April 20, 2017
Exclusive

28-Light Shootout: The Most Comprehensive LED Test You've Ever Seen

This LED shootout compares 28 lights, ranging in price from $599 to $10K. It also introduces the science of LEDs.

[Editor's Note: No Film School asked Timur Civan to conduct an exhaustive LED shootout.]

Last year, I compared 11 LEDs on this site, merely out of curiosity. I was looking to buy LEDs, but, having used some of the older classic LED panels, I was concerned about the quality of color; LEDs at the time were not known for color accuracy. There were tons of reviews online, but very few direct comparisons between the brands, and not much actual data about the color quality available. So I borrowed a C700 from my friends at Sekonic and went to B&H—the only place with a showroom full of LEDs—and I started metering the lights. I purchased the brand I thought performed the best and shared my findings here.

LED lights for filmmaking are trying to achieve a few things: match Tungsten, match daylight, or create RGB blends to for pure colors. 

This year, as a follow-up, I wanted to go a bit more in-depth to seek accurate results with a wider variety of lights. Not everything on my wishlist made it to the set, but we got a wide range of lights from all price points, covering the major manufacturers—both legacy and newcomers.  

What to Expect

In this post, you'll watch videos of the tests and then have the chance to read some of my thoughts on each light. The videos are self-explanatory, for the most part, but here is a quick breakdown:

  • The Main Test  

The first video starts with the control light for each group, daylight and Tungsten. Then, it cycles through the different lights, one after another, full screen with the C700 readings displayed on the right side, so you can see what is happening with regards to the spectrograph and magenta/green readings.

  • Direct Comparison

The second video is a direct, side-by-side comparison of each light versus the control. For example, the daylight or Tungsten control shot will be on the left side of the screen, and the LEDs will cycle through as they did with the first video, but this way you can directly compare the color quality versus the control group. This is where you really see the differences.

  • Color Correction

The third video is similar to the second, but this time, halfway through each LED clip, the LED images will be manually white balanced with the dropper tool using the gray card to neutralize the color. This third video is intended to show you the possibilities of what the light can look like when simply corrected to neutral in post. At the top of the "data" section, you can download R3D stills of each light as well as the Waveforms and RGB data sheet if you want to examine anything more closely. Your eyes will be the best judge overall. Do your best to watch on a calibrated monitor.

The Lights

Here are the lights we tested (in ascending price order). More specs for each are listed at the bottom of the post:

  1. The Sun - Free
  2. ARRI: 1k Tungsten - $599
  3. Aputure: Light Storm LS-1 - $695
  4. Ikan: Lyra LB10 Bi-Color - $799
  5. Westcott: Flex X-bracket Bi-Color - $899
  6. Hive: Wasp 100-C Omni - $999
  7. Westcott: SkyLux - $999.90 
  8. Dracast: LED1000 PRO Bi Color - $1220.63
  9. Rayzr: 7 - $1,269
  10. Litepanels: Astra BiColor - $1,350
  11. Quasar Science: QLED - $1400 ($250 per bulb, we had an array of 4 with Housing)
  12. F&V: Ultra Color K8000 2x1 BiColor - $1,679
  13. Light and Motion: Stella 10,000 LED - $1970.80
  14. Fiilex: Q500-AC 5" Fresenel - $1,995
  15. BB&S Lighting: PipeLine Remote Phosphor 4' LED Bank - $1,997.45
  16. Cineo: Maverick  (Tungsten & Daylight units): $2,150,50
  17. Zylight: F8 LED Fresnel : $2,250
  18. BB&S Lighting: Area 48 - $2,269.95
  19. ARRI: L5-DT Daylight - $2,338
  20. Kino Flo: Select 30 DMX - $2,386.88
  21. Mole Richardson: Junior LED 200W Tungsten DMX - $2,409.95
  22. Fiilex: Matrix Bi Color Panel Light - $2,495
  23. Hive: Wasp Plasma Par - $3,279
  24. Outsight Creamsource: Doppio+ Bender Bi Color 2x1 LED - $3,695
  25. Cineo: HSX Color Tunable Fixture - $3,735.33
  26. Kino Flo: Celeb 401 DMX - $4,628.95
  27. ARRI: Skypanel S60-C LED Softlight - $5,625
  28. ARRI: M18 HMI - $10,515

The Meter (and how to read it)

The top row, CCT, is your measured color temperature. The CCi reading below that is the magenta or green correction level. It's telling you though an internal lookup table what gels or camera filter to add to correct it, not what is there. For example, if it reads .7M, it is essentially saying the image has a bit of green, and will tell you what filter you need to correct it on the filter page (the closer to "0" the better). None of the lights tested went beyond 1M/G on the meter. That means very small corrections—less than an 1/8th +- G. Just keep in mind: the smaller the number, the better.  

Next up: the Spectrograph, a graph plotted on two axes. The horizontal is the wavelength ranging from 380nm (nanometers) to 780nm. The vertical height is the saturation of that specific wavelength. Theoretically, a perfect light source would be a full rectangular block of color, fading slowly through the rainbow from blue on the left, to red on the right. Real life doesn't quite work that way, so you will see waves, spikes, and bumps in the curve indicating that certain wavelengths are being represented differently. The atmosphere, pollution, and water vapor all affect the sun's rays, hence the slight warbling of the graph. I suspect that if you took the C700 into the vacuum of space, the whole graph would be a rectangular block of solid spectrum.

We are interested in the quality of "white light" color being output by an LED system, and how closely it can mimic Tungsten or daylight sources.

The last thing to look at is the CRI, which is not critically important here. I go into why a bit later, but essentially, the very nature of the way CRI registers color is that it averages the illumination on the 15 color chips you on the bottom bar graph. It doesn't really measure the way LEDs work. All LEDs have a discontinuous spectrum that can fool a CRI meter. That said, it's not altogether useless—just not refined. 

The Videos

The Method

LED lights for filmmaking are trying to achieve a few things: match Tungsten, match daylight, or create RGB blends to for pure colors. In this examination, we are interested in the quality of "white light" color being output by an LED system, and how closely it can mimic Tungsten or daylight sources.

I set out to achieve this by shooting two control groups: daylight and Tungsten. In the frame is a standard color checker chart, gray card, white card, and three live models. Each model represents a skin tone you will likely encounter on various sets: one dark-skinned model, one middle-toned model, and one fair-skinned model. This variety of standard charts and models will show you how color rendition can be affected by the discontinuous spectrum of LEDs.  

We pointed an ARRI 1K fresnel at the scene. Then, we again set up the charts and models under the noon midday sun with a southern exposure and metered the color temperature of both scenes with a Sekonic C700-U Color Meter. The Tungsten unit metered at 3165K. We rounded up to 3170, because the digital target temperature in the meter only operates in increments of 10 degrees Kelvin. The sun metered 5437K and I rounded up to 5440K as a target temperature on the meter.  

On the slate, we recorded the control color temperature permanently for ease when making the comparison. Essentially, you can tell the meter what color temperature you are trying to hit, and it will give you a sense of how far off the mark you are in terms of temperature and magenta/green shifts. This is useful because it shows you how much green or magenta you need to add to the light to correct it to your target reading. (This information will be displayed on the meter under the CCi reading.)  

We wanted the LEDs to be used as they would in real life—either as a supplement or in some cases the primary lighting source in a scene.

With this baseline data, we wanted the LEDs to be used as they would in real life—either as a supplement or in some cases the primary lighting source in a scene. This means that you want your LEDs to match other sources so that there are no strange color shifts that can poison the look of a scene or create an unnatural skin tone. Using the target color temperature we selected on the C700-U,  we used all of the tools on the LEDs lights themselves to dial them in to match the color temperature as closely as possible. This way, the LED had a fighting chance to perform at its best.  

Some LEDs represent daylight as 6000K on their dials and Tungsten as 2900K—rather far off from the actual 3165K of a real Tungsten light or the 5437K of the actual sun. However, using the bi-color abilities, you can tune the color temperature to your liking. Most units allow for fine-tuning, others are rougher, and some are fixed at one color temperature. Turning the dial all the way to blue doesn't necessarily make daylight; sometimes you need cooler or warmer to match what's actually in the environment. This is why the meter is so important: it allows you to dial in as close as possible color wise.  

At this point, once we knew the actual control color temperatures, we took the camera, a RED Weapon Helium 8K, set to RedLogFilm/DragonColor2, and manually white-balanced in RCX from a still frame off of the white card. Our DIT Thomas Wong took the RedLogFilm image into Resolve and build a simple LUT that made black touch black and white touch white without touching any other color settings. That newly corrected image was the baseline from which the different LED units would be compared; it was completely neutral (manual white balance), accounting for any color from the lens, a Sigma ART 85mm F1.4, the sensor of the camera, OLPF, or anything else.

The Goal

There will be no winners picked here. Different lights have different uses, price points, and form factors. This is simply a way to see how the different lights look side by side. Some people prefer the color of one system over another. Sometimes a light's features are what's important. The perfect LED that's nearly $5000, 4' wide, and 25 lbs isn't going to help the action cinematographer documenting a Mount Everest ascent; he or she may need the self-contained, battery-powered lights that are the size of a soup can and can be thrown into a backpack before heading off for adventure. This test is intended to help you find what works for you and what you shoot. 

The Data

You can download a data packet here that contains LederScope RGB Waveforms, the RGB values in IRE datasheet, and RED Weapon 8K R3D stills so you can see how they grade. These are in 8K RedcodeRAW 5:1 compression and the native RedLogFilm. The LUT used on set can be found here.

In our previous test, we relied on the CRI rating and the visual look of the light hitting the scene. In this test, we rely a bit less on CRI and more on the DIT's analysis using a Leder Scope, spot-metering the white card. What you are looking at in the data below are the three RGB points on the white card that make up "white" on a color image. The values represent IRE on a scale of  0-110 IRE. The exposure on the white card was about 72IRE, with some fluctuation.

We put less emphasis on CRI because LEDs have a discontinuous spectrum. Much like with fluorescents, the spectrum will have sudden and drastic gaps and spikes.  

Each light had a different falloff pattern. Our exposures were set to the 18% Gray chip in the middle of the scene. The different falloff patterns led to slight variances in the final exposures of the white chip on the left of the frame. The three RGB levels, when blended, will make up the normal "white light" color image. What you are looking for here is a similar to daylight balance between three colors. The readings for natural daylight, for example: R: 70.5IRE, G: 71.5 IRE, B: 75.1 IRE. This is the baseline: red being weakest, green slightly above red but overall matched pretty closely, and blue having a distinct strength. You are looking for a similar pattern from the LED: Red/Green pretty close to each other with a strong blue a few points higher to balance them out.

The readings on the LEDs were taken after all the white balances had been done for the control groups, daylight and Tungsten. The white card was reading roughly 72 IRE, with the Gray card reading 50 IRE after the LUT was applied. Bear in mind, if either the Red or Green seem to overpower the other in a way that breaks the pattern established by the Tungsten or daylight, it's going to affect the image. A strong green channel or a weak red will look green. A weak green channel looks magenta. A weak red will make the image feel cool; a strong red, warm. The three work together to create the colors we see in an additive color method. It's all about balance. There is a chart below that lists all the RGB values for each light.

These three numbers speak to the true sense of accuracy in the recorded image. If the LED can put out numbers similar to the sun, it will look similar in color quality. We put less emphasis on CRI (Ra rating on the meter) because LEDs have a discontinuous spectrum. Much like with fluorescents, the spectrum will have sudden and drastic gaps and spikes that can be too small be detected on the spectrograph, without specialized, high-sensitivity equipment.

These gaps and spikes, depending on where they line up on the spectrum, can fool a CRI meter into giving a high CRI reading, simply because the gaps happen to not be where the CRI detection chips are looking. CRI was initially an 8-color detection pattern. Then, seven additional chips were added to increase its accuracy. But in the electronic nature of LED, and how suddenly and sharply its spikes and gaps can appear, they can sometimes go undetected or over-represented in a CRI reading. It's still okay for a ballpark judgment, but I wouldn't use that as the one and only determining factor.

Using the Leder Scope method, in conjunction with our eyes, and to some extent CRI, you get a fuller picture of what the light is doing on a technical and perceptual level. In this case, we trusted a DIT, his eyes, and the data from his Leder Scope in conjunction with the other pieces of information the C700 gives us: magenta/green shifts, and accuracy to color temperature in Kelvin. 

Daylight Leder Scope readout.
You are looking for how closely each light mimics the daylight control. (*Quasars are out of order because they were last minute arrivals and tested 2nd to last)
 
Tungsten RGB comparison Data.

The Controls

1: Daylight: The sun 

The sun. Fusion based light source. It's really bright.

We shot at about 1PM on a cloudless, clear day in NYC, facing southeast so the sun was at an angle matching that of the LED units. It metered at 5437K. As you can see, it has a rich, fully saturated (the vertical vector on the Spectrograph) range of color. It has heaps of every visible wavelength. It is the purest form of light we have. All three models look perfect. Their skin tones are beautiful, the white in the shirts, white card and gray all look great, not to mention the richness of the color checker.  

The sun is essentially completely neutral. Only at dawn or sunset does the angle of the atmosphere affect the color of the sun and cause it to be dramatic. What is important here is the saturation of the blue and red channels on the spectrograph. There is so much light on those ends of the spectrum, and that's what gives the feeling of rich colors. The spectrum is fully balanced. Take note: the LEDs just can't reproduce the density of light in those ranges. This is why there will always be an overall warm feeling from the LEDs, regardless of manufacturer. The blue end of the spectrum is just so strong. 

Distance to Chart: 93,300,000 Miles.
RGB Values: R70.5 G71.5 B75.1 This is the baseline for daylight.

2. Tungsten: Arri 1K - $599

ARRI 1K

All Tungsten lights are black body radiators (a material that emits no light, and as it gets hotter and hotter, the heat begins to convert to light). The hotter the object, the bluer it becomes. This is the origin of the Kelvin scale for color temperature. Kelvin being a measurement of heat, the color temperature correlates exactly to the temperature the Tungsten would have to be to emit that color light. The Tungsten filament, when at 3200 degrees Kelvin, is making balanced light.    

If you could get the Tungsten filament up to 5600 degrees Kelvin, the light would appear to match daylight. The filament in the 1K is burning around 3165K according to the meter, and as we all know, Tungsten lights get hot. The density of the red side of the spectrum you are seeing is the heat, the infrared spectrum gradually turning into visible light. This is why its spectrograph reading is so straight; it's essentially beginning to produce all the colors of the spectrum, but its physical temperature isn't high enough it make all of them. The consistency is remarkable because it's completely organic—just a function of heat turning into light.

Distance to Chart: 12'6" to achieve 50 footcandles (fc)
RGB Values: Tungsten; R69.3 G71.2 B73. Notice the nice, smooth steps between the levels. The blue is stronger the green middle and red weakest, because it's the opposite of the strengths of the spectrum of actual Tungsten. Look at the Spectrograph: it's countering the red, which is strongest, and boosting the blue channel, which is weakest, to maintain balance. 

The LEDs 

Here is my take on each light, based on the tests. The daylight examples are on top; Tungsten below. For a reference on how bright the lights are, I noted their distances from the chart to achieve 50fc, our at-key exposure for the camera set up at 250ISO and an F2.8 on the lens. The greater the distance from the card, the brighter the light. Most lights average 8-10 feet undiffused. Some were almost triple that. 

1: Aputure Light Storm - $695

Daylight
Tungsten

The Light Storm seems to be doing a decent job, considering the price. Mild magenta tint on the daylight example, with the Tungsten example feeling a bit green. It has an interesting physical design. The body is built like a heat sink—thin, but sturdy. This is a good case to talk a bit about how CRI can be somewhat misleading. The 95+ CRI the meter registered should mean a very accurate light. I found the green and magenta to be more visible in this example than some of the other lights that scored lower CRI ratings.

Distance to Card: 8'6" 50fc
RGB Values: Daylight: R73.4 G72 B74.4. The slightly lower G value, is why magenta seems to creep into the image. Tungsten: R71.5 G72.9 B74.2 

2: IKAN LYRA LB10 - $799

Daylight
Tungsten

The Lyra LB10 is a decent performer—somewhat magenta overall, but is consistent about that across daylight and Tungsten. It's a simple LED panel with silver barn doors. This can intensify the light output somewhat but makes it difficult to control the spill. 

Distance to Chart: 7'6" to 50fc
RGB Values: Daylight: R69.7 G68.7 B72.5. Slightly weak G channel gives it the magenta tint. Tungsten: R67.3 G68.2 B73.1 It's a ttiny bit weak on blue, hence the warmth.

3: Wescott Flex X-bracket - $89

Daylight
Tungsten
This is a great example of the CRI issue. I think these lights look really good: neutral, pleasing skin tones, no major shifts in color. Yet, notice the CRI numbers. Lower than the visibly green and magenta Aputure Light Storm.

Distance from Chart: 6'0"
RGB Values: Daylight: R68.8 G68.6 B72.5 - Pretty good. Notice the lack of Magenta/Green shift, because the R/G values are so similar. This is a well-balanced light. Tiny bit light on the blue, hence the bit of warmth, but that's not a huge deal.  Warmth is easier to correct than M/G. Tungsten: R69.5 G70 B74.1 Again very well matched R/G values.   

4: Hive Wasp 100C - $999

Daylight
Tungsten
The Hive Wasp is unique in that it's a chroma-based LED. Like an RGB light, it can create pure chromatic light, however it functions differently; instead of having red, green, and blue LEDs inside, it has amber, sapphire, red, cyan and lime. (It's almost like CMYK vs RGB). It does have some flaws when producing white light, as you can see in the spectrograph above, it has a pretty steep yellow spike and almost no red at all. That said, it's more or less designed to be used for creating beautiful color washes. Again, no winners here. The 100C may not have the best color rendition to the eye, but it probably makes beautiful and subtle colors for special effect. 

Distance to Chart: 6'0" to 50fc
RGB Values: Daylight: R63.4 G66.6 B67.2. You can really see the lack of blue channel and the yellow spike come together to bring out the yellow in the image. Tungsten: R68.4 G69.8 B66.8. Again, the lack of blue channel isn't there to counteract the yellow spike.

5: Wescott Skylux - $999

Daylight

A nice-looking light. Its physical design resembles a photo strobe. I imagine it's aimed at photographers looking to incorporate motion because it enables them to continue to use their existing modifiers without having to re-purchase everything for film-style workflow. This light is daylight only. It is reminiscent of the Westcott Flex. The look is consistent.

Distance to chart: 7'0" to 50fc
RGB Values: Daylight: R68.3 G68.7 B71.7. Solid performer overall.  

6: Dracast LED1000 PRO - $1220

Daylight
Tungsten
I think the Dracast has very solid performance in the daylight range, with a tiny bit of magenta. The Tungsten is a tiny bit green. You can see the green spike in the spectrograph on the Tungsten image. On the daylight image, the two bumps are closely matched in vertical height, so they balance the color. The levels were a bit thin overall because of the falloff on the white card in the RGB readings below. The Dracast has a natural tendency to spot. 

Distance to Chart: 7'3" at 50fc
RGB Values: Daylight: R69.1 G68 B:70.5. The slightly weak green channel is the culprit for the tiny bit of magenta shift. Tungsten: R66.7 G68.7 B:68.9. You can see how the image takes on a bit of green with a weaker red channel. 

7: Rayzr 7 - $1269

Daylight

This light came as a bit of a surprise to everyone at the test. It has great neutral color, is about a $1000, and is one of the brightest LEDs I've used. My gaffer Gary Wilkins calculated that this 300w daylight LED was as bright as a 575W HMI. A 575w HMI is over $7,000! The output per dollar ratio is absurd. Notice the distance from the gray card: 24 feet! The light was practically in the street while testing it. It is three times brighter than the average LED panel, but in a fresnel form factor, meaning it can be spotted or flooded. It impresses on the spectrograph as well. The Well-balanced performance, intense output, and low relative cost make me think this will be a very popular light for years to come.

Distance from Card: 24'0" to 50fc
RGB Values: Daylight: R69.1 G69.5 B72.3  Good numbers.

8: Litepanels Astra BiColor - $1350

Daylight
Tungsten

The Astra is an industry standard for a reason: great color performance, good price, and extremely flexible. Considering that it is also extremely bright but can still run on standard camera batteries just checks all the boxes of a classic. Minimal M/G shifting, and nearly spot on color temperature really makes you appreciate how far the Litepanels company has moved forward in its technology.

Distance from Card: 15'0" to 50fc
RGB Values: Daylight: R72 G72 B73.8. Picture-perfect R/G balance, though a touch warm with a strong R/G and a tiny bit underwhelming blue channel. No M/G shift at all. Tungsten: R68.4 G70 B69.6. I think in this one case, there may be a spike somewhere that isn't represented by the meter's spectrograph. The lack of red and blue should mean a bigger shift towards green, however the image just looks great. That said, the meter wants .1M to correct out the green, yet sees a 98.5CRI; that's one point lower than the sun, the highest CRI I've seen from a LED, period. Yet the RGB values seem to tell a bit of a different story. 

 9: Quasar QLED -  $250 per bulb 

Note that the price depends on how you use them. In our configuration of four, it was roughly $1400 for all four lamps and the Kino bulb holder.

Daylight
Tungsten
The Quasars were a bit of a dark horse. I only became aware of them in the last couple months, but the internet has apparently gone Quasar crazy; they are all I hear about on social media. Because they arrived on the day of the test towards the end, on some of the graphs and tables, they are sort of tacked on the end. If you notice that they seem to be out of order, that is why. They are essentially LED Kino Flo replacement bulbs. However, they don't require a ballast or even any special electrical needs; 12v and you are good to go. Most people who use them seem to have clever DIY style housings for them. One thing I noticed immediately with the Quasars was that in daylight, the skin tones of all three models seemed to look quite good, though the light was warm overall. All three models stood out in a good way: glowy and looking healthy. In Tungsten, they are rather magenta. They are bright individually, and as you start stacking them up, they really develop a lot of punch. 

Distance to Chart: 10'0" to 50fc (4x quasars)
RGB Values: Daylight: R68.8 G69.8 B70.5. The lack of blue is likely why they feel warm compared to daylight. Tungsten: R65.2 G65 B72.2. The spectrograph seems to indicate that there is a lot of color coming out of these—the band is wider than most lights. Overall, the magenta and cool tones are as a result of a lack of R/G end of the spectrum. 

10: F&V Ultra Color Soft K8000 2x1 panel - $1679

Daylight
Tungsten
Full disclosure: I own these lights. I love them. I chose them for a reason, so take this with a grain of salt. In my opinion, the color is excellent. Looking at all these images, one thing I notice about the F&V with regards to color, aside from all three models having lovely skin tones, is that it seems to show subtle colors in the cheeks and lips. Many of the other LEDs lose the subtle definition here. These are completely soft lights; there is a milk glass-style diffuser built in. The 2x1 size K8000 and 1x1 K4000 don't necessarily need any diffusion for a soft look. However, it does make them a bit harder to control. Also, when dimming or changing color temperature, they make a "loud enough to be noticeable" high-pitched whine that settles once the adjustment is finished. 

Distance to Card: 9'0" to 50fc
RGB Values: Daylight: R69.9 G69.5 B71.8. Great balance. A tiny bit warm with the lean blue channel. Tungsten: R67.3 G68.5 B70.7. Mimics the Tungsten slow climb in even steps pattern, hence the performance.

11: Light and Motion Stella Pro 10,000 - $1970

Daylight

The Stella is an abnormally compact LED light. The light quality is substantially green; however, this is one of those instances where a small flaw in light quality can be overlooked when taking into account the other features the light has to offer. The Stella is incredibly small—about the size of a soup can—yet despite its size, it is quite bright. I can see myself taking this light on trips to remote locations. They are daylight only. 

Distance to Card: 13'0" to 50fc
RGB Values: Daylight: R65 G69.2 B69.6. The thin red channel gives the impression of green in the image, but it's somewhat compounded by the fact that this light is unadjustable for color temperature. Unless we white balance to its color temperature, we won't see its true potential. Watch Video 3 up above to see it corrected in post to neutral white.

12: Fiilex  Q500 AC - $1995

Daylight
Tungsten
I happened to like the Fiilex Q500 not only because it looked good, but it was also a fresnel.  That gives you some nice options to send the light where you want it.  I noticed it has an interesting blue spectrum: it has two spikes, whereas most LEDs have only one. I think what is happening is that they are achieving the good color by having two sets of LEDs meshing together that are tuned to fill out the spectrum. It has a good density of color across the board.  

Distance from Card: 10'0" to 50fc

RGB Values: Daylight: R68.4 G71.6 B72.5. A tiny bit thin on blue gives it a touch of warmth. Tungsten: R66.9 G68.3 B67.7.  Definitely feeling a bit of warmth here.

13: BB&S Pipeline 4' 4bank -  $1997.45

Daylight

This is the first of the remote phosphor lights, a newer technology in film lighting. Essentially, a group of LEDs in the body of the unit illuminate a front plate coated with phosphorescent chemical. The excited phosphorescent panel is what emits the light you see in the scene.  They have some benefits, as the panel is capable of creating light that is less prone to spiking and has a more natural response. They do, however, have fixed color temperatures. If you want to change to Tungsten, daylight, or anything in between, you need a different phosphorus panel. I think they look pretty darn good—a little green, but not overwhelming. 

Distance to Card: 10'6"

RGB Values: Daylight: R67 G71.9 B72.1. You can see the stronger green channel here slightly overpowering the red channel. 

14: Cineo Maverick - $2150

Daylight
Tungsten
The Cineo Maverick is another remote phosphor light. Since we know the limitations of the single-color temperature per panel, this is no different. The look overall is about equally as good as the other remote phosphorus systems, but this daylight system produces different color temperatures. They have a good output and despite the color temperature shift, the color quality is pleasing—especially in the third video that shows the corrected images. The Tungsten panel looks great, color temperature-wise it's spot on, and the color quality is very good. 

Distance from Card: 7'0" to 50fc

RGB Values: Daylight: R67 G70 B73.8. Nice values spread. Tiny bit green is likely in the signal, but it is hidden by the warm color shift in color temperature. Tungsten: R66.8 G69.1 B70.2. Similar assessment as the daylight panel. 

15: Zylight F8 Fresnel Daylight - $2250

Daylight

The Zylight F8 is a very interesting light. It's a 7" Fresnel LED, but it has some interesting features. Instead of existing in a cumbersome housing, its fresnel lens is attached to a rubber bellow that extends forward to spot/flood the throw. This is only possible because of the low heat output of the LED. This means when put into its smallest configuration, its can pack flat and take up less than half the space of a traditional fresnel. Also, the 7" fresnel gives it the beam characteristics of a much larger light; it performs with the same soft-edged beam that we are all used to with 1K and 2K Tungsten fresnels. While it doesn't have the punch of those lights, it does have the light quality in terms of beam and falloff. (A newer, more powerful version is apparently in the works.) It has decent performance; some green creeps in, but not an abnormal amount. It's great for documentary, as it can be packed so small but still give you the throw of a full-sized light. 

Distance from Chart: 10'8"

RGB Values: Daylight: R68 G72.5 B74.6. The stronger green channel produces the slight green in the light.

16: BB&S Area 48 - $2269

Tungsten
The Area48 units are also remote phosphorus. In daylight, it looks pretty good, if not a touch green. In Tungsten, a bit warm compared to the ARRI 1K. The phosphor panel came in at 3093K. Spectrum-wise it is respectable, with a hint of green. I think when white balanced, it will fall into line nicely. (Nothing 1/8 CTB won't fix.) They are widely used and respected for good reason.   

Distance from Chart: 9'0" to 50fc

RGB Values: Daylight: R66.1 G70.5 B71.5. Sometimes the RGB seems to put up numbers that would indicate a significant green or magenta shift, but visually, it does not. This may be due to discontinuous spectrum anomalies. Considering the green channel is 4.5IRE stronger than the red, I would think the green would be stronger in the image. Yet, somehow it seems to affect certain lights more than others; the green doesn't seem to be that noticeable. Tungsten: R69.4 G69.7 B66.1. The low blue channel, indicative of a warmer image, definitely shows itself here. 

17: ARRI L5-DT - $2338

Note that this is mislabeled in the videos as an L5-C.

Daylight

This is the daylight-only version of the RGB chroma light, the ARRI L5-C. It's brighter than the L5C, and we were not necessarily looking to test chromatic RGB LEDs. The spectrograph shows a good strong dose of blue, green, and yellow, but lacking a bit in orange/red.   However, I feel the light has a distinct yellow/green cast. It is a fresnel-based light but was not as bright as I expected, considering the size and price tag. The Rayzr7, at half the price, was three times brighter and produced better color to my eye.

Distance from chart: 9'0"

RGB Values: Daylight: R66.9 G70.1 B69.8. Here, you can see how low the red channel is.

17: KinoFlo Select 30 - $2386

Daylight
Tungsten
Remarkably good visually—look how well the daylight and Tungsten-balanced images match. I don't see any major visible shifts, though the meter does detect some magenta in the Tungsten image. This light is the crew favorite for color.  You will notice that in the third video, where we white balance the images, the Kino Select 30 hardly changes. It's just locked in from the start. Its only flaw is that it isn't terribly bright. It needed to be 7' from the talent to register exposure at F2.8 @250 ISO. That said, most cameras shoot 800 ISO comfortably, and some far higher. I don't think it will be a huge problem for the run-of-the-mill interview or studio shoot. 

Distance to Chart: 7'0"

RGB Values: Daylight: R69.9 G71.4 B71.5. The balance is good; the touch of green comes from the slightly weaker red channel. Tungsten: R71.3 G72.3 B71.4.

18: Mole Richardson Junior LED 200w - $2409

Tungsten

This Tungsten-only 200W LED from Mole Richardson continues the tradition of top-quality, studio-ready lighting. The quality of the beam and ergonomics are what you would expect from a Mole light, but along with that came a light that was rather large. That said, it has a definite green tint to the light it throws, and it was a bit underwhelming on brightness. It doesn't flatten the skin tones like some of the greener lights can, but I think there would need to be some color correction on set by way of 1/8th minus green if mixing units between Tungsten and the Junior 200 on set. If you exclusively used the Junior 200 and its family of lights, it wouldn't be a worry, because the changes you would apply in post or through manually white balancing the camera would affect all the lights globally. That said, even on the CRI section, you can see the lack of red channel. We did not get to test the daylight version; I'm hoping that has a more balanced color output.

Distance from Chart: 13'4"

RGB Values: Tungsten: R68.9 G71 B67.6. The green channel is definitely the strongest of the three.

19: Fiilex Matrix - $2495

Daylight
Tungsten

The Fiilex line of lights both share a similar trait: they have two spikes in the blue section of the spectrograph. Considering how good they look, I think it reaffirms my theory that they are using "dual band" LEDs to achieve a balanced look to their lights. One little bit of comedy: the light has a very unorthodox form factor. It strongly resembles a large bluetooth speaker so much so that one of the grips thought it was one from behind and asked if he could put on music. This doesn't change the fact that it outputs great light. Based on the spectrograph, it seems a little more orange and red than most LEDs, while outputting a lot of blue and green. This is why it performs so well. 

Distance to chart: 11'0"  

RGB Values: Daylight: R67.5 G71.4 B71.7. Tungsten: R67.7 G68 B69.5.

20: Hive Wasp Plasma - $3279

Daylight

The Hive Wasp Plasma Light is not an LED, but I still think it belongs here. I see this test not only as a test of LED technology, but all the cutting-edge technology available today. The Wasp "bulb" is a tiny bead of glass that looks like a fuse you would put in your car. In the middle is a bubble, and in that bubble is a gas that gets excited electronically by the ballast.  Once it reaches a certain state of excitement it outputs a bright blast of light. Much like daylight, on the spectrograph, you can see tons of blue channel saturation, red saturation, and a nice rich amount of color pretty much across the spectrum. There are spikes, however, and I'm certain they will effect the image somewhat, but remember the daylight sample had spikes, too. It does have a cool overall feel, but so does the sun! Once this technology matures a bit, it may be a potential replacement for HMI. It is more efficient, though at the current time, larger units are not available because of the technology limitations. It's a peek into the future of film lighting!

Distance to Chart: 16'0"

RGB Values: Daylight: R67.2 G70.4 B73.7 

21: Creamsource Doppio+ Bender

Daylight
Tungsten
The Creamsource Doppio+ physically looked and felt like a luxury item. It is incredibly well made and aesthetically pleasing. Performance-wise, it's good, with some warm tones in the daylight mode and relatively neutral colors on the Tungsten end. There was a bit of magenta when used in Tungsten mode. It is a pleasure to use and manipulate because it's engineered so well, but it is a bit on the heavy side (solid metal construction). Also it's worth mentioning that it's a 2x1 style panel and has a lot of power. It's similar to the Astra in output, though the Atra is a smaller form factor.

Distance to Chart: 15'6"

RGB Values: Daylight: R69.9 G73.2 B71.5. Tungsten: R69.4 G71.1 B72 

22: Cineo HSX - $3735

Daylight
Tungsten
The Cineo HSX another remote phosphor system and has similar performance and limitations as the rest of the remote phosphor lights.   It's a "2x1" style wide rectangle, though the actual surface area of the light emitting panel isn't as large as a 2'x1' standard panel. Tungsten performance was very nice, while daylight was a bit warm. Skin tones really held up well. 

Distance from Card: 13'0"

RGB Values: Daylight: R70 G69.3 B72.1. Tungsten: R68.4 G67.5 B72.6

23: KinoFlo Celeb 401 - $4628

Daylight
Tungsten

The Kino Flo Celeb 401 isn't the brightest light ever made, but there is just something about the way Kino Flo designed or sourced their LED diodes that makes them look right. They are completely neutral to my eye. All three skin tones look exactly as they should, and the models look bright and healthy. I may not buy them, but I would rent them with no hesitation and know that no matter what I'm using— whether it's Tungsten, the sun, or HMI—I would have little problem matching the look from these units. They do have a bit of magenta, but as I mention in the video, I find a slight bit of magenta when coming from an LED seems to make it feel neutral, as it combats the natural green spikes.

Distance to Chart: 10'0"

RGB Values: Daylight: R69.1 G70.9 B71.2. Tungsten: R69.4 G70.3 B72.8.

24: ARRI Skypanel S60-C : $5625

Daylight
Tungsten
The ARRI Skypanel S60-C is an RGB-style light. Notice how, in the Tungsten sample, like the Hive Wasp 100-C, also a Chromatic LED light, they have very similar Spectrographs. (I'm talking about the sharp yellow spike.) Perhaps it's a side effect of making "white" light by blending the other chromatic LEDs. I felt the Skypanel struggled a little in the Tungsten range. Meanwhile, the daylight end of its capabilities seemed to be performing very well. I'm curious why the spike is so prominent in Tungsten but is much better controlled in the daylight mode.  

Distance to Card: 15"

RGB Values: Daylight: R69.6 G69.7 B71.2. Tungsten: R71.1 G71.1 B73.3 

25: ARRI M18 HMI - $10,515

Daylight

This is the darling of the film industry in NYC. You can strike it off any standard plug, in any building, anywhere in the city. It's the brightest and generally most ubiquitous light in indie cinema. As you can see, the M18 has lots of small spikes, but an overall richness of saturation from blue all the way to red. (The Hive Wasp Plasma had a similar spectrograph reading; must be part of the nature of arcing electricity or plasma gasses.) This light obviously is the one that matches daylight the best, as it has a curve that resembles the sun very closely.

Distance from card: 19'6" (with 4 stops of steel scrims in the head)

RGB Values: Daylight: R71.2 G73.7 B79.8

Conclusion

After seeing so many lights in action and having examined the results, I feel have an understanding of what to expect from not only certain manufacturers, but also from certain technologies. Whether it is the convenience of a lightweight BiColor 1x1, the raw output power of some of the bigger fresnel units, or the need to produce pure saturated colors, LEDs are certainly up to the task. I hope this gave you a window into what is out there available in the industry and how it can help your productions.

The Team

Finally, I want to thank our sponsors for this test: SekonicB&H Photo VideoRED Digital Cinema, NoFilmSchool, and, last but not least, my producer, Ab Sesay. Without them, this wouldn't have happened. Thanks also to the whole team that was involved:

Your Comment

34 Comments

This is exactly what I needed! Thank you so much! I live by the the skypanels I'm surprised by Ikan honestly! Does anyone know when they're finally updating the sun? The first model came out 4.5 billion years ago.

April 20, 2017 at 1:50PM

0
Reply
avatar
Julian
Director
331

Yes, there are plans for the sun to go nova in another 5 billion years and swallow up the Earth, so you might have to quickly go run'n'gun because there is no second take.

April 20, 2017 at 4:45PM

0
Reply
Saied M.
1169

All too dear for me. How do they stack up with say cheap 600 LED ones?

April 20, 2017 at 4:07PM

0
Reply
Donald McPherson
hobbyist
169

Check out Caleb Pike's DSLR video, he covered cheaper LED options a few weeks ago. Here's a link to a whole $500 kit, also has more videos covering much cheaper LEDs as well - thttps://www.youtube.com/watch?v=O7vz2ONG5nk

April 20, 2017 at 7:34PM

0
Reply
Caleb Owen
Director/Editor
157

First off this is awesome, thanks a lot for the info, but I have to ask why the Digital Sputniks are left out of these LED comparison tests. The company argues that only the Red Blue Green areas of the spectrometer graph is important since thats all the camera sees, and they match the camera's sensitivity to color rather than attempting a high CRI. It would be great to see how they compare since more and more big movies are opting for their system.

April 20, 2017 at 4:08PM, Edited April 20, 4:08PM

0
Reply
Indie Guy
648

We weren't able to get one. We tried.

April 20, 2017 at 4:40PM

0
Reply
avatar
Timur Civan
Director of Photography

I know the DS guys in LA. Im sure they can get you one for this comparison test. I know OFFHOLLYWOOD in NY has DS also and possibly can lend you the lights for the Comparison or you just rent from them for a day.

But give DS a Call! Here is their LA office # (818) 262-9284

would love to see this comparison!

April 20, 2017 at 6:04PM

0
Reply
Mike Mack
DMX TECH
166

Ok so my specialisation is stills/motion beauty photography and motion. I learnt my craft assisting some pretty spectacular photographers on the world stage where you would alter chemistry or labs to process your film. In this digital age I have become fixated about sensor colour rendition of skintone. I've shot with every sensor out there, stills and motion. I've never liked LED lighting for skin tone and this test proves it. Take a screen shot of the 'sunlight' daylight. Stick it up on your screen. Look at the whites (shirt) - they look amazing! Now keep that screenshot on your screen and run the video...... aside from the Kino Flo (and that is not great) they all look positively shocking compared to the ARRI 1K. I'm sorry but actually ALL of those LED look like sh1t! Forget the LED's and just take it on the chin, for now, that HMI or Tungsten remains the no1 light choice of source despite the weight. We are a long way off LED becoming usable. (just to repeat, forget the grey cards, light meters etc - just look at the 'whites' in the daylight then look at them with the LED - they just look like, murky, muddy sh1t!). Particularly in the 'whites' - get the 'white's' wrong and every other colour will be wrong!

April 20, 2017 at 4:34PM, Edited April 20, 4:40PM

0
Reply

So I've now just watched the 'Daylight' direct comparisons where they are trying to grade to the 'daylight sample'............Everything I said is true. (aside from Kino Flow - which was not ideal!)

April 20, 2017 at 5:03PM

0
Reply

The sunlit sample as shown above does not show neutral whites. This shot has a significant blue cast. The whites and the overall color balance is about 15 points higher in the blue channel than the red or green channel (as measured in Photoshop). Once this sunlit shot is neutralized, it looks dramatically much closer to the daylight LED samples. The sun still looks better, but the LED's are not as dramatically off as is suggested when comparing them to this sunlit sample with a blue cast.

April 21, 2017 at 2:33PM, Edited April 21, 2:33PM

0
Reply
avatar
Benton Collins
Camera aimer
291

Just to add to the discussion, daylight by nature is a mixed-source light.

The direct sun rays are one color temperature, but the ambient rays are another. People often forget about the giant blue dome above our heads, a.k.a. the sky, which acts as a giant softbox. So the shadows are always blue in comparison to the direct light.

A single direct light in a studio will never achieve that kind of "daylight" quality until you bounce some blue into the shadows.

I wish the testers had built a "room" with flags/floppies with a small window to allow direct sunlight in and block the ambient blue light.

April 21, 2017 at 10:21PM

0
Reply
Aaron Lochert
Director of Photography
167

All it takes is to white balance the scene and the blue cast in the direct light goes away. But yes, some blue is always present in the shadows in natural outdoor sunlit shots. I'm actually surprised that proper white balancing was not done in the sunlit control sample. It's dramatic when it's done and the skin tones look so much better.

April 21, 2017 at 11:45PM

0
Reply
avatar
Benton Collins
Camera aimer
291

Pretty disappointing showing from the new Hive 100-C. And the results look way off from what Hive is advertising. I hope they comment on these results to calm the nerves of backers of their Kickstarter who are still waiting for lights...

April 20, 2017 at 4:42PM, Edited April 20, 4:42PM

0
Reply
Kyle
96

Hi Kyle,

Jon Miller here from Hive Lighting. So first just want to say we were very grateful to be included in the test, and we rarely comment on these tests as they are completely independent and for better or worse we want everyone to judge our products without undo influence from us.

But we are surprised by the results here, having said that I think the test sort of misses the point of what we hope we created with the 100-C. We aren't just a "white light" LED. The point of the 100-C is that there is no single white light setting. With the ability to fine tune every part of the color spectrum, you don't ever need to have any kind of distinct spike in yellow, or blue etc. In fact you can adjust the color temp and add points of blue or yellow, red or green etc. as you see fit both to your eye, your camera and your meter. The point is you have total control to correct to any skin tone and camera sensor. If you turned on your light and your camera and your light, found yourself getting this specific result and disliked it, a couple dial spins later and you can have totally different output.

As for the output level, we assume it was either dimmed, or through some lens and diffusion combination as that is completely different from our photometrics, both in house and independently tested.

We do have independently verified test results from Light Lab in Anaheim California we will put up on our website and Kickstarter.

Again we still think this was a cool test, we are glad to have been a part of it, and really appreciate the time and effort of NFS, Timur and everyone who worked on this, it just might not have been the best test for what our light does well

- Jon Miller, Hive Lighting

April 21, 2017 at 10:34AM

0
Reply

That's super reassuring to hear!

Yeah, I wasn't too worried about their idea of color accuracy because I know it can be tuned to whatever you want, but I was confused by the output. Thanks for the clarification!

Hopefully we hear back from NFS about what caused the output to be strangely low in their testing.

April 21, 2017 at 4:17PM, Edited April 21, 4:17PM

0
Reply
Kyle
96

I'm looking into it Jon. Trying to figure out why the output was low.

April 21, 2017 at 10:54PM

0
Reply
avatar
Timur Civan
Director of Photography

Nice, but you forget to include THELIGHT's VELVET Light & Power Series. I am sure that is impossible to include everything, but it is a standard here in europe. It will be great to find them (and also other brands) next time.
Great job!
David García (Videographer). BARCELONA, Spain

April 20, 2017 at 5:26PM

0
Reply

Hi Tim - Thanks for such an in depth article. I noticed you referenced the Cineo HSX as a Remote Phosphor fixture. This is actual inaccurate. Can you please update to reflect the product properly. Please visit our product page for details: http://www.cineolighting.com/hsx/

Thank you!

April 20, 2017 at 5:33PM, Edited April 20, 5:33PM

0
Reply
Ashley
79

I will address this tomorow. Thank you for bringing it to my attention.

April 21, 2017 at 12:26AM, Edited April 21, 12:26AM

0
Reply
avatar
Timur Civan
Director of Photography

Fantastic test, extremely thorough and allows us to look at real world results. Thank you for putting all the work in. The results are interesting to say the least.

Litepanel Astra looks like an even better value now. (Wish they'd come out with an Astra 2x1) KinoFlo Select 30 looks really great, as does the F&V K8000 and the Celeb. Westcott Flex did better than I expected, I have several of those and they're awesome for travelling.

April 20, 2017 at 6:48PM

0
Reply
David K
86

Audio drops out in LED Shootout 2017: Direct Comparison at this point - https://youtu.be/bhX-qaYKZ0k?t=7m15s

Really enjoyed the commentary, seeing the differences really helps when thinking about light sources and is making me reconsider a few things. Thanks!

April 20, 2017 at 7:50PM

0
Reply
Caleb Owen
Director/Editor
157

Look at that. Will have NFS reupload the video. Must have glitched.

April 21, 2017 at 12:28AM

0
Reply
avatar
Timur Civan
Director of Photography

Thanks for the video!

April 20, 2017 at 8:50PM

0
Reply
avatar
Faris Beitar
Film Director
94

My favorite light (which I personally prefer over the Litepanels Astra Daylight) wasn't included in the test - it is the "Aputure Light Storm LS 1/2".

It is not just a smaller version of the "Aputure Light Storm LS1" - it uses completely different LED Chips - it produces beautiful skin tones and is softer than most other LED lights.

I purchased 3 of them because of their small size, easy rigging with the lightest Manfrotto 5001B Nano Stands and I can fit 3 of these lights with all accessories (except the Nano Stands) as well as 3 V-Mount batteries into a Pelican 1560 Case w/Padded Dividers - very portable.

April 20, 2017 at 10:14PM

0
Reply
Gomali
7

Wow! So in depth! Thank you so much! I wonder, did you test any with COB led's like the Aperture COB120d or COB120t. They are much like the ones you mentioned that resembled photography studio strobes.

April 21, 2017 at 1:00AM, Edited April 21, 1:00AM

0
Reply

Looks like it was the Westcott SkyLux LED Light .

Wonder how the Aputure lights compare. COol that they have a fresnel attachment and seems like a very versatile light.

April 21, 2017 at 1:03AM

0
Reply

I really appreciate the time put into making this test - really thorough and helpful.

But it does seem like you've made it hard to compare the daylight tests by eye when you've shot the control against a white background, and the LEDs against a black background. That difference in settings makes it impossible to compare skin tones, because our brains perceive the shades in context.

eg. https://files.brightside.me/files/news/part_4/41705/41005-5-650-4e56912f...
Cover the dividing line with your finger to see what I mean!

April 21, 2017 at 6:27AM

0
Reply
Andy Toovey
Director/DP
79

Thanks for the comparison! Answered a lot of my questions. Was definitely surprised to see the S60 did so poorly with settings set to zero. I'd be curios to see how the G/M adjustment affects the spectrograph. Quasar, HSX, Select 30 - Wow! Great rendition.

April 21, 2017 at 12:51PM, Edited April 21, 12:51PM

0
Reply
avatar
SpacemanSteve
Director of Photography
131

The sunlit sample as shown above does not show neutral whites. This shot has a significant blue cast. The whites and the overall color balance is at least 10 points higher in the blue channel than the red or green channel (as measured in Photoshop). Once this sunlit shot is neutralized, it looks dramatically much closer to the daylight LED samples. The sun still looks better with much cleaner whites, but once this sunlit shot is white balanced for neutral whites, the LED's are not as dramatically off as it appears when comparing them to this sunlit sample with a blue cast.

April 21, 2017 at 6:59PM, Edited April 21, 7:32PM

0
Reply
avatar
Benton Collins
Camera aimer
291

Whoa! Lots to digest, so this is a quick note to say "Thanks!" This is going to be a great learning resource.

April 22, 2017 at 12:40AM

0
Reply
avatar
Nikhil Kamkolkar
Writer/Filmmaker

Amazing test! The Kino Flo Select and the Cineo and the F&V are my favorites! Great work! Thank you thank you - this saved me a bunch of money and time!

April 22, 2017 at 9:27PM

0
Reply
avatar
Ed David
Director of Photography
1472

thanks for taking the time to do this review. For my own taste I'm more interested to see how the Apurture 120d, Came-TV Bolten series ( I noticed B&H have 100watt versions listed now) and other Fresnel and Par Lights perform. Panels seem to be a pain to pack for location work. Plus I like bouncing light.

April 23, 2017 at 9:31AM

0
Reply
jeff
grip Electric
88

Woww... Really appreciate this detailed Study.
Thanks.
Looking for a hack to shape/ control LED lights.

April 24, 2017 at 4:41AM, Edited April 24, 4:41AM

0
Reply
avatar
Sameir Ali
Director of Photography
491

Good test y'all. We need more of this on NoFilmSchool!

Question: Who is selling the Razyr 7 for that price? I don't see it with major US retailers, and no one that cheap.

April 24, 2017 at 4:20PM, Edited April 24, 4:20PM

0
Reply

The US distributor was recently announced as Zylight and they were indeed showing the Rayzr7 at their NAB booth. However, Zylight won't be selling them direct until sometime this summer. If you need one sooner, I suggest ordering direct from the Rayzr7 website. That's how I got mine a couple months back, but I did have to pay import duties to DHL on top of the purchase price plus international shipping. Despite the extra cost and hassle to get it, I'm totally loving that light. It's mind blowing to have output roughly equivalent to a 575watt HMI without the heat and weight. It runs so cool that you can gel it by just gaff taping a little circle of gel right to the face of the fresnel. Crazy.

April 30, 2017 at 3:29AM, Edited April 30, 3:30AM

0
Reply
Jamie LeJeune
Director of Photography
186

Thank you Mr. Civan and company for this most excellent in-depth analysis!
Regarding the Stella Pro 10000 (and the other "bare bulb" units with small form factors), did you use any light modifier on the unit to focus or concentrate the output?
Some of the units you reviewed (e.g., the Stella and the Westcott) have very wide beam angles when used without any modifiers, so obviously their output jumps significantly when a fresnel or similar is attached... Might be nice to know whether they were tested bare bulb or modified.

June 6, 2017 at 11:39AM

0
Reply
J Drake
Photographer
1