On this page
Last updated July 27, 2004
Before diving into the challenges of flower photography, let's size up our quarry by examining a few flower photographs—some good, some not so good. I took all the samples below with one or another digital camera, some with more understanding and care than others. Click on any thumbnail to see an 800x600 or 600x800 version.
As you've just seen, flowers can be tough subjects.
They're also irresistible, so let's see what can be done to overcome the challenges. But before tackling that in the next section, let's visit the thorny issue of color accuracy.
Color accuracy is a constant frustration in flower photography, and digital and film cameras struggle equally in this regard. The narrow-band off-primary colors flowers have evolved to attract pollinators and discourage predators and freeloaders please the eye but tend to resist depiction as linear combinations of primary colors other than those used by the human eye.
Eyes, color films and digital cameras all sample the spectral content of incoming light using different primary colors, and all make different inferences regarding off-primary inputs. Spiky flower reflection spectra have a way of unmasking the substantial variations in peak and shape found in the tri-stimulus spectral response curves characteristic of
When discrepancies appear, we naturally regard the eye's color rendition as the correct one. But short of duplicating the eye's spectral responses in detail, cameras are doomed to render narrow-band off-primary flower colors differently.
According to the informative Kodak Technical Data bulletin Why a Color May Not Reproduce Accurately, blue morning glories, gentians, and ageratum flowers are cases in point. Along with some common fabric dyes, these flowers exhibit high reflectances at far red wavelengths. Common light sources like the sun and incandescent lamps provide a lot of far red to reflect, but human eyes are much less sensitive to it than color film and digital cameras. Photographs overly reddened by this effect are said to show anomalous reflectance. In a related phenomenon, chlorophylls reradiating in the far red due to UV fluorescence can over-redden or neutralize plant stems and leaves, but chlorophyll turns out to be rare in true flower petals.
Automatic brain-eye system color adjustments far beyond the sophistication of camera firmwares also contribute to the color " inaccuracies" found in camera images. Other color space issues no doubt enter the fray as well.
How Accurate Do Flower Colors Need to Be?If you're producing a mail-order rose or sweater catalog, color accuracy between subject and final output is a must. For most other purposes, you have some wiggle room. In fact, your viewers probably won't know or care if your flower colors are off a bit—as long as they remain credible and pleasing to the eye.
Color fidelity between your monitor and printer is always important to guarantee predictable output from post-processing, but that's a different issue.
The guidelines below are based on experience and carefully cross-checked reading, but I strongly recommend heavy experimentation with your own gear early in the flower game:
The importance of the first 4 items in this table—lighting, exposure, composition and motion—can't be overemphasized. Beyond working hard to avoid obviously phony colors, I no longer sweat the color accuracy.
Overexposure is a good way to lose the subtle details that so enrich the visual experience of flowers. The right balance between detail and exposure is often best found through spot metering and bracketing. This is especially true for light-colored flowers, which usually require some positive exposure correction (EC) but can easily end up blown out if too much EC is applied. Medium-toned flowers are generally less challenging with regard to exposure, but my C-2020Z's eagerness to oversaturate reds of any tone in bright sunlight keeps me on my toes.
The series of white cala lily exposures in the table below illustrates the exposure-detail trade-off and the value of bracketing. All shots were spot-metered on the white petals at ISO 100.
Bottom line: When in doubt, bracket. Include overexposed settings for light-colored flowers and underexposed settings for dark ones.
Many flowers, particularly yellow varieties like dandelions and the day lilies shown below, display prominent markings visible only in the near ultraviolet (UV-A) band. They target these air traffic control signals at specific pollinating birds and insects equipped with the UV vision needed to see and follow them. UV markings and unique narrow-band visible colors are but two of the many tricks flowers have evolved to attract effective pollinators selectively, without advertising to predators and free-loaders.
A suitably equipped digital camera can provide a window into this fascinating unseen UV world, and flowers make for ready and rewarding UV subjects. For an introduction to the practical aspects of digital UV reflection photography, see the dpFWIW UV imaging section.
The long exposures typical of digital UV work make bright sunlight, still air and good camera support critical ingredients in UV flower shots. (Note that this bright sunlight recommendation does not extend to visible light flower photography, as explained above.) The monopod used in all the 18A + hot mirror samples above was clearly not steady enough.
I had to zoom out to get a larger aperture on the 18A + hot mirror cala lily above—hence the change in framing.
For a much more detailed look at UV flower photography, see Bjørn Rørslett's fascinating site, Enter the Unreal World of Ultraviolet Colour Photography.
Luckily, verifiable IR and UV contaminations aren't easily detected in digital flower photographs taken in visible light, even with a relatively IR- and UV-sensitive camera like my Oly C-2020Z. Why that's so isn't entirely clear, but ultimately, the dominant visible light signal must overwhelm the IR and UV signals coming out of the CCD.
The human eye is blind to near IR (NIR, 700-1100 nm) and near UV (UV-A, 320-400 nm) light and has only limited sensitivity to the far reds and violets near the boundaries of the visible band nominally at 400-700 nm. These invisible wavelengths are all abundant in sunlight, and NIR is abundant in incandescent sources as well. Ordinary films tend to have limited UV sensitivity and negligible IR sensitivity, while digital cameras tend to be just the opposite. In theory, at least, UV contamination should be more problematic in visible light images on the film side, while NIR should be more problematic on the digital side.
Since flowers are commonly highly reflective in the UV-A, the NIR or both, you might well expect big trouble in digital flower photographs due to direct IR contamination or to a lesser extent, UV contamination. Indirect contamination might come via fluorescence of flower pigments across the visible-NIR boundary. In this scenario, absorbed visible or UV-A photons excite the emission of NIR and far red photons that in turn stimulate the camera more than the eye. Chlorophyll a and b fluoresce strongly in just this manner, absorbing UV-A through blue-green photons (375-430 nm at full width half maximum) and emitting deep red photons (640-680 nm at full width half maximum), as illustrated in Figure 1 of SCUFA's informative online PDF. Since film and digital cameras are both more sensitive to far red than the eye, chlorophyll mediates a UV contamination of sorts—one resembling the artifact of anomalous reflectance. True flower petals don't usually contain much chlorophyll, but other fluorescent flower pigments might contribute. Sepals masquerading as petals (as in the yellow flannel bush blooms above) might contain enough chlorophyll to participate in such UV contamination.
Here's a crude demonstration of UV fluorescence in catnip.
Don't Worry, Be Happy
So much for theory. In practice, I've been hard pressed to find compelling evidence of either IR or UV contamination in my digital flower photographs. I certainly haven't detected any visible benefit from my Heliopan 8125 IR/UV cut filter, but the 8125 turns out to be only a partial defense against near IR and UV-A contamination. Its complete transmission spectrum clearly shows that the 8125 passes some very, very near IR and some very near UV-A as well. A high quality hot mirror filter might be helpful on the IR side, but I doubt that it would be worth the effort on most higher-end consumer-grade digital cameras.
(See also the home page links.)
Blacklock, Craig and Nadine, Photographing Wildflowers, Voyageur Press, 1987.
Enter the Unreal World of Ultraviolet Colour Photography—Bjørn Rørslett's fascinating, copiously illustrated and comprehensive film-oriented UV photography site, with many UV flower shots.
Kodak Q-60 Color Input Targets—a PDF document detailing the use of this compact but pricey color reference. (Right-click on this link and select "Save Target As..." to save the a copy of the PDF file on your computer.)
Spellenberg, Richard, National Audubon Society Field Guide to North American Wildflowers, Western Region, Chanticleer Press, 1998.
Why a Color May Not Reproduce Accurately—a Kodak Technical Data bulletin.
Unless explicitly attributed to another contributor, all content on this site © Jeremy McCreary