High Output AC LEDs

An LED consists of a chip of semiconducting material doped with impurities to create a “p-n junction.” As in other diodes, current flows easily from the p-side, or anode, to the n-side, or cathode, but not in the reverse direction. As illustrated below, when the opposing electrodes of the p-n junction have different potentials, electrons fall into the lower energy level, releasing energy in the form of a photons or light. LEDs, by nature, require direct current (DC) with low voltage, as opposed to the mains electricity from the electrical grid that supplies a high voltage with an alternating current (AC).

LED lights used in motion picture lighting applications fall into a category of LED technology called AC LED lighting. The term AC LED lighting refers to illumination generated by High Power LED (HPLED) light engines supplied with a sinusoidal AC voltage source—typically the utility line voltage (e.g., 120 V in the U.S., 100 V in Japan, 220 V in Europe). AC LEDs present many advantages over incandescent light sources including lower energy consumption, longer lifetime, improved robustness, smaller size, faster switching, and greater durability and reliability. For these reasons, but principally because of its high luminous efficacy, AC LED lighting has tremendous potential to become the dominant type of lighting in motion picture production. However, they are relatively expensive and require more precise current and heat management than traditional motion picture light sources.

One of the key advantages of LED-based lighting is its high efficiency, as measured by its light output per unit power input. White LEDs quickly matched and overtook the efficiency of standard incandescent lighting systems. In 2002, Lumileds made five-watt LEDs available with a luminous efficacy of 18–22 lumens per watt [lm/W]. For comparison, a conventional 60–100 W incandescent light bulb produces around 15 lm/W, and standard fluorescent lights produce up to 100 lm/W. In September 2003, Cree, Inc. introduced a white LED light giving 65 lm/W at 20 mA, becoming the brightest white LED commercially available at the time, and more than four times as efficient as standard incandescent lights. In 2006 Cree, Inc. demonstrated a prototype with a record white LED luminous efficacy of 131 lm/W at 20 mA, which is even better than standard fluorescent lights. Since then, LEDs have been proven (in lab tests) to surpass the assumed theoretical limit of 200 lumens per watt. For instance, Cree has claimed to have a laboratory prototype LED achieving 231 lumens per watt (typical of high power LEDs the correlated color temperature was not great, it was reported to be 4579 K.) However, these efficiencies are for the LED chip only, held at low temperature in a lab. In a lighting application, operating at higher temperature and with drive circuit losses, efficiencies are much lower. In 2009, a United States Department of Energy (DOE) test of commercial LED lamps showed that average efficacy was about 46 lm/W at a time when manufacturers were claiming levels of 105 lm/W (e.g. the XLamp XP-G LED chip pictured below.)

Cree's high-power LED XLamp 7090 XR-E Q4

Such exaggerated claims on the part of manufacturers is common when a new technology comes to the market. Why should the manufacturers of LEDs be any different? In the marketing of their products, manufacturers of LED luminaries put a little spin on the scientific data, which has a tendency to cloud not only the issue of lumen efficiency, but also, as we shall see, of color rendering, color temperature, useful life, and finally energy savings. While there is truth in the claims as they pertain to single LED diodes, as the US Department of Energy (DOE) study (more details below) clearly demonstrates, it has turned out to be much more difficult for fixture manufacturers to realize anything close to the kind of lumen efficiency, color rendering, and lamp life within the framework of a practical light fixture that the LED manufacturers publish for their emitters. Some fixture manufacturers have made a substantial commitment in R&D and have made enormous progress in addressing some of these issues. For other manufacturers the key to success has been to downplay the limitations and keep the price of their fixtures low. For this reason, to pick the right LED luminary for a particular job it helps to have a little knowledge of the technology.

Manufacturers have come up with many ways to use LED technology in illumination devices. The common types are pads, small panels, lens lights, ring lights, larger panels, and RGB washes. In addition, there are small theatrical lights or entertainment venue lights, and wide variety of architectural lights in various shapes and sizes. Where it is beyond the scope of this article to analyze the advantage and disadvantage to each, I would suggest the 4th Edition of Harry Box’s “Set Lighting Technician’s Handbook” for a comprehensive survey of some of the most promising fixtures currently on the market. I will concentrate instead on the inherent advantages and limitations of this new technology in motion picture applications. Without a doubt LEDs have become one of the most efficient light sources available (Plasma Lamps being the other.) But, before the full potential of High Power AC LED lighting can be realized for motion picture lighting applications, LED manufacturers must overcome some key barriers: color rendering, cost, power quality and versatility. Let’s now explore the issues surrounding each of these in more detail.

The Color Rendering/Cost Trade-Off

A primary problem for manufacturers of LED fixtures for motion picture applications is how to manage the color spectrum of their fixtures. As we will see, this is a complex issue and how it is addressed ultimately determines the suitability of the fixture for motion picture lighting applications. At this point in time in the manufacture of LED luminaries for motion picture applications there exists a trade-off between color rendering and cost. How a manufacturer trades one off for another depends on what approach the manufacturer takes to creating "white light."

There are two primary ways of producing high intensity white-light using LEDs. One is to use individual LEDs that emit three primary colors — red, green, and blue — and then mix all the colors to form white light. The other is to use a phosphor material to convert monochromatic light from a blue LED to broad-spectrum white light, much in the same way a fluorescent light bulb works.

Phosphor White LEDs

A "white' LED comprises a die (diode chip) that creates blue light. A portion of this blue light is used to activate a phosphor layer. Once activated, the phosphors create light of other colors that blend with the blue light to produce a fuller spectrum white light much in the same way a fluorescent light bulb does. In effect, a fraction of the blue light undergoes what is called a Stokes shift and is transformed from shorter wavelengths to longer. Called remote phosphor technology, a primary benefit to this approach is that the color spectrum of a given LED can be manipulated to a desired "white light" by applying phosphor layers of distinct colors. Depending on the chemistry of the phosphor used, the color balance of the resulting light can be correlated to daylight, or stretched closer to a tungsten color balance. The cooler white LEDs use semi-transparent phosphors so the blue "pump" color comes through. In contrast to blue LEDs, the warmer white LEDs are opaque to the pump color, and are therefore much lower in efficiency. This Stokes shift process reduces the total output , so there is a tradeoff in lumen output with warmer color temperatures and broader spectrum white LEDs.

Even though phosphor based White LEDs have a lower efficiency than normal LEDs due to the heat loss from the Stokes shift (as well as other phosphor-related degradation issues), the remote phosphor method is the most common method for making high intensity white LEDs for a broad range of applications including home, architectural, industrial, and motion picture lighting. The reason for this is that the design and production of a light source or light fixture using a monochrome emitter with phosphor conversion is simpler and cheaper than a complex RGB system.

The less expensive motion picture LED lighting instruments are affordable because they use the same White Phosphor LEDs as those mass-produced for home, architectural, and industrial lighting applications. Stamped from the same semiconductor wafer as the chips used in these other applications, what differentiates the chips used in motion picture lighting instruments is the higher tolerances for flux (output), color (CCT and Green/Magenta bias), and forward voltage used in the sorting of chips during a process called "binning" (illustrated below.) Put another way, they are simply the cream skimmed off the top of a production run.

Given the irregularities inherent in the manufacture of the semiconductor wafers from which White Phosphor LEDs are stamped, the LEDs in a production batch are all slightly different. In a mechanized testing procedure, they are sorted and grouped together into bins according to their flux and color. Binning has been refined over the years, and these days the tolerance of the best binning systems allow barely perceptible differences between LEDS from a selected bin. The difference in color between two sources is quantified using what is called the "MacAdams ellipse." A MacAdams ellipse defines the distance at which two colors that are very close to one another first become distinguishable to the human eye as different colors. As illustrated below, for a given point of color on the chromaticity diagram, the MacAdams ellipse defines the contour around it, where the colors that surround the point are no longer indistinguishable from that of the point.

Binning has now been refined to the point where LED manufacturers can now bin their LEDs to separate them to within two to four MacAdams "steps." In other words, the LEDs within a single bin may be as much as four distinguishable separations apart from one another, where one MacAdams step is not visible, two to four steps is barely visible, 5 or more is readily noticeable. Each LED manufacturer has its' own unique bin parameters (one has 15 distinct color bins for "warm white" alone.) It is also generally true that the tighter the bin parameters (fewer MacAdams Steps), or the higher the flux (output), the higher the cost of the LEDs (as much as 2x-3x.) And, to complicate things even further, LED suppliers charge a high premium to fixture manufacturers who buy from just one bin.

To be cost effective fixture manufacturers must mix LEDs from a range of bins. This has prompted some manufacturers to create recipes using computer modeling. For example, for each LED from a slightly green bin, the recipe compensates with an LED from a magenta bin. This way the fixture designer can assure that the sum balance of the light from all the LEDs in each fixture falls within their prescribed parameters for color balance and green/magenta shift. For a daylight-balanced LED fixture a well designed system can yield a light fixture having very respectable tolerance of +/- 50K.

Fixture manufacturers for motion picture lighting generally rely on just such careful binning practices to provide as accurate and consistent color performance as they can. But this approach to white light has an inherent limitation: regardless of how tight the bin parameters are set, by their nature the spectral distribution of Phosphor White LEDs is less than optimum for motion picture lighting applications.

Spectral Power Distribution

The spectral power distribution of a lamp indicates how much energy is present in each part of the spectrum. Natural sunlight and incandescent lamps have a continuous spectrum. Discharge lamps, like HMIs, have a discontinuous spectrum with a series of emission lines at different wavelengths. To fill in the missing colors, they contain additives that contribute the absent parts of the spectrum and make it continuous. Given how they produce white light, Phosphor White LEDS have a unique discontinuous spectral quality that is unlike that of natural daylight, HMI, or incandescent light. In the case of 3200K White LEDs, the phosphors added shape the spectral distribution by enhancing certain colors in the spectrum to simulate the spectral distribution of incandescent light. As a result, the spectral distribution of Phosphor White LEDs resembles a series of peaks and valleys. If you look at the spectral power distribution of a white LED (below left) you will notice a big spike at about 465nm (the blue LED) and a broader bump between 500 and 700nm produced by the phosphors. Even though the spectral power distribution has these peaks and valleys, the human eye perceives the light as white light.

The curves shown above are the spectral power distributions for the three approaches to producing white light that we will discuss as well at that of an incandescent lighting instrument. While the discontinuous spectral distribution of Phosphor White LEDS (far left) may appear white to the eye, and the color of objects illuminated by it appear natural to the eye, to film emulsions and digital imaging systems designed to reproduce accurate color under continuous spectrum light sources (like daylight or incandescent lamps), the color of the same objects will appear unnatural on screen. That is, the hue of an object being illuminated by this "apparent white light" can be drastically different than expected when it appears on the screen. For example, below is a "Macbeth chart" contrasting the resulting effect upon different color swatches of studio tungsten light and a representative Phosphor White LED lighting instrument.

Split Macbeth chart: each color patch shows the visible effects of studio tungsten light in the top half of the patch, and a representative Phosphor White LED lighting instrument in the bottom half.

A common test chart used for assessing color performance of motion picture imaging systems, the chart above would be more accurately called a "split Macbeth chart" because each color patch shows the visible effects of the two light sources – studio tungsten in the top half of the patch, and the Phosphor White LED lighting instrument in the bottom half. Although your computer display is not likely to be a calibrated reference monitor, the wide variations in color patch hue caused by the discontinuous spectral distribution of the Phosphor White LED lighting instrument should be readily apparent.

What accounts for these results? First, as you can see from its' spectral power distribution above, Phosphor White LEDS, compared to continuous light sources, have no output at wavelengths shorter than about 425nm, which means that violet colors don't render well. Second, there is minimal output in the medium blue-cyan-turquoise range from about 465-510nm, which is why the aqua-type colors don't render well either. Lacking these complementary colors within the spectrum, skin tones and warm, amber-yellow colors don't stand out. Third, with the long-wavelengths cutoff in the high-600 nm range, pinks, reds, oranges, and other long wave-length colors tend to look a little dull under Phosphor White LEDs, compared with how they look under continuous spectrum light sources (daylight, HMI, Tungsten) which extend all the way out on the long-wavelength end. Finally, as you can see from the gray scale at the bottom, this particular Phosphor White LED Luminary has a overall magenta bias. It is a common mistake to think that a custom camera white balance can correct for the deficiencies of LEDs. While you can white balance out/time out this magenta bias in digital video cameras/digital film intermediate, the camera/timer is not able to replace the parts of the spectrum that are missing all together. And since gels only rebalance the spectral distribution of a light source by passing the wavelength of the color that they are, gels cannot correct for these deficiencies either because there is not light of those wavelengths to pass in White Phosphor LEDs to begin with.

Left: Tungsten source, Right: White Phosphor LED source.


This inability of LEDs to render color is very visible in tests recently performed by The Academy of Motion Picture Arts and Sciences (AMPAS) as part of their "Solid State Lighting Project Technical Assessment" (see below for details.) In one (above) a model was photographed wearing a dress that had a number of different blue/cyan tints. Footage was shot with both a true tungsten source and a White Phosphor LED source. The tungsten-lit footage displayed all of the subtle differences in blue/cyan tones in the fabric, while the LED-lit footage, lacking cyan output, showed just a nice blue dress, without the same richness of hue. You can also see above that, absent cyan, the skin tones don't stand out because that complementary color (cyan) within the spectrum is not present. Since the light doesn't put out much cyan, the camera/film simply can't record it. And, as Cinematographer Daryn Okada, ASC, discovered the hard way, color gel packs, camera white balance, or digital intermediate timing can't bring it out if it isn’t there to begin with.

Like many of us, Daryn Okada uses LEDs as "touch up" lights to add a little something where key lights don't cover. Needing to touch up a face on one talent mark, he once hid a small LED unit behind a chair, to add some glow to an actress's face when she reached a mark where the keys had fallen off. "The manufacturer claimed the unit to be a 'tungsten LED source,'" he recounts. "She stopped right in the doorway, where I had this LED, and looked fine. But when I got the dailies back, her face was totally magenta." What's worse, Okada says the image could not be repaired in post, because there wasn't enough of the right color of light in the scanned negative for a color timer to bring out.

Had Daryn Okada been shooting with a digital video camera, he would have noticed the off color of the LED source immediately. But, given the limited spectral output of LEDs, his ability to remedy the problem would have been limited. If he white balanced the camera for the LED source, the background of the room beyond the doorway that was lit by tungsten lights would turn very green. In a mixed light situation such as this, the only alternative is to match the LED source to the prevalent tungsten source with a custom gel pack on the LED head. But, since gels rebalance the spectral distribution of a light source by passing the wavelength of the color that they are, gels cannot correct for these deficiencies either because there is not light of those wavelengths to pass in White Phosphor LEDs to begin with. In other words, White Phosphor LEDS are so deficit in certain parts of the color spectrum that by the time you came up with a color gel pack to match them to a continuous light source like a tungsten light, the LED panel would put out very little light with all those gels on it.

To understand why this is so, we need only look at a similar situation: the conversion of tungsten light to daylight using full CTB gel. As you can see in the Spectral Power Distribution graph of tungsten light above, tungsten light is so deficient in the blue part of the spectrum that it takes a quite saturated blue gel to balance it to daylight. In fact, the transmission coefficient of full CTB gel is only 36%, which means that it passes only 36% of the source. That is why gelling tungsten lights is a very inefficient way of creating a daylight source (a tungsten 1000w gelled with CTB becomes a 350W daylight source.)

A gel pack that would match a White Phosphor LED to tungsten light would have to include violet to extend its' spectral output below 425nm. It would have to include medium blue, cyan, and turquoise to fill in the missing wavelength from 465-510nm. Finally it would have to include pink, red, and orange to extend its' spectral output beyond its' 600nm cut-off. All of these gels would have to be quite saturated, since there is very little, if any, output of these wavelengths in White Phosphor LEDs to begin with. Imagine how much light you will get out of a LED light panel with such a gel pack (LED light panels put out barely enough to begin with, and have no output to waste to such accurate color correction.) Since, under most circumstances it is simply not feasible to completely match LED sources to tungsten sources with a gel pack, in mixed light situations such as these you are left, without recourse, with the off color generated by LEDs.

If the actress in Daryn Okada's shot were the model in the blue/cyan dress above, you can imagine what would happen when she stepped onto the mark lit only with the LED source by looking closely at the contrasting photos above. Since, under the circumstances, Daryn Okada could not white balance for the LED source (and it would not be feasible to match the LED source with a gel pack) the rich blue/cyan hue of her dress in the left photo would turn into the simple blue of the right photo. The vibrant skin tone of the left photo would turn into the flat skin tone of the right photo, and it would have an overall magenta cast to boot. This example clearly demonstrates that White Phosphor LEDs, simply by nature of their discontinuous spectral distribution, cannot accurately reproduce colors on screen under all circumstances no matter how tightly their bin parameters are set.

While they are less than perfect at reproducing parts of the color spectrum, the color rendering of Phosphor White LEDs may be adequate in certain situations. For a specific application, say where lights must be operated off of batteries, a LED fixture offers the unique advantage of greater power efficiency over conventional lights, which may out weight its shortcomings in color rendering. However, a Phosphor White LED is clearly not the best choice in applications where color rendition is critical (food/product shots) or mixed with a uniform continuous light source, such as a studio lit with tungsten fixtures, where its color deficiencies will be quite noticeable and unacceptable in comparison.

The OSRAM KREIOS LED chip

Since, the color spectrum of LEDs can be manipulated by blending phosphor layers, it is possible to tune their output specifically for the color sensitivity of film emulsions and digital imaging systems. The motion picture lighting division of OSRAM did just this in the KREIOS LED stage light module that Mole-Richardson uses in their MoleLED 12 Pack.

The OSRAM Kreios stage light LED module


The KREIOS stage light module (pictured above) is a metal core circuit board with 20 high-output blue LEDs each topped with a remote phosphor dome. The phosphor domes are an OSRAM proprietary design, which are blue light activated to produce light that is a very close approximation of Tungsten and Daylight. While, remote phosphor technology is commonly used to extend short wavelengths to create a fuller color spectrum in Phosphor White LEDs, OSRAM was the first LED manufacturer that set out to use remote phosphor technology to match the spectral sensitivity of Tungsten and Daylight balanced film stocks.

The spectrum distribution of the Daylight balanced MoleLED clearly shows the blue light emitted by the InGaN-based LED (peak at about 465 nm)
as well as the more broadband light created by phosphors.


And, with the help of Kodak, Panavision, Technicolor, and Shelly Johnson ASC, multiple film tests were shot with the KREIOS stage light module in order to verify the color generated by the MoleLED on set would translate to the color viewed in film dailies. This attention to detail and strict testing has resulted in a LED fixture that almost seamlessly mixes with existing Tungsten Halogen light sources, natural daylight, or any daylight balanced light source.

The spectrum distribution of the Tungsten balanced MoleLED clearly shows the blue light emitted by the InGaN-based LED (peak at about 465 nm)
as well as the more broadband light created by phosphors.


Even though OSRAM was successful in using remote phosphor technology to closely match the spectral sensitivity curves of Tungsten and Daylight balanced film stocks, this approach has several drawbacks. Since the OSRAM KREIOS stage light module is specifically designed for motion picture lighting applications, and not mass produced like Phosphor White LEDS, they are considerably more expensive – making the MoleLED 12 Pack (pictured below) one of the more expensive LED light panels on the market.

The MoleLED 12Pack

And, given the seemingly unavoidable energy loss in the Stokes shift, the remote phosphor approach is not the most efficient method to achieving white light with LEDS. However, much effort is being spent on optimizing these devices to higher light output and higher operation temperatures. With development ongoing, the efficiency of phosphor based LEDs is sure to rise, but without a mass market for remote phosphor LEDs tuned to our specific needs, there is not much hope of the prices coming down significantly.

Since, market competition often focuses on brightness (lumens/watt), some manufacturers opt for brighter LEDs with unappealing color characteristics, while others select LED chips with better color characteristics and rely on good optical design and drive electronics to produce reasonably good output. With all these factors in play, it would be quite misleading to judge the quality of an LED fixture strictly on the basis of brightness. Manufacturers using remote phosphor technology are forced to balance these competing priorities, and each approaches it in their own way.

Tunable Multi-emitters

A more efficient approach to generating white light with LEDS, and one resulting in more lumens/watt, is to use tunable multi-emitters. These lights use multiple LEDs that each emitt light at different wavelengths. The simplest form of a multi-emitter instrument is a "tri-color panel" using red, green and blue LEDs, also known as an RGB LED. While the RGB approach is capable of delivering not only a vast array of colors as well as tunable white light, the CRI (Color Rendering Index) is quite low. The typical color rendering achievable for such a system ranges between 40 and 60 depending on the targeted color temperature (3200K or 5500K.) Motion picture lighting typically requires a CRI in excess of 80, and higher the value the better. Therefore, an RGB tunable emitter may be well suited to color washing applications in stage lighting, it is less suitable for creating white light for motion picture lighting applications.

A variation of the multi-emitter technique that is better suited to motion picture lighting uses "cool white" and "warm white" LEDs within the same fixture, and provides a control to mix the colors as needed. The idea is that by mixing the two sets of LEDs, the user achieves a nominally correct CCT for any color temperature from tungsten to daylight. While this is an intuitive solution to color balancing that one would think eliminates the need for fractional CTO and CTB color correction gels to achieve an intermediate color balance, this approach does entail a compromise non-the-less.

If on the chromaticity diagram above, you were to plot the color point of two illuminants, all the colors that are possible by mixing the two colors of light will be located on the straight red line drawn between the two points. However, the line (black line above) that would be charted by heating a black body radiator (as it turns red, orange, yellow, white, and finally blue as it is heated) is not a straight line, so it is not possible to create light that remains neutral in terms of their green/magenta shift, while mixing only two colors.

Another variation of the multi-emitter technique that is even better suited to motion picture lighting uses additional LEDs that emit additional wavelengths to create a closer match to natural daylight or studio tungsten light. For instance, the addition of an amber emitter (creating an RGBA system) significantly enhances the CRI of the light generated while still maintaining the desired color point. The primary advantage to RGBA systems is that color flexibility is still achievable, but more importantly for motion picture lighting applications, it is possible to walk the white point along the black body line while still maintaining a CRI approaching 90.

RGBA emitters do however present LED lighting manufacturers with some design challenges. Delivering a consistent color across the beam without color fringing or shadowing can be a complicated task. Reducing the optical source size of the LED or LED system can improve the color mixing. One approach is to create a multichip LED (like the one pictured above), locating the individual colors as close as possible and therefore reducing the mixing distance. Although various multi-chip packages exist, few exist that enable both high flux density (output) and color mixing. Most multi-chip packages are limited to a low drive current, reducing the flux density and output. Alternatively one may use multiple discrete high output sources, but this increases both the optical source size and color mixing challenge. Such fixtures typically place a diffusion material immediately over the LEDs to help blend the discrete colors, but this comes at a loss of output.

Another problem is that, while it is relatively easy to put a dimmer on an LED, and blend two different color LED chips to achieve variable color mixing, as we saw above it is quite a different matter to track the color so that it remains on the black body locus at every point from daylight balance to tungsten balance. Maintaining a specific color temperature at a high CRI while dimming is made even more difficult by virtue of the fact that temperature in the LED changes when they are dimmed. Change in temperature shifts output wavelength as well as efficiency, and different LED chips change efficiency at different rates and at different temperatures. For these reasons, a more complex approach to dimming is required in order to control all these factors.

Some manufacturers start by incorporating more colors of LED chips into the fixture in order to counter the shift in green/magenta bias as the color tracks along the black body locus. The chips are controlled by microprocessors using algorithms to continuously regulate the intensities of the different colored LEDs so that the mixed light gives the desired color temperature at any dimmer setting. Using multiple LED colors within a white-light fixture increases the total color space achievable by the light, and in this way greatly improves on the limited color rendering of Phosphor White LEDs when used alone. Using this approach also allows the manufacturers to calibrate each fixture at the factory to compensate for any difference in the color output of the particular LEDs. Without a doubt, multi-emitter LED fixtures require a complex set of design considerations to deliver consistent color. These fixtures require intelligent drivers that can ensure Color and CRI consistency over time and temperature through predictive means.

The better multi-emitter designs, like the new Arri L7 LED Fresnels (see below for more details) incorporate a color-feedback system of self-monitoring sensors to ensure stable color across a range of output levels, as well as correcting changes in performance caused by ambient temperature and component aging, which ensures consistent color temperature. These units use an LED array optimized for film and video image capture, with standard color temperature presets. These arrays can create broad-spectrum white light and, unlike conventional multi-emitter LED lights, the color temperature remains consistent throughout the full range of dimming, ambient temperature and life of the unit. The drawbacks to this approach are that they are expensive (they require micro-processers) and their spectral distribution is still discontinuous. Even though they emit additional wavelengths to create a closer match to, say, tungsten lighting, as we can see by their spectral power distribution graph below, the resultant curve is very "spikey". This has the same noticeable side effects in color reproduction as discussed above. Like remote phosphor LEDs, tunable multi-emitter designs, simply by the nature of their discontinuous spectral distribution still cannot reproduce colors on screen with complete accuracy.

Unfortunately it is very hard to judge with your eye how an LEDs spectrum is going to line up with the sensors in the camera. As we have seen, eye response and camera response can be quite different. Additionally common color meters, like the Minolta III F, are completely useless with LEDs. The meter makes its calculation of the color temperature based on an assumption that the light source has a continuous spectrum. Color readings of an LED have been shown to be misleading for both correlated color temperature and green/magenta shift. To make matters worse the CRI ratings published by manufacturers can be misleading. As we have seen, both the remote phosphor and multi-emitter approaches to generating "white light" offer the possibility to tune output to a desired end. Where the CRI index indicates the ability of a light source to reproduce only 8 colors faithfully (a different 8 colors are used in Europe), it is possible for LED luminary manufacturers to tune their output to the limited color range of the CRI color scale below.

Also bear in mind that the CRI Index was not designed for photographic purposes, but simply to provide a reference scale for general illumination. It is as follows:
    - CRI 90 - 100: Retail (merchandise, artwork) and work spaces (design) where faithful color rendering is critical.

    - CRI 70 - 90: Most office, retail, school, educational, medical, and other work and residential spaces where good color rendering is required.

    - CRI as low as 50: Industrial security and storage lighting where color fidelity is not important.
As you can see the CRI index makes no mention of photographic reproduction. It is simply because there is no means by which to measure for photographic purposes the discontinuous output of LED lights (CCT is for continuous spectrum black body radiators only) that LED manufacturers use the CRI index. The upshot is that it is possible for LED manufacturers to tune the output of their luminaries to obtain high CRI ratings and deliver good color rendering to the eye while delivering generally poor color reproduction on the screen.

For these reasons, the best approach to judging the color of a LED luminary is to shoot tests. A side-by-side comparison using a color chip chart and a full spectrum light source, will clearly demonstrate how the camera system will respond to a specific LED fixture. To see such camera tests of wardrobe, set, and make-up for the different approaches to achieving white light with LEDs that we have discussed, use this link to the Solid State Lighting Project Technical Assessment generated by the Academy of Motion Picture Arts and Sciences.

Lumen Maintenance/Depreciation

How well a lamp maintains its lumen output over time is referred to as lumen maintenance. Greater lumen maintenance means a lamp will remain brighter longer. The opposite of lumen maintenance is lumen depreciation, which represents the reduction of lumen output over time.

A number of factors effect lumen maintenance/depreciation. In the case of discharge lamps (HID or HMI) and incandescent lamps (Tungsten), the repeated cycles of heating and cooling that their quartz envelope goes through, cause the molecules to lose their crystalline structure. As a result, the quartz envelope over time loses some of its transparency and becomes more opaque. This process is called devitrification. As the quartz envelope devitrifies, it blocks more of the light and prevents it from escaping from the envelope, contributing to the depreciation of the lumen output. Another cause of lumen depreciation in discharge lamps (HID or HMI) and incandescent lamps (Tungsten), is that metal from their filament or electrodes evaporates and condenses on the inner wall of the quartz envelope, causing blackening. This also contributes to the loss of light output.

A third factor that effects lumen depreciation is the intended use which is reflected in the bulb design. One way to think about HMI lamps is as a metal halide type lamp with a rock-n-roll life philosophy of "live fast, die young." To increase output, improve luminous efficacy, and color rendering, HMI lamps are designed with comparatively very short electrode gaps. However, their increased luminous efficacy places an increased load on the bulb wall which accelerates the devitrification of their quartz envelope; which, in turn, leads to increased lumen depreciation. In other words, their brilliance comes at the expense of lamp life. Over their relatively short life of 500-750 hrs, it is not uncommon for HMI lamps to lose 15-20 percent of their initial lumen output before they fail all together.

Evaluating the lumen maintenance of an LED luminary is more complex because in many ways it is more complex than traditional fixtures. An LED is an electromechanical system: in addition to the essential light emitting source, an LED luminary also includes a provision for heat transfer, electrical control, optical conditioning, mechanical support, and protection, as well as aesthetic design elements. Because the LEDs themselves are only one part of this elaborate electromechanical system, the affect of these other components must be taken into account when determining the lumen maintenance of an LED luminary.

For instance, while LEDS do not radiate heat like a tungsten filament, half or more of their input energy may be converted to heat in current carrying components. This heat must be conducted away from the diodes for an LED to operate efficiently. This situation requires a heat conducting assembly, be it a passive heat sink or active fan cooling, that will operate reliably over an extended period of time - most do not. For proper operation, the power supply and electronics of an LED luminary must provide a well-controlled DC drive current and other control features - most begin to fail long before the rated life of the product. Any optical components must also be able to withstand years of exposure to intense light and possibly heat without yellowing, cracking, or other significant degradation. Reflecting materials need to stay in place and maintain their optical efficiencies. Since it is nearly impossible, even in the best designed LED luminaries, to completely protect against system degeneration of the type described above, the lumen maintenance of an LED luminary is significantly less than that given by manufacturers for a single diode. For example, recent tests from the Caliper program (US Dept. of Energy) suggest that the lumen output of many LED luminaries depreciate to less than 50% after only 500hrs (see page 27 of the summary for examples of lumen depreciation after about 500 hours .) While, this might not be representative of every product, the vast majority of those that were tested fell far short of the manufacturers claims of lumen maintenance.

Color Shift

In addition to losing light output, an aging luminary, also exhibits color shift. With HMI bulbs, color temperature varies significantly with lamp age. A new bulb generally will output at a color temperature close to 15,000 K during its first few hours. After this short burn-in period, the color temperature reaches its prescribed value of around 5600 K or 6000 K. With age, however, its' arc length becomes larger as more of the electrodes burn away. Greater voltage is required to sustain the arc, and as voltage increases, color temperature decreases proportionately at a rate of approximately 0.5 -1 Kelvin for every hour it burns. Which means that by the time it fails, after approximately 750 hours of use, its' color temperature will be between 5250 and 4875K.

Determining the color shift of LED luminaries is also complicated. That is because color stability is not exclusively determined by the performance of the LED diode. Other factors that contribute to color shift in LEDs include LED design, materials, manufacturing processes, optics applied to the LED, and the temperature and time the LED operates. For instance, the increase in junction temperatures typical of LEDs over time can cause color shifts. Likewise, optical components in the LED luminary may discolor, crack, or significantly degrade after extended exposure to the intense light the luminary generates. Environmental conditions (including air quality) may cause materials in optical components to deteriorate. Finally, luminary design my create non-uniform color characteristics such as halos or yellowish, bluish, or greenish hues around the edges of the beam, and these color characteristics may vary over time for the reasons mentioned above.

A new technology also requires a new method of measuring shifts in color. Since LEDs are not "black body" radiators (that turn red, orange, yellow, white, and finally blue as they are heated), the Correlated Color Temperature (CCT) ratings in Kelvin are not adequate to describe color shifts in LED luminaries over time as it is in conventional lamps. This is another reason why the LED industry uses, instead, the "MacAdam ellipses" discussed earlier.


That samples in the Caliper study cited above (one is illustrated above) exhibited color shifts greater than a 36-step MacAdams ellipses and a 30 percent drop in lumens within 3000 hours, suggests that that LED luminary reached it's usefulness to serve even as a set practical for motion picture lighting long before 3000hrs (remember one MacAdams step is not visible, two to four steps is barely visible, 5 or more is readily noticeable.) Unfortunately, the manufacturers of LED luminaries for motion picture lighting purposes do not publish specifications for color shift over time; probably because, as the results of the Caliper study cited above suggests, they are significant and would appreciably shorten their claim of 50'000 hours of "lamp life." The better LED luminaries for motion picture lighting applications (the Arri L7 and Gekko Kezia) compensate for color shift with active microprocessor controlled color management that involves self testing using internal color sensors (see below for more details.) However, sensors and controls may themselves shift over time and affect color - so even such methods of color management are not full proof.

Lamp Life

In light of what we now know about LEDs, claims of 50'000 hr lamp life made by manufacturers like Litepanels should be taken as the marketing hyperbole that they are. Determining the "life" of an LED luminary is a very complex matter and how to do so is still being hotly debated within government and industry regulatory bodies. Until a standard measure of LED lamp life is settled upon, extravagant claims of 50'000 hrs should be taken with a grain of salt.

For conventional lamps (HID, HMI, Incandescent, Fluorescent), the method for determining "rated average lamp life" is well established and easy to calculate: it is simply the point at which half the lamps cease to emit light. For LED luminaries it gets quite a bit more complicated for several reasons. One complicating factor is that LED luminaries are made up of multiple components and usually have no replacement parts. That means, even though the individual LEDs in a luminary may be rated to last for 50'000 hours, its' actual life will fall well short of that mark if one of its other key components fails sooner. For example, HMI ballasts are manufactured to outlast their lamps many times over, because those lamps are relatively short lived and easily replaced. But at this relatively early stage in the development of LED technology, there's very little data available to confirm that LED drivers will last as long as the expected life of a LED diode that carries a 50'000 hr rating.

Another complicating factor is that LEDs have no filament or electrodes to burn out and thus generally keep on producing light, although at declining levels and a gradual shift in color. And, since a well designed LED luminary has a comparatively very long-rated life (as conventionally defined), over which its' lumen output drops continuously, it also has an appreciatively greater lumen depreciation over that life than does an HMI lamp. To extend the analogy used above, if HMI lamps are "Live Fast, Die Young" Rock-n-Rollers, LED Luminaries would be suburban Dads working 9-5 jobs. With their best years behind them and a few more pounds around their waist, they now spend their weekends driving the kids to soccer practice in a minivan, and usually fall asleep by 10pm on a Saturday night watching TV. Given their continuous lumen depreciation and color shift over time, it is clear that there comes a time when, like a suburban Dad, a motion picture lighting LED luminary has surpassed it working life and should be retired. Since it won't burn out in its' prime, like an HMI lamp, how do we determine when a light has surpassed its' useable life. Clearly, this new technology requires a new approach to determining useable "lamp life" than that used for conventional lamps like HMIs. For instance, how useful is a manufacturer's "rated average lamp life" of 50'000 hours when testing has found that complete LED luminaries can depreciate as much as 50% in just 500 hrs. Whatever the stated lifetime of any lighting product, it must reflect a meaningful statistical measure of the performance of a given fixture design. Clearly, in the case of LED luminaries to be meaningful "lamp life" must include not only the median time to failure of the array of diodes under normal operating conditions, but "failure" must also be defined as unacceptable lumen depreciation for the particular application rather than complete failure to light.

One such alternative rating system for LED Lamp Life proposed by the Department of Energy (DOE) LED Lifetime and Reliability Working Group denotes "B" and "L" factors; where B represents the interval of time it takes for half of the diodes to "fail" (called B50), and L is the lumen performance level defining a "low-light failure". For example, the LED luminary that depreciated 50% in just 500 hrs in the Caliper study above would have a B50/L50 Lamp Life rating of 500 hrs. In other words, after 500 hrs, half of the diodes will have considered to have "failed" because their output dropped to 50 percent (low light output.) In this rating system B represents the time interval in hours in which a percentage of diodes have failed, where L represents failure as defined by an unacceptable lumen performance as a result of lumen depreciation.

One benefit to this rating system, is that each industry can determine a "lamp life" that is meaningful to its' application. For example, a B50/L50 rating of 500hrs (a drop-off of 50 percent of the diodes (B50) to 50 percent of their original value (L50) in 500 hours) may be perfectly acceptable in warehouse illumination. In more demanding applications, like home illumination, more demanding criteria would be required in determining lamp life for it to be meaningful. Most LED manufactures for home illumination use L70, or the point where lumen output has declined by 30 percent from initial output, to define the end of an LED's life because the average human eye can't detect decreases in light levels up to that point when the light is used intermittently. Once lumen maintenance has depreciated below that point, we begin to perceive the LED luminary is not as bright as it was when we first bought it and so become dissatisfied with its' use for home illumination. Since an L of 70 (30 percent lumen depreciation) would be clearly unacceptable in more critical applications, such as traffic signal illumination, an even more stringent definition of low-light failure would be chosen. Whatever level of lumen depreciation is chosen for low-light failure, to be meaningful to it's users, it should be in line with existing lamp technologies used in that industry.

In the case of motion picture lighting, I would argue that nothing short of a B50/L85 criteria would be appropriate and meaningful, since that has been our experience with both HMI and Tungsten Lamps (as noted above the rated average life is the interval in which 50% of lamps fail and with an average lumen depreciation of 80-85%.) If we adopt this criteria for motion picture LED lighting luminaries, the rated lamp life in hours would be the interval in which a drop-off of 50 percent of the diodes (B50) to 85 percent of their original value (L85) occurs. Without a doubt, this is a more meaningful criteria by which to judge the lamp life of an LED array for motion picture lighting applications. Unfortunately, motion picture lighting manufacturers have not adopted this rating system - possibly because it would expose their exaggerated claims of 50'000 hrs as nothing more than marketing hyperbole. Short of hard data, how can we estimate the useable life of an LED light-panel? One way would be to apply a B50/L85 criteria to the Caliper test results for similar style architectural LED luminaries - after all their diodes are cut from the same semi-conductor wafer and arrayed in similar style housings. Applying the B50/L85 criteria to similar White Phosphor Array Type LED architectural luminaries suggests that the useable lamp life of similar type of LED luminaries designed for motion picture lighting applications is probably no more than 1500 hrs. To put that in perspective, it is the equivalent of burning through two HMI globes. Since these types of LED fixtures have no interchangeable parts that can be replaced after reaching low-light failure, after 1500 hrs the fixture can only be thrown away while an HMI head can be lamped with another bulb.

The Power Quality of AC LED Lights

Despite improved color rendering and significantly higher energy efficiency than incandescent lighting, the power quality of AC LED lighting has been a much less compelling story because of its reliance on Switch Mode Power. As was true of early electronic HMI ballasts, manufacturers of AC LEDs will have to address the relatively poor power quality generated by the Switch Mode Power Supplies (SMPSs) used in AC LED lighting ballasts before AC LED lighting can replace incandescent lamps in studios and on location sets powered by generators.

High Power AC LEDs use separate power supplies because they require more precise voltage/current management than traditional motion picture light sources. Unlike incandescent light bulbs, which illuminate regardless of the electrical polarity, LEDs will only light with correct electrical polarity. When the voltage across the p-n junction is in the correct direction, a significant current flows and the device is said to be forward-biased. If the voltage is of the wrong polarity, the device is said to be reverse biased, very little current flows, and no light is emitted.

.

While LEDs can operate on direct alternating current, this approach is unsuitable for motion picture lighting applications because it will cause flicker in the image. Since an LED only lights when forward-biased, when powered directly with alternating current they will only light with positive voltage, causing the LED to turn on and off at the frequency of the AC supply. As in the case of magnetic HMI ballasts, these pulsations of light will lead to flicker in the image unless both the power supply and camera shutter are tightly regulated. For this reason, LEDs used as a light source for motion picture production require direct current (DC) be applied to their diodes. To operate on AC mains power, LEDs need not only some type of AC-to-DC converter but also additional regulation of the DC to the diodes.

A small voltage change results in a exponentially large change in current.

LEDs require additional regulation of the DC to the diodes because the current drawn by LEDs is an exponential function of voltage. As illustrated above, a small voltage change results in a large change in current. It is therefore critically important that the right DC voltage be provided to the diodes. If the voltage is below the threshold, or on-voltage, no current will flow and the result is an unlit LED. If the voltage is too high, the current will go above the maximum rating, heating and potentially destroying the LED. To make matters worse, as an LED heats up, its voltage drop decreases, further increasing current. For these reasons, High Power AC LEDs require a high degree of power conditioning unlike incandescent light sources.

Left: The Cool Lights LED 600 Fixture. Right: Schematic diagram for an LED string with series resistor and linear voltage regulator.

Existing AC LEDs employ additional power conditioning in the head that is either a DC-to-DC Switch Mode Power Supply type (SMPS) with “constant current regulation”, or a Resistor Type with “linear voltage regulation.” While a resistor wired in series with a string of LEDs permits a linear voltage regulator to stabilize the LED current, this approach (used by Cool Lights and illustrated above) has several drawbacks. First, considerable energy is wasted in the series resistors. Second, the linear voltage regulator that converts the supply voltage to the desired voltage for the LED strings maintains a constant output by wasting excess electrical energy by converting it to heat. As such, this approach is highly inefficient and not ideal for battery operation. The same function is performed more efficiently by using DC-to-DC switched-mode power supply (SMPS) in conjunction with a constant current regulator in the light head.

Left: The Litepanel 1x1 LED Fixtre. Right: Schematic diagram for an LED string with constant current regulator.

The SMPS/constant current regulator approach, in contrast, regulates output voltage by rapidly switching a pass transistor (typically between 50 kHz and 1 MHz). In this approach (used by Lite Panels, Mole, & Zylites and illustrated above) voltage regulation is provided by varying the ratio of “on” to “off” time of the transistor. Since transistors have no resistance when "closed" and carry no current when "open" almost all the input power is delivered to the load; no power is wasted as dissipated heat. This higher efficiency is the chief advantage of switch-mode power supply when compared to a linear power supply that dissipates excess voltage in the form of heat to regulate its’ output. In addition to enabling the total LED string voltage to be a higher percentage of the power supply voltage, resulting in improved efficiency and reduced power use, the highly regulated power provided by SMPS also stabilize the light output of High Power AC LEDs over the wide range of voltages provided by batteries as they discharge.

Figure 1(a): schematic diagram for a SMPS type AC-to-DC Converter that converts sinusoidal AC voltage to DC voltage to drive a LED.

Since, LEDs used in motion picture lighting require direct current (DC) be applied to their diodes, to operate them on AC mains power requires some type of AC-to-DC converter. Again, because of their higher efficiency AC-to-DC switch-mode power supplies are almost universally used for this purpose.

Litepanel 1x1s use the Cincom TR70A24 SMPSs Type AC-to-DC Converter (above right)
which boasts a line regulation of +/- 1%, load regulation of +/- 2%, and an efficiency of 84%..

But, as Figure 2(a) below illustrates, the SMPSs used in AC LED ballasts can draw a very distorted current, and can result in current that is significantly phase-shifted with respect to the sinusoidal voltage waveform. For instance, the AC power supply that Litepanels uses for their 1x1 panel arrays have a Leading Power Factor of 0.54 and generate high harmonic distortion (THD upwards of 68.1%). As such, the AC power supplies of LEDs can have an adverse effect on power quality similar to that of CFLs described above.

Figure 2(a): Voltage and Current waveforms generated by SMPS type AC-to-DC Converter used to drive AC LEDs.

Like Fluorescent and HMI electronic ballasts, the power quality of SMPS-based AC LED ballast can be improved, but this comes at the additional cost of adding a power factor correction module to further condition and control the current drawn by the load. Unfortunately, the manufacturers of High Power AC LED Light fixtures for motion picture lighting applications generally do not give Power Factor specifications for their products. But, if the present state of Fluorescent and HMI lights is any indication, it is probably the case that the less expensive LED lights are not Power Factor Corrected, while the more expensive ones are Power Factor Corrected.

Distribution of harmonics generated by the power supply of the Litepanel 1x1 LED Fixtures.
Note: predominance of the 3rd, 5th, 7th, and 9th harmonics that don't cancel on neutral returns


Versatility and Control

The final issue LED Manufacturers will have to address, before LEDs will be widely accepted in motion picture production, is their lack of versatility and control. The drawback to existing LED light panels is that their light falls off very rapidly and is hard to control. These characteristics make LED light panels only suitable as Key sources in documentary interview set-ups where the Keys are typically positioned close to the interview subject. In that capacity LED light panels (with heavy diffusion) can generate a wonderful soft light that wraps around the interview subject without wilting them. However, in dramatic set lighting, where Key sources must be capable of throwing a distance, LED light panels have only limited applications as fill sources. The broad soft light they put out drops off too rapidly, and is too difficult to control, for them to be effective as a Key or Backlight source in dramatic set lighting. However, progress is being made in the development of a LED light with the versatility and control of a traditional Fresnel instrument. In April 2010, Litepanels introduced a prototype for their Sola-Series LED “Fresnels” at NAB. While definitely a step in the right direction in developing a production LED Fresnel light, the Litepanel Sola fixtures still don’t quite combine the advantages of LED illumination (cool-burning, energy-efficient) with the characteristics of a traditional Fresnel fixture.

The 75W Liepanel Sola 6 LED "Fresnel"

With the Sola fixtures, Litepanels has not overcome the basic problems of LEDs discussed above. For instance, Litepanels claims the 75W Sola 6 (pictured above) has the output equivalent to a 650W Tungsten, but comparing the photometrics published on their website to those of an Arri 650 Fresnel, the Arri has nearly three times the output of the Sola 6. Litepanels doesn’t give CRI ratings for the Sola Fresnels on their website, but when asked they say the CRI is in the 80s – which is still rather anemic compared to other light sources. And, even though, Litepanels has improved the Power Factor of the Sola Fixtures over that of the 1x1s, at about .75 (the 75W head draws a maximum of 100 Watts according to their technical specs.) it could stand further improvement (a Power Factor Corrected HMI has a Power Factor of .98 and Tungsten lights have unity power factor.) And, while the Sola 6 has an impressive spot to flood range (10 to 70 degrees), spot/flood capability is not the only characteristic that makes a Fresnel light versatile. Of equal importance is the ability to render clearly defined shadows and cuts. The ability of Fresnels to render crisp shadows make them ideal for creating gobo effects like window or branch-a-loris patterns. And, the ability of Fresnels to render clearly defined cuts enables their light to be precisely cut to set pieces and talent. Finally, Tungsten & HMI Fresnels have sufficient output that the crispness of their shadows or the hardness of their cuts can be varied by simply adding one of a variety of diffusion material to soften their output if desired. These are the characteristics of traditional Fresnels that make them extremely versatile, that the Sola "Fresnels" have not been able to emulate. To understand why it is so hard for LED manufacturers to produce the qualities of a true Fresnel, let's look at those qualities in more detail and how they are achieved in traditional Fresnel heads.

The Quality of Fresnel Instruments

A Fresnel lens can be regarded as an array of prisms arranged in a circular fashion, with steeper prisms on the edges and a nearly flat convex lens at the center. The prisms near the center of the light source act as "dioptric" lenses that magnify and concentrate the output of the lamp filament. At the same time, the multiple prisms mounted around the periphery of the lens (above, below, on one side and the other of the filament), act as "catadioptric" lenses that collect and intensify the light and redirect it in the same plane as the dioptric lenses towards the center. In this fashion, a Fresnel lens bends the light of a source into a column of nearly parallel rays as shown in the illustration below.

Not only does a Fresnel lens refract the diverging rays of light emitted by a point source (lamp filament) into a highly collimated beam of light, but moving the point source toward the lens floods the beam - increasing its spread and decreasing its intensity. Moving the point source away from the lens spots the beam, making it narrower and more intense.

At full flood, the beam is relatively even across a broad sweep (it has no central hot spot), then falls off quickly toward the edges, making for a very even field. As the lamp is spotted in, the rays become less divergent, more nearly parallel. The beam narrows and gets brighter at the center, falling off rapidly on either side. At full spot, the usable portion of the beam is narrow, about a 10 degree angle.

A common misunderstanding is that the reflector collimates the light of a Fresnel head. In fact, the purpose of the reflector is to double the intensity of its' output. When the light-emitting filament of the bulb is placed near the center of curvature of a spherical, concave polished mirror reflector, the reflecting surface creates an image of the filament. That image is located in the same plane, but slightly displaced from the filament itself. This has the effect of doubling the amount of light forward projected from the locale of the lamp filament

In other words, without the reflector, "this reflector light" (the dashed lines in the illustration above) would have been lost in the back of the lamp housing. With a reflector, these rays of light are collected and sent back to their point of origin where they emanate forward, parallel with the direct rays of light from the filament (the solid line in the illustration above), towards the back of the Fresnel lens where they are together collimated by the lens (for this reason the filaments of the bulbs used in Fresnel heads are designed with an open geometry to minimize blocking of the retro-reflected light - making them not quite an ideal point source. Now that all the light that emanated forward and back, emanates forward from a single point within the fixture (the filament and its mirror image), the light projected forward is doubled. The efficiency of this lamp/reflector design, the collimated quality of it's light output, and the ability of the Fresnel lens to focus the rays of light quickly and easily to obtain a desired intensity or beam width, is what makes the Fresnel head one of the most versatile fixtures to work with.

For example, a Fresnel head in spot position will "throw" light a great distance, meaning it will illuminate a subject to the same brightness at a much greater distance. This particular characteristic of Fresnel heads is very useful when they have to be used at a distance. But, Fresnel heads are useful in the making of motion pictures not only because of its ability to focus the beam brighter than a typical lens. What makes them incredibly versatile is that in flood position they also generate light that is crisp and has a relatively consistent intensity across the entire width of the beam of light.

Fresnels end up used in flood position a lot because that is where it creates its hardest, most delineated shadows (the flood/spot mechanism is often used for little more than to fine-tune intensity.) The more spotted in the fixture, the less sharp the shadow lines appear. In full spot position, rays from the Fresnel travel more nearly parallel, but some converge slightly and cross one another. This creates fuzziness to shadows cast from an object. If one wants to project a sharp shadow or make a pattern (e.g. the classic gag below of venetian blinds cast on a wall), one would want to use the light at full flood.

Another common application of a Fresnel in Flood is to create a consistent wash of light over a large area with multiple fixtures. For example, say that, to light a large room, several lights are required because the spread of one is not large enough. Fresnel instruments spaced evenly apart along one wall of the room can do the job nicely, if the lamps are set at full flood, and the edge of the beam of each light (the 50% drop-off point) is overlapped slightly so that it feathers into that of the next. Overlapping the beams in this fashion will create an even 100% intensity seamlessly across the whole room.

The highly collimated light of a Fresnel head will enable you to use barn doors to place the edges of the beam of light so that the light does not spill onto adjacent talent or set elements (if they are very reflective.) By closing the two large leaves of a barn door into a narrow slit you can make a narrow slash of light. The slash can be horizontal - for an eye light for example - or turned diagonally to make a slash across the background. When a very confined, narrow, circular beam is desired, replace the barn doors with a snoot. Snoots come in different aperture sizes so that you can adjust the beam width. You might use a snoot, for example, to light a set with small pools of light - lighting tables in a cafe for instance. As an added bonus, it is fairly easy to calculate the intensity and beam diameter of a Fresnel because it has a focused beam (doing so for soft-lights is not so straightforward.) It is for these reasons, that the Fresnel head is still probably the most commonly used fixture in motion picture production.

LED Fresnels

There are two big hurdles to manufacturing a LED Fresnel head of a practical size - say, the equivalent of a 650W Tungsten Fresnel. First, Fresnel lenses require a point source of light. Second, they are not very efficient and so require a lot of light. Unfortunately, at this time there is no single-diode LED with sufficient output to serve as a point source for a Fresnel lens, nor will there likely be one in the near future. To understand why that is, let's look at these hurdles in more detail.

At 15 lumens/watt, a 650W tungsten globe will generate 9'750 lumens. But, given the doubling effect of the Fresnel head's highly polished spherical reflector, nearly 19'500 lumens is forward projected onto the back of the Fresnel lens from a relatively localized source (given the open geometry design of the bulb's tungsten filament it is not quite a point source.) If we take even the brightest LED that exists today, the newest addition (as of 5/11/2011) to the CREE XLamp LED family that in lab tests generated 231 lumen/watt, it would require an 84.4 Watt LED of this efficiency to match the forward projection of that achieved by a 650W tungsten filament in front of a highly polished spherical reflector. Even if Cree can maintain that efficiency in what is presently the largest commercially available single-die component LED available, 10Watts, we are still looking at an LED array of at least nine LEDs to equal the forward projection of a 650W tungsten filament in front of a highly polished spherical reflector.

Another reason we are not likely to see a single diode LED with sufficient output is that it is highly unlikely that Cree can maintain that level of efficiency in practical application as the efficiency was optimized at 350mA in the lab test, and had a correlated color temperature of 4500K. By the time that technology is scaled up to a 10W single-die component LED, and phosphors are added to extend the color temperature to 3200K, you can be sure the end product will generate no where near the same lumen/watts. Put in perspective, we are still a long way off from having a single-die LED with sufficient output and correlated color temperature to match a 650W tungsten filament in front of a highly polished spherical reflector.

To make matters worse, the inefficiency of Fresnel lenses increase as the source of light diverges from that of an ideal point source. For example, given their open geometry, tungsten lamp filaments don't nearly approximate a true point source. As such light rays enter the prisms of a Fresnel lens at different angles (angles of incidence), and because some undergo multiple reflections or refractions or are totally internally reflected for this reason, not all of them emerge on the other side of the array. The width of the vertical step between grooves in a Fresnel lens also block light. This loss does not exist for rays parallel to the optical axis of the grooves. But, for rays making a large angle (20 degree or greater) with the optical axis, the loss can be significant. The more the source departs from an ideal point source, the greater the angles of incidence, the more light that is blocked by the steps. The more the light source approaches a point source, the more the light output favors the central portion of the lens, which has greater transmittance. Since there is no single-diode LED with sufficient output to serve as a point source for a Fresnel lens, by default LED "Fresnel" fixtures have to use an array of LEDs. Far removed from an ideal point source, an LED array not only increases the inefficiency of a Fresnel Lens (requiring a larger array) but also changes the quality of its light output from the collimated source that is desired.



The ARRI L7 LED Fresnels

Regardless of the difficulty, Arri seems to have developed a true LED Fresnel in their new L7 series LED Fresnel heads. At NAB in 2011 Arri introduced the first of their L-Series LED Fresnels: the L7-D, L7-T and L7-C. All three models share the same basic housing and the same 7" Fresnel lens, and all have output comparable in intensity and quality to a conventional 1K Fresnel.


As you can see in the pictures above, that compare the output of the L7 Fresnel to an Arri ST-1 Quartz Fresnel, the L7 Fresnel has clear and defined shadow rendering capability like that of the ST-1 Quartz Fresnel. And, as the pictures below demonstrate, the L7 Fresnel has a spot to flood range similar to that of the ST-1 Quartz Fresnel and excellent field homogeneity in both flood and spot.


And, just like the ST-1 Quartz Fresnel (pictured below), the beam of the L7 Fresnel (pictured above) is easily controlled with barndoors - enabling the light to be precisely cut to set pieces and talent (see far right photos above & below.)


And, given the discernable amount of light the L-Series Fresnel prototypes threw in a show demonstration video from IBEC in the fall of 2010, on what appears to be a 6x6 Ultrabounce rigged 20’ overhead, and under the high ambient light levels of the show hall, it seems the production model L7 Fresnel has more than enough output to waste some to diffusion and color gel if one so desires (a shortcoming to most LED panels is that they have barely enough output – and certainly none to waste.)


Since the L7 fixtures are in the proto-type stage (production models are due soon), Arri is being very reticent about how they are able to accomplish the characteristics of a true Fresnel from a multi-emitter driver. To speculate, I would have to guess they are using a simple lens system - possibly like the one illustrated below.

By moving a Fresnel lens forward and back in relation to an assembly consisting of an LED Array and a Condenser Lens with an extremely short focal length, the focal point of the Condenser lens moves in relation to the Fresnel lens much like the lamp/reflector assembly of a traditional Fresnel Head. This would have the effect of spotting and flooding the light output of the LED Array much like a Fresnel spots and floods the output of a point source. Of course this is nothing but speculation, until we are able to get our hands on an actual production model of the Arri L7 LED Fresnel.

Where the L7 models differ is in terms of color temperature. The D model outputs a daylight-equivalent 5600 K, the T model a tungsten-equivalent 3200 K, and the top-of-the-range C model offers total color control. With the L7 series, Arri achieves the best color rendition I have seen yet from a multi-emitter fixture. By blending color with a highly sophisticated LED engine, the L7 series is able to overcome the generally poor color rendering capabilities of other LED fixtures (both remote phosphor and multi-emitter.) Both the 3200 K and 5600 K color temperature models offer a CRI and CQS greater than 90 so skin tones, costumes and scenery appear more life-like. The L7-C's fully tuneable white light can be adjusted for different skin tones, camera sensors and mixed-light environments, while specific color shades can be matched through full gamut color mixing. Unlike other LED fixtures, this level of color control does not involve compromising the quality of the light field: the L-Series is unique in combining uniform light and single shadow rendition with absolute control of color attributes.

Split Macbeth chart: each color patch shows the visible effects of studio tungsten light in the top half of the patch, and a representative multi-emitter LED lighting instrument in the bottom half.
Note: this is not the L7 but results typical of the best of the multi-emitter LEDs up to the release of the L7s.

An added benefit to using a color blending multi-emitter LED engine is that the mix of different color emitters can be adjusted to compensate for the inevitable color shift and diminished output of the LEDs with age. Using an internal optical sensor, the L7's firmware performs this calibration function at switch-on thereby assuring consistent realistic color rendition throughout the fixture's life and between fixtures - which means there won't be a variance in color between fixtures when talent walks out of one key and into another, or when using multiple L7s to create a wash up a cyc or backdrop.


There are two alternative cooling systems: one passive and the other active. The passive cooling system was designed for broadcast studios. It incorporates no moving parts or fans and is therefore completely silent. The active cooling system was designed to provide a more compact and lightweight option for location work. It uses an extremely quiet (<20 dB) fan and weighs 10lbs less than the studio version.


The location fixture carries an IP54 rating for weather resistance which means that it is protected from falling rain and splashing water, and that the internal electronics, optics and LEDs are protected from dust, dirt and humidity - making it a very robust fixture that will stand up to the rigors of location production.


All the L7s feature Power Factor Correction with a near unity Power Factor of .91. Which means that the 200W fixtures will draw no more than 1.98A at 120V (220W) and cause virtually no Harmonic Distortion. Since it creates virtually no line noise, you will be able to power nine 200W L7s on the 20A circuit of a portable generator without a problem.

To assure that they are not quickly rendered obsolete by the rapid advances being made in LED chip efficiency, the Arri L-Series LED Fresels are designed to be an expansible platform, with replaceable parts, that can incorporate future developments in LED technology. Not only, do the heads allow for the incorporation of more efficient LED chips when they become available (or when the lumnen output of the original ones drop), but the light engine is also fully upgradeable, ensuring that the fixtures can take advantage of technology advances as they happen. To accommodate future control protocols (such as ANC), their firmware is also upgradeable through the USB port on the rear of each unit. They will also be compatible with planned future optic accessories that will expand the L-Series versatility. Able to incorporate future developments in LED technology, the expansible platform of the L7s ensures that they will have a long useable life and so will assure a return on investment in them. Given the rapid pace of LED Chip development, I can't think of another LED fixture that won't be obsolete in a year or two.


Given the output, the clear and defined shadow rendering, the excellent field homogeneity and the color rendition demonstrated in the show video, it is evident that Arri has finally engineered the first true production LED Fresnel light (see the 2010 IBEC Show video below or the NAB 2011 Show Demo for more details (the 2010 IBEC Show video below demonstrates the light quality better, I think.))

__________________________________________________________________


About the Author

Guy Holt presenting to the Electrical Department of IATSE Local 481 as part of the
"Advanced Power and Generation for Set Lighting Technicians Seminar"

Guy Holt has served as a Gaffer, Set Electrician, and Generator Operator on numerous features and television productions (for a partial list of credits see his imdb listing). Guy Holt presented on Harmonics to the Electrical Department of IATSE Local 481 as part of the "Advanced Power and Generation for Set Lighting Technicians Seminar" offered by Russ Saunders of Saunders Electric (the provider of power generation services for the Academy Awards since 1952 and a recipient of a technical Emmy). Here is what industry leaders have to say:

Guy Holt is "among the 1% of film technicians world wide that truly understand the dynamics
of power generation and Harmonics."
- Russ Saunders, Saunders Electric

Guy Holt's other credentials include:

    - IATSE Local 481 Certified Gaffer
    - IATSE Local 481 Certified Generator Operator
    - IATSE Local 481 Certified Lighting Balloon Operator
    - Certificate Holder of the MQ Power "MQP Special Generator (Crawford) Technical Service Seminar"

Guy Holt is also the owner of ScreenLight & Grip, a lighting and grip rental company in Boston, MA renting Honda, MQ, and Crawford generators for motion picture production for 18 years. Inquiries can be sent to the attention of Guy Holt at rentals@screenlightandgrip.com

Guy Holt participating in a panel discussion as part of IATSE Local 481's Advanced HMI & LED Lighting Seminar (Fred Horne, Former Arri Northeast Sales Rep pictured left)


__________________________________________________________________


Each component of our

"HD Plug & Play Package"

is available for

Rental or Sale

Our HD P&P Pkg. capitalizes upon the following technological advances in lighting and production technology by
Mole Richardson, K5600, Power to Light, Kino Flo, and Honda.
    - More efficient and compact HMI, Quartz, and now Fluorescent Par Lights.
    - Brighter and more efficient "short arc" HMI bulb designs.
    - 120/240V Electronic HMI ballasts with Power Factor Correction.
    - Cool burning, energy efficient Fluorescent Lighting Fixtures.
    - Innovative new rigging hardware.

These technological advances in lighting design now enable the use of bigger lights, or more smaller lights, off of standard wall outlets or off a new generation of quieter and more powerfull Honda Inverter Generators. Click here for more detailed information on our HD P&P Pkg.