CRI of White LEDs

[SemiMan]

The author of that post says that cyan is missing, but a relative comparison of LED spectrum and 3K tungsten spectrum shows similar cyan output. The peakiness on either side of the cyan is likely more of an issue. One of the problems is that the demosaicing of the camera is not tuned specifically to LED spectra. This can be done to greatly improve the color rendering under "traditional" LED light sources ... but that is more a discussion for consumer level devices like cell phones where this is done for LED flash. There is quite a bit of research that was done around this when LED flash came out for camera phones .... and why some of them looked really terrible at first beyond just being dark.

The latest high CRI LEDs, at least compared to incan, are not deficient on a relative basis, and their blue peak is greatly subdued. I would suggest looking at the 90+ CRI Luxeon Rebels, or the Xicato Artist Series modules. I would call them somewhat the state of the art in best spectrum at reasonable cost points. Sharp's 90+ CRI Zenigata has tons of cyan, but still has a somewhat significant blue peak. Interestingly enough, Sharp's have quite a smooth spectrum other than the blue peak which would indicate the blue peak could be filtered to have an overall smooth curve. I have not done this, but I looks possible on paper.

Semiman

 

[idleprocess]

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.

Someone tell that to B&H, whose catalogs feature a wide variety of LED video lights.

That forum post focuses on how film reacts (digital seems to be steadily grabbing market share and can do quite a bit of correction both before and after-the-fact), and also couples ancillary characteristics of specific products - such as PFC - with the underlying light source technology, which is a bit odd in a fast-changing field.

 

[SemiMan]

'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.

In many situations, like you, I would prefer those Bridgelux LEDs. But not all, and perhaps that is the other variable for preference. It is application and environment dependent. If I was doing any tasks where fine visual acuity was important, I am going to go towards the cool whites as long as the CRI is not terrible. I would often feel that way about general daytime tasks as I need the blue in the cool LED to stimulate the circadian rhythm. My kitchen has high CRI cooler white 4000K. Its pleasant for me and others that occupy my kitchen at times. My shop and lab are lit with daylight high CRI .. yes fluorescents. Bedrooms are 2700K ... they are places to relax (except for the cool white reading light), and living room is 3000K ... mainly for relaxing but mixed. Bathrooms are mainly 3000K, but think I am going to change the ensuite on the master to 4000K. All the 2700 and 3000 in my house is 80-85 CRI CFLs and/or LEDs.

I do outdoor commercial lighting and in those instances would take 65-70 CRI 5000K over 100CRI 3000K any day. 3000K at low lux levels suck when you are trying to accomplish something ... at least that is my opinion

Semiman


There was a study a number of year ago that compared perceptions of 70-75CRI cool white against warm white. For that study, there was actually a preference for the cool white.

Not sure that latitude matters to color temp preference but I liked the theory. North America tends towards warmer whites at least for interior. Asia tends toward cooler whites. Europe somewhere in the middle .... all at the same latitude.

 

[UnknownVT]

Someone tell that to B&H, whose catalogs feature a wide variety of LED video lights.
That forum post focuses on how film reacts (digital seems to be steadily grabbing market share and can do quite a bit of correction both before and after-the-fact), and also couples ancillary characteristics of specific products - such as PFC - with the underlying light source technology, which is a bit odd in a fast-changing field.

found this under Color Rendering Index on Wikipedia:

Film and video high-CRI LED lighting incompatibility
Problems have been encountered attempting to use otherwise high CRI LED lighting on film and video sets. The color spectra of LED lighting primary colors does not match the expected color wavelength bandpasses of film emulsions and digital sensors. As a result, color rendition can be completely unpredictable in optical prints, transfers to digital media from film (DI's), and video camera recordings. This phenomenon with respect to motion picture film has been documented in an LED lighting evaluation series of tests produced by the Academy of Motion Picture Arts and Sciences scientific staff.[30]​

 

[SemiMan]

I would say that the color balancing issue is not unique to LED but applies to fluorescent as well for cameras. One of the big difference is the range of experience with fluorescent. I was involved in some early camera phone stuff with LEDs. For bayer filter sensors, properly rendering with LED required demosaicing algorithms that were specific to the spectrum of the LED. This could significantly improve color accuracy. As color of any object is dependent on the lighting used, really the goal is to accurately render colors under what the light is "supposed" to be, not necessarily what it actually is. High end cameras used to be mainly 3 sensors, one per color, but really high end stuff now is actually single sensor, so the whole color of the filter, demosaicing, etc. comes into play.

The bandpass filters on the sensors and/or emulsions does not need to match the light source though. What it needs to match is eye response such that what is rendered by the camera matches what the eye would perceive. If you achieve that, you can adjust the end result. Unfortunately depending on the medium, both in film and digital, that may not be the case. The RGB on RGB sensors generally does not match the eye response. That is if RGB is used. Complimentary colors are also used as they allow significantly more light to reach the sensor. This throws the whole eye match out the window.

Complex stuff, but not unique to LED specifically.

 

[teddoman]

I'm a photography enthusiast, and first time poster, so please forgive any errors in advance. I must say this is an incredibly useful discussion. Obviously, the people in this forum know a ton about lighting. From the photography perspective, I'm just looking for continuous full spectrum lighting that give my photographs proper white balance and color rendition without unpredictable color spikes. However, after reading this thread, it feels like the folks in this forum haven't really come to a consensus on whether that is even possible with products on the market today.

If price is not a major issue (within reasonable limits), is there any continuous full spectrum lighting that can be bought today, and is definitively this type of lighting and not just falsely marketed as such? I just want to know that if I'm going to drop a few extra bucks to get the right lighting, that I'm not just fooling myself. I guess I don't want to rely on the manufacturer's marketing, since as has been discussed at length in this thread, CRI is not really an accurate measure, nor is "full spectrum" a regulated term so any manufacturer is free to use it and cannot be held accountable for it.

I have read that halogen lights are essentually continuous full spectrum lighting, but those generate heat and burn a lot of energy. Is there a more environmentally friendly alternative, if price is not a major issue?

 

[Anders Hoveland]

From the photography perspective, I'm just looking for continuous full spectrum lighting

It is very rare, but there actually are two incandescent and fluorescent products that have spectral graphs that almost exactly match natural sunlight. Truelite brand fluorescent tubes, and SoLux brand halogen reflectors.
However, you need to use some caution, virtually all the other products I have researched that claim to have "natural light" are exaggerated or deceptive.

For a simple solution, the best fullest spectrum of light can be obtained by using energy-saving halogen replacement bulbs, together with a small ammount of blueish white LED light.

There actually are two LED products that have high 92+ CRI at higher color temperatures but they are very expensive and have to be special ordered. These are the Richardson Company MoleLED series, and Selador Desire series, the latter of which just consists of multiple different frequency chips in the same spotlight. The main reason LED spotlighting has not been more widespread in the film industry so far is that they have done a poor job at properly illuminating deep red and cyan (teal) colors.

 

[neutralwhite]

adding to this topic, I never knew the CRI of the MCE 4000CCT LED is just 75CRI, and the Nichia 219 is like 93 HCRI.
so obviously, its best to go for the Nichia in this 4000 right?.

btw my pd32ue is also 75cri.

 

[SemiMan]

I'm a photography enthusiast, and first time poster, so please forgive any errors in advance. I must say this is an incredibly useful discussion. Obviously, the people in this forum know a ton about lighting. From the photography perspective, I'm just looking for continuous full spectrum lighting that give my photographs proper white balance and color rendition without unpredictable color spikes. However, after reading this thread, it feels like the folks in this forum haven't really come to a consensus on whether that is even possible with products on the market today.

If price is not a major issue (within reasonable limits), is there any continuous full spectrum lighting that can be bought today, and is definitively this type of lighting and not just falsely marketed as such? I just want to know that if I'm going to drop a few extra bucks to get the right lighting, that I'm not just fooling myself. I guess I don't want to rely on the manufacturer's marketing, since as has been discussed at length in this thread, CRI is not really an accurate measure, nor is "full spectrum" a regulated term so any manufacturer is free to use it and cannot be held accountable for it.

I have read that halogen lights are essentually continuous full spectrum lighting, but those generate heat and burn a lot of energy. Is there a more environmentally friendly alternative, if price is not a major issue?

What are you lighting with continuous sources or is this just a model light? Most of the 3000K 90+ CRI lights will give you great colors and a little emphasis on the blue end may help at the capture end of things and can always be reduced later. Generally you need a strobe to stop motion even portraits and to get low ISO/ low noise. I am not on the latest CS edition from Adobe, but I think it can do custom mapping from light sources.

If you can find a light with Xicato Artist series LEDs of Bridgelux 90+ CRI units (can't remember the series) then that is as smooth as you will need.

Semiman

 

[AnAppleSnail]

adding to this topic, I never knew the CRI of the MCE 4000CCT LED is just 75CRI, and the Nichia 219 is like 93 HCRI.so obviously, its best to go for the Nichia in this 4000 right?.btw my pd32ue is also 75cri.

It depends on the wavelengths you want to perceive. If you don't need reds, you won't miss the high CRI and will appreciate the bonus lumens.

 

[Esko]

It is very rare, but there actually are two incandescent and fluorescent products that have spectral graphs that almost exactly match natural sunlight. Truelite brand fluorescent tubes, and SoLux brand halogen reflectors.
However, you need to use some caution, virtually all the other products I have researched that claim to have "natural light" are exaggerated or deceptive.

There is a bunch of manufacturers that make so-called full spectrum fluorescent lights in daylight CCT and cri range of 95-98. For example, Phillips and Viva-Lite. I once saw a project where some amateur video photographers made a big videolight from some 10-15 full spectrum fluorescent tubes and a sheet of corrugated iron. It looked pretty nice.

There actually are two LED products that have high 92+ CRI at higher color temperatures but they are very expensive and have to be special ordered. These are the Richardson Company MoleLED series, and Selador Desire series, the latter of which just consists of multiple different frequency chips in the same spotlight. The main reason LED spotlighting has not been more widespread in the film industry so far is that they have done a poor job at properly illuminating deep red and cyan (teal) colors.

What is your source concerning MoleLED? The leds used in MoleLED seem to be rather ordinary blue leds with phosphors and the spectrum seems to be rather traditional, too. On the other hand, it seems to be somewhat close to the spectrum of high cri Nichia 219. I'd still pick Selador, it is more versatile and seems to cheaper, too.

 

[teddoman]

It is very rare, but there actually are two incandescent and fluorescent products that have spectral graphs that almost exactly match natural sunlight. Truelite brand fluorescent tubes, and SoLux brand halogen reflectors.
However, you need to use some caution, virtually all the other products I have researched that claim to have "natural light" are exaggerated or deceptive.

Thank you Anders. Today it requires so much knowledge to be an educated consumer. I can't imagine how regular consumers deal with an issue like this. Here I am posting in a forum for lighting enthusiasts, and I myself am coming from the photography perspective!

The Solux in particular seems to have some great testimonials from photographers and museums. It comes in different types though. Is there any reason from a photography perspective or from a human health perspective to use 3500 vs 4700 vs 6000? Would all of these be equally "full spectrum" Solux bulbs?

For a simple solution, the best fullest spectrum of light can be obtained by using energy-saving halogen replacement bulbs, together with a small ammount of blueish white LED light.

What is it about the halogen replacement bulbs that require a blueish white LED light? Are they insufficiently "full spectrum"?

What are you lighting with continuous sources or is this just a model light? Most of the 3000K 90+ CRI lights will give you great colors and a little emphasis on the blue end may help at the capture end of things and can always be reduced later. Generally you need a strobe to stop motion even portraits and to get low ISO/ low noise. I am not on the latest CS edition from Adobe, but I think it can do custom mapping from light sources.

If you can find a light with Xicato Artist series LEDs of Bridgelux 90+ CRI units (can't remember the series) then that is as smooth as you will need.

Thank you. I don't know if you had a chance to read this entire thread but there are a bunch of people who know a lot more than me about lighting theory that are trying to debunk high CRI lighting as being falsely advertised as "full spectrum". So I'm less likely to rely just on a CRI number.

I am looking for bulb solutions for home photography. With the right lighting, I may not have to do all kinds of post processing adjustments that would otherwise be required.

 

[AnAppleSnail]

The reason to add blue or red light is to cover a wide spectrum. For better or worse, CRI declares that the perfect ideal is a black body radiator. Never mind perceptual studies that show white-perception being nearly always off a black body curve (which of these lights is white? Mhmm.)

The trouble is, a black body radiator emits a bell-curve-like spectral power distribution. So one end will be weak, depending on the CCT you aim for. Most other light sources have a weak end, which can be propped up by added light at that wavelength. So halogens at low CCT need more blue photons. And so on. This is why I prefer to mix CCT for viewing lights. Of course, for photography it's easiest to match all CCT and bump saturation selectively as needed.

 

[neutralwhite]

what would be the point then in say having an MCE LED at 4000k with a CRI being at 75, than just better using a Nichia 219 at like 93CRI?.
my PD32UE is 75cri, so what's so good about the old MCE anyway then if that too is 75CRI.

which would be better in a flashlight?. certainly the 219 right?.

 

[Esko]

What is it about the halogen replacement bulbs that require a blueish white LED light? Are they insufficiently "full spectrum"?

The full spectrum is just biased towards the warm colors like in all warm white incandescent bulbs (even if the cri is ~100).

Of course, since the Solux bulbs (the higher CCT neutral ones) cost less than $10 each and the spectrum and the consistency will be better, there is no real reason to make a mixed light source.

Is there any reason from a photography perspective or from a human health perspective to use 3500 vs 4700 vs 6000? Would all of these be equally "full spectrum" Solux bulbs?

Full spectrum yes, but with different intensities in both warm and cool ends of the spectrums. For photography purposes? I'd choose the most balanced 4700K (there are Solux bulbs with more than three CTT choices available). Human health perspective? No reasons, unless you are going to use them in your ambient home lighting.

 

[SemiMan]

It is very rare, but there actually are two incandescent and fluorescent products that have spectral graphs that almost exactly match natural sunlight. Truelite brand fluorescent tubes, and SoLux brand halogen reflectors.
However, you need to use some caution, virtually all the other products I have researched that claim to have "natural light" are exaggerated or deceptive.

For a simple solution, the best fullest spectrum of light can be obtained by using energy-saving halogen replacement bulbs, together with a small ammount of blueish white LED light.

There actually are two LED products that have high 92+ CRI at higher color temperatures but they are very expensive and have to be special ordered. These are the Richardson Company MoleLED series, and Selador Desire series, the latter of which just consists of multiple different frequency chips in the same spotlight. The main reason LED spotlighting has not been more widespread in the film industry so far is that they have done a poor job at properly illuminating deep red and cyan (teal) colors.

As has been pointed in the past Anders, Solux bulbs DO NOT match natural sunlight, they match the blackbody curve at 3000, 3500, and 4000K within the visible light spectrum AND sunlight when it is at these color temperatures which is for a short period of time in the morIning and evening DOES NOT match the blackbody curve. That said, as a continuous spectrum source for photography, that should be good. However, keep in mind for photography, as they go up in color temp, they go down in efficiency as they are using a filter to accomplish this.

Of course, generally for photography, except for taking pictures of small items, a continuous light is rarely used in the studio beyond a modelling light and you use strobes for the actual photos so I am not sure what you are trying to accomplish? Can you give us some more insight?

Teddoman, in terms of regular consumers, they would not be able to tell the difference unless told exactly what to look for between a 95CRI somewhat peaky spectrum and 95CRI smooth spectrum. Actually without reference colors even those that are quite experienced with lighting will not be able to readily tell the difference between incandescent and 85CRI LED ... or even a good CFL for that matter.

If you add "blueish white" LED to halogen, you will end up with a peaky spectrum ... just like you were trying to avoid.

 

[Esko]

Of course, since the Solux bulbs (the higher CCT neutral ones) cost less than $10 each and the spectrum and the consistency will be better, there is no real reason to make a mixed light source.

However, keep in mind for photography, as they go up in color temp, they go down in efficiency as they are using a filter to accomplish this.

It looks like I didn't quite realize how much the efficiency drops. (link). Should have read the specs in more detail.

Mixing ordinary halogen bulbs with blue or cool white leds might have an advantage in reproducing reds (and blue) but one would still be lacking some output in cyan and deep blue range.

Of course, generally for photography, except for taking pictures of small items, a continuous light is rarely used in the studio beyond a modelling light and you use strobes for the actual photos so I am not sure what you are trying to accomplish? Can you give us some more insight?

Teddoman, in terms of regular consumers, they would not be able to tell the difference unless told exactly what to look for between a 95CRI somewhat peaky spectrum and 95CRI smooth spectrum. Actually without reference colors even those that are quite experienced with lighting will not be able to readily tell the difference between incandescent and 85CRI LED ... or even a good CFL for that matter.

This is also true. Perhaps some extra information would be useful.

 

[teddoman]

I reread this thread in full again. This is a lot of info to digest! I think I "got" a lot more on the second pass, particularly in studying the links on CCT, the Kruithof curve, and googling blackbody radiators. Thanks all for your patience. Now I have a few keeper points:
1. All objects emit light radiation and are natural blackbody radiators
2. CCT measures the radiation level of each blackbody radiator
3. Artificial light sources mimic natural blackbody radiators imperfectly. The spectrum of colors may not be smooth and follow the curve of a natural blackbody radiator. The colors graph often has big spikes.
4. The color spectrum of natural daylight, from some of the graphs I looked at, is not a straight line. It's a curve with a small peak in the center, and the entire curve is upward sloping, lower at the warm colors and higher at the cool colors.
5. Light emitted by natural blackbody radiator may not always be pleasing to the eye. For any CCT, the Kruithof curve shows that pleasing light colors are obtained only within a certain range of illuminance levels. But the Kruithof curve was found empirically and hence may vary from individual to individual.

Human health perspective? No reasons, unless you are going to use them in your ambient home lighting.

Yes, this is EXACTLY my purpose, ambient home lighting. It's winter now, there was a post in this thread on seasonal affective disorder, my wife has at times had trouble sleeping, my 1 year old is not a great sleeper, and dealing with 2 kids has thrown my wife and I's sleep cycle rhythms off in the last few years too. So yes, whether it's proven or not, I would like to mimic natural light as much as possible, to capture any ancillary benefit regarding seasonal affective disorder and circadian rhythms. (Err on the side of health and good sleep.)

As has been pointed in the past Anders, Solux bulbs DO NOT match natural sunlight, they match the blackbody curve at 3000, 3500, and 4000K within the visible light spectrum AND sunlight when it is at these color temperatures which is for a short period of time in the morIning and evening DOES NOT match the blackbody curve.

Thanks for this clarification. I think my summary at the top of this post are now consistent with what you are saying here.

So technically, to mimic real daylight, we would actually have to have a lighting solution that changes CCT throughout the day, peaking midday just like natural sunlight. Instead, what we actually have in our homes and offices is a solution at one fixed CCT, which the sun gives us just one or two times during a normal day. So we just have to choose, do we want it to always be a cloudy morning? Or do we always want it to be a sunny day in the shade? etc etc. From a human health perspective, I wonder which choice is better. This sort of reminds me of typical space shows/movies where the computer changes the time and it goes from looking out the window at nighttime to suddenly looking out the window at sunny daylight.

Of course, generally for photography, except for taking pictures of small items, a continuous light is rarely used in the studio beyond a modelling light and you use strobes for the actual photos so I am not sure what you are trying to accomplish? Can you give us some more insight?

I am not talking about studio photography here, which is perhaps what you are thinking of. That is an entire model of photography that requires different lighting. Think candid photography of authentic moments, rather than formal portrait photography in a studio.

I am trying to achieve a home lighting situation where flash of any kind is not required for home photos of my kids in action. This gets into the aesthetics of photographs, but essentially if ambient lighting is sufficient, some photographers prefer to avoid flash. While sophisticated flash solutions can make the lighting seem more natural, I am trying to craft a solution where flash simply becomes unnecessary. Again, it's just my personal photography aesthetic inclination (but one I know a segment of photographers share).

However, keep in mind for photography, as they go up in color temp, they go down in efficiency as they are using a filter to accomplish this.

It looks like I didn't quite realize how much the efficiency drops. (link). Should have read the specs in more detail.

Ok, so the higher CCT Solux bulbs are basically filtering out a lot of the spectrum to achieve the blackbody curve at the higher CCT so you'd need more bulbs to get the same illuminance levels?

I assume the 4700 on Esko's chart looks so flat because it's a log chart, but that the shape of the 4700's color spectrum should be the same as the others.

Related to the efficiency issue is the Kruikof curve. If I am going to use a cooler CCT, the Kruikof curve requires higher illuminance (and energy usage) in order to achieve pleasing light colors (assuming the Kruikof curve is true). A warmer CCT allows more energy savings because pleasing colors can be achieved at lower illuminance.

 

[neutralwhite]

thanks, nice post, so if you do suffer from SAD (seasonal affective disorder ), is the best room light say, a 4000k one, or a 5000k one?.
i thought 5000 would be daylight, to keep your moods happy. bright day, rather than sunset like.
thanks.

 

[AnAppleSnail]

BRIGHT, first. Then good color rendering and decent CCT. I get along just fine with 4000K. Other people use pure blue light with regular home lighting.

Edit: how bright? I've seen behavioral studies get strong effects up to 2500 lux indoors.

Edit 2: 500-25000 lux. That's a wide range from office bright to 25000 lumen per square meter.