High CRI lighting

[Eildosa]

Hi,

Some time ago I was making a 3D scanner and people told me that I should use high CRI
lighting in order to get nice texture (skin) colors.

As you can see high CRI led have the full color spectum :
http://www.yujiintl.com/img/graphics/delta-comp960.jpg

And pictures feels "richer" :
http://www.hisemicon.com/uploadfile/...4143110160.jpg

Since camera RGB sensor are similar to our eyes (it's not just RGB, it captures a little bit of what is around the main wavelength) : https://goo.gl/kd85a1

The theory make sense.

HOWEVER!

My computer/camera screen used to show me the image is using monochromatic RGB subpixels!
How come I could see a "richer" picture?

Thanks.

PS : yes taking a picture with and without high cri lighting would have awsered my awnser, but I never had the money to buy high CRI leds.

[Manfred M]

Nice words, but no data to help anyone analyze what is going on here isn't going to get you anything more than educated guesses. The metadata on the file would be useful as a start,

If I were to make an informed guess, the wrong colour space was likely assigned somewhere in your workflow. It could be other things as well, but if this is happening in-camera, check you colour settings. Any raw data that is displayed on your camera is converted to a jpeg for display on the screen,

[Eildosa]

Nice words, but no data to help anyone analyze what is going on here isn't going to get you anything more than educated guesses. The metadata on the file would be useful as a start,

If I were to make an informed guess, the wrong colour space was likely assigned somewhere in your workflow. It could be other things as well, but if this is happening in-camera, check you colour settings. Any raw data that is displayed on your camera is converted to a jpeg for display on the screen,

Well I was hoping someone on this forum would have already tried and could show me a picture with and without high CRI lighting.

[Inkanyezi]

For rather obvious reasons, I have not made such comparison images, but it is a bit more complicated, as CRI is not really consistent over the entire visible range for different light sources. The CRI for sources of 5000 K and above is "standard daylight", but under 5000 K it is black radiator (incandescent). Those spectra are very different, and the below 5000 K colour temperatures have very little of the blue to violet tones and are also deficient in green, whereas red is over-represented compared to the daylight spectrum.

So for best colour rendition as with daylight, no incandescent or low colour temperature lamp should be used, but only 5000 K or above, with a CRI higher than 90.
My favourite lamp is fluorescent colour 954, but others prefer 960 or 965. The first digit is CRI as a single digit (9 > CRI 90 and 8 > CRI 80).
The two digits after is colour temperature in K/100 - 954 has a colour temperature of 5400 K.

There are always slight differences between different sources, even of the same manufacturer if different batch, and there is colour drift with time for all fluorescent sources, including LED. (All modern white LED sources are fluorescent.) With more than one source, always use lamps from same batch and of same age.

I think Philips or Osram might have comparing images on their websites for different CRI lamps.

[xpatUSA]

Well I was hoping someone on this forum would have already tried and could show me a picture with and without high CRI lighting.

To the original poster:

First let me respond by saying that lamp manufacturers can adjust the peak emissions from their lamps to get a high CRI "score" while still having peaks and gaps in the lamp's spectral emission. For example, I have an overhead Philips fluorescent that claims a CRI of 98 !! Pretty good, eh? But here's it's spectral emission graph:



Source: http://kronometric.org/phot/lighting...8-Brochure.pdf

Just look at those peaks - so right for perfect skin tones .

OK, the graph is radiometric, not photometric i.e. not lumens, but still . . .

So much for the value to us of "CRI" . . . (there's a later measure called "Ra" but, for some reason, manufacturers are sticking with CRI, I can't imagine why ).

https://en.wikipedia.org/wiki/Color_...ex#Test_method

I could take pictures for you of a Macbeth card 1) under that light and 2) under a pair of 3500K LED floodlights but I'm not sure what that would tell you, to be honest.