Sunlight, Cool LED, Warm LED

Fellas,

I mentioned this concept in another thread (https://www.candlepowerforums.com/posts/3225632&postcount=50) but just finally got around to grabbing the picture I wanted. I am interested in thinking about cool vs. warm LEDs and how this really compares to actual sunlight... I thought, wouldn't it be cool to directly compare a beamshot of the sun?!

Behold! Here is a pic of the mid-afternoon sun (~3pm) shining through a window onto a white wall on the L; the Quark Mini 123 (Cool) in the middle and the Quark Mini 123 (Warm) on the right. Both lights are on High. The camera is a Canon S90 set to manual:


Now the same image (a few minutes later, note the movement of the sun), now a few steps down:



A few thoughts...

First, wow, is the sun bright! I wonder what the lux would have read on each of these (note: I was about 4 feet from the wall with the Quarks)...

Also, compared to sunlight, the difference between the warm and cool LED is pretty much nothing (compare that to my beamshots in a dark room here:

From my thread Quark Mini 123 Cool vs. Warm here: https://www.candlepowerforums.com/threads/255553

The technique I used is frought with problems and very uncontrolled--ideally, I would use a mirror to reflect the sunlight into a truly dark room and then compare it to a number of lights, including HID and incans! That is my dream... But this is still cool! It's neat to really think about the ultimate "High CRI" light...

I would love to hear people's thoughts about this... and would LOVE to see similar photos! Thanks guys!! :twothumbs

I have noticed this in an exercise I do to compare all my lights to sunlight. I compare light beams to sunlight in shadow close to a window and in those conditions all LEDs turn more blue. The purest white cool tint LEDs turn very blue and neutral LEDs seem much closer to sunlight, but still have a hint of blue. I have no really warm lights so I don't know how close they get to more even light.
I have not compared to direct sunlight, that is an interesting result. I think direct sunlight is warmer than shadow.
To me, neutral lights are much more natural in general use. In hiking and mountain biking where depth percetion is key, neutral has it all over cool tints even at reduced output. The comparison to shadow sunlight is a simple way to compare tints for dominant color in the spectrum.

I think the intensity of the sunlight is causing a problem. We don't think of moonlight as being warm. It seems bluish, but it has the same spectral distribution as sunlight. The big difference of course is that it is a lot dimmer.

I think some part of the perception of the LED lights as being very blue might be due to their relative low intensity. It might be an idea to try the same thing with a beam of moonlight and some LEDs on low.

applevision said:

First, wow, is the sun bright! I wonder what the lux would have read on each of these (note: I was about 4 feet from the wall with the Quarks)...

Click to expand...

Yea, more photos like this would be interesting including the moon idea. Thanks for putting these up.

I think that moonlight does actually exhibit a slight spectral shift from the blackbody radiation of the sun, since it's not bouncing off of a perfectly reflective surface. IIRC, moonlight is around 4100K while sunlight is about 5700K. Fairly big shift, and would explain why moonlight is "cooler".

.
Finding the perfect LED tint is a "pigment of the imagination".
.

That's a good one!

But the quest for "Better" is worthy - and never ends.

march.brown said:

.
Finding the perfect LED tint is a "pigment of the imagination".
.

Click to expand...

csa said:

I think that moonlight does actually exhibit a slight spectral shift from the blackbody radiation of the sun, since it's not bouncing off of a perfectly reflective surface. IIRC, moonlight is around 4100K while sunlight is about 5700K. Fairly big shift, and would explain why moonlight is "cooler".

Click to expand...

Wouldn't that suggest moonlight is "warmer" ? :thinking:

The sun is bright indeed. Test is not accurate until the sun is 4 feet from the wall like the lights. LOL.

Cheapskate said:

I think the intensity of the sunlight is causing a problem. We don't think of moonlight as being warm. It seems bluish, but it has the same spectral distribution as sunlight. The big difference of course is that it is a lot dimmer.

I think some part of the perception of the LED lights as being very blue might be due to their relative low intensity. It might be an idea to try the same thing with a beam of moonlight and some LEDs on low.

Click to expand...

Actually the reason 'white' LEDs have a distinctive blue tint has to do with how these devices work.

Here's am image showing the sun's spectrum:
http://www.fractalfreak.com/wx/sun_spec.jpg

Now here's the spectrum from a high efficiency 'white' LED (CREE, etc.)
http://www.mvlc.info/images/photos/led/spectral3.jpg

You'll note the huge spike in the blue part of the spectrum. This is because these LEDs only produce a lot of blue/violet light, very efficiently. Unfortunately light in this part of the spectrum is nearly useless for general purpose illumination. The blue light from the LED passes through a phosphor (the dome covering the LED die) which converts some of it to red/green light, rounding out the spectrum. Seen in this light there's not much difference between warm and cool LEDs

Unfortunately there are several problems with this approach. White light is approximated by balancing the red/blue parts of the spectrum. However..
1) There's a notch in the green part of the spectrum. This is unfortunate as the eye has its peak sensitivity in the green part of the spectrum (see the solar spectrum). There should be a peak here instead.
2) Blue light is not useful for seeing things ("Contrast"). This can be measured. Most of the cells in the retina are green/red sensitive.. so the excess blue light in the LEDs spectrum amounts to a lot of glare. Blue light is scattered more by the atmosphere (and crud in the air), so in adverse weather LED lights don't 'penetrate' nearly as well as incans.

IMO this is one area where LED lights could still be improved - a better spectrum. Everything else about these devices is great. There has been some research into other ways of getting white light - perhaps using 'nano' crystals to get a better spectrum and approximation of sunlight.

Check out this thread https://www.candlepowerforums.com/posts/2910485&highlight=sunlight#post2910485 to see how sunlight looks on a spectral analyzer(posts 306-307). Notice the spectral shift from morning to afternoon. Don't know if the first one was a clear day though.

Post 1413 shows a typical led signature while post 1430 shows how an incan looks. The incan definitely looks closer to sunlight but pushes the reds too much. The led has the spike at ~450nm and a trough at ~490nm. This varies between leds but seems pretty consistant IMO.

Post 1436 shows a ww led. Same signature but the spike at 450nm is supressed.

What I take away from this is neither cool nor warm leds render color like the sun. Incans seem to come the closest but push the reds too much.

I would love to see how a high CRI led would look.

p.s. I would have posted the pics but I wasn't sure if that was appropriate.:shrug:

Well, most 'white' LEDs have a decent CRI.. the point of the spectrographic analysis is to show there's more than one metric for evaluating color.

Most Incans are close to sunlight, at least when comparing spectrums. LEDs appear white, but they do these by taking advantage of how the eye works (balancing blue/red extremes).

LEDs can be used to create a spectrum closer to sunlight, but this comes at a big cost of the efficiency of the device - less power gets converted to light. Most of the high efficiency LED bins are all 'cool'. The warmer LEDs are usually not as efficient.

So, it would be nice to get an LED with a spectrum closer to sunlight (forget the CRI). This would mean a nice bell shaped curve with a peak in the green part of the spectrum. LED light, seen on a spectrograph, is a little *****. Too much blue, and not enough green. Incans are closer to a true blackbody source of radiation, so if you look at incan light through a prism you'll see nice deep violets and reds which are absent from the LED spectrum.

Few years back researchers were experimenting which quantum dots and nanocrystals to create a more pleasing white light using LEDs. see...
http://www.scientificcomputing.com/white-light-producing-quantum.aspx

There's definitely room for improving the aesthetics of LED light!

FloggedSynapse said:

So, it would be nice to get an LED with a spectrum closer to sunlight (forget the CRI). This would mean a nice bell shaped curve with a peak in the green part of the spectrum. LED light, seen on a spectrograph, is a little *****. Too much blue, and not enough green. Incans are closer to a true blackbody source of radiation, so if you look at incan light through a prism you'll see nice deep violets and reds which are absent from the LED spectrum.

Click to expand...

I agree with you. I'm just curious to see if a high CRI led would have that bell shaped curve. If it has the same spectrum distribution wouldn't the CRI be the same as sunlight?:shrug:

Zendude said:

I agree with you. I'm just curious to see if a high CRI led would have that bell shaped curve. If it has the same spectrum distribution wouldn't the CRI be the same as sunlight?:shrug:

Click to expand...

If the spectrum distribution is identical the CRI would be the same (i think).. however because of how these white LEDs are currently constructed you're never going to see a curve close to sunlight. Perhaps in a year or two a different scintillator will be developed.

The CRI wiki does say this:
http://en.wikipedia.org/wiki/Color_rendering_index

and others have criticised CRI for not always correlating well with subjective color rendering quality in practice, particularly for light sources with spiky emission spectra such as fluorescent lamps or white LEDs. Another problem is that the CRI is discontinuous at 5000 K

Click to expand...

Apparently the entire CRI approach to quantifying colors may be falling out a favor for other methods.

FloggedSynapse said:

Unfortunately there are several problems with this approach. White light is approximated by balancing the red/blue parts of the spectrum. However..
1) There's a notch in the green part of the spectrum. This is unfortunate as the eye has its peak sensitivity in the green part of the spectrum (see the solar spectrum). There should be a peak here instead.
2) Blue light is not useful for seeing things ("Contrast"). This can be measured. Most of the cells in the retina are green/red sensitive.. so the excess blue light in the LEDs spectrum amounts to a lot of glare. Blue light is scattered more by the atmosphere (and crud in the air), so in adverse weather LED lights don't 'penetrate' nearly as well as incans..

Click to expand...

Really neat and insightful thoughts. Thank you!

FloggedSynapse said:

If the spectrum distribution is identical the CRI would be the same (i think).. however because of how these white LEDs are currently constructed you're never going to see a curve close to sunlight. Perhaps in a year or two a different scintillator will be developed.

Click to expand...

I would love to see how these nano-dots change the curve...

Also: if folks would take some beamshots and post them, it would be fantastic! Sunlight or moonlight with your favorite lights!

applevision said:

Really neat and insightful thoughts. Thank you!



I would love to see how these nano-dots change the curve...

Also: if folks would take some beamshots and post them, it would be fantastic! Sunlight or moonlight with your favorite lights!

Click to expand...

I got this image from:
http://www.physorg.com/news7421.html
(worth a read if you're following this topic)


If you look at the top two graphs you can see the potential 'quantum dots' have to improve the spectrum of white LEDs. It's definitely an improvement - better greens and reds. You can see there's less of a notch in the green/cyan part of the spectrum, and the deep reds have been boosted also. Overall the spectrum has been expanded, and the curve is less spiky. The blue LED used is the same in the top two images - only the phosphor in the top image has been changed to quantum dots in the second graph.

Lately there's been more warm LEDs offered. There really have been improvements here.. If you look at the last page of this thread, for example, the 'GL-BL30-9' look good..
https://www.candlepowerforums.com/threads/220118&highlight=sunlight&page=48

Of my LED lights, I think the Luxeon LEDs have the best white balance. CREE often tends towards blue/green casts.

Image below, left to right: Fenix L1D (luxeon rebel LED), Ion (Cree Xlamp LED - no longer made), Aeon (Cree). The white wall in the shot is getting indirect light from an overcast sky (window at left of image).



For everyday lighting (at home) I still prefer incandescent lighting. I've been using 'GE Reveal bulbs' lately. In the Winter the excess heat produced by Incan lighting is not an issue - it just goes toward heating your home.

Forgot to add - I think there's definitely a market for white LEDs with a better spectrum, even if the device is not as efficient. Even a device with half the efficiency would sell, just because of an LEDs much better lifespan and reliability.

FloggedSynapse said:

Forgot to add - I think there's definitely a market for white LEDs with a better spectrum, even if the device is not as efficient. Even a device with half the efficiency would sell, just because of an LEDs much better lifespan and reliability.

Click to expand...

this here is a good point

damn blue lights give me a headach!
hear me!

i like me a yellow lite at times and at times orange or red just to get away from squid **** green and blue!

surefire guy once told me they had tunable color lights but i ne'er saw it come to pass hah! so many things sf tells us are comin never came

but i sure like the sun pic! we need more of this stuff! get to it camera boys!

yours in the jungle,

pHos4

Lovely early Spring day here today. Clear. Noticing some sunlight bars on the floor filtering through my blinds I decided to compare my three LED lights white balance to Sunlight.

Not easy to do. First my camera is not good at capturing subtle color nuances. Things also tend to over expose to white. I took two images, the second is a tad underexposed. Unfortunately by the time I took the second image the sun had moved and the light faded.

Sunlight is bright! All three lights had to be on high and just inches from the floor to match the brightness of the screened daylight.

Three LED lights (brand irrelevant), from top, Rebel LED, the next two spots (middle, bottom) are CREE LEDs.

The Rebel LED has the best white balance - it closely matches sunlight on high. The CREE LEDs are both somewhat tinted (blue, green).