The color rendering index (CRI)

[Panama Hat & Camera]

I read somewhere that the light from incandescent bulbs reproduce better the colors than fluorescent lamps. I became a little confused because the light from incandescent bulbs are yellowish - reddish (color temperatures from 2500 K to 3500 K) and the light from daylight (5000K) fluorescent lamps seems to be more natural than light from incandescent bulbs.
I had already noticed that some RAW photos taken with light from some fluorescent lamps did not present natural colors, even correcting the white balance. The situation were far worst whith light from mercury vapour lamp and from low-pressure sodium lighting. I did a fast research about the qualities of various kinds of light and I discovered that the color rendition also depends of the spectra of the light source.
The sources with a continuous spectra (like the sun and the incandescent bulbs) reproduces better the colours than the sources with a "discontinuous" (lower level of relative power compared to daylight - except for a few peaks) spectra (e.g. mercury vapour lamp and low-pressure sodium lamp) and I learned about the color rendering index.
The color rendering index (CRI) is a quantitative measure of the ability of a light source to reveal the colors of various objects faithfully in comparison with an ideal or natural light source. Numerically, the highest possible CRI value is 100, and would only be given to a source identical to standardized daylight or a black body (incandescent lamps are effectively black bodies), dropping to negative values for some light sources.

Summarizing and concluding: an ideal light source for a correct color rendering should have both a color temperature similar to daylight (but this can be corrected by the white balance) and a high CRI value.


https://en.wikipedia.org/wiki/Color_rendering_index
http://light-measurement.com/spectra/
http://www.topbulb.com/color-rendering-index
http://lowel.tiffen.com/edu/color_te...mystified.html
http://www.fullspectrumsolutions.com/cri_explained.htm

[xpatUSA]

Summarizing and concluding: an ideal light source for a correct color rendering should have both a color temperature similar to daylight (but this can be corrected by the white balance) and a high CRI value.

Unfortunately, all is not gold that glisters:

For example, the Philips TL950 fluorescent tube claims 5000K and a CRI of 98! But here is the spectral output:



On the other hand, my LED floods are 3500K with a CRI of 85 and a smooth spectrum but they are quite easy to custom white balance effectively.

Less scrupulous manufacturers can adjust the phosphor's peaks to give a good CRI but not the best of lighting.

Not saying that Antonio's research was invalid, just saying that his rule-of-thumb does not fit all cases.

[Manfred M]

Antonio - with a very limited set of exceptions (expensive tubes designed for photography), fluorescent lights are usually regarded as the worst light sources for photography. The reason is shown in the graph that Ted posted. Fluorescent tube emit specific wavelengths of visible light when they are excited by UV. Different mixes of phosphors are responsible for the performance of individual models of tubes and they all tend to have spikes at specific wavelengths.

Sunlight and even tungsten lights exhibit curves that are typical of black body radiators.

[William W]

. . . the Philips TL950 fluorescent tube claims 5000K and a CRI of 98! . . .

Key word.

I have never bothered looking into it - is there an industry standard to test and allocate (and interrogate claimed) CRI ? or is it awarded to oneself by oneself?

WW

[xpatUSA]

Key word [claims].

I have never bothered looking into it - is there an industry standard to test and allocate (and interrogate claimed) CRI ? <>

WW

Here y'are, Bill, it's CIE Publication 13.3:

http://www.cie.co.at/index.php?i_ca_id=981

Basically the color accuracy of 8 color patches is compared to that from tungsten light (illuminant 'A') and some number up to 100 derived therefrom.

Hence my earlier speculation re: "tuning" phosphors to get a good score.

I understand that the metric has been improved but, for some reason , lamp manufacturers appear to be sticking to CRI like glue . . .

More here about LED lighting CRI:

https://en.wikipedia.org/wiki/High_CRI_LED_lighting
.

[DanK]

Hence my earlier speculation re: "tuning" phosphors to get a good score.

This is a specific instance of what is called "Campbell's Law" in the social sciences, named after Don Campbell, the psychologist who was one of the founders of the scientific study of program effectiveness. He wrote in 1975:

The more any quantitative social indicator is used for social decision making, the more subject it will be to corruption pressures and the more apt it will be to distort and corrupt the social processes it is intended to monitor.

In other words, people distort their behavior to maximize gains on the measure rather than improvements in the larger whole the measure is supposed to represent. In this case, manufacturers know what parts of the spectrum are sampled and can tune their bulbs to do well in those sampled ranges. The same phenomenon appears anywhere one looks: in health-care quality control, the Soviet manufacturing system, and in teaching to the test in schools--not to mention the Volkswagen emissions-control scandal, which was only one of several such cheating scandals over the years.

In addition--Ted, you may be able to either correct me or fill in details--I have been told that the CRI index undersamples the red portion of the spectrum, perhaps because the manufacturers of florescent bulbs had difficulty producing light in that range. There is another index (R_9? R_19?) that measures that range.

I have had reasonable luck adjusting WB for LEDs with CRIs down to 80, but I avoid florescents like the plague.

[xpatUSA]

In other words, people distort their behavior to maximize gains on the measure rather than improvements in the larger whole the measure is supposed to represent.

Not unlike some photographers (elsewhere, of course) who run their masterpiece through Topaz six times and then claim that their lens is the sharpest of it's kind on the planet!

In addition--Ted, you may be able to either correct me or fill in details--I have been told that the CRI index undersamples the red portion of the spectrum, perhaps because the manufacturers of [fluorescent] bulbs had difficulty producing light in that range. There is another index (R_9? R_19?) that measures that range.

Sorry, Dan - I'm no expert on CRI, having only had a passing interest long ago when a Sigma SD9 was turning red into brown whereas a Nikon D50 was not. Part of the problem was indeed the Sylvania Mini-Craft 5000K CFL but part was also the early Foveon F7 sensor.

I have had reasonable luck adjusting WB for LEDs with CRIs down to 80, but I avoid [fluorescents] like the plague.

For table-top work (mainly eBay stuff) I use two LED floods successfully - but I also leave the Philips "98 CRI" 17W 24-inch overhead tube on, out of sheer laziness. My lux-meter tells me that the floods are the dominant lighting by far, so it doesn't bother me at all.

[William W]

Thank you gentlemen.
You both are on a roll.
Making coffee now - preparing to read in detail.
Thanks again.

Bill

[Panama Hat & Camera]

Here y'are, Bill, it's CIE Publication 13.3:

http://www.cie.co.at/index.php?i_ca_id=981

Basically the color accuracy of 8 color patches is compared to that from tungsten light (illuminant 'A') and some number up to 100 derived therefrom.

Hence my earlier speculation re: "tuning" phosphors to get a good score.

I understand that the metric has been improved but, for some reason , lamp manufacturers appear to be sticking to CRI like glue . . .

More here about LED lighting CRI:

https://en.wikipedia.org/wiki/High_CRI_LED_lighting
.

Ted,
Thank you very much for your clarifications.
I take my photos mainly under natural light and I've never thought before about color rendition for artificial sources of light (my only concern about colors was the white balance).
I wrote this thread because I've never read (or I don't remenber of having read) before about that kind of quality of a light source (the ability to reproduce the colors of various objects faithfully in comparison with an ideal or natural light source) in my books of photography or in photography sites (including the CiC).
I'm always learning.
Antonio.

[Panama Hat & Camera]

This is a specific instance of what is called "Campbell's Law" in the social sciences, named after Don Campbell, the psychologist who was one of the founders of the scientific study of program effectiveness. He wrote in 1975:



In other words, people distort their behavior to maximize gains on the measure rather than improvements in the larger whole the measure is supposed to represent. In this case, manufacturers know what parts of the spectrum are sampled and can tune their bulbs to do well in those sampled ranges. The same phenomenon appears anywhere one looks: in health-care quality control, the Soviet manufacturing system, and in teaching to the test in schools--not to mention the Volkswagen emissions-control scandal, which was only one of several such cheating scandals over the years.

In addition--Ted, you may be able to either correct me or fill in details--I have been told that the CRI index undersamples the red portion of the spectrum, perhaps because the manufacturers of florescent bulbs had difficulty producing light in that range. There is another index (R_9? R_19?) that measures that range.

I have had reasonable luck adjusting WB for LEDs with CRIs down to 80, but I avoid florescents like the plague.

Dan,
I think that the Campbell's Law is part of the our world (unhappily).
Cheers,
Antonio.

[Panama Hat & Camera]

Thank you gentlemen.
You both are on a roll.
Making coffee now - preparing to read in detail.
Thanks again.

Bill

Bill,
Thank you for your vote in my thread (classifying it as a Helpful Post).
Antonio.

[Panama Hat & Camera]

Not unlike some photographers (elsewhere, of course) who run their masterpiece through Topaz six times and then claim that their lens is the sharpest of it's kind on the planet!

Ted,
What is this Topaz? A kind of sharpening software for photos?
Antonio

[xpatUSA]

Ted,
What is this Topaz? A kind of sharpening software for photos?
Antonio

It's a Company that makes photo-enhancement plug-ins and also stand-alone applications.

https://www.topazlabs.com/store

I used the name as a general name for that kind of product, in a similar way to "Google" or to "Hoover".

Never used it so I can not tell you anything about it, sorry.

[William W]

I used the name as a general name for that kind of product, in a similar way to "Google" or to "Hoover".

Hoover (google topaz). . . "hoover" is and the others . . . broadly / loosely are: an eponym

[William W]

. . . Thank you for your vote in my thread (classifying it as a Helpful Post).

In addition to learning much about CRI - I have had a lot of fun learning some off the mainstream Economic Theory . . . amazing what we read here.

WW

[george013]

I remembered some other discussion years ago so I knew where to look for. Look at next link, special to the images. https://photo.stackexchange.com/ques...ith-color-cast
Solution is a long shutter speed.

George