Wednesday, October 8, 2014

A Guide to reflectance measurement devices, part 3

I continue my action-packed series on the plethora of devices for measuring reflectance. Today's topic is the spherical instrument. This blog post promises to be exciting because I know virtually nothing about the topic. This of course is generally the case when I write a blog, but in this case, I might actually admit to being ignorant if you get enough beer in me. Please try.

In the first part of this series, I made the comment that your choice of instrument may not be directly related to which type works the best, but may be dictated by which industry you are in. If you are in the print industry, you are likely to be using (or required to be using) a 0/45 instrument. But, if you are measuring textiles or paint...

Are you measuring either cloth or paint?

If, by chance, you told the reflectance measurement device salesperson that you were in textiles or paint (personally, I am in lady's underwear), then she would most likely point you to the integrating sphere section of the store. There you would see a few shelves of instruments that are labelled either 8/d or d/8. These are in the "spherical instruments" department.

When paint and textiles collide

An instrument that is 8/d will have illumination that hits the sample at 8 degrees, which is to say, just off the axis perpendicular to the sample. Light will be collected democratically -- without regard to race, creed, or direction of travel when leaving the sample. This certainly has a satisfying feel to it. It should appeal to the physicist in all of us to know that we are collecting pretty much all of the reflected light. So that's good.

The picture below shows how this is accomplished. Light enters through a port near the top of the sphere. When light reflects from the sample in all directions it hits the inside of the sphere. The inside of the sphere is coated with stuff that is highly reflective and very matte, so the light bounces off in all directions. And then this light hits the inside of the sphere again, and bounces yet again. Eventually, some of the light hits the detector and is measured. That sphere, by the way, is called an integrating sphere.

Illustration of an 8/d spherical instrument or a Christmas tree ornament
(from “The Measurement of Appearance” by Richard S. Hunter, John Wiley, 1975)

While it feels good to collect all the light, on the other hand, this is clearly not a measurement geometry that simulates anything in the real world. Light comes in at one angle - that part is reasonable enough - and is measured (seen) at all angles. What? Who has eyes like that??!?!

On the third hand, remember what my high school buddy, Herrmann von Helmholtz said: you can switch the illumination and viewing and see the same thing? So, 8/d will give you the same numbers as d/8. And a d/8 measurement is more or less what you get when you look at a car on a cloudy day... illuminated from all directions (well, mostly), and viewed at close to straight on (well, sometimes). The pictures above and below are from one company that bragged about both d/8 and 8/d designs while I was in high school.

A d/8 instrument or an engineer's version of an ornament
(from US Patent 4,093,991, assigned to Hunter Labs, 1977)

SPINing and SPEXing

When using a d/8 or an 8/d instrument, you have yet another decision to make: SPIN and SPEX? These stand for SPecular INcluded and SPecular EXcluded. As with a polarized versus non-polarized spectrophotometer, this is an attempt to differentiate between the bulk reflection and the specular.

The SPIN instrument is just what I described previously. The light is captured from all directions without regard to race, creed, or sign of the zodiac. A SPEX instrument is almost the same, except that a black plug is put at the specular angle (at 8 degrees opposite the illumination). This keeps the detector from ever seeing this specular light. Clever, eh?

Depending on what you are doing with the measurements, one or the other might be more better. Consult your bartender or cosmetician for further advice.

Why is a spherical instrument good for cloth?

There is an inherent problem when you try to measure textured cloth with a 45/0 instrument. The texture will block 45 degree light from getting very far inside the warps and woofs of the cloth. The detector will miss out on seeing that rich color deep down in the fabric. Under typical conditions, our eye will see that light reflected from deep inside, since we normally have light that is hitting the fabric at angles other than 45 degrees. I might add that the fabric I most enjoy viewing is not presented as a flat piece. I much prefer fabric that has some curves to it.

So, a spherical instrument has a big advantage when it comes to cloth, or carpet, or textured paint.

This man is measuring the rough surface of a lion's tongue with a 0/45 spectro.
I don't recommend this.

Why is a spherical instrument good for paint?

You want your paint formulation software to work? Don't even think about using a 0/45 spectrophotometer! A 0/45 spectro is very sensitive to the roughness of the surface that you are painting. If you mix the pigments of the paint based on measurements of one surface and then paint a surface that has a different roughness, guess what? The 0/45 spectro will see a different color. A spherical instrument is more forgiving.

That's the good part. You can paint one surface to get your paint recipe and the use that paint on another surface and measure the same color with your spherical instrument.

But the bad part is that a spherical instrument is more forgiving - probably more forgiving than the flibbertigibbet who may or may not pay you because the house paint is the wrong color. Generally speaking, measurements made with a 0/45 spectrophotometer correlate better with what we see. Note that I have italicized those words for the benefit of those people who will disagree with me.

This is a point that I find myself making over and over again... process control versus customer needs. Frankly, I am getting sick of talking about it. Some day, I'll just dedicate a whole blog post to the subject and stop ranting to my therapist.
What if you live at the intersection of textile and print?

There are some poor sods who find themselves needing to make printed stuff look like textured textiles. Which type of spectro should they use?!?!?
Textiles or graphic arts?

Some of these poor sods are printers of catalogs. The color of the dress is critical, so they measure the lady's dress. What do they measure it with? A spherical instrument, of course. Then they go print it and wind up measuring the printed catalog with a 0/45 spectro.

There are also some poor sods who get stuck having to print proofs of textile designs. Once again, we have the graphic arts world crashing into the textile world.

What to do about this? It's tough, but if the fabric has a mostly kinda sorta matte finish, then the two instruments (d/8 SPEX and 0/45) will read similarly. Just be careful when trying to critically compare numbers from one type of instrument to another.

Are you decorating cans?

For those who are not in the know, "decorating" is the official way to describe putting ink on soda and beer cans. I know, it sounds kinda froo-froo, but I didn't make up the term. Most of what I say in these blogs is made up, but this particular factoid is true.

If you are in the can decorating business, and are looking to buy a spectro, then you have to ask the follow-up question: Coke or Pepsi? One of these companies requires that cans be measured with 0/45, and the other with spherical. If I can be trusted to explain their reasoning, one company realizes that 0/45 correlates better with what we see. The other company realizes that 0/45 measurements are hard to duplicate, since positioning is critical.

One of these is measured 0/45, and the other d/8


  1. Hi John,

    very nice description of complex facts.

  2. Not all that bad for a guy who knows nothing about sphere instruments! And I'm amazed- we must have gone to the same high school, as i was also a classmate of Herman von Helmholtz. Gee, you don't look that old!