Our first impression of the test: 

The  Softlights 27-600K seems to have one of the best results in both Daylight and Tungsten. Daylight reading for TLCI is 99. 

The curve is very close to the real daylight curve - and the only LED in our test with a color spectrum below 420 Nanometer. 

From a Colorist´s POV; 
Cam Ozkilicci ı Senior Colorist

I had a look at the LED tests but needed the time to actually write down my thoughts. Firstly, I have to say that this is a great test to show the subtleties between the various LEDs. 

 

I tend to look at this from two perspectives (possibly even three), the first one being what is actually recorded by the camera, and secondly, how closely setups look when they are matched as close as possible in grading. There is also another point of view which is basically that numbers don’t lie, but our eye is really the final judge.

 

Looking at things from the first point of view, which is what the camera records and how even both sides of the image are under daylight, I decided to view it with a standard ARRI R709/K1/S1 LUT. Even though the camera doesn’t record with a LUT, it is certainly what the DP/director view it through. I personally think that the best results are Juliette and ARRI Skypanel.

 

I find that Juliette is nice, and reflects nicely on caucassian skin tones, the issue I have however, is basically how there is a visible difference between the gray backdrop. The right side shows what daylight is like and the left, I assume, is what the light source generates. Juliette has a slightly more magenta tone compared to the daylight coming in, which is why on Scandinavian skin tones, the results feel more pleasing. You can also notice this slightly on darker skin tones. One side of the black actor is slightly greener compared to the other.

 

Looking at the Skypanel and daylight, I find that it gives a more ‘even’ result. Skin tones might not be as pleasing for our girl as it goes a little green, but the overall colour balance of the shot is more levelled. You can observe this on scopes too. I think the ideal would be to have an LED source that is something perhaps between ARRI and Juliette… :) Again though, this really is dependent on what/who you are shooting. 

 

The other LED sources have issues that are very similar but greater in the amount of perceivable difference. 

(...) 

Candlelight tests are tough to judge given the fact that differences in shadows play a greater role than daylight tests. Overall, I feel like the same observations for daylight apply for candlelight. 

create your own Jpeg samples

If you don´ have the possibility to compare the original ARRI RAW files, we have created some JPEG stripes of test-pictures for you to test with your own eyes. There will always be a possible error when converting from color space to color space / ARRI RAW to Jpeg. The conversion is uniform for all test light units. 

- close up of the difference in contast, color dept from some selected high quality LED lights. Softlights is #20.