The flowergirl was a ham.

The truth is that event photography is hard.  Wedding photography as a subset of event photography is an incredibly hard thing to learn on your own and is really best left to people who have solid experience doing it. 

You really need a second-nature understanding of flash photography (which I still don’t fully grasp) and a complete command of your equipment in order to perform well under the strenuous conditions of a wedding.  Realistically, I probably spent about as much time trying to figure out how to change the settings on the camera and flash equipment as I did actually shooting pictures.  In addition, you need to be able to pose groups of people on the fly, anticipate moments before they happen and keep a mental checklist of the photos you think the couple will want as you go.

I missed few great moments, many of my shots were less then stellar, and I spent about a solid month working in photoshop before I was able to hand over a CD of the best images.  The couple ended up mostly satisfied with the results, though they were slightly annoyed with the time it took me to complete the post-processing work. 

I never intended to accept responsability for being the sole photographer at someone’s wedding, and I regretted almost every moment of the experience, but I did learn a number of things as a result of completing the task.

1. Equipment does matter.  Anyone that tells you “The photographer makes the shot not the camera” is lying to you.  Unlike day-to-day shooting, event photography is not an arena where you can mitigate the faults of your equipment by shooting to your strengths. 

If your lens is slow at long distances, you’re not going to be able to work without raising your ISO significantly.  If your camera system isn’t intuitive and easy to operate, then you’re going to miss great moments while you fiddle with controls.  If your flash has a slow recycle time, then you’re going to have a lot of unusable shots where your flash didn’t fire and even more shots that you’re going to miss while you’re waiting for the flash to charge. 

Most of all, people aren’t going to put a genuine moment on hold and wait patiently while you fiddle with your equipment.  If you’re not able to grab a shot as it’s happening, you won’t ever get the opportunity to get it again.  Case in point, to the left is the shot I captured right after the bride smashed the grooms face with cake.  If you miss it because your flash wasn’t ready, then you missed it.

2. Shooting RAW is the only way to go.  Sure you need a ton of cards because you can rack up over a thousand frames pretty easily, but even the best JPG compression is unforgiving in low light and high-ISO situations.  RAW won’t save a hopeless shot, but it will give you the latitude to rescue a bunch of marginal ones.  I admit that I bought the arguments professed online by interweb blowhards that JPG was “good enough” and that the storage space required for RAW images didn’t justify the minimal increase in quality, but after switching over to RAW shooting with Aperture 2, I’ve been shocked at the differences in processing latitude afforded by RAW over JPG. 

Even images that were saved in the highest JPG compression quickly block-up in the toe of the curve, showing the ugly compromises of DCT compression.  By contrast, images saved in RAW format have huge amounts of latitude to pull up the toe and bring the shoulder of an images curve back into range.  If I’d shot this wedding project in RAW instead of JPG, I would have cut my processing time in half.

3. Regardless of your ability, your time comes at a cost.  This is a tricky concept for amateurs like me to understand, but it’s important to realize that even if you don’t think your photography worth charging someone for, your time is finite and thus has an inherent value to you.  There’s nothing wrong with offering your services  for little or no fee, but you should be realistic about what it is costing you and limit the level of your obligation according to the amount of time you’re willing to part with.

Although shooting this wedding did only take a single day of my life, I did not consider in advance the 20+ hours I spent after the event color correcting, applying curves adjustment layers, retouching, cropping, and sharpening all the images I captured.  Even when performed during your spare time, 30 hours of your labor is a tangible expense that takes away from other aspects of your personal life.  It’s important to be realistic about what you’re agreeing to provide someone and how much of your time it will take to finish the job.

4. Sometimes you have to see what you do wrong to see what you do right. The shots that didn’t work will be obvious and easy to criticize, but it’s important to look for patterns within the shots that did work.  Are all of your best shots taken at a similar distance?  Do they share similar lighting or framing?  Do they capture specific candid moments or are they staged and posed?  Recognizing these patterns can often lead you to find strengths in shooting you didn’t know you had.

Not frame worthy, but quirky enough to like

What was plainly obvious in the massive number of shots that I captured at this particular wedding is that my strengths are in mid to close range candids.  My posed formal shots of the bridal party were acceptable, if challenging to visualize and achieve, but the candid shots of the bridal party taken before the staged session was far more successful.

The long range shots of the bride and groom’s first dance did the job, but the close up shots of guests kicking it on the floor were far superior.  This contrast showed me something that I was coming to understand about my own photography and encouraged me to explore more photojournalistic close-range candids.  Eventually, this exploration developed into a specific technique that I’ve developed that has gradually become a “signature” part of my photographic style…something I’ll delve into for my next post.

So, although I regretted taking the job for the unwanted responsibilities it placed on me along with the amount of time it removed from my life, I was able to turn it into a learning process that would better prepare me for similar event photography situations in the future. 

These shots here are acceptable, but to truly understand what great wedding photography looks like, you’ve got to see the work of a great wedding photographer.  One of the best out there today is Ryan Brenizer from New York.  Take a look at his flickr.com page and drool.

I don’t mean to be critical of the site for propagating slightly inaccurate information because the post is very good information for the amateur photographer, however I believe that they oversimplify and repeat some misconceptions about light color that serve only to obfuscate one of the most fundamental aspects of image capture.  Camera manufacturers are somewhat complicit in this simplification of lighting color by using Kelvin numbers as white balance settings, but it’s important to remember that “Color Temperature”, “Color Balance” and “White Balance” are all different things.

As the article states (emphasis included from the original post):

“Light comes in a variety of colors. It’s not just “white”. Tungsten light bulbs, for example, are yellow/orange in color. Fluorescent bulbs are somewhat green. Daylight is blue. We know this is the case because of how these types of light reflect off a white surface… and we call this color cast the color temperature of the light.”

Color Temperature is essentially a shorthand measurement of spectral radience based on Planck’s Law.  Although it’s used to describe nearly all light, the color temperature scale relates only to incandescent sources.  Arc and gas discharge sources like fluorescent, sodium vapor and HMIs do not incandesce and the color temperature scale does not apply to their illumination.  A Correlated Color Temperature, used for non-incandescent sources, is generally used in the same descriptive terms as traditional Color Temperature (though as I note in this post, the color rendering of each lamp can be widely divergent from incandescent sources).

What’s confusing in understanding Color Temperature is that it’s a measurement of the emission of light, NOT the reflection of light as the article suggests.  Color balance and white balance are electronic adjustments performed to the captured measurements of reflected light intended to achieve a perceptually neutral image. 

In this sense, white balance CAN be described as the color of light reflecting off a white surface, however Color Temperature cannot be described the same way.  Furthermore, in many situations, adjusting a neutral white point can still leave color casts in neutral mid-tones which affect color balance.  In an 8-bit colorspace, there are only 256 values per channel to describe a color,  providing a very tiny range of adjustment when limited to the white values of an image.  This simplification of white balance as an overall color balancing tool can lead one to believe inaccurately that in-camera white point adjustments will reliably result in balanced color across the entire range of tones.  In practice, measured adjustments of neutral black, mid-tone grey and white points will often be needed to produce a truly neutral image.

In summation: 

• Color Temperature is a descriptive term relating to the emission of a given light source.
• White Point/White Balance is an electronic adjustment of RGB values to neutralize the light reflected from a white surface within an image.
• Color Balance is the overall neutrality of an image exhibited by neutral values in pure black, mid-range grey and white points of a given image.

Intertwined, yes, but all different concepts. 

For a much better explanation of basic color temperature, I’d read this article here.

Want a low-priced CFL bulb?  No problem, but don’t expect color rendering to be very reliable.  Want a cheap bulb housing?  No problem, but be prepared to work with some crappy mounting options, limited flexibility and a 6′ cord.  

I’d already decided that I needed to revisit my lighting situation at a later date, but the weak clamp on my ghetto lighting setup ensured that date would come sooner then I expected.  I was just getting used to working with the quirky color range of this particular lamp when the short cord conspired with the slippery clamp to bring the whole setup crashing to the floor at my bare feet.  Below is one of the better examples I’d been able to capture for the fashionable foodie (recipe not yet posted).

Subtle, colorful and diverse in tone.  This shot above looks very much like natural daylight, but is the result of a bare CFL in an aluminum clamp reflector with a foamcore bounce standing upright on the opposite side of the bowl.  It took quite a bit of futzing to get the reflection of the lamp out of the shot, but I was finally able to situate the clamp light in a place where it was not reflecting back into the lens.  I think I ended up using the other CFL as a sort of backlight, but I can’t remember the exact setup.

So, I’m down to one bulb and I’ve suffered some expected mercury exposure.  Remember folks, these things are filled with toxic gas, so they’re not exactly as “green” as we’re lead to believe.

Next project will be to build a more robust lamp with some specific criteria.

1. The lamp must be robust in nature.  It should be able to be moved without falling apart, and should be able to survive a minor bump or fall.
2. The lamp must feature high CRI bulbs to provide more reliable color results.
3. The lamp must be a more diffuse source then a bare CFL.
4. The lamp must be more flexible in positioning then my ghetto lights.

Let the low budget lighting adventure continue…

Knowing that the end goal is shooting some of the food we cook for my wife’s site The Fashionable Foodie, I decided that I would set a few parameters before shopping.  Since we’re both new at this, I wanted to minimize as many variables as possible.  We have very little experience plating, styling, photographing, lighting and setting scenery, so I thought it would be best to cut out table settings entirely.  I decided to begin by building a table-top cyclorama to shoot all the subjects on a plain white background for simplicity’s sake. 

I knew that I’d be able to obtain a workable sweep by purchasing some simple poster board at the craft store, but when I arrived in the posterboard aisle, I discovered that they also sold incredibly thin sheets of dry-erase plastic.  I was intrigued with the idea of shooting on a semi-reflective surface and it was only $1.99 a sheet, so I picked up a single sheet to play with.  I also grabbed a sheet of 20×30 white foamcore board for $1.50 and a sheet of white posterboard for 50¢.

In walking out of the store, the wind started to blow and the incredibly slick sheet of dry-erase plastic slipped out from my stack of items and I ended up stepping on the sheet.  Grrr.  I dented one side of it up pretty good, but pride prevented me from going in and buying another sheet.  Next stop was the Home Depot.

Note: the dry-erase sheeting comes with a gigantic 2″x4″ sticker on one side.  I ripped the sticker to shreds trying to get it to peel off before just cursing the whole thing.  Your experience might differ from mine, but that sticker ain’t going nowhere.  I just have to photoshop that part out if it makes it into a picture.

I already had some negative assumptions about Compact Fluorescent Lights (CFL) before I arrived.  What people don’t understand about them is that unlike a conventional fluorescent bulb, the ballast has to be built into the base of each bulb in order to make it compatible with existing Edison screw-socket lamps.  In a standard commercial fluorescent lighting fixture, the ballast alone is at least twenty bucks, in a film-quality fluorescent light, the ballast is at least half the total cost of the fixture.  The economies of scale dictate that the quality of ballast included in a disposable household fluorescent bulb can not be equal to the quality of ballast in a more robust fixture.

But what’s a ballast anyway?  Fluorescent lamps are really just a gas-filled tube that has an arc of electricity running through it, but there’s a negative differential resistance within the lamp.  This means that while lot of current is required to ignite the arc, after the current begins to flow, the resistance drops and allows even more current to flow.  This is where the ballast steps in to regulate the flow of current through the lamp so that it doesn’t run away.  Also, the larger the tube, the more voltage is required to ignite the lamp from a cold start and ballasts also have a system to ignite the lamp as well as regulate the voltage after ignition.

Even though I did not think highly of the CFL system for photographic use, I did not really want to get into the process of building a fluorescent fixture from scratch.  I decided that I would buy the best two CFL bulbs I could find and see what my results turned out to be.  

Entering the aisle of Home Depot where they keep the bulbs, I was a bit shocked to see that nearly the entire aisle was taken up by a single brand of CFL bulbs – “n:vision“, a brand I’d never even heard of before.  This certainly would work to streamline my choices, as I couldn’t really find any competing manufacturers products on display.  This kind of ubiquitous primo product placement made me wonder if n:vision might be the Home Depot store brand, but I could find no solid leads on my iPhone while standing in the aisle.

I did discover a promising series of traditional bulbs from Phillips under the “Natural Sunlight” product line.  These bulbs were marked with a corelated color temperature of 5000 K and a CRI of 92.  While the color is on the warm side of daylight, the CRI is incredibly high for a household grade fluorescent tube.  That being said, I’d already decided I was going to go with a CFL solution so I made a mental note and selected a CFL bulb instead.

I settled on the n:vision “daylight” bulbs.  I found a two pack of the bulbs for $8.97 and just decided that if they didn’t work out, they’d be fine to use in any other lamp in the house.  The package was listed as “100w equivilant” bulbs, 27 watt usage producing 1400 lumens.  In tiny print on the back, they listed “5500°K” with no CRI number.  I assumed this would not be a high CRI lamp, but decided to see what it looked like anyway.

The fixture I set out to mimic is the Lowel Ego light, but in examing the lamp’s design, I couldn’t understand why they were using a giant panel of diffusion in front of an already diffuse source.  Obviously, you greatly increase the surface area of your light by shooting it through diffusion, but I assumed you’d also be losing a significant amount of your already diffuse source when shooting it through more diffusion.  Because of this assumption, I decided to purchase two individual fixtures and use the lamps in an open-face situation.

Click to see larger

I found two low quality clamp lights that would fit the bill perfectly.  The two fixtures, both from Commercial Electric, were simply a plastic edison screw-socket fitted around an aluminum reflector, wire clamp and a 6′ cord.  I picked up one in an 8 1/2″ size and one in a 5 1/2″ size.  

Getting everything home, I decided to setup in the kitchen.  I taped one end of the plastic dry-erase sheeting to the tile wall about a foot up and let the sheet drop naturally into a sweep.  I screwed the CFLs into the fixtures and noticed the first problem: they stick out the front about an inch on each fixture.  Guess I won’t be able to lay these units face-down.  I started positioning them and immediately noticed my second problem: the clamps won’t grip to anything that isn’t shaped like a board.  The back of a chair worked perfectly, however the rounded top of our paper-towel holder caused the unit to droop and slip under the weight of the CFL.  Positioning these lamps was going to be a pain.

I fired up the lights and noticed the third problem.  When placed in these aluminum reflectors, the diffuse light of a CFL becomes much more directional then I expected.  (This is a “duh” moment for me.)  Where I expected the slightly unnatural diffuse omnilight of a Kino-Flo, I instead saw a very directional, if diffuse source…perhaps similar to an inkie shot through some frost.  Diffuse, but still very directional and casting some noticeable shadows on the subjects.

Click to see larger

Again, this is a point when I realize that shooting small objects is harder then I already assumed it to be.  I positioned a tomato on the acrylic sheeting and tried to start lighting it.  With such a small object, I found that I could light it pretty well, but I had trouble blowing out the white background and still keeping the subject properly exposed.  What I really needed was a way to position a light overhead, but given the constraints of the kitchen and the clamp lights, that wasn’t going to happen.

The CFLs are far more intense then I expected them to be and the clamp fixtures were really not suited to support the weight of a CFL bulb, perhaps those crazy Lowel cats were onto something when they designed that Ego light.

Pictured below is the tomato captured in the first lighting test.  I spritzed it with a bit of water to give it some much needed definition.  There’s just some white balancing in Aperture 2.0 and a little bit of exposure control.

Click to see larger

Decent for a first attempt.  A nice shadow at the foot of the tomato, a bit of a trailing shadow to the upper right side.  Some good specular highlights in the water and relatively good color reproduction.  The bad?  Well, the cyc isn’t lit evenly under the two source lights.  The sides of the tomato wash out, diminishing definition and separation from the background and the whole scene has a certain “tone” to it even after color correction.

In investigating further, I could not find reliable information about the CRI of the n:vision bulbs, but after shooting with them for a few days, I came to believe that they’d probably rank in the low 80′s.  There are some noticeable color casts in the green and somewhere in the blue spectrum that are very difficult to completely eliminate.  Pictured below is a completely uncorrected shot of some steamed clams.  To my eyes, there is a noticable yellow/green cast to the pale tones within picture.  The white point is not far off, but you can see some distortion in the yellow garlic and the coral tones of the shells.  The distortion seems to enhance the green within the yellows and mute the corals by diminishing the magenta tones – exactly what you’d see with a low CRI Cool White bulb, just to a lesser degree.

Click to see larger

For bright and vibrant colors, the color cast doesn’t seem to be very much of an issue, but for very light and near-white colors it is far more of a problem.  In the final picture that was used for this particular recipe, I struggled for a good period of time trying to remove the tarnished green hue from the garlic butter.  I don’t think I was fully successful in the end result.

Conclusions?

I’ve determined that it’s not really possible to build an acceptable lighting setup from your run of the mill big-box hardware store.  At least, not in the CFL realm.  

Manufacturers are generally not posting CRI numbers on their CFL offerings which generally indicates to me that they’re not going to be very good at color reproduction.  High CRI compact fluorescent lamps are sold online by many sites that offer “full spectrum” lighting, but you’re generally not going to find those same bulbs in your neighborhood hardware store.

Additionally, when placed in an aluminum reflector, the CFL lamps are far more directional then I’d expect.  If I were to redo this whole setup, I’d probably build something closer to the Ego setup with a large panel of diffusion to make a larger emission source.  The clamp lights are just too directional to be used as a single source, and too limited to be used in a multi-point lighting situation.  I plan to revisit this project and construct a ghetto version of the Ego lamp using some full-spectrum bulbs in the future and I’ll post my results when I get around to it.

Despite the obvious deficiencies, I was able to build a very affordable lighting solution that, although imperfect, allowed me to continue shooting food without relying on any natural lighting sources.  In the future, I’ll revisit this setup and see what I can do to improve functionality and color fidelity, but for the purposes at hand…it will do.

Totals:

• n:vision “daylight” 100w 2-pack: $8.97
• 5.5″ Clamp light: $6.47
• 8.5″ Clamp light: $7.47
• 1 Sheet acrylic dry erase sheet: $1.99
• 1 sheet white posterboard: $0.49
• 1 sheet 20×30 white foamboard: $1.50

• Total: $26.89

Well, they’re not deliberately trying to be deceitful, what they’re using is a Correlated Color Temperature or CCT.

Effectively, the rating listed on the packaging of a given bulb indicates the incandescent color temperature that the light will most closely mimic.  Because our eyes are incredibly good at adapting to different qualities of illumination, this correlated color temperature measurement is essentially “close enough” for every day life.  For example, “Warm White” bulbs are listed with a CCT between 2800 K to 3500 K, and for most purposes are similar in color tone to common tungsten-based interior lighting.

An important side note to consider is that the common tungsten color temperature of 3200 K is generally only found in photographic quality quartz tungsten lighting.  The standard 60 watt light bulb used in many homes produces much warmer illumination in the range of 2800 K, which is why color film balanced for Tungsten can still appear quite warm when exposed under lower wattage household lights.

So while this CCT rating is acceptable for general lighting purposes, it does not convey the true color properties for photographic purposes.  The illumination produced by a fluorescent light is a result of a gas discharge reaction producing ultraviolet radiation that causes phosphors deposited inside the lamp to fluoresce in the visible spectrum.  While the light produced by these lamps within the visible spectrum and appears to our eyes to be white light, the actual spectrum is discontinuous in nature.

When viewed on a spectrograph, Tungsten and Daylight illumination both appear as a curve across the visible wavelengths – essentially, it is a rainbow that progresses smoothly through all the visible wavelengths.  Conversely, fluorescents and other vapor-discharge lamps produce illumination that shows huge peaks and valleys across the visible spectrum.  While our eyes are able “sum” these peaks and valleys into something we can perceive as white light, imaging chips and film will more accurately capture the higher intensity in narrow bands of the visible spectrum in reflected colors.

What does that mean in plain English?  Essentially, the light produced by a fluorescent bulb can have intense slivers of certain colors along with incredibly dim expanses of other colors in-between.  While the light may appear white, when it strikes a colored object, the color that is reflected from that object is tainted by these deficiencies in the spectrum.  For example, certain Caucasian skin colors when photographed under a Cool White fluorescent source can often appear sickly and pasty.  This is caused by the huge green spike inherent to the light source reflecting from the surface of the skin and negating a large portion of the magenta spectrum present in normal skin tones.  In plain English, it makes people look more green then they really are and it’s the magenta in Caucasian skin that provides the “healthy glow”.

So from this, we can see that Fluorescent lighting provides a challenge when it comes to reproducing color, but where there is a challenge – there is often a half-assed solution already provided to the marketplace.  Welcome the CRI number.

CRI stands for Color Rendering Index and it is a measurement used by lighting manufacturers in an attempt to describe how closely a fluorescent light can mimic daylight in reflected color reproduction. A CRI number of 100 (which is essentially impossible) would indicate that a discharge lamp perfectly equals daylight.  Film generally requires a CRI of 95 or above to avoid any color casts, and any light below 90 will most likely exhibit distorted reflected colors when compared with daylight or tungsten.

It’s important to note that what we’re talking about is distorted colors, not distorted white points.  While you can use software to correct distorted white point captured under fluorescent lighting, adjusting the white point does not balance the distorted colors within the scene.  If an image exhibits a greenish color cast from a fluorescent lighting source, you can easily correct the overall white point, but what you can’t do as easily is correctly re-introduce the individual colors that were muted or enhanced under the discontinuous spectra of the fluorescent illumination.

For purposes of photography, it’s worth the effort to seek out the highest CRI fluorescent lamps to ensure that you’re starting out with the highest baseline color fidelity so that more of the actual color is recorded correctly in the initial capture.

So, now that we understand what a CCT measurement and a CRI number mean, we can drive out to Home Depot and see if we can build a suitable lighting solution from off-the-shelf parts.

Up next, adventures with CFLs.

Since I’ve challenged myself to try and create a lighting solution similar to the Lowel Ego light, understanding some basic fundimentals of fluorescent light will greatly help in choosing an “off the shelf” system that will produce optimal results.

The color of visible white light is commonly described using a standard of Color Temperature measured in degrees over the Kelvin scale.  In photography, we are generally divide light into the two ranges of “Tungsten” at 3200° Kelvin and “Daylight” at around 5600° Kelvin, but since those measurements only apply to incandescent lighting, all too often the concepts of the Color Temperature scale are not fully understood.  At it’s basic level, color temperature is exactly as the name describes: color of the electromagnetic radiation (light) similar to that emitted by a pure black object heated to a given temperature.

Color | Temperature.

Woah, woah, woah…what’s this black body?

Well, in physics, they visualize color temperature through heating a “theoretical” black body radiator.  If you want to go steady with a physicist, I’d that you read up on Planck’s Law, but if you’re an average Joe like me, I’d suggest that you just try and absorb these CliffsNotes.  I say “theoretical” because this object does not actually exist, it’s an ideal form of matter that absorbs all electromagnetic radiation and reflects nothing.  When this black body is heated it gradually begins to emit thermal radiation.  As the temperature of the body increases, that object begins to incandesce, emitting radiation as visible light. 

An instant analogy are the coils you see in any electric oven, space heater, or toaster: at room temperature, the coils appear close to black, but as they heat up they become incandescent, emitting some of that thermal radiation as a warm red light.  For an ideal black body radiator, as the temperature reaches 3200° Kelvin, the light that is emitted is visually similar to the color of light emitted by a photographic quality tungsten-quartz lamp. As that temperature approaches 5600° Kelvin it emits light similar to average daylight.  So, the color temperature scale is a relative or comparative scale that indicates the visual similarities between this black body radiator and a given light source. 

Why is this important when considering fluorescent lights for photography?  Well, it’s important because fluorescents are NOT incandescent lights.  Unlike tungsten lighting, there is no solid filament within them that is being heated to produce illumination through incandescence.

Since Color Temperature is the measured color of incandescence, it does not apply to discharge illumination sources such as fluorescent lights.

Wait, what?  Yes, that’s right.  If you see a specific color temperature on the packaging a fluorescent light, you can be assured that the measurement is a lie. 

Next, we’ll start to explore the magic of the Correlated Color Temperature (CCT).

**I updated this post this morning to correct a few spelling errors and add in wikipedia links.  I’ll try a to have a little less vodka before posting next time**