CRI of White LEDs

[HarryN]

Less discussed, but equally important in our perception of which "white light" is most pleasing, are our genetics.

If you compare a picture taken with traditional Kodak vs Fuji film, it jumps right out at you how virtually impossible it is to find one solution for everyone.

 

[UnknownVT]

Less discussed, but equally important in our perception of which "white light" is most pleasing, are our genetics.

It's not just genetics but personal preference and where we come from and our "conditioning".

Not everyone will conform to the norm although the Kruithof curve is empirical -
ie: can be confirmed and reproduced by experiment -
individuals can have different results - eg: jtr1962 above from this thread.

This can make any visually based assessment of CRI kind of suspect -
eg: what happens if it's a panel of LED marketing people?
or people who may be color blind?

Where's the proposed standard for controls and controlled testing and even a mere mention of eye/brain adjustments to different lighting levels and environments?

 

[Curious_character]

As UnknownVT and others have pointed out, a lot of the problem is due to the way our brain processes visual input. We internally declare any of a wide range of spectra to be "white" after being exposed to them for even a fairly short period. I first became aware of this a long time ago when we had chosen a very slightly off-white color to paint a room -- the paint mixer added just one glurp of green to a can of white. I was painting in a ceiling corner, where I was surrounded by the color. After some time, I stepped down, turned around -- and everything except the wall had a distinct pinkish hue. My brain had decided that the wall was "white", so everything else was pinkish. This is happening to me a lot now with the various lights in my house.

If you shine a white (say, hot incandescent) light through an object, say a thin piece of fabric, the object absorbs some wavelengths more than others, resulting in what we perceive as colored light. The spectrum that gets through probably has a bunch of lines of various wavelengths and intensities, and ranges of wavelengths where there isn't much being reflected. As it turns out, you can make a light consisting of just three spectral lines and, by adjusting only their amplitudes, have it look like exactly the same color as the complex spectrum from the filtered white light. The spectra are very, very different, yet they look just the same to us. This is, of course, the basis of color television, which produces nearly any color by combining just three. (As I recall from long ago, you can't quite imitate all possible colors, but you can get very nearly all by careful choice of your three primary colors. It seems our eyes and brains can be fooled pretty easily.

The problem is that if you shine the two apparently identically colored lights on something of another color, you'll get different results. If, for example the three-color light is made of pure red, green, and blue, and the object reflects none of those exact wavelengths, it'll look black. The filtered light might contain some of the wavelengths the object reflects, so the object will be some color and not black. This happens with "white" lights, particularly so with fluorescents which tend to have spectra consisting of a number of discrete lines rather than a more continuous, filled-in spectrum. And that's why an RGB LED light might look white (or any other color we want by adjusting the three LED intensities) but won't render colors anything like a heated filament (black body) light.

c_c

 

[kitelights]

I just made a comment in another thread about production LED headlamps not being suitable for surgery.

I don't own any high CRI LED lights, but I do own some 'neutral' and 'warm' LED lights and none of them will show the distinction of redness on human tissue, such as a sore throat or an infection that I had on my shin.

They seem to do fine for me to distinguish colors in other applications, but not tissue. Can one of you gurus explain the reason to me, but so that I can understand it?

I seem to remember that Don put out a few specialty high CRI lights and Baby Doc was a tester who reported that the light worked well for him in medical applications. I had the impression that this was the springboard for the current demand for these higher rendering LEDs.

 

[jtr1962]

I just made a comment in another thread about production LED headlamps not being suitable for surgery.

I don't own any high CRI LED lights, but I do own some 'neutral' and 'warm' LED lights and none of them will show the distinction of redness on human tissue, such as a sore throat or an infection that I had on my shin.

They seem to do fine for me to distinguish colors in other applications, but not tissue. Can one of you gurus explain the reason to me, but so that I can understand it?

I seem to remember that Don put out a few specialty high CRI lights and Baby Doc was a tester who reported that the light worked well for him in medical applications. I had the impression that this was the springboard for the current demand for these higher rendering LEDs.

The reason has to do with the phosphors used. "Regular" neutral and warm non-high CRI LEDs simply have a higher ratio of phosphor to primary blue emitter output. This alters the color balance so that our eye perceives the light as warmer, but doesn't add much on the far red end. The relative size of the phosphor hump is higher, so there is a little more deep red there, but no much more than cool whites. Incidentally, triphosphor fluorescents use the same trick, simply altering the ratios among the red, green, and blue phosphors to achieve the desired color temperature. End result is that low color temperature triphosphor fluorescents are lacking in red relative to a blackbody emitter.

High-CRI LEDs are a different animal. They use an entirely different phosphor which has a lot more output in the deep red area ( at the expense of efficiency ). Although the light color on a white wall might appear the same as a non-high CRI LED of the same color temperature, the spectrum is much closer to that of a blackbody. When I shine one of my lights modded with a SSC high-CRI neutral white emitter outdoors in the backyard, it's like night and day compared to the cool whites. Indoors the difference isn't as great ( I still prefer cool-white indoors ) but outdoors among grass and trees the high-CRI light gives three dimensions to whatever you're lighting. That was the inspiration which led me to make a reading light using 5 of these high-CRI neutral white SSC LEDs ( my only complaint is I wish the color balance was closer to 5000K-5500K ). They should do just as well lighting human tissue. Also, from what I understand the Nichia 083 is slightly better than the SSC high-CRI, although it is a bit harder to obtain.

 

[McGizmo]

Kitelights,
Consider this graphic representation of Nichia's regarding CRI:

Specifically consider R6 thru R10 and R12. R9 is probably the most significant when it comes to typical white LED's and to what extent is the CRI R9 effected by the CCT's offered in white LED's?

I would guess that R12 sucks for LED's compared to the incandescent "standard" established by incandescent (black body) for two reasons. Black body sources are not relatively strong in the low wave lengths and LED's are overly so. My gut tells me that indirect sunlight can not match the CRI-12 of an incandescent source; especially if there is plenty of blue sky above.

 

[blasterman]

If you compare a picture taken with traditional Kodak vs Fuji film,

Fuji dye coupling technology allows for a longer saturation ramp with a more density via more sloped integral resulting in more saturated colors without blocking. The sacrifice is density range. Kodak films are just the opposite in that they have very linear saturation ramps and high degrees of density, but poor saturation handling, aka color blocking. Take a picture of something very red or orange with a classic Kodak slide (or print) film and it will appear one solid shade of color. This is a big reason Velvia practically destroyed Kodak's entire slide division by itself. So, there is a measurable difference in Fuji emulsions -vs- Kodak, although Kodak has done much better to fix this problem (for those who still shoot film).

I actually ran into an interesting requirement that proved that cold-white, or simply older tech emitters weren't good enough to distinguish colors correctly. That application was building custom LED light-bars in a very busy night club where drinks are set and the LEDs are primary source of color. The problem is that because of the lack of decent red and amber cool white LEDs, or older Luxeons simply couldn't distinguish certain types of whiskeys or beers. They all looked the same in the window, and this really caused some issues.

I experimented with half a dozen types of neutral and warm white emitters to fix the problem, and was about to resort to using high CRI Seouls when I stumbled on some older 1watt warm-white Cree P3s with that distinct 'rose' hue, and they do an amazing job on distinguishing drinks from each other. The difference is actually rather dramatic. I've not found this tint in newer Crees, but newer Bridgelux warm-whites are close.

 

[Nil Einne]

It's not just genetics but personal preference and where we come from and our "conditioning".

Not everyone will conform to the norm although the Kruithof curve is empirical -
ie: can be confirmed and reproduced by experiment -
individuals can have different results - eg: jtr1962 above from this thread.

This can make any visually based assessment of CRI kind of suspect -
eg: what happens if it's a panel of LED marketing people?
or people who may be color blind?

Where's the proposed standard for controls and controlled testing and even a mere mention of eye/brain adjustments to different lighting levels and environments?

Indeed differing preferences is one important thing that seems to be often missed. It always irks me when people say 'cool white' is 'unnatural' as I don't consider there's strong evidence for this. Obviously it's questionable whether incandescent light bulbs, or even camp fires etc can be considered 'natural'. The most 'natural' thing is very little light at night so the ideal colour temperatures of relatively bright lights at night doesn't seem something you can necessarily get from the patterns of the sun. (Of course I dislike many common uses of the word 'natural' anyway.) Perhaps we've been dealing with camp fires for long enough that we've evolved to prefer that sort of light, perhaps not, ultimately without good evidence it's still primarily theoretical. In any case, modern lighting is a lot brighter then that.

The circadian rhythm issue is perhaps one of the few interesting things I've read whenever this 'natural' light idea comes up. However light which is good for your circadian rhythms (and that may very well be no artifical light source, i.e. just the sun during daylight hours and at night moon if present and stars) may not necessarily be light you prefer. The other thing which interests me is the rods & cones issue but it too appears to be primarily theoretically.

And as for the anecdotal evidence and studies, and here I get to what I began with. Although I've never found a great ref (http://www.archlighting.com/industry-news.asp?sectionID=1341&articleID=460610 is one of the best I've found) I personally believe (and do have some personal knowledge here coming from Malaysia originally) it's generally accepted that colour temperature preferences in parts of Asia, what parts is one of the things that I've never really read studies of but I suspect most of SEA and East Asia at least, differs from that of much of the West. In particular cooler whites are usually preferred there. Most of the anecdotal evidence and studies are primarily coming from Western cultures (the museums ones I've seen mentioned have been in the west and while I haven't looked at the studies from what I've read it doesn't seem like it was possible for them to try and determine if there was specific difference in visitors from some places and in any case a visitor to a foreign visitors and in any case a tourist's preference may vary from an average person from those places anyway.)

So the question becomes, why is this? While this could be genetic, getting back to my early points I don't think there's sufficient evidence that we can say any of the preferences are strongly genetic or innate. They could very well all be primarily cultural. In fact in some ways it reminds me of the way some people argue a female preference for pink is probably innate because it appears so widespread despite the fact we have some evidence it wasn't even always the case and pink may have actually been a masculine colour 80 years or so ago.

While I admit I haven't really looked into the research available that well, it seems to me there is plenty of obvious areas of research here. I've already mentioned that it doesn't even seem to have been well studied how the colour temperature preference vary from country to country but another obvious study would be comparing people of different ancestoral groupings in the same country (since some argue there's a genetic component).

P.S. Just for clarity, I have no problem with someone saying something seems unnatural to them. Obviously people are entilted to their individual preference. What I is the idea that cooler whites lights are in some ways inherently 'unnatural' or at least more so then warmer light ones.

 

[McGizmo]

I think there is strong evidence for subjective and cultural preference in regards to color temperature and even color itself. In previous threads on these and related subjects it was mentioned that as we age, our eyes respond differently and even get yellowed. It was suggested that some of us older folks would be inclined to prefer a cooler CCT as we inherently warm the light as it is filtered through our eye's lenses.

Another aspect of the subjectivity I think has substance is that of the terms of warm and cool light and how they relate to our present environment. In tropical and warm climates I find it no surprise that a cool CCT might be preferred and in the higher latitudes or in winter, a warmer CCT would have preference.

If you are freezing your heinie off and look in one direction and see a cool moon reflected off the snow and in another direction a warm fire reflecting off the snow, which will draw you to it?

On the other hand, if you are in sweltering in a humid tropic evening and see a cool moon reflecting off a lake or ocean and in another direction see a cooking fire's warm light, which would draw you near in hopes of perhaps some welcome relief?

In temperate zones where one is neither cold or hot, I suspect one would be most indifferent to the CCT of light. :shrug:

We have evolved to use ambient light in order to see and perceive. We have also used fire at night to see as well as cook (at anytime of day). We have learned and adapted with light as a significant part in our ability to perceive and function. There are likely inherrent biases and instinctual responses based on visual clues and perceptions and how they relate to light sources as well as time of day.

One thought I just had would be of primitive man, our ancestor, sound asleep under the stars. If all of a sudden his eyelids were bathed in a bright cool light this would be from a rising full moon and not likely to cause alarm or necessarily wake him. On the other hand, if all of a sudden his eyelids were bathed in a bright warm light, either day break has caught him by surprise or worse, a fire is present. In either case, it would make sense that this light would initiate a waking process.

The significance of color rendering is a function primarily of the need to identify and differentiate among objects and landscapes based on their color. In some cases this may be critical and others of no import. Our eyes have evolved based on natural light which is for the most part based on a full spectrum as represented by a black body source. It would be curious to know if primitive man shared the same visual spectrum as we do now or if perhaps he was more perceptive to say the IR end or even the near UV. I wonder if the proportion of rods and cones or their predominance has changed any over our history or if there are differences among genetic backgrounds. The need for glasses for instance does not seem to be consistent among contemporary races or perhaps that is a false perception on my part?!? Is there a "universal eyeball" that would make sense as a base and is there a universal, optimal, artificial light source? I kind of doubt it. :shrug: :thinking:

Regardless, we are no longer limited to only black body sources and to what ever extent they may fall short of an ideal, we may now be able to improve and enhance our visual perceptions.

 

[blasterman]

Obviously people are entilted to their individual preference

'Preference' is fine, provided were talking about light sources with similiar CRI indexes.

What I have a problem with is people pulling the 'preference arguement' and then justifying cool-white emitters simply because they have a technology fetish with LEDs.

Low CRI, cool-white emitters might be better than cheap daylight CFLs, but they are still the 'fast food' of lighting technology. Basically, over-marketed, over-convenient, and just pimped on the public because they are in too much of a hurry to really stop and look at the difference.

The difference in color rendition and livability of Cree cool-white R2s -vs- drastically less efficient warm-white Bridgelux in my living room is drastic. I honestly don't care if the Crees are 2x as efficient. The color sucks, and is devoid of entire bands of spectrum our visual cortex spent millions of years evolving to see. This is the type of light thats good for the laundry room, or a porch light, or a college kid in his dorm spending 99% of his time on a computer.

 

[jtr1962]

Although I've never found a great ref (http://www.archlighting.com/industry-news.asp?sectionID=1341&articleID=460610 is one of the best I've found) I personally believe (and do have some personal knowledge here coming from Malaysia originally) it's generally accepted that colour temperature preferences in parts of Asia, what parts is one of the things that I've never really read studies of but I suspect most of SEA and East Asia at least, differs from that of much of the West.

I've said as much many times in response to the idea that "people prefer warm colors at home". Not true when you look at the world as a whole, and not even really dependent upon climate. Also, anecdotally when I go for walks I see more and more 5000K and up lighting compared to just a few years ago. Now that more people are using CFLs, they also realize they're not restricted to the 2700K of household incandescent. Many are choosing much cooler lighting as a result.

So the question becomes, why is this? While this could be genetic, getting back to my early points I don't think there's sufficient evidence that we can say any of the preferences are strongly genetic or innate. They could very well all be primarily cultural.

I tend to agree. We haven't lived long enough under artificial light ( that includes fire ) for it to have shaped us genetically. Those kinds of changes take millions of years. Any preferences are largely to do with cultural expectations. If mainly genetics were involved, I would expect to see very little variation across countries in the type of preferred lighting. And yet that's not what we see. When you get huge general variations in preferences, such as 2700K-3500K in the suburban US, perhaps 4100K in US cities, 5000K in Japan, as high as 7500K in a few other places, this is telling me it's largely culturally-based. I doubt there's any significant differences due to race. NYC for example is a mix of races, and yet I don't see much variation in preferred lighting color based on neighborhoods. The increased use of hgiher CCT seems to be citywide, although it might be a little less prevalent in poorer neighborhoods where CFLs aren't as readily used due to higher initial cost.

While I admit I haven't really looked into the research available that well, it seems to me there is plenty of obvious areas of research here. I've already mentioned that it doesn't even seem to have been well studied how the colour temperature preference vary from country to country but another obvious study would be comparing people of different ancestoral groupings in the same country (since some argue there's a genetic component).

Such studies would be interesting although I doubt we'll find a genetic component to lighting preferences. For example, I prefer anywhere 5000K and up, same as in many Asian countries, but I'm not Asian. I know some Asians who can't stand higher CCT lighting. Same thing with every other race or ethnic background. I think the best answer is lighting preference is largely influenced by culture, but even within a culture there are outliers ( like myself ).

P.S. Just for clarity, I have no problem with someone saying something seems unnatural to them. Obviously people are entilted to their individual preference. What I is the idea that cooler whites lights are in some ways inherently 'unnatural' or at least more so then warmer light ones.

The interesting thing here is first time I saw a white LED, I felt it was more natural than any other artificial light source I had seen up until then. High-CRI, high CCT fluorescents weren't bad, but the LED just seemed better. This was before CPF and before I learned about spectral distribution. Now I know why the LED felt more natural. There's probably a good case to be made that a continuous spectrum will feel "right" more often than a discontinous one, even if the CRI numbers are the same. This is why we need a better measure of light quality than CRI.

 

[UnknownVT]

I seemed to have missed this in the Wikipedia entry on CRI -

" New test color samples
As discussed in (Schanda & Sándor 2005), (CIE 1999) recommends the use of a Macbeth (now X-Rite) color chart owing to the obsolescence of the original samples, of which only metameric matches remain.[13] In addition to the eight ColorChart samples, two skin tone samples are defined (TCS09* and TCS10*). Accordingly, the updated general CRI is averaged over ten samples, not eight as before. Nevertheless, (Hung 2002) has determined that the patches in (CIE 1995) give better correlations for any color difference than the Macbeth chart, whose samples are not equally distributed in a uniform color space."

 

[Canuke]

Such studies would be interesting although I doubt we'll find a genetic component to lighting preferences. For example, I prefer anywhere 5000K and up, same as in many Asian countries, but I'm not Asian. I know some Asians who can't stand higher CCT lighting. Same thing with every other race or ethnic background.

This would likely be correct. The lens of the eye has a yellowish cast, which filters out a little blue and UV, and this filtering tends to get stronger with age. There is a similar layer of yellow overlaying the fovea. Variations of the strength of these filters, due to age, genetics, nutrition etc. are likely the driving factor for color preferences.

 

[Anders Hoveland]

This is an interesting article about CRI and LED light:
http://www.cinematography.com/index.php?showtopic=51884&st=20

Apparently regular white LED's are not suitable for illumination in the film industry because of their inferior color rendering.

 

[RoGuE_StreaK]

We get it, you don't like LEDs. Why do you bother hanging out here then?

 

[Anders Hoveland]

We get it, you don't like LEDs.

I am very enthusiastic about LED's. Did you not see my thread? http://www.candlepowerforums.com/vb/showthread.php?344630-My-100-Watt-LED-(pictures)

I am just more enthusiastic about building LED's for novelty than actually lighting my home with them. I am anxiously waiting for full spectrum white LED bulbs to become available. Only then I will start replacing more of my lighting with LED's. I have already experimented with several different types of LED bulbs but have mostly been unsatisfied with the quality of their light in many places.

 

[AnAppleSnail]

Anyone still shooting film emulsions is kind of crotchety. Digital filming has done a lot, including full-reel color correction. I already correct for CCT and spectrum gaps with still photography.

 

[blasterman]

I am just more enthusiastic about building LED's for novelty than actually lighting my home with them

Then go off and do that then.

The dedicated commercial fixtures we discuss in the fixed lighting forum such as the ones Cree makes do nothing less than surpass halide, fluorescent and incan light sources in terms of efficiency and especially color. You are basing conclusions on bad science and what's available at Walmart. Replace that crappy, 100watt contraption you made with a decent 4000k Bridgelux and you woulnd't be making these comments.

I already correct for CCT and spectrum gaps with still photography.

For fun, browse through the continuous lighting options at B&H or another supplier and you'll see an endless product stream of essentially garbage. Many of the lights are still based on 5mm cold-white LEDs...basically the same lights made for DJ and dance floors, but marketed towards video enthusiasts. So yeah....it's no wonder the cheaper lights tick off the guys in post because they can't get the color right. Digital cameras based on bayer sensors (which are like 99.99% of them) have an issue with red contrast in fine detail. When the digital camera is adjusted for color balance with light using cold white LEDs the lack of red starts causing all kind of issues. You also have the problem that dSLR video shooters think they're now Steven Spielberg when they are in fact shooting with 4:2:0 compression and throwing most of the color information away, so cheap light sources don't make the problem any better. Still, the myth is that a 5000-6000k light source with a CRI less than 70 is still better because it matches the "color of daylight".

 

[bshanahan14rulz]

This post is old. The post Anders linked to is also, relatively speaking, old. Like B-man said, those lights they use don't use special LEDs or anything, so just because their crappy lights don't have good spectral content does not mean that you can generally say that LEDs do not have a good CRI. We now have high cri LEDs that aren't stuck in the warm white region, and maybe someone will be more concerned about making a good video light than about saving money and actually put these to good use.

 

[Anders Hoveland]

The post Anders linked to is also, relatively speaking, old. We now have high cri LEDs that aren't stuck in the warm white region,

That article specifically mentioned cyan frequency light, showing how that lady's dress appeared muted under LED illumination. Even the latest high CRI white LED's are still deficient in cyan frequency light, and, for example, the Phillips L-prize "enhanced spectrum" LED bulb and recessed lamps do absolutely nothing to solve this issue.