White vs High CRI

53 posts / 0 new
Last post

Pages

wle
wle's picture
Offline
Last seen: 1 day 17 hours ago
Joined: 01/07/2015 - 13:49
Posts: 2564
Location: atlanta ga
White vs High CRI

So

What is the relation between ‘‘high CRI’‘, and “that light color looks perfectly white to me”?

Is it possible to just have one of those 2 things, more of one than the other, or are they basically the same thing?

Seriously

I have some LED bulbs that appear to me as perfectly balanced, white light.

Yet I think of “high CRI” as “expensive”, and I have never paid over $3 for an LED bulb.

wle

"You never have the wind with you - it's either against you, or you're having a good day."
    Daniel Behrman, "The Man Who Loved Bicycles".
It never gets easy, you just go faster.   
-Greg Lemond.
       ,ø¤º°`°º¤ø¸,ø¤º°`°º¤ø¸,ø¤º°`°º¤ø¸

LuxWad
LuxWad's picture
Offline
Last seen: 18 hours 52 min ago
Joined: 08/07/2021 - 13:22
Posts: 68
Location: USA

There is basically no correlation between color rendering and whiteness.

Consider this web page: BLF is mostly just white. That is produced by your computer screen using just three monochromatic colors: Red, green , and blue. The three of these together can produce “white”, while being extremely low-CRI – try comparing light from a phone or computer screen to a white LED and you will see this in action.

My Review Channel - https://www.youtube.com/luxwad

spamyak
Online
Last seen: 3 min 22 sec ago
Joined: 10/23/2021 - 16:01
Posts: 19
Location: Georgia (US)

CRI – the color accuracy of light as compared to a “black body” emissions source (like an incandescent lamp or the sun), where a black body source is 100 CRI. A low CRI light will not show certain colors accurately or brightly.

color temperature – the amount of yellow-orange or blue visible in the light source measured in Kelvin. a 60W incandescent lamp heats up to approximately 2700K, thus the color temperature of the light it produces we call 2700K. this light is considered “warm” though the number in Kelvin is rather low. a Xenon arc lamp glows at around 6200K producing a blueish light quite similar to daylight at noon, and we call this “cool” despite its higher Kelvin temperature value.

tint – a measure of the “greenness” or “rosiness” of the light. perfect tint appears neither green nor magenta and is exactly aligned to the black body locus (see below) whereas green tint (often seen as undesirable here) is any point above the line and rosy tint (sometimes desired here) is below the line

If you want a “perfectly balanced” white light, you may not need high CRI depending on what you are looking at. But if you want colors to “pop” and be rendered accurately as if in sunlight, you’ll want a higher CRI. You’ll almost certainly want a light with “perfect” tint as close to the black body locus as possible, and your preference for color temperature will depend on context—where are you using the light and what color temperature will your eyes be adjusted to? 4500k is a good medium value, but if you are using it outside near 1700K low pressure sodium street lights it will look quite cool and if you are exiting your kitchen where you have installed 6500k cool white bulbs it will look warm. If it is the only light source, your eyes will tend to adjust and the color temperature you see will look pure white.

fogofwar
fogofwar's picture
Online
Last seen: 4 min 56 sec ago
Joined: 08/19/2021 - 05:38
Posts: 130
Location: Shenzhen, China

Yep, there is no correlation and maybe negative correlation depending on your definition of “white.” It sounds like your definition of “white” is a high color temperature above 5000k. On the other hand, lower color temperature lights look “yellow” and tend to have higher CRI. This is because the red part of the spectrum is the most inefficient to produce and lower color temperature lights need less red to be close to the black body line.

People on here and Reddit especially try to chase high CRI but sacrifice too much in color temperature in my opinion. I’ll take high CRI, high color temperature over the same CRI but low color temperature any day. You have more color separation at high color temperature if CRI is constant.

Another way to say this is that 100 CRI at 1800k will still make everything look yellow-orange. Or a more extreme example is 100 CRI peaking in the infrared will make everything invisible.

Rayoui
Rayoui's picture
Offline
Last seen: 42 min 28 sec ago
Joined: 08/06/2019 - 00:38
Posts: 580
Location: Portland, OR

I don’t think low CCT is a “sacrifice” so much as a preference. I prefer warmer temps because I think cooler temps make things look cold and sterile. I’d rather have a 3500K 80 CRI than a 6000K 90 CRI.

wle
wle's picture
Offline
Last seen: 1 day 17 hours ago
Joined: 01/07/2015 - 13:49
Posts: 2564
Location: atlanta ga

OP here

what i consider a good white, is usually rated 3500-4000K

should that be high CRI, or not?

i guess what i don;t get is, why are the ‘other colors’ necessary, since our eyes only respond to 3 broad peaks of wavelength?

since white has all colors, if something appears white, why would it not be ok for the human eye to see all the colors the eye can see (which isn;t many)?

does a 100-CRI light appear to be a perfect white, or something else?

"You never have the wind with you - it's either against you, or you're having a good day."
    Daniel Behrman, "The Man Who Loved Bicycles".
It never gets easy, you just go faster.   
-Greg Lemond.
       ,ø¤º°`°º¤ø¸,ø¤º°`°º¤ø¸,ø¤º°`°º¤ø¸

spamyak
Online
Last seen: 3 min 22 sec ago
Joined: 10/23/2021 - 16:01
Posts: 19
Location: Georgia (US)

an emitter producing light exactly matching the wavelengths your eyes respond to in correct proportion would produce a perceptually perfect white light but I believe the light reflected back from objects you shine it on could look strange and “wrong”

spamyak
Online
Last seen: 3 min 22 sec ago
Joined: 10/23/2021 - 16:01
Posts: 19
Location: Georgia (US)

if you want to see examples of 100 CRI light just look at any clear glass incandescent or halogen light or unobscured sunlight. they can appear to be “perfect white” depending on context. perception is subjective.

dondizzurp
Offline
Last seen: 2 days 57 min ago
Joined: 11/12/2021 - 02:38
Posts: 27

CRI is for vibrancy and colour accuracy.
Kelvin is for warm or cool temperature. Basically a level of whiteness from yellow to pure white to blue.
Natural white light is somewhere around 4500k.

If you point your 2700k High CRI flashlight at a red object, you’ll without a doubt know that it’s red.
It will have a warm tint to it, but you can clearly tell it’s red. It won’t look orange, pink, maroon, yellow, or anything else but red.

If you point your 2700k low CRI flashlight at the same object, not only will you have a very warm tint but you won’t be able to accurately identify colours.

They are not mutually exclusive. You can have high or low CRI with warm or cool temperature. It’s just more common to see it with warmer tint.

Rayoui
Rayoui's picture
Offline
Last seen: 42 min 28 sec ago
Joined: 08/06/2019 - 00:38
Posts: 580
Location: Portland, OR

wle wrote:
i guess what i don;t get is, why are the ‘other colors’ necessary, since our eyes only respond to 3 broad peaks of wavelength?

since white has all colors, if something appears white, why would it not be ok for the human eye to see all the colors the eye can see (which isn;t many)?

I believe you are talking about the difference between emitted and reflected light.

A source that you look at directly, like a TV screen, can get away with emitting only three wavelengths because you are looking directly at them. It can mix those three colors to trick your brain into thinking it’s seeing an intermediate wavelength. However, when light reflects off an object, it may not reflect 650nm and 530nm light the same way it would reflect, say, 580nm light. If a source only emits three individual wavelengths, objects will appear to have strange and altered coloration when viewed in the reflected light from said source.

wle
wle's picture
Offline
Last seen: 1 day 17 hours ago
Joined: 01/07/2015 - 13:49
Posts: 2564
Location: atlanta ga

Rayoui wrote:
wle wrote:
i guess what i don;t get is, why are the ‘other colors’ necessary, since our eyes only respond to 3 broad peaks of wavelength?

since white has all colors, if something appears white, why would it not be ok for the human eye to see all the colors the eye can see (which isn;t many)?

I believe you are talking about the difference between emitted and reflected light.

A source that you look at directly, like a TV screen, can get away with emitting only three wavelengths because you are looking directly at them. It can mix those three colors to trick your brain into thinking it’s seeing an intermediate wavelength. However, when light reflects off an object, it may not reflect 650nm and 530nm light the same way it would reflect, say, 580nm light. If a source only emits three individual wavelengths, objects will appear to have strange and altered coloration when viewed in the reflected light from said source.

that doesn’t sound right

but it’s original, at least!

"You never have the wind with you - it's either against you, or you're having a good day."
    Daniel Behrman, "The Man Who Loved Bicycles".
It never gets easy, you just go faster.   
-Greg Lemond.
       ,ø¤º°`°º¤ø¸,ø¤º°`°º¤ø¸,ø¤º°`°º¤ø¸

wle
wle's picture
Offline
Last seen: 1 day 17 hours ago
Joined: 01/07/2015 - 13:49
Posts: 2564
Location: atlanta ga
spamyak wrote:
an emitter producing light exactly matching the wavelengths your eyes respond to in correct proportion would produce a perceptually perfect white light but I believe the light reflected back from objects you shine it on could look strange and “wrong”

WHAT!WHY?

"You never have the wind with you - it's either against you, or you're having a good day."
    Daniel Behrman, "The Man Who Loved Bicycles".
It never gets easy, you just go faster.   
-Greg Lemond.
       ,ø¤º°`°º¤ø¸,ø¤º°`°º¤ø¸,ø¤º°`°º¤ø¸

xevious
xevious's picture
Offline
Last seen: 9 hours 19 min ago
Joined: 02/27/2013 - 21:55
Posts: 2664
Location: Hoboken, NJ USA

So essentially, Hi-CRI is not bound to a specific color temperature, but to a range usually below 5000k?

When a temperature rating is stated for a given emitter, that’s the emitted light, not the temp that will be reflected?

spamyak
Online
Last seen: 3 min 22 sec ago
Joined: 10/23/2021 - 16:01
Posts: 19
Location: Georgia (US)

Consider temperature to be totally independent from CRI. An 1800K sodium vapor street lamp is very slow CRI (notice that you cannot easily distinguish colors by this light) whereas a candle at that same color temperature is very high CRI. Likewise there are (a few) 6500k >90 CRI emitters and many with very poor CRI.

Temperature measures essentially the ratio of yellow-orange to blue light you perceive. CRI measures the rendering of each individual color. This image comparing a high CRI (left) and low CRI (right) light source may help

Warmer color temperatures will have a peak further to the right, cooler will have a peak further to the left. This is independent of the overall shape/smoothness/completeness of the spectrum which is more or less what CRI measures.

Rayoui
Rayoui's picture
Offline
Last seen: 42 min 28 sec ago
Joined: 08/06/2019 - 00:38
Posts: 580
Location: Portland, OR

wle wrote:
Rayoui wrote:
wle wrote:
i guess what i don;t get is, why are the ‘other colors’ necessary, since our eyes only respond to 3 broad peaks of wavelength?

since white has all colors, if something appears white, why would it not be ok for the human eye to see all the colors the eye can see (which isn;t many)?

I believe you are talking about the difference between emitted and reflected light.

A source that you look at directly, like a TV screen, can get away with emitting only three wavelengths because you are looking directly at them. It can mix those three colors to trick your brain into thinking it’s seeing an intermediate wavelength. However, when light reflects off an object, it may not reflect 650nm and 530nm light the same way it would reflect, say, 580nm light. If a source only emits three individual wavelengths, objects will appear to have strange and altered coloration when viewed in the reflected light from said source.

that doesn’t sound right

but it’s original, at least!

Nothing original about it. Ever try to illuminate a room with an RGB LED?

NeutralFan
NeutralFan's picture
Offline
Last seen: 13 hours 2 min ago
Joined: 03/20/2014 - 19:22
Posts: 1588
Location: Wisconsin, USA

To me, perfectly white means sunlight at midday which is 5700K.

And then for high CRI you need to find LEDs that have high CRI at that CCT.

For me, I have a flashlight that has quad Nichia E17A SM573-B12-R9080 5700K LEDs that seems to fit your criteria.

I’d rather use my flashlight around the house than turn on the lights.

Unheard
Unheard's picture
Offline
Last seen: 2 hours 30 min ago
Joined: 01/16/2019 - 11:38
Posts: 2275
Location: Germany

Our vision adapts, and so there’s a broad region in the spectrum we can perceive as white. In the presence of two light sources, the reflection of one will appear white and the other not. If pointed to a white wall only, one light source appears rosy and the other greenish.

Smile, you cannot kill them all.

KeepingItLight
KeepingItLight's picture
Offline
Last seen: 17 hours 53 min ago
Joined: 06/15/2015 - 14:52
Posts: 760
Location: California

One easy way to learn about high CRI LED lights is to get your hands on a one (a bulb or flashlight), and shine it on some light colored wood furniture or flooring. Oak is a good choice.

Then compare against a low CRI light with the same Kelvin rating.

The difference is easy to see, and will probably surprise you. It sure surprised me. That's why I pay the extra bucks for high CRI.

For most rooms in my house, I use "neutral" 5000K bulbs, but, as you can see from the posts here, many folks prefer something warmer. Once again, you gotta try 'em to find out what you like.

 

 

JaredM
JaredM's picture
Offline
Last seen: 6 hours 3 sec ago
Joined: 10/31/2011 - 13:33
Posts: 2363
Location: Pittsburgh, Pennsylvania

I’m always perplexed at people that ask a question about something they admit they don’t know about (implicit and explicitly), and THEN come back at people providing answers with attitude and dismissing / discrediting the information.

Rayoui is correct, as are others here. Maybe in the situation where you don’t understand an answer, you could try asking a follow up question or for clarification instead of dismissing folks.

There are many layers to this topic, and a requirement of nuanced thought/understanding of the interdependent variables. Absorption, reflection, and emission are maybe terms to research and gain understanding of before continuing here.

CNCman
CNCman's picture
Offline
Last seen: 4 hours 42 min ago
Joined: 04/07/2018 - 10:10
Posts: 1227
Location: Mississippi Gulf Coast

Very interesting replies. I am still trying to fully understand it. Maybe someone can explain this. I build / mod mostly throwers for fun.
I had a MF04 with a XHP 70.2 P2 1A 6500K ( 70 CRI ) that burned up, so I replaced it with a XHP 70.2 N4 5000K ( 80 CRI ). I was pleasantly surprised at the down stream visibility increase because the reflection was less and this helped to see clearly. Lumens appeared less but I think it was because there was less reflection. Is this true ?
.

kennybobby
kennybobby's picture
Online
Last seen: 1 min 25 sec ago
Joined: 05/10/2017 - 09:13
Posts: 990
Location: huntspatch, alabama

@CNC, was the reflected light maybe causing a glare or wash-out that distorted your vision, and with “less reflection” the long distance viewing was clearer?

The whole idea of CRI is related to how well you can see the true colors, like if you spotted someone walking out at a distance was their jacket brown, black or navy blue.

The temperature in Kelvins is related to the overall average tint color of the light, with the higher numbers toward blue-white and the lower kelvins toward orange-red.

It’s hard for me to see the real colors of things when the light is biased toward red, could be worn out eyeballs…

[edit] improper use of the technical meaning of “tint”

Now i used to think that i was cool,
drivin' around on fossil fuel,
until i saw what i was doin',
was drivin' down the road to ruin. --JT

jon_slider
jon_slider's picture
Online
Last seen: 47 sec ago
Joined: 09/08/2015 - 12:20
Posts: 5214
Location: Central North America

fwiw, Tint is not the same as Color Temperature

.

and white is not just one Color Temperature, warm white is 3000K and lower.. Cool White is 5000K and higher.. 4000K is Neutral White

Neutral Tint is not the same as Neutral White..
Neutral Tint is actually all along the Color Temperature line, neither green, nor magenta

there is a LOT more info on Color Temperature and CRI, in this thread, includes photos… scroll down to posts by BrokenRecordBot

Streamtronics
Streamtronics's picture
Offline
Last seen: 5 hours 9 min ago
Joined: 01/02/2019 - 19:14
Posts: 98
Location: Germany

wle wrote:
WHAT!WHY?

I'll try to dumb it down a bit. If you mix red, green and blue light, you get white. So far so good. Imagine an object which only reflects yellow light, but no red and green. In sunlight this object will look yellow.

But in the RGB mixed light it would look dark and dull, because there's no yellow light to be reflected back.

Maybe, depending on the actual spectral reflectance of the object, it might even appear more orange or more greenish than it would be under sunlight, even if the RGB light source looks exactly like sunlight to your eyes. This is the effect of low CRI.

This is also true if you compare a high CRI and low CRI lightsource. They can have the same CCT (color temperature, for example 2700K for incandescent) and tint (green/magenta shift, pretty much perfectly neutral for incandescent) and look exactly the same to your eyes on a "perfectly white" surface, but once you introduce real world colors (inks, pigments etc.) the colors you see will be different.

Oli
Offline
Last seen: 12 hours 36 min ago
Joined: 08/03/2018 - 21:03
Posts: 579
Location: United States

NeutralFan wrote:
To me, perfectly white means sunlight at midday which is 5700K.
I’ve seen lower numbers than that quoted. But it’s not entirely relevant. How often do you hold your flashlight directly overhead pointing straight down? How do you like that midday sun when the ground is covered in snow? When you’re holding a flashlight at night and light is reflecting back at you off of trees and or other surfaces that’s a whole lot different than the sun being straight over your head. Cooler temperatures including 5000k and above are going to reflect more light back at you off of objects than warmer temperatures. That affects your ability to see past the close objects in your light path. So we can’t compare the big flashlight in the sky to the little one in your hand.

You can't compare the big flashlight in the sky to the little flashlight in your hand.

jon_slider
jon_slider's picture
Online
Last seen: 47 sec ago
Joined: 09/08/2015 - 12:20
Posts: 5214
Location: Central North America
Streamtronics wrote:
If you mix red, green and blue light, you get white. So far so good. Imagine an object which only reflects yellow light, but no red and green. In sunlight this object will look yellow.

But in the RGB mixed light it would look dark and dull, because there’s no yellow light to be reflected back.


Maybe, depending on the actual spectral reflectance of the object, it might even appear more orange or more greenish than it would be under sunlight, even if the RGB light source looks exactly like sunlight to your eyes. This is the effect of low CRI.


This is also true if you compare a high CRI and low CRI lightsource. They can have the same CCT (color temperature, for example 2700K for incandescent) and tint (green/magenta shift, pretty much perfectly neutral for incandescent) and look exactly the same to your eyes on a “perfectly white” surface, but once you introduce real world colors (inks, pigments etc.) the colors you see will be different.

Excellent Post!

one of the very best, explanations I have seen

Happy Holidays!

> But in the RGB mixed light it would look dark and dull, because there’s no yellow light to be reflected back

same thing happens with Red, and this is The BIGGEST difference between Low CRI and High CRI:

things that reflect red, will look Red, only if the LED produces Red, which Low CRI does not

if the light is Low CRI with Negative Red output, then the red object will look brown

here is one example:
High CRI and Low CRI Tomato sauce
.
(guess which is which)

extra credit quiz
which one tastes Redder? LOL

KeepingItLight
KeepingItLight's picture
Offline
Last seen: 17 hours 53 min ago
Joined: 06/15/2015 - 14:52
Posts: 760
Location: California

jon_slider wrote:
fwiw, Tint is not the same as Color Temperature

.

and white is not just one Color Temperature, warm white is 3000K and lower.. Cool White is 5000K and higher.. 4000K is Neutral White

Neutral Tint is not the same as Neutral White..
Neutral Tint is actually all along the Color Temperature line, neither green, nor magenta

Very useful info from Jon Slider. I tend to agree with most of what he has to say. I’ve learned a bunch from his posts over the years.

One minor difference between him and me (which we have discussed in pms) is that I go with the photographer’s perspective on what makes neutral white.

Take a look at the image above. If you draw an oval around the white central region, I want to be on the black-body radiation line right in the middle of the white area. Roughly speaking, that puts me somewhere between the dot at 4000K and the dot at 6500K, closer to 5000K than 4000K.

jon_slider
jon_slider's picture
Online
Last seen: 47 sec ago
Joined: 09/08/2015 - 12:20
Posts: 5214
Location: Central North America

CNCman wrote:
I had a …6500K ( 70 CRI )

I replaced it with a … 5000K ( 80 CRI ). I was pleasantly surprised at the down stream visibility increase
.

sounds like the Higher CRI helped create better color separation… imagine how much better it could get with a 90+ CRI LED Wink

have some fun with this Low CRI vs High CRI comparator:
https://techneesh.com/assets/juxt/cri1.html

Firelight2
Firelight2's picture
Offline
Last seen: 5 hours 49 min ago
Joined: 04/08/2011 - 15:17
Posts: 4745
Location: California

I think the perception that you must sacrifice “white” to get “high CRI” comes in part from the available selection of LEDs.

  • Many people think of “white” as “cool white”. This is a misconception. A color that looks pure white under one lighting condition may look too warm or too cool in another. Human eyes have an auto-white balance which tends to correct for color temperature, to an extent.
  • Most high CRI leds were only available in warmer color temperatures. This is true for XPG, XPG2, SST-20, XPL-HI. Nichia 219B at 4500K is a bit of an outlier on the cooler side, but has low output.
  • It is only recently that high-CRI has become commonly available in some cooler tints. LH351D at 5,000K for example. But even then, these high-CRI LEDs tend not to be perfect, often featuring a greenish tint.
  • Result is some people associate “high CRI” with “not white LED. This despite the fact that color temperature and tint (which together determine how white an LED appears), are separate factors from CRI.
pinkpanda3310
pinkpanda3310's picture
Offline
Last seen: 5 hours 19 min ago
Joined: 12/28/2013 - 08:45
Posts: 3038
Location: Perth

@jon_slider regarding the tomatoe sauce – I thought red was more to do with R9 which is not part cri calculations…?

  

Firelight2
Firelight2's picture
Offline
Last seen: 5 hours 49 min ago
Joined: 04/08/2011 - 15:17
Posts: 4745
Location: California
pinkpanda3310 wrote:
@jon_slider regarding the tomatoe sauce – I thought red was more to do with R9 which is not part cri calculations…?

I thought R9 is part of CRI calculations.

It’s just that CRI is calculated based on the average score for a whole range of colors, including R9. If all the other colors score high and R9 scores low, an LED may still rate as high-CRI, because the average is high. However, without R9, reds will still look washed out or the wrong color.

I think this is why the top-rated leds for CRI on BLF are typically also the ones with decent R9: SST-20 and Nichia 219B.

EDIT: see Luxwad’s explanation in post 31. My explanation in this post is wrong.

LuxWad
LuxWad's picture
Offline
Last seen: 18 hours 52 min ago
Joined: 08/07/2021 - 13:22
Posts: 68
Location: USA

CRI is an average score of specific (consistent) shades of color. R9 is the ninth color used, which happens to be red:

Note that each color is “Rx”, and has a rating from 0-100. Really bad LEDs can actually have negative scores, usually in R9.

However “CRI”, which is usually CRI (Ra) only covers R1-R8, so R9 is not included in that measurement. The extended CRI (R96a) covers R1-R15. In either case, a low measurement in one area may not be enough to rate an LED as low-CRI, since it’s an average value.

There are other scales, such as TLCI (Television Lighting Consistency Index), that use more & different colors.

My Review Channel - https://www.youtube.com/luxwad

Pages