Mixed Lighting Illustrated . .

[xpatUSA]

. . in the "unreal world" of product photography where you can have any lighting you want :-)

I'm in the habit of using LED lamps, an overhead fluorescent and a MiniMag (for highlights) - all at the same time (gasp!). Now, I'm the proud owner of a Sigma Flash thingy which fits nicely on top of the SD10. Yet another color temperature beaming down on my subjects. So, I set up this rig to see if that could work.



The backside of a gray card at left to reflect any stray flash. The flash aimed at the wall and some tracing paper at right for a bit of flash diffusion. Do not laugh at my studio . . . nor at my hand-held picture quality, we're not testing for acutance today.

The plan was to take shots under several kinds of lighting and then one with them all. In PhotoShop, to do the white balance and then compare the Red, Green and Blue color hues of each shot and also to list the theoretical (D65) hues of the Card (don't leave home without it).

So . . .

Overhead lamp:



LED:



Flash:



All together:



The rather unsurprising results, in degrees of hue according to Adobe PSE6:

Lamp, Red, Green, Blue
tube, 17, 100, 217
LED, 21, 99, 238
flash, 344, 99, 220
all, 345, 97, 224

card, 357, 111, 237 (theoretical under D65 lighting).

From which we learn that colors don't look too bad under varied lighting but we also learn that white balancing is no fix for imperfect lamp spectral emissions. Witness the brown red under LED lighting as opposed to a much redder red with the flash, and so on and so forth . . .

And, for my purposes, a bit of a flash is a very good thing!

[Manfred M]

The missing red component in the light is really well demonstrated in your shots Ted. I'm a bit surprised that the green isn't higher in the in the tube shot, like the blue component is in the LED shot.

A nice demo of where theory and reaility meet!

[xpatUSA]

The missing red component in the light is really well demonstrated in your shots Ted. I'm a bit surprised that the green isn't higher in the in the tube shot, like the blue component is in the LED shot.

A nice demo of where theory and reaility meet!

Thanks Manfred,

The tube is fairly expensive Philips TL950, 5000K and a claimed 98% CRI. The spectrum is better than most but far from perfect. There's a peak at 530nm but it's quite narrow which lessens the effect on broader-band reflections.



Here's my LED lamp:

[AB26]

Ted a dumb question. Have you measured the intensity of each light source on the subject? Or does that not matter?

[xpatUSA]

Ted a dumb question. Have you measured the intensity of each light source on the subject? Or does that not matter?

Not a dumb question at all! I'll try to explain why it didn't matter for the purpose of the test:

I was interested in the differences in color, by which I meant hue, as shown by a color wheel (an RGB color wheel, not the kind used for mixing paint). Here's an RGB representation:



The hue is often quantified as an angle going clockwise around the above figure. Thus pure red is 0 degs (or 360) and orange is 22-1/2 degs, green 120 and blue 240. The other two components of the HSV or HSB color representation have to do with the "shade" and brightness of the color, but I was only interested in the hue - meaning "how red is the red?". As you can see in the above diagram, the hue under the LED lamp could be better described as "orange"!

Also note that the Macbeth card primary colors are not exactly the above 0, 120, and 240 degs, even under perfect (D65) lighting!

A measurement of the "intensity of each light source", taken literally i.e. candelas per square meter at all surfaces of each lamp, would be quite difficult to do. And it would not account for light paths, diffusers, reflections, etc. Having got that out of my system, I hope no confusion is caused thereby.

[xpatUSA]

From which we learn that colors don't look too bad under varied lighting but we also learn that white balancing is no fix for imperfect lamp spectral emissions. Witness the brown red under LED lighting as opposed to a much redder red with the flash, and so on and so forth . . .

OT, but I'll mention here that there is fix for those folks who are bothered by this kind of color inaccuracy. It takes time and patience. I have lots of time but very little patience!

1) Download Adobe's RAW to DNG Converter
2) Download Adobe's DNG Profile Editor
3) Shoot a RAW pic of your handy-dandy Macbeth card
4) Convert the pic to a DNG
5) Open the DNG in the Profile Editor
6) In the Editor, you can go to a page, click each corner patch and tell it to make a Camera Profile (automaticamente!)
7) Adjust, if needed, in the Editor home page and save the new Profile.
8) Now you can re-open your original RAW or the DNG in ACR and select your super-accurate profile, don't know about CS or LR though.

For it to work with a RAW file, the profile has to be the proper ACR folder for the camera - which can vary, I'm sure. And it doesn't work with Sigma X3F files, only the DNG you make from one. Ho hum. Good old Sigma, they do like to make it hard on us dedicated Foveonistas . . .

Now you know why orangey red's look a lot better to me these days ;-)