows and highlights were adjusted slightly so the shadow and highlight points on the histograms would match up well.  The files were cropped to the target, and processed with the default Adobe camera profile.  The Working Color Space is Adobe1998.

Remember.  The primary objective of the entire science and technology of CCD development and Color Management is to replicate the human visual experience of color.  Color management tools’ job is to take these colors we see above and try to correct them to match the known colorimetric values of the target.  They can do that a lot better and bring the colors closer to what we perceive if they have a starting point closer to where they’re trying to end up.  If they don’t have the colors, they have to make them up, or stretch the colors they do have into the space they need to be.

The conclusions are pretty obvious to the eye, but even more so when you look at the histograms.  Each light source is indeed different.  The intensity of each of the color patches on the Colorchecker is rendered in a slightly different way in some cases, in others, the differences are pronounced.

So, yes, each is different.  That begs a few questions.

Which one is more “accurate”?  That question will send you down another vortex of Color Theory.

Which one is more faithful to what the artist saw?  That question is simple.  The light source that the artist used when creating the work can only be the most faithful to the artist’s vision and intent.


A painter’s view of light, color, and their importance in reproduction of paintings:

Warren worked closely with me in developing the technique I use for Fine Art Reproduction…  inspiring me to develop and refine the process and the tools – the X-Y Easel.

We’ve seen some samples of the difference between standard copy-board lighting and what is pretty typical for an artist’s studio. Let’s take it a step further and look at what we can do using basic lighting techniques.  If you’ve been the victim of one of my lighting classes, this is going to look familiar.


Painting by Nellie Ashford:  “Best Friends”

This is an example of what you’d typically see in an artist’s studio or a gallery.  It’s lit with a “Main” that’s a halogen spot from above, and has all the characteristics of that type of lighting – lots of texture, rich colors, deep shadows, bright highlights.  There’s a very slight amount of cooler fill, spilling from the surrounding walls.  This is “hard-edged closed shadows”, that is, the shadows have a hard edge, but are closed up, or dark.


This is the same light, but simply adding fill.  “Hard-edged open shadows”.  Take a good look, the shadows themselves still have a hard edge, but they’re now brighter, and show color and detail.  This is pretty typical of what you’ll see from a modern artist’s studio, often a combination of North light and halogen spots.  The fill in a studio like this is often far more cool that what we’re seeing here, though.


These last two examples are lit with a very soft, very diffuse light source.  Typically a “North Light Studio” without any additional spots would look like this, but the first one has no fill in the shadows.  Thus, it’s called “soft edged, closed shadows”.  We still get a fair amount of texture and the appearance of relief, but the effect is less dramatic, more flat, with the overall contrast lower and less saturated colors.  It’s more forgiving, but has far less impact.


Finally, we take soft, diffuse lighting and add fill.  This almost completely flattens out the work, showing very little texture, very diffuse colors, and very low contrast.  You’d use this if, for example, you had an original that had major surface flaws that you didn’t want to call attention to.  This is how an artist’s studio would look if they had North light, and very bright walls, floors and even ceilings.  It would be a fairly rare thing to see an artist working under these conditions.

Go ahead and click on each image to take a good close enlarged look.

These are what I consider to be the four basic building blocks of lighting.  What gets interesting is where you start working with degrees of these techniques – starting with hard-edged closed lighting and adding just a little fill…  controlling the degree of softness in soft-edged closed shadows – or any combination of types.   Not only do I get 4 basic flavors, but I can mix and match and hit somewhere in between.

This is true photographic lighting, and why it’s so powerful in redering artwork in a way that’s faithful to the artist’s vision.  It’s possible through no other process – no scan system or copyboard lighting can achieve these kinds of basic results, and certainly not in any combination.

In my previous post I argued that your light source dictates the range of color you have to work with.  The simplest way to prove that is to show you, so I’m using a tool called ColorThink, which maps the colors in an image on a conventional gamut map.

I shot a ColorChecker under mixed lighting with North light and tungsten, as I would a painting in my client’s studio.  I also shot it under so-called “full-spectrum” fluorescent.  Here’s how they compare:

Think “boxes of crayons” here.  These gamut maps show how many crayons you have to work with after all’s said and done, once the file is processed and in Photoshop format.  Even with the small range of 24 color patches in the ColorChecker you can see that “full spectrum fluorescent” doesn’t give you as many colors to work with as daylight, or daylight with tungsten.

Imagine what this would look like when you’re looking at the full range of colors in a painting.

These were shot RAW, processed in Adobe Camera Raw using the X-Rite Passport profiling system to achieve optimal color rendering.  They were shot with the Nikon D5000 and the 55mm Micro Nikkor 3.5.

For art reproduction, I’ve put a lot of stock in past posts on using the same lighting that the artist uses in creating the work.  Here’s why.

Take a look at these charts showing the output of various light sources on the visible spectrum, from the Olympus Microscopy Resource:

Note a couple of things.  The curve of daylight – or Noon Sunlight – is full, and smooth.  The curve of tungsten is similarly smooth, but not as full.  Every other artificial source is comprised of extreme peaks and what are, essentially, gaps.  The white LED curve is the only source that is even remotely close to daylight.

Now, here’s the comparison between what’s now referred to as “full spectrum” fluorescent, speficially, from the NLPIP Lighting Resource Center:

Two conclusions here…  first, “full spectrum” is marketing BS, and not supported by the data.  Second, fluorescent sources, of whatever type, have dramatic peaks in their output.

Now, let’s consider the process that gets color to our camera.  Light falls on our subject.  Our subject reflects that light, but, by virtue of the idea of “metamerism” doesn’t reflect necessarily the same wavelengths that fall on it.  (Look metamerism up, if you think it’s simply a color shift that your prints make under different light sources.  Besides being misleading, in a lot of cases the term is used flat-out wrong.)  Our lens takes that reflected light and filters out some of it – even the best lenses – and our sensor filters the light again, breaking it up into R, G, and B values as well as luminance (measured with a second G channel).

The goal of Color Management is to get your colors, after this long trip and the many transitions, to a known, and predictable place.  We want stuff we see as red to be red once they get to our computer, then finally to our print.  Most importantly, we want a neutral point – gray – that remains gray throughout the process, and maps all out colors from that starting point.

So here’s the process.  The artist is working away in the studio, with that nice broad spectrum of daylight streaming in from the North light studio window and skylight.  As is often the case in modern studios, there are halogen or tungsten down spots shining on the work too.

The work reflects this broad spectrum in a fairly complete way, but with it’s natural limitations of the pigments being used in the paint.  Let’s pick on purple.  The pigments may not really reflect the exact wavelengths around that color accurately…  but, it will reflect them in a consistent way.   And that is what the artist sees.

If you light the work with a source that doesn’t “have” that range of wavelengths, the one around that color purple, for example, how can the work possibly reflect that wavelength?  Those gaps in the fluorescent chart are what I’m talking about here.  Every peak and gap represents colors that are too strong, or simply not there, in the starting point of your color journey from source to print.

The goal of Color Management is to adjust for those gaps and peaks, but, it can only do so much.  In the case of the purple, our system is going to measure our purple patches, see they’re reading as very low, and adjust them by literally adding value in a simple lookup table.  It’s not working from actual data, it’s simply looking at some numbers, seeing they’re low, and adding to them to smooth out the curve that it’s looking to build.

Good Color Management, especially camera and scanner input profiles and the RAW processing “profiles” of the X-Rite Passport system can get you consistent results from a consistent light source.  It can iron out specific characteristics of your particular lights and lenses and sensors, and bring them into a range of the industry standards.  It can’t, however, make a perfect match between a light source that has a full and complete spectrum and one that has dramatic peaks and gaps in it’s spectral response.

For a nice, simple visual demonstration of this, see my post here: Light Sources and Color Spectrum, Simply.

Which begs the question…  if you’re trying to reproduce artwork, and want to be faithful to the visual experience of the artist, then why would you start by limiting your palette of illumination?  You are, at the very beginning of your process, assuring that you will never have the same colors to work with as the artist had.

Here, for reference, is the NLPIP Lighting Research story, the last page, which has comparison graphs.  Read the whole post, it’s very, very interesting.

I’ve talked about this in class, and this is simply the coolest example of it I’ve ever seen. The punch line is that the green color and the blue color are actually the same. Don’t believe it? Open it in Photoshop and read it, using “Info”, or just use DigitalColor Meter on your Mac.

The site it came from is pretty cool, too, Color Illusion 12, here.

Here’s an interesting story of applying processing standards to streamline production:

Ross-Simons, one of the country’s most successful fine jewelry retailers, faced a staggering challenge. They support 14 retail locations, an online store (named a “Top 500” site by Internet Retailer Magazine in 2005), and a quarterly catalog, first mailed in 1981, that now tops 60 million catalogs mailed all over the world every year.

With all of this size and scale, they are selling the highest quality jewelry. Color, size, cut and polish are all critical to the customer, and Ross-Simons needs their photography to show it all, accurately.

While some companies can have in-house photo studios, the sheer number of products in the Ross-Simons catalog and limited time-frame means that multiple photographers all over the country are all working on various stages of the projects. Jay Dunn, as VP of Creative, was seeing a huge degree of variation in the photography coming in from the studios. Considering each studio was using different cameras and different practices in processing and delivering the files, it’s no wonder. In an industry where the difference between a fine gem and an average stone can be a few points of color, the images had to be spot-on, regardless of their source.

“In our last catalog run we spent over 600 hours for post-production Photoshop time in color adjusting file standardization and retouching. We really felt that we could cut that in half if we could somehow standardize the Color Management and processing…It’s not that we’re unhappy with the photographers’ work. In fact, we feel we’re partially to blame. We just have never been able to tell them what we want.” – Jay Dunn, VP of Creative at Ross-Simons

Jay had worked with Michael Oh and the Tech Superpowers team for help deploying a new hi-res design workflow in Jay’s previous position at Brookstone, but this time the project was even more complex, and an entire series of catalogs depended on results.

Huge Challenges

This was a huge problem that had many challenges. The first one was simply to isolate each of the factors that was causing variations in output.

First, the photographers used different cameras – a Leaf Aptus75, a Valeo 22, a Sinar 54, and even a Nikon D200, with software that was just as varied. Since there were over six different makes and models of cameras, lenses and lighting, Ted Dillard, head of TSP Imaging Services, had to synchronize the color rendering of each camera to match the others. In this case, it wasn’t so much an attempt to match the cameras to any “industry standard”, more that they needed to match each other.

Second, Dillard and the Ross-Simons teams recognized that the problem wasn’t simply the photographers: a complete end-to-end, or “Capture-to-Press” solution, was needed. So collaboration and agreed standards were key. In this case, the best way to make sure that this was reasonable and understood was to bring all of the vendors together – 24 people in all – at Tech Superpowers to review process, standards, and best practices in a full-day meeting… to form a consensus.

Third, once this consensus was reached, Ross-Simons needed to give the entire team a set of guidelines – from exposure, capture and processing settings to scaling, sizing and color management standards – right out to prepress and proofing, including a communication “loop” from the press back to the photographers. Ted was able to create a capture, RAW-processing and color management workflow that worked from end to end, and establish lines of communication to reinforce, and correct, the process, during the process.

“Ross Simons’ problem was very common, but we had an unusual opportunity to create a new solution. Rather than apply a fix after the fact, by trying to profile the cameras – a notoriously inaccurate and ineffective approach – we elected to go to the RAW files and standardize the processing at the capture level. Each studio had a set of guidelines for file delivery, as well as individualized processing settings to assure one camera would look like the next, regardless of the make, model, lens or lighting used.” – Ted Dillard, Imaging Services Manager

Huge Results

Using our experience and training in RAW file processing as well as our considerable experience with the individual digital camera systems, TSPIS was able to minimize the differences in color, contrast, and look between all of the cameras, and a standard of file quality and specifications between all the studios.

“When I reached out to Tech Superpowers, what I needed was the insight of a professional photographer, fused with the knowledge of the digital and technological advances, to create a cost-effective, efficient, multi-user, multi-city, photography and asset management workflow.

They far surpassed any expectations I had… [and] engineered a strategy that allowed photographers and color houses in seven different cities to align to standards and protocol that created speed-to-market and cost advantages worth large dollars to our organization.” – Jay Dunn

Given the cost of a trained Photoshop artist, cutting 300 hours for each one of four catalog runs per year… well, that’s a solution that you can take to the bank.