Color rendition beamshots: incan, GD, Rebel, Nichia 083, Seoul high CRI, and Cree 5A

[mudman cj]

Re: Color rendition beamshots: incan, GD, Rebel, Nichia 083 (high CRI), Seoul high CR

Huh, I had no idea that someone else was using an M&M character for color rendition. :shrug: They are little chocolate candies with a thin candy shell that provides a little texture, color, and prevents the chocolate from melting by the heat of your hand. It is a widely recognized product here to say the least. I chose it because I was looking for reds and greens. But I have noticed that LEDs can sometimes render certain shades of green and blue poorly while rendering others well due to 'gaps' in the spectral output. I think when I get the 5A Cree in hand I may modify the scene to include more types of green and also browns to begin to convey the differences between these light sources with respect to rendering objects typically found in nature.

 

[Cemoi]

Re: Color rendition beamshots: incan, GD, Rebel, Nichia 083 (high CRI), Seoul high CR

I just ordered a Dereelight drop in that supposedly contains a Cree 5A

Thanks and sorry for asking, but I didn't know what a "Dereelight drop in" meant in your former post. So I searched CPF, and I suppose you are talking about the last item on this page?
But $35 is too much since I only need the PCB-mounted emitter.

 

[mudman cj]

Re: Color rendition beamshots: incan, GD, Rebel, Nichia 083 (high CRI), Seoul high CR

Yes, that is what I ordered. I don't know of another source for them right now. There were 20 of them that were sold in the marketplace, but they are all gone. You could keep your eye open or try to PM some of those individuals that bought more than 1 and may not have used them all yet. You could also PM the person that ordered the lot of them for more info.

 

[tsl]

Re: Color rendition beamshots: incan, GD, Rebel, Nichia 083 (high CRI), Seoul high CR

As I said above, I just ordered a Dereelight drop in that supposedly contains a Cree 5A with Q3 brightness bin. I say supposedly because until I saw that the brightest bin I had heard of in 5A was a Q2. But either way I think I will use it.

Have you received the Q3 5A yet? I'm interested to see how it compares to the others.

 

[mudman cj]

Re: Color rendition beamshots: incan, GD, Rebel, Nichia 083 (high CRI), Seoul high CR

No, but it shouldn't be long now. In addition to the Dereelight drop in I also ordered a Fenix TK20, and unlike the Dereelight I know the TK20 has already shipped.

 

[mudman cj]

OK, I got in my Fenix TK20 with Cree 5A emitter today, and took another set of beamshots. This time I added a few more objects and tried harder to get even exposures. If anyone would like to make these into a slideshow, be my guest.

As before, I will start with incan:

And now the Seoul high CRI (4000K CCT):

This one is the Cree 5A (and since the TK20 has a tighter beam pattern, I had to place the light further away to illuminate more of the scene):

And now for the Nichia 083 high CRI LED (this one was closer because the light is more floody):

On to the Rebel 100 (again closer because of the more floody beam):

And finally the Golden Dragon in my LS20:

After playing with the TK20 for a bit I noticed some differences in color rendering compared to reference light sources, but as others have said, the 5A bin Crees are quite a nice compromise in brightness for color rendition compared to other LEDs. IMO this makes it a great choice for outdoors lights. I find that indoors it has a somewhat yellow/green characteristic that can be seen a little on a white wall but more so on skin. The 5A also lacks a bit in red and orange which makes orange objects appear more yellow. By comparison, the Seoul has bountiful orange and red which brings out the pinks in skin much more than the TK20, but my particular Seoul has a yellow/rose tint to it. The Seoul also tends to yellow things up a bit, while the Nichia 083 gives a higher color temperature version of high CRI that is quite tint neutral and just great at close range.

 

[jirik_cz]

Nice beamshots :twothumbs

 

[LukeA]

These are nice beamshots. They let me know that anyone who claims that LEDs can't render colors either A) has never seen a WW LED or B) is seeing what they want to see.


I hope you don't mind that I made an animated gif with the Cree 5A and incan shot.

 

[mudman cj]

Not at all, though it would be even better if I could have managed to support the TK20 on the other side of the room for that shot since the change in camera angle is apparent: I moved the camera location to avoid hard shadows.

I would like to see the incan in an animated gif with the Seoul. Those would be the closest comparison shots IMO, and the Seoul was closest to incan in real life too (not that incan should be considered ideal mind you). This gif would help prove the point that LEDs have finally learned a new trick.

 

[LukeA]

High CRI SSC and incan:

I wish GIF was updated to support more colors. It isn't 1987 anymore.

 

[Outdoors Fanatic]

Great thread!

Big thanks to the OP.

 

[OrlandoLights]

Thanks for doing this thread, Mudman, it's just what I needed in finding a light that rendered colors realistically. Or as I think of it in my non-scientific way, a beautiful beam.

 

[glyphin]

Excellent comparison of CRI in the second series, Mudman - thanks! The Nichia is clearly the best at rendering color. I wonder about an array of these LEDs...

 

[kongfuchicken]

Thanks a ton for doing this mudman. It's still pretty new to me how well these high cri leds seem to work. I'm definitively going to look into this.

The only thing I noticed in your comparison is that the Black&Decker toolbox might possibly be made of fluorescent plastic material which means it'll probably shine bright orange even if it were illuminated by a monocromatic blue light... Not that it seems to throw off the effectiveness of the test at all.

 

[mudman cj]

Thanks a ton for doing this mudman. It's still pretty new to me how well these high cri leds seem to work. I'm definitively going to look into this.

The only thing I noticed in your comparison is that the Black&Decker toolbox might possibly be made of fluorescent plastic material which means it'll probably shine bright orange even if it were illuminated by a monocromatic blue light... Not that it seems to throw off the effectiveness of the test at all.

Well, I don't have a monochromatic blue light, and I don't want to enter into a debate on the subject, so I guess I will just have to take some more beamshots.

This time I tried to focus on different orange objects, and I realized that there would be some added benefit to using objects that are somewhat standard; so that anyone could obtain a similar object for reference. This is why I used some more fruit in this one. I wish I had an eggplant, but just grabbed a few purple items to show how purple can look quite different. The cardboard in the background is another example of how browns look different as well. Oh, and this time I added my L4's Lux V to the mix.

Incan:

Seoul high CRI 4000K CCT:

Cree 5A:

Nichia 083 high CRI:

Rebel 100:

Golden Dragon:

Lux V:

 

[kongfuchicken]

Mudman, I honestly have no clue whether the box is fluorescent or not but my laundry detergent bottle, the tag on my bag, the handles of my scissors and the traffic cone in the trunk of the car are and I never noticed until i played with a blue laser with the lights off =P

I certainly didn't intend my observation to be of any criticism to your test (which is eye-opening to me) and now I feel guilty for troubling you into reloading the whole thing... sorry and thanks for doing this!


In any case, the washed out colors effect is more than obvious to me in every shot past the rebel100 one. I'm stunned that the effect can be recreated on a picture displayed on an lcd screen when I thought it were something with my eyes being below average or something the likes...

Well I'm certainly impressed and I look forward to the many weeks of tinkering and rebuilding all my lights with these new leds.

 

[kaichu dento]

Nice color shots Mudman! Every time I see new comparison shots showing the Q3 5A up against anything else the more I want them in every light I own! Can someone put one in my Titan for me?!? If so, please PM me! :wave:

 

[mudman cj]

kongfuchicken - I didn't take your comments as a criticism, so please don't feel guilty. I thought that you made a valid point that I simply was not equipped to address, so rather than wonder if the effect you mentioned was influencing the results or not, I thought I would take it as an opportunity to further demonstrate differences between these light sources.

I wish there were a good way to demonstrate the relative ease with which these different light sources allow one to hike through the woods for example. There are definite differences in depth perception and situational awareness for lack of a better term between these light sources IMO. And these differences are really what matters most to me, not accurate color rendition.

As others have said, sometimes more light output can make up for these other differences, so you might be disappointed by changing the LED in certain lights to one of these alternatives with lower output. Just something to keep in mind. I would say that the Cree 5A is largely a no-brainer because the output is only 15% less than cool white LEDs, but when you consider using the Seoul high CRI LED keep in mind that the output is on par with a Lux III. And the Nichia 083 has a similar Lux III effieciency but with lower power handling its more like a Lux I, and you can't expect the Nichia 083 to give hardly any throw. I use one with a Q3 reflector for a nice beam that throws to maybe 50 feet and thats it. And when I use throw in this context I mean I can still identify objects at that distance: this does not necessarily conform to any accepted definition of throw.

 

[Opto-King]

Different Kelvin?

For me it seams that all LEDs used in the "CRI experiment" does NOT have the same CCT (Kelvin & X,Y) is this correct?

Since this will also affect the resoult a lot as well as the amount of lumen pushed out of the LEDs to the objects.

 

[mudman cj]

I know where you are going with that. CRI is only defined for light sources having the same CCT. The light sources used in these beamshots did not all have the same CCT. But I am not trying to measure CRI here - this was not intended to be nor labeled as a "CRI experiment". And yes, the light intensity is also important when comparing things of this nature.

I have made an attempt to adjust the position of the lights to achieve equal intensities. Did I succeed? No. I realize that the intensities as well as beam shapes are not equal.

What I have tried to show is roughly how these different light sources render objects. And by that I mean how they appear to my eyes. The beamshots do not perfectly show what I see when I use these lights, but they give a pretty good idea of the difference. That is all. I never claimed that these beamshots show or prove that some light sources have a higher CRI. I could care less about the CRI value in some sense, because CRI is defined off of a black body curve, and black body radiation is not the ultimate standard IMO. I am after a light source that renders objects similarly to sunlight. And the spectrum of light reaching our eyes from the sun is not black body radiation.