Sticky for Neutral White flashlights possible?

[1970MP]

It seems that fans of neutral white light are few. But, I think we feel strongly that a neutral color temperature is an essential characteristic of a handheld flashlight. Personally, I will not even entertain a purchase of anything less.
In the spirit of trying to be helpful to myself and other fans, I would like to request a sticky be placed in the LED Forum for flashlights with a neutral white chip die. I will start with my neutral white flashlights that I have and like.

1) Olight Warrior X Pro 2,000 Lumens 21700
2) Olight M2R Pro Warrior 1800 Lumens 21700
3) Olight SMini Baton / Titanium 500 Lumens CR123
4) Thrunight TN12 1000 Lumens 18650
5) Manker E21 180 Lumens AAA x 2

 

[flashflood]

"Neutral" has been eclipsed by "high-CRI" over the last 5 years, and we are seeing more focus now on R9 in particular, as well as specific CCT/CRI combinations. You probably want something more like HCRI and/or R9080 threads.

 

[Jean-Luc Descarte]

Another relevant bit is the bands of color temperature. What is neutral to you could be warm to some and cool to others.

 

[1970MP]

Another relevant bit is the bands of color temperature. What is neutral to you could be warm to some and cool to others.

Sounds like you don't really care about the subject. Why even reply? Neutral is pretty globally recognized as ~4500 - 5500 kelvin, as opposed to warm at 4,000 or cool at 6000+
But, since there are zero positive replies, I guess I'm by myself, although I frequently see posts regarding requests for neutral.

 

[1970MP]

"Neutral" has been eclipsed by "high-CRI" over the last 5 years, and we are seeing more focus now on R9 in particular, as well as specific CCT/CRI combinations. You probably want something more like HCRI and/or R9080 threads.

Well, you are way ahead of me. I don't care about labels, I just want something in the 45k kelvin to 55k kelvin range. I had no idea about high-CRI, R9 or CCT/CRI. Thanks, I'll research those.
Just looked up high-CRI to find out that it is different than color temperature.
Just looked up R9 which stands for rendering of the color red. RA was mentioned and seemed correlated with CRI (color rendering index)
I have NEVER seen a flashlight description or technical detail for color rendering index, much less one with 80% to 90% or high CRI. An article listed 8 supposedly high CRI flashlights, but none were more than 300 lumens.
I think I have reached a point of too much knowledge being a bad thing.
Years ago, I saw an ad for Eagletac flashlights. Two products were compared. The difference was that one was cool and the other was daylight temperature. The daylight temp showed colors much better than the cool light. I have since paid attention and compared lights for myself. I find that I see colors more "correctly" with a daylight temp. There was never any mention of how high the CRI was. I imagine high-CRI costs a bunch, being new terminology. Kelvin temps seem to be no more expensive regardless of being warm, neutral or cool.

 

[bykfixer]

Decoding the Color Temperature Chart: Kelvin Scale Explained

See how to use the Kelvin scale and color temperature chart effectively. Learn to gauge light warmth or coolness in this informative guide.

www.shutterstock.com

Daylight is in the cool range.

Flashlight makers have developed a knack for promoting things that may or may not be true and throw out charts and graphs to prove the point. These days they toss around the R9 value stating the higher the better.
Now there is some truth to that. Using a boosted R9 makes reds show up better and that leads us to think we see other colors better because we do. Boosted red causes tricks to take place due to contrast between colors that a boosted red exagerates.
However the correct color is not what is seen using a boosted R9 even though many have been led to believe that. Now if a persons eyeballs judge colors better with a 3000k artificial light, great. If it's 4200k, good deal. Some prefer 5000, others 6000. What matters is what temp works best for the individual.

The reason there are no temp stickys is because of the arguments that would result.

 

[Duster1671]

As a counterpoint to jon_slider's post, I regularly use low CRI lights and don't have any problems figuring out which colors are which. Yes, things look better and more accurate under a high CRI, but imo the importance is sometimes grossly overstated. Especially when the hairsplitting between 9050 and 9080 starts...

 

[flashflood]

As a counterpoint to jon_slider's post, I regularly use low CRI lights and don't have any problems figuring out which colors are which. Yes, things look better and more accurate under a high CRI, but imo the importance is sometimes grossly overstated. Especially when the hairsplitting between 9050 and 9080 starts...

Nothing is better at auto white balance than the visual cortex. I use a Fenix TK75 quad XM-L for walking the dog, but for roasting coffee, I use an Acebeam EC65vn quad Nichia 519a.

 

[aznsx]

Decoding the Color Temperature Chart: Kelvin Scale Explained

See how to use the Kelvin scale and color temperature chart effectively. Learn to gauge light warmth or coolness in this informative guide.

www.shutterstock.com

Daylight is in the cool range.

Flashlight makers have developed a knack for promoting things that may or may not be true and throw out charts and graphs to prove the point. These days they toss around the R9 value stating the higher the better.
Now there is some truth to that. Using a boosted R9 makes reds show up better and that leads us to think we see other colors better because we do. Boosted red causes tricks to take place due to contrast between colors that a boosted red exagerates.
However the correct color is not what is seen using a boosted R9 even though many have been led to believe that. Now if a persons eyeballs judge colors better with a 3000k artificial light, great. If it's 4200k, good deal. Some prefer 5000, others 6000. What matters is what temp works best for the individual.

The reason there are no temp stickys is because of the arguments that would result.

Byk: You don't usually confuse me, but today you did. It's a valid assumption that I'm no SME on this (aka: I don't know jack). I believe this thread subject is about color temperature, but since it has diverged to include CRI, here's why I'm confused:

In the case of my current favorite LED for correctly rendering my natural surroundings (as determined thru some fairly extensive personal human testing) [not instrumentation], which I think is called a 'SST-20 R9080 4000K': I thought that the 'standard' CRI number (which I think I've read excludes R9) is 90+, and the R9-specific rating is 80. The bar-graph type CRI charts for many other LEDs including some 'HCRI' ones usually shows the bar for R9 far below all the other bars (basically a spectrum 'dropout' or trough), whereas the chart for this SST-20 shows the R9 bar being almost equal to the other bars (although still short of them).

That doesn't seem 'boosted' to me, compared with and with reference to the other wavelengths / spectral ratings on the chart. Instead the R9 bar seems 'almost up to but not quite equal to' the other bars on the chart - certainly not boosted in terms of being exaggerated above the others - it's still below them.

Are you saying you think the R9 being (almost) 'equalized' to the other levels is 'boosted', and the LED with the R9 at a much lower level than all the other bars on the chart are more 'natural' (ie: un-boosted)?

 

[bykfixer]

I'll try to explain what I meant AZ but in an over simplified manner that may cause pixel counters a nervous breakdown.

The graph you speak of indicates how well an R factor shows up. In the old days the R9 value was really low because at that time LED lights sucked at getting colors correct.

An old LED chart would look like this.
Example would be my old SureFire G2 with its P60 sized LED was so awful it made my red pickup look purple-ish gray. The goal at the time it was new was bright, bright, brighter.

Some LED's like the Seuol were fairly neutral tinted in that they were on the warm end of the cool spectrum. I'm guessing 5200 or so because I don't know the actual number. Pleasing versus that awful baby blue beam of the day. But again they were not really good at making it so what you saw appeared as acurate as sunlight even though the tint was similar. Low R9 numbers on charts and graphs and low CRI numbers (compared to today).

Time passed and companies like Nichia created an LED that could produce a beam that could cause my faded red pickup to look brand new fire engine red. Hence "boosted R9"……many touted the "rosey" look to the 219 variants. I liked the 219c for its more "golden" look.

To me it's all about the R13. Fleshtone. Get that right and the rest follows. In the meantime the industry still touts R9 as the magic ingredient of what makes a good CRI or not.

These days the manufacturers are getting LED's closer and closer to natural light thanks to a variety of reasons. And neutral tones are very popular with the masses at big box stores. Natural as in what appears as a neutral tone. Not cool, not warm. For years we used what were called "soft white" light bulbs in our home. That seems to be the way the flashlight industry is heading.

Here's an article that may help.

CRI, R9 & R13: What's the difference? - Lightology

1

www.lightology.com

 

[idleprocess]

Parametrek has a tint filter (LED color) if you'd like a reasonably up-to-date listing of what's on the market.

Daylight is in the cool range.

It is, but context matters; when your entire field of view is being flooded with sunlight in the 100,000 lux range - or about 1000 watts per square meter - your pupils snap nearly shut and your white perception is very different from typical flashlight intensity orders of magnitude less only partially illuminating a scene.

 

[bykfixer]

Agreed and on the flip side if a bunch of lux from a flashlight is focusing on a small area colors may appear washed out from over exposure.

 

[Duster1671]

@jon_slider - haha that's quite dramatic.

It's just a matter of practicality for me. Outside of internet flashlight nerdery (in which I am a regular and enthusiastic participant), color accuracy is just not that important.

It's sort of like...when I need to do a brake job on my car it doesn't really matter if I'm using Snap-on wrenches or Home Depot Husky wrenches. They both turn the bolts just fine.

 

[twistedraven]

Everybody knows that peak flashaholic is posting pictures of wood-tones and flesh-tones lit using flashlights to farm reddit karma. There is no higher purpose to the hobby.

 

[jeffsf]

Good to hear that at least some manufacturers are going for reasonable CRI rather than just lumens of central beam power. At least for us, we need to be able to tell the difference between a tan-colored and a gray-colored bird. For that, not only the old-school CRI, but also R9 provide some insight into if a light will be able to easily do that or not. R9 isn't anything magical, getting high-efficiency and reasonable red content at a low price have been challenging for emitter manufacturers.

Arguments aside, coming back and looking at lights after a few years since the Cree emitters seemed to be the only decent ones around for color rendering, at least a sane discussion of emitter options would be helpful.

 

[idleprocess]

@jon_slider - haha that's quite dramatic.

It's just a matter of practicality for me. Outside of internet flashlight nerdery (in which I am a regular and enthusiastic participant), color accuracy is just not that important.

It's sort of like...when I need to do a brake job on my car it doesn't really matter if I'm using Snap-on wrenches or Home Depot Husky wrenches. They both turn the bolts just fine.

Same. Some >99% of the time I need to be able to see at all, and be able to distinguish colors sufficient to recognize objects. Even ~20 years ago in the days of unpublishable CRI, wild swings in the Luxeon tint lottery, and Nichia angry blue white LED was sufficient.

 

[aznsx]

I'll try to explain what I meant AZ but in an over simplified manner that may cause pixel counters a nervous breakdown.

Thanx byk! I think I might have been confusing myself (which has been known to happen:-(. I think I get where you were coming from.

Well I've done all the hard work, which was coming up with an emitter that performs exactly as I feel it should in illuminating exactly the things I need to illuminate, in the most accurate possible way (among the emitters I've experienced to date), and in a far superior way to anything else I own or have ever owned. Contrary to my normal methodology, I determined the answer to my question without fully understanding and analyzing the question, which is 180-out for me - but it has worked beautifully!. Some day I'll do a short post about the experience. Now just for grins in my spare time, I can look at 'data' (retroactively) to try to understand WHY it's performing so outstandingly, although it's largely academic at this point I guess.

I believe this chart is at least roughly applicable to the emitter I'm using and referring to (although perhaps not exactly, keeping in mind that I don't know what I'm talkin' about here). If I look at this and ask myself: If part of the superb performance is due to the effect of one of these 'R' numbers, and if I had to choose whether that would be a) R9, or b) R13.... I guess I'd have to say: Can I go with c) - both of the above??

 

[flashflood]

Thanx byk! I think I might have been confusing myself (which has been known to happen:-(. I think I get where you were coming from.

Well I've done all the hard work, which was coming up with an emitter that performs exactly as I feel it should in illuminating exactly the things I need to illuminate, in the most accurate possible way (among the emitters I've experienced to date), and in a far superior way to anything else I own or have ever owned. Contrary to my normal methodology, I determined the answer to my question without fully understanding and analyzing the question, which is 180-out for me - but it has worked beautifully!. Some day I'll do a short post about the experience. Now just for grins in my spare time, I can look at 'data' (retroactively) to try to understand WHY it's performing so outstandingly, although it's largely academic at this point I guess.

I believe this chart is at least roughly applicable to the emitter I'm using and referring to (although perhaps not exactly, keeping in mind that I don't know what I'm talkin' about here). If I look at this and ask myself: If part of the superb performance is due to the effect of one of these 'R' numbers, and if I had to choose whether that would be a) R9, or b) R13.... I guess I'd have to say: Can I go with c) - both of the above??
View attachment 27926

That looks like a Nichia 519a. The tell is the R9 (red) of nearly 100, but R12 (blue) of 80. It's weird because deep blue is the one thing all LEDs make a ton of.

 

[twistedraven]

Don't forget that CRI is how a particular CCT compares to a reference black body illuminant of the same color temperature. The Ra bar graphs only tells you how accurate or inaccurate each particular color swath is rendered compared to the hypothetical black body illuminant of the same color temperature. We only know from that graph is that R12 (deep blue) is only 80% accurate to whatever the hypothetical black body illuminant renders deep blue as-- the inaccuracy could either be from desaturation or oversaturation. Color vector graphs of said emitter would tell you exactly if the deep blue is being oversaturated or undersaturated.


TLDR, CRI is a measurement of color accuracy, not color saturation. This is why you see most conventional blue leds with large blue spikes score poorly in R12, where as violet leds with very small blue spikes score very highly in R12.

 

[aznsx]

That looks like a Nichia 519a. The tell is the R9 (red) of nearly 100, but R12 (blue) of 80. It's weird because deep blue is the one thing all LEDs make a ton of.

Sorry FF. I was referring to the specific LED I mentioned and discussed in Post #9, but should have repeated that for clarity. The chart is from a test of a:

"Luminus SST-20-W 4000 K CRI95"

Since (unlike me) many here understand complete emitter specifications, it was specifically stated to be"

"SST-20-W40H-A120-J4FB4 VH"

The test (bare emitter / nominal drive) was posted in another forum in the past by a (evidently fine) person who goes by 'maukka' of Finland (who hopefully will not sue me for pasting that screen shot in this post. Full credit and thanks to 'maukka'!

I do not have the FULL specification of the specific LED(s) used in my human testing, but they were simply identified by the flashlight distributor who provided them as (as I recall): SST-20 R9080 4000K. They may or may not have the full specification / part number for those devices.

Hopefully the two are comparatively close relatives. I don't know the full definition(s) of the full specification(s), so as I mentioned, I'm speculating that they're "roughly" comparable devices, but without the full nomenclature on my emitters, there's an assumption involved. Hopefully someone will let me know if / how much that speculation might be in error.

Just to re-state something: My subjective human (non-instrumented) testing results are (tentatively) conclusive, and I'm merely speculating as to which (if any) objective measurements might possibly account for the device's excellent performance, and specifically possibly 'R' values, and specifically possibly those of R9 and R13. No other factors are being considered, expressed, or implied, so the scope of my post is intentionally quite limited (as appropriate for my knowledge level. Again, I'm essentially (speculatively) reverse-engineering test results and speculating about potential causal factors, rather than starting with specifications (data), and trying to establish the value / effects of those data in actual testing. It's the reverse approach of that taken by many people, but is interesting, because my subjective testing was essentially unbiased by the hard data - essentially almost like a 'blind' test. There were other differences in my approach, but I won't get into all that here.