The only thing missing from that video is a second Sekonic device to see how much variation you get with them.
My Opple II is all over the place on measurements of CCT and duv.
I’ve measured >100 different LEDs and here are my conclusions:
CCT of emitters measures low. The higher the CCT, the further off it seems to be.
Examples:
-XPL HI 5A bin emitters measure 3600 to 3800k
- XPL HI 3A bin emitters measure 4600 to 5000k
- XML2 2A bin emitters measure 5000k to 5300k
- XHP50.2 6500k emitters measure 5750k
- XPL HI 1A bin emitter measures 6000k
Low CRI emitters measure much higher duv by 0.005 to 0.008.
The Opple II measurements are fairly repeatable and useful for relative measurements of like emitters (similar CRI).
So, for example, it works for comparing different Nichia 219 LEDs to each other. The relative CCT and duv match what I see.
It does not work for comparing different LED types with much different CRIs.
Two very similar looking Samsung LH351D emitters, one 90 CRI and one 70 CRI, measure much different duv (–0.0012 for high CRI, 0.0064 low CRI).
The low CRI looks slightly more pink.
I don’t know if there this is a difference between the II and III versions, or I got a bad one?
Anyone else test the II version?
FB
thanks for sharing your tests
it does not make sense that 0.0064 looks more pink than –0.0012
did you make a typo, or did you mean that?
if possible I would be interested in seeing a photo of those two lights, with white ballance set to 5000k.
fwiw, here is the data from the Cheule video:
.
it looks like CCT and CRI are quite consistent across the 3 devices…
the DUV is more variable, in some of the data comparisons I see Opple reading about 0.0024 lower than the sekonik, or in the case of lights w High DUV, the Opple can read about 0.0024 high…
otoh, there are many examples where the DUV values are much closer. See the chart and form your own conclusions
I posted these images from my Opple. Both lights have a fantastic -duv feel to them and the test bares that out. These were both tested at full power.
Best sst-20’s @ 4000k
Even mix of sw-35 and sw-45k, non-ramping.
I find it to be a very usable tool, wish I could afford something that really broke out the discrete Rx values, but ain’t gonna happen! Works better than using a FOV WB filter and histogram on my 5Dii, at least easier.
Not a typo.
I have several examples of this, and agree that it doesn’t make sense. That is why I posted it. Most reports on the Opple are very favorable for a cheap device. My experience hasn’t been so “rosy”.
All my low CRI lights measure very high duv. Many of them are XPL HI 3A and 5A bins and do not appear green (that’s why I have so many 3A and 5A bin lights).
I also have a lot of Nichia and Samsung high CRI lights. Most are good tints to my eyes. All measure varying degrees of - duv. The biggest deviation vs my eyes happens with the highest CRI LEDs I have: SST-20s. It really seems that my Opple sample’s results are heavily biased by CRI.
I watched the Cheule video and studied the data. I’ve read and watched everything I can find on the Opple. In Cheule’s data, I also see the Opple reading higher duv on low CRI and lower duv on high CRI LEDs, just to a lesser amount than my device.
Again, I have the Opple Light Master II (not III). Maybe my device is different? Maybe they tweaked the calibration of the III to be closer to reality?
Just reporting my experience. I’m sure other members have a low CRI and High CRI lights with similar Tint and CCT. Compare them using the Opple.
I’ve never bothered to post images… maybe I should.
FB
I am not sure that +duv will always appear green. Looking at binning charts, it also appears that they could appear yellow and I suspect would photograph yellow as well. After dark this evening I can test XP-L hi 3D in my FW1A and XP-L hi something warmer in my GT micro. I to have noticed that several do not look green, I do select for that behavior in non-Nichia emitters. Will post the results later.
The perception depends on ambient light. A sw45k D220 appears green in the light of a sw45k D200.
Quite correct. Most of the 45k’s I have in my stash are D180 and D200 and they make anything look green!
Before the Opple, that is how I compared tints, side by side in a darkened room.
I’m comping low CRI lights to LEDs like Nichia 219b.
Another example…
Nichia 219b 4000k 9080 from azhu (earlier batch than he has today): Opple: CCT: 3678, duv –0.0053.
Nichia 219b 4000k 9050 from Illumn: Opple: CCT: 3527, duv: –0.0037.
Cree XPL Hi 5A3 from KaiDomain: Opple: CCT: 3741, duv: 0.0073.
They all appear very similar in a dark room shining on a white wall. All in 1AA lights on high (LT Tool).
I can see the slightly different CCT, but they all have a slightly pink tint. I would put the duv (by eye) between the two 219b, closer to the 9080.
I picked these to 219b LEDs because they are very close visually to this XPL HI.
FB
In follow up to xp-l hi performance testing… Both CRI at full power were right around 67.
XP-L Hi @ 5000k, should be a 3A
XP-L Hi @ 4000k, this one is really nice, but I did not build it so I am not sure exactly what bin.
Thank you mattlward,
I plugged those numbers in to online calculators and got:
XP-L Hi @ 5000k, should be a 3A: CCT 5055k, duv 0.0014
XP-L Hi @ 4000k (looks like 5A to me): CCT 4225k, duv 0.0009
For comparison:
I’ve measured (6) XP-L Hi 3A LEDs with Opple duv readings of 0.0060 to 0.0085
I’ve measured (10) XP-L Hi 5A LEDS with Opple duv readings of 0.0062 to 0.0090
I can only conclude that my Opple misreads low CRI LEDs quite severely. It seems to be much more reasonable with high CRI LEDs.
FB
I do not use dark rooms for visual and photo comparisons.
I do use a dark room for Opple readings.
I find it helps me to see Pink or Green Tint much better if I use Daylight white balance. That includes comparing lights with my eyes, when my brain is using daylight white balance.
here are some of the techniques that work for me.
1. take the photo during the day near a bright window, not in a dark room, and not at night when my brain is adapted to warm house lights
.
the reason we can see the hosts, is because of the daylight in the room, and that daylight is telling my auto white balance iPhone what to do about the Pink Tint of the sw45k LEDs
I find Pink Tint is Most Visible during daylight white balance. It is NOT noticeable at night when Im Incandescent adapted, and my brain sees sw45k as slightly blueish… ie cooler white than the 3000k Ambient source.
2. Shine lights on a white wall or piece of printer paper or something white, during the day, not in the dark.
3, include a cool white light in the lineup, to create a daylight white reference for brain and camera.
.
folded printer paper makes for a very simple white wall studio…
4. My grey silicone soldering mat gives me some nice results, when not using white walls or white paper…
.
that is a 3000k 219b, photo taken during daylight, to demonstrate the warmth, and pink tint, of the LED.
Specs, meters and opinions are helpful; but I wish I could also find a large poster-sized inexpensive "color chart" or "color wheel" which would "challenge my lights" so I could more easily visually observe and compare with my own eyeballs their ability or inability to illuminate the various colors.
It would need to be composed of many colored boxes that touched each other (with no other visual border between the colors) and contain enough strategically arranged colors so that the boundaries between the adjacent colors would be more difficult to identify for the lesser lights.
I should just start collecting "color chip samples" at various paint stores and paste them in an overlapping pattern on a large poster board for a test chart.
something like this?:
https://www.amazon.com/Pixel-Perfect-Camera-Color-Correction/dp/B07VCTX9SC/
.
Something more like these images, but without any white borders between the colors, so I can visually decide if I can actually see a distinct difference in the colors where they touch each other:
I like lower CCT lights (especially indoors) for their soothing illumination, but I have noticed that for some colors, it is more difficult for me to see the difference between different shades of the same color with the lower CCT lights (even those with high CRI) and I would appreciated a test chart that allowed me to investigate these difference for all colors.
I hope you find a tool that works for you.
I agree that 3000k High CRI does not necessarily show colors as well as 4500k High CRI
besides the difference in CCT, and Tint DUV, the CRI R9 (red) output makes a big difference in how realistic some colors look.
so a 9080 LED such as 219b or SST-20, can do a better job of showing the Red of a Tomato, than an LH351d 9050 that has lower Red output than 9080 LEDs.
Found these on AmazonUS from a third party seller:
Largest Size (36 x 54 inches) is $27 ($15 + $10 shipping + $2 sales tax)
Best Deal is for 24 x 36 inch size at $18.50 ($14 + $3 shipping + $1.50 sales tax)
Most Durable is for Laminated 24 x 36 inch size at $22 ($17 + $3 shipping + $2 sales tax)
But I am not sure if reflections from the glossy laminated surface will be a problem.
Those are off-set printed posters, so the “colours” are made up of mixed dots of just three dies (cyan, magenta, yellow). That does not represent how reflections from much more complex “real world” objects look like, in other words: they may alter the spectrum in a different way than a similar looking colour in a printed poster.
To check tints and colour rendering, I have a couple of objects in the house that I shine at, the best are different types of wood, my skin, an orange, some blue things etc, with some training this gives away CRI. Estimating CCT is very doable too, it requires a second known light source of known colour temperature, such as daylight or an incandescent lamp, or both, for comparison. Duv is the most difficult, if it is very far from the BBL you may spot it directly, but for any more normal tint you need a known reference light of similar CCT. And you can be easily fooled, i.e. if the well-known Nichia 219B V1 sw45 R9080 (very rosy tint far under the BBL) is used as reference, any completely ok tint looks awfully green.
I once owned the HKS charts, made from the actual paints. Those can be useful I imagine. As Jos said, make sure you don’t buy CMYK Charts.
Over they years, I have used Lego blocks and a couple of other items for their color intensity,
sw-45k D180
sw-403 unsure of brightness bin, but 2017-18 stock.
