I have this Photography light, It is a battery powered 120 LED array, And the specifications on it do not say the CRI value of it, Which is one of the most important things with light.
I was wondering of a way to measure the CRI value of it. The most I could think of is doing the CD spectrometer trick.

Measure the exact spectrum and then calculate.

Hypothetically the first you may do using a CD as a diffraction grating and a camera (with just one of the LED's ON or from a greater distance), that is the easy part. What makes it difficult (and why I wrote the "hypothetically") is how to calculate the spectrum from the image you get. Most likely you will need two images with the measured LED ON and OFF to cancel out the other light sources (subtracting these two will give you just the measured light source; assume the camera setting remains the same), but then deembeding the pictrure is the part my math is too short to solve (the light source is of a nonzero size and nonuniform in intensity, the grating only "smears" it according to the spectra)...
Well, I write this not because I want you to do that, but because someone may find it a good inspiration for a problem to solve (as a SW project)...

For the second part the equations should be in the literature, it is included in a W bundle with practically all color measurement instruments...

Sounds like a plan, But you can only dim the whole array, not individual LEDs.
Could you measure the CD in a very dark room to not have to calculate that difference?
Now I just have to find an instruction manual.
Edit: I found some instruction manuals, But they only told you how to use the meter, Not calculate CRI.
These meters also measure CRI, And it's even labeled as "entry-level", But at a price of $2500(60000 CZK), it's anything but "entry-level".

There is no relation between the CRI and light quality. A light source can have high CRI and yet harash quality light (LED) and can have low CRI and good light quality (Halophosphors fluorescent and deluxe white MV).

CRI is something I've never really "got". Triphosphor 827 has large gaps in its spectrum, and the Mercury HPL4 lamps with ~60% CRI have a typically sparse Mercury fluorescent spectrum. Yet halo phosphors have a broad spectrum which looks almost continuous (with peaks) and incandescent, while accepted to be complete, renders reds strongly and mutes blues by virtue of its CCT. So the whole thing has me a bit stumped!

It's because the traditional CRI value uses some specific pastel colors, And is considered obsolete(It's the same situation with QUERTY keyboards.).

There is no relation between the CRI and light quality. A light source can have high CRI and yet harash quality light (LED) and can have low CRI and good light quality (Halophosphors fluorescent and deluxe white MV).

I really love Halophoshosphors, Because of the "Retro" look to it.
I also think the trick to having good CRI and light quality is to have a good CCT, Brightness and spectrum.
(97+ CRI LED, I actually have never seen one in person, But the spectrum generally looks good.)
Maybe I will contact the company to directly see what the CRI is.

There is no relation between the CRI and light quality. A light source can have high CRI and yet harash quality light (LED) and can have low CRI and good light quality (Halophosphors fluorescent and deluxe white MV).

The relationship is very strong, but still it is indeed by far not equal.
Big part of that is, what the "quality" really technically means about the light. By definition, the the quality is the ability to perform in a given task. That means, than the meaning of "quality" is for each task different. And not always it is possible (or really known how) to translate into a set of exact technical requirements.
The CRI definition is one way how to translate it, but it matches only some uses - mainly the general illumination, with expected use range of 20 to about 70 (so the CRI ranking matches pretty well the ranking of the perceived light color quality). That means most of modern light sources are way above the upper limit, so outside the range the CRI was designed for. So no wonder, the CRI ranking does not match that well with the CRI value - because they all are in the "very good" range. So in other words the HPS has worse light color than fluorescents pretty matches with the HPS having lower CRI than the fluorescent. But with the fluorescent of CRI80 vs LED with CRI85, both are quality color light sources (compare to e.g. the HPS), but which one looks better does not correspond to the CRI value, comparing such high quality sources is not what the CRI system was designed for.


Pastel papers and calculating CRI:
The pastel papers are nothing else than a color filters, which reflect the different wavelengths with different ratio, so characterized by their reflective spectra. The spectra of each of the color paper is standardized, so it is well known.
Now when you want to match how well the color of some light source renders the given paper is, you illuminate the paper with the reference light (black body, same temperature; so spectra mathematically known) and compare the reflected brightness and color shift with the measured light.
In math that means calculating the reflected spectrum - at each wavelength multiply the intensity of the measured light source with the reflectivity of the test paper. On the result calculate the brightness (correlation with the eye sensitivity) and color coordinates (position on the "horseshoe" color diagram). Then do the same with the reference light. And then the deviations directly say, how well that color is rendered: No deviation means 100%, no resulting reflected light or loss of color from it means 0%
And then repeat for all of the defined "color papers" and calculate the average per the standard.

For letting only one LED: That is necessary to ease the deembedding of the final picture - because the pattern created by each LED would be shifted from each other, while the camera will see the mix of all together. It will be then very complicated (if even not possible; the spectra images are shifted and then overlaid) to deembed the spectra from the overall image. Having only spot of zero size means you get directly the spectrum, having the source brightness homogeneous or a kind of ball smears it, butt still when there is one peak, it is not that complex to extract it back. But when there is some repetitive pattern, it will be very hard to find an algorithm, which won't get confused by that...

Awesome post, But I just got some news!
The person that I contacted said that they do not have it measured right now, But next week they will have measured it.
(Translation: They are going to measure it next week.)
Also, How would you get a black body for 3500+K light if most things at that temperature would melt?

Also, How would you get a black body for 3500+K light if most things at that temperature would melt?

The "taking papers" and "black body" was meant take their spectra, I was describing what I think the software would have to do, all the manipulation was intended to be done just by the calculation on a computer.
And the blackbody spectra you may calculate without any measuring, it is basic physics...
All the spectra measurement I was describing would just need the user to take two pictures, one with the light ON, one with the measured light OFF (to have background calibration), maybe one of the light source (to have clue about it's exact pattern shape).
Then the rest (all the processing I was describing above) will be done by the software...
The result will be just any ordinary CD, digital camera and a computer will be sufficient to measure light spectrum...

The thing is, I may come up with some method how to mathematically do the most of things above, but I'm not able to turn that into a computer program...
So I just voiced the ideas I had here, maybe someone will be interested in it, pick the idea and turn it into a real, working program (standalone, Matlab, Excell, whatever...)...

Sounds good, But you would have to compensate for the direction the CD is facing, And that all cameras are different.
If we made this into a real thing, It could be marketed as "Affordable Spectroscopy".

Sounds good, But you would have to compensate for the direction the CD is facing

Of course. But the cameras do see colors, so even when they are not able to see the spectra, it is enough to restore the orientation vector (going radially, across the grooves in the CD), as well as provide the basic calibration (how long distance corresponds to given difference in wavelength,...).

And that all cameras are different.

Well, they are somewhat, but they can not be that much, because they can not distort the colors.
There are more severe problems: The ambiguity of the picture interpretation, mainly when the light source shape is not known and/or complex, limited resolution due to the same cause,...
It is supposed to be a "cheap, readily available" tool, not any accurate lab grade equipment.

If we made this into a real thing, It could be marketed as "Affordable Spectroscopy".

That is definitely the main idea, an affordable spectrum analysis for enthusiasts like the ones gathering here on LG...

Quote from: wattMaster on May 13, 2016, 07:05:20 AM

Sounds good, But you would have to compensate for the direction the CD is facing

Of course. But the cameras do see colors, so even when they are not able to see the spectra, it is enough to restore the orientation vector (going radially, across the grooves in the CD), as well as provide the basic calibration (how long distance corresponds to given difference in wavelength,...).

Quote from: wattMaster on May 13, 2016, 07:05:20 AM

And that all cameras are different.

Well, they are somewhat, but they can not be that much, because they can not distort the colors.
There are more severe problems: The ambiguity of the picture interpretation, mainly when the light source shape is not known and/or complex, limited resolution due to the same cause,...
It is supposed to be a "cheap, readily available" tool, not any accurate lab grade equipment.

Quote from: wattMaster on May 13, 2016, 07:05:20 AM

If we made this into a real thing, It could be marketed as "Affordable Spectroscopy".

That is definitely the main idea, an affordable spectrum analysis for enthusiasts like the ones gathering here on LG...

Except that the complications we described is likely the reason nobody has done this.
Maybe another setup is to just have a mount for the camera and CD, And you just supply those and have a more predictable pattern.

CRI values in photography are meaningless. I think you want to know the CCT or color temp so you can match it with your camera °K setting(?).

Light a scene w/your LED, start taking pictures w/camera at different °K settings. Use whatever is nice.

CRI values in photography are meaningless. I think you want to know the CCT or color temp so you can match it with your camera °K setting(?).

Light a scene w/your LED, start taking pictures w/camera at different °K settings. Use whatever is nice.

Well, CRI describes, how accurately are the colors of the real scene rendered. The colors we see depends on to what extend is which receptor stimulated, so indeed for the color perception it does not matter how exactlythe spectrum looks like.
The thing is, when you want to have certain color object, it may be so it reflects all longer than 650nm, or as well when it reflects one narrow gap around 700 and some around 600nm. The reflectivity will be so, both will have the same color when illuminated using continual spectrum (e.g. an incandescent light source), because the reflected light stimulates the eye in the same way.
But when your source has just one line at 650nm (and others at 500nm and shorter, so it still forms a white light), it will display the first object well (the wide spectrum with of the reflectivity will reflect the 650nm well), but the second will be dark (it just does not reflect the 650nm). Here it does not matter how the reflectivity spectrum would look like as light spectrum, it has to first match with the light source regardless what the eyes see. And that is, where the real spectrum becomes important and what the CRI is aimed for.

And if you are about to taking photos, the same applies, the only difference is it is not about sensitivity of eye receptors, but the camera sensors (or even the film).
And by the way these are quite different from the eye. The whole chain is calibrated so the resulting picture looks the same with common natural objects, but with some spectrum shape the calibration is quite off and so make it responding far different - a very typical example is the light from a MV lamp - normally it looks white, but camera pick it as greenish. The reason there is the fact, the eye spectral sensitivity shapes of the "red" and "green" receptors are different from the cameras, so while the MV does stimulate the "red" in a human eye quite strongly, it does not in the cameras. So in fact the "camera CRI" (so how well it renders colors on the picture taking equipment) actually differs from the "hu8man CRI" (how we see it). That is the main reason, why the special stadium lamps for quality color TV broadcast appear to have not so high CRI rating - they are just optimized for the cameras and not the human eye...

And if you are about to taking photos, the same applies, the only difference is it is not about sensitivity of eye receptors, but the camera sensors (or even the film).
And by the way these are quite different from the eye. The whole chain is calibrated so the resulting picture looks the same with common natural objects, but with some spectrum shape the calibration is quite off and so make it responding far different - a very typical example is the light from a MV lamp - normally it looks white, but camera pick it as greenish. The reason there is the fact, the eye spectral sensitivity shapes of the "red" and "green" receptors are different from the cameras, so while the MV does stimulate the "red" in a human eye quite strongly, it does not in the cameras. So in fact the "camera CRI" (so how well it renders colors on the picture taking equipment) actually differs from the "hu8man CRI" (how we see it). That is the main reason, why the special stadium lamps for quality color TV broadcast appear to have not so high CRI rating - they are just optimized for the cameras and not the human eye...

I would think that high-end cameras would be calibrated/designed to more accurately take pictures like the eye.
I think it comes down to the color filters cameras use, There are really only 2 common filters, Bayer and Fujifilm.
The Fujifilm filter boasts to have better colors, And more accurate rendering.

I would think that high-end cameras would be calibrated/designed to more accurately take pictures like the eye.
I think it comes down to the color filters cameras use, There are really only 2 common filters, Bayer and Fujifilm.
The Fujifilm filter boasts to have better colors, And more accurate rendering.

Well, that works if there are filters.
The thing is, the color borders with the camera filters tend to be rather sharp cutting, but in the eye they are way smoother and that is quite hard to emulate.
But mainly the new CMOS sensors do not use any explicit color filters at all (except to block the off-visible wavelengths). Their sensor elements use stacked junctions in different depth (Nplus/Pwell/DeepNWell/Psubstrate; before the shutter window begins, they are precharged to a reference voltage so all are in reverse, then when the shutter window starts, the precharge switches are switched OFF, so the light created current discharges the junction capacitances, after the exposure the voltages are measured via an ADC, then corrected for the capacitance nonlinearity), each sensing different parts of the spectrum (and their spectral sensitivity overlaps). Then the resulting signal is recalculated for the standardized RGB values.
The reason is to suffice with the most common CMOS manufacturing process without any added steps (all the layers are already present, originally designed for the basic NMOS and PMOS transistors). The other benefit is, such concept allows the complete pixel area to sense all color components, so one pixel is really a full color pixel, no smaller subpixels or so.
And here is not much room for physical adjustments, so the color recalculation matrix is just optimized for the most common scenes (that is, why there different scene settings even when the CCT and exposure settings are identical - it is to use a matrix better optimized for different light sources)