I was the gaffer on a job recently with a portrait-type talking head shot. There was a discussion with the DP about what key light to use. We ended up using an LED light. Common these days as LEDs are rapidly become the go-to light on sets.
I wanted to use a tungsten light; specifically a Rifa 88. It is a cheap, junky (okay that is a bit harsh, but it is definitely not high-end) light that has a beautiful, soft, warm look. I love the Rifa 88 as do many DPs. I am not surprised when I'm on a big job, with a big DP from New York or LA, and they ask for a Rifa.
We ended up not using a Rifa on the talking head shot. It looked great but I would have preferred using a tungsten light. They are my favorite for rich skin tones and colors. Tungsten lights are just a better light source. It's science!
Here are readings from my Sekonic C-800 color meter from a classic tungsten source, a par 64:
 |
| Tungsten Par 64 CCT |
Typical of a tungsten light, even though it is technically a 3200K color temp source, it reads slightly warmer at 3076K.
(I think it is funny that my meter is suggesting I add 1/8 CTB to "correct" the color temp. I don't think any gaffer in his right mind would correct this.)
Here is the CRI:
 |
| Tungsten Par 64 CRI |
Impressive numbers at 99.6 as would be expected. That's about perfect!
I know CRI is an imperfect measure of a lights quality, but I often use the CRI graph as a quick reference. More about this when I write about the next light, an LED.
Here is the TM-30 color vector of the tungsten light. This is a more accurate measure of the light quality:
 |
| Tungsten Par 64 TM-30 |
Again the tungsten light has impressive scores as to be expected. The Rf and Rg scores are perfect indicating that this source will accurately represent colors. The delta UV value is nearly perfect.
The most interesting reading is the spectrum graph:
 |
| Tungsten Par 64 Spectrum Graph |
You see half of an almost perfect black body curve which is to be expected from an incandescent source. The curve is very smooth with all visible frequencies (colors) present. The Sekonic reads a little past each end of the visible spectrum. The visible spectrum is approximately 380 to 750 nanometers. Interestingly, the blue end continues into the ultra-violet region. At the other end of the spectrum, most of the energy is towards the red with the curve appearing to peak at around 750 nanometers. The other half of the black body curve doesn't show on this graph since this is well beyond the visible spectrum.
We can extrapolated that this light source is rich in the infra-red spectrum with about half the energy this light is emitting being not being visible. We feel it as heat.
The presence of all color frequencies in a smooth curve along with rich deep red colors is why tungsten lights give better skin tones and colors.
Here are reading from an LED light; an Aputure 300X set to 3200K:
 |
| LED Aputure 300X 3200K CCT |
Like a true tungsten source, the color temp is a bit warmer than 3200K. The deviation is slightly more than the tungsten light, but not unacceptable.
Going back to physics class; a tungsten source emits light via incandescence. Incandescent sources emit light (electromagnetic radiation) following a "black body" curve. The significance of this is the color temperature reading is accurate.
Note the Sekonic doesn't read out "color temperature" but correlated color temperature (CCT). Simplified, you can think of CCT as comparing a light source to the closest equivalence on the black body curve.
LEDs emit light via electroluminescence. This type of source does not follow a black body curve. LED manufactures attempt to simulate the black body curve of incandescence light sources, but it is imperfect.
(LED manufactures have improved the quality of LED light dramatically over the past 5-10 years. Years ago, I never wanted to light people with LEDs becasue skin tone and color quality was poor. It is not that way anymore. I routinely light with LEDs and the color quality is good.)
So the reading of the 300X of 3062K has to be taken as an approximation of the color temperature. It's important to realize that 2 lights can have the identical CCT but look very different.
A common mistake I've seen over the years is DPs/gaffers attempt to match light by CCT only, but this is fundamentally flawed.
Here is the CRI graph of the 300X:
 |
| LED Aputure 300X 3200K CRI |
It has an impressive 97.8. There are some deficiencies at R9 and R12 though the numbers are still good.
The R9 value is often deficient with LEDs. This is particularly a problem with lower cost lights. This deficiency in the R9 value, red, is the reason many LEDs have a green cast.
The CRI value is an inaccurate way judge the quality of a light source but it is useful tool estimating the quality of the light. The CRI graph is a quick way to check for deficiencies with lights.
Here is the TM-30 color vector, which is a more accurate measure of light quality:
 |
| LED Aputure 300X 3200K TM-30 |
These are very good reading for an LED, but there are some deviations which are to be expected. The Rf value is very good but off a little. It is telling me that color reproduction will be slightly off, but subtle. The Rg value is almost perfect. On the graph we can see the slight deviation from ideal that the CRI graph had shown. There is a slight desaturation in some colors most notably in the reds.
There is deviation from white with a delta UV value of 0.0022. While this is still considered good it is almost 1/8th correction gel off.
(0.0024 would need 1/8 magenta to correct the green deviation from white. Minus 0.0024 would need 1/8th green gel to correct.)
This agrees with the CCT screen which indicated no color correction for the light. (This is a bi-color light; a trade-off of this feature is the delta uv value will be different at different color temperature settings.)
Here is the color spectrum graph:
 |
| LED Aputure 300X 3200K Spectrum Graph |
To me this is the most dramatic visualization of why the tungsten light is a better source than the LED.
Overall, this is a good graph for an LED. You can see the engineers of this LED light have worked to simulate an incandescence source with an electroluminescence source. They have done a good job. The graph is relatively smooth without sharp spikes in any color, though there is a little dip-and-spike in the blue. This is a common feature of LED lights.
Looking at the red end of the spectrum, there is a dramatic difference from the tungsten light. There is a steep drop-off from the deep red colors going into the infra-red. Similarly at the blue end of the spectrum, while much less dominate than the red, the levels falls-off abruptly towards deep blues.
Here are the 2 spectrum graphs side-by-side:
 |
| Tungsten Par 64 and LED Aputure 300X 3200K |
Side-by-side you can clearly see dramatic differences between the tungsten source and the LED. At the extreme ends of the visible spectrum are the most significant differences. This is why tungsten light is still the best option for the best skin tones and color fidelity.
PS: I was recently reading about a newer LED light--I have forgotten the manufacturer--that has added deep red emitters to better simulate a tungsten source. LEDs are continuing to evolve with better color quality.
PPS: When I started this post, I thought is was going to be a quick post that was going to take me 20 minutes to write. Wrong! It took a few hours. Even though this is a brief exploration of tungsten lights vs. LEDs, it quickly became a rabbit-hole of technical subjects.
PPPS: Here is a great article exploring TM-30 in an understandable way:
https://tuckerd.info/06/what-is-tm-30/
PPPPS: I love tungsten lights. It is great to have a client walk onto a set and marvel at a tenner (10K fresnel). Lights like these conjure up Hollywood. They are what many people think a set should look like.
Of course, tungsten lights have major drawbacks. Two of the biggest are power consumption and heat. Depending upon where you are shooting, this can be a big issue.
It is easy to use large tungsten heads in a studio. There is power and air handling to deal with them. On location, it becomes more problematic.
We've all heard the stories about a large tungsten light being placed under a sprinkler head. Then the shoot became a disaster when the sprinklers went off! (Thankfully has never happened to me.)
With the rise of LEDs, I've seen a drop in the use of tungsten lights which is unfortunate since they still provide the best color reproduction and fidelity. Even in studios, I see more LEDs used. Some studio have even replaced space lights with LED lights. Certainly there are many advantages to the LEDs, but the color quality is not as good. As a gaffer, I work to achieve the best possible image. Tungsten is still my first best option.
PPPPPS: I may be setting a new post-script record today.
You may be wondering, where is the Spectral Similarity Index (SSI), one of the the newest measures for comparing light quality.
My new color meter, the Sekonic C-800 will read SSI. Unfortunately, I haven't had the time yet to learn SSI. I don't know how to use it yet but is high on my list of things to learn.
It is a great time and a hard time to be a gaffer. The technology is advancing so fast it is hard to keep up!
PPPPPPS!: I own a small package of tungsten lights. A sign of changing times is how often I've have been renting these lights. Last year, 2020, I had few rentals for my tungsten lights. My Arri kit didn't get rented once.
Update 2/17/21:
I didn't go past the first LED light I had at hand for this example of tungsten vs LED. Here is another example of an LED light; an Astera Titan Tube. The Titan Tube is a "ringer". It has great color quality.
Here is the CCT with the color temperature set to 3200K:
The CCT is almost perfect; better than an actual tungsten source. Tungsten sources tend to be warmer.
Here is the CRI:
Again, very impressive numbers.
There is an universe of different types of LED lights. The 300X (I first compared) is a true bi-color light. It uses two sets of LEDs to achieve tungsten and sunlight. It blends the two sets of LEDs to achieve different color temperatures in-between. The biggest downside of this feature is the delta UV will change at different settings.
Typically it is worst at the extreme ends of the color range and the middle. The manufacturer designs the light to be best at the most common color temperature of 3200 and 5600K. It is a compromise balanced with the benefits of higher light output and the cost of the light.
The Titan Tube uses a different system of 5 sets of different color emitters; red, green, blue, mint, & amber. It has a sophisticated "light engine", a fancy name for it's computer and programing calculating the mix of colors required to achieve the set color temperature. This is a more accurate system than a true bi-color light. The result is an LED light with outstanding color quality. There is a compromise with this system too; it has less output than a true bi-color with more cost.
(It would be interesting to compare a 3200K only LED light. Ideally, this type of light should have the highest output with the best possible light quality.)
Here is the Titan Tube's TM-30 results:
Once again, very impressive results. The number which really catches my eye is the close to perfect delta UV value. It is almost identical to the tungsten light! This is an LED light which will have excellent color reproduction.
Finally, here is the spectrum graph:
The graph shows a smooth simulation of the black body curve a tungsten light would have with the typical LED dip-and-spike in the blue.
Here are the 3 spectrums side-by-side:
You can see some of the advantage the Titan has over the 300X; while both are excellent LED lights, the Titan Tube has a slightly wider and fuller spectrum. With both lights there is an abrupt fall-off at each end of the spectrum, particularly in the reds. The tungsten light will still have the best skin tones and color reproduction becasue of this. It's science!
PPPPPPPS: Here is an interesting site I stumbled across where you can convert CIE 1931 coordinates (x,y) to a delta UV value:
https://www.waveformlighting.com/tech/calculate-duv-from-cie-1931-xy-coordinates