Saturday, March 16, 2013

Highlights and Specularity, Part 1

This photograph shows three spheres with varying surfaces. The one on the left is matte, the one in the middle is glossy, and the one on the right is highly reflective. 

On the matte sphere there is no highlight. On the glossy sphere in the middle, the highlight is clearly apparent. The mirror-like sphere on the right also has a highlight, which is really a reflection of the light source. The light source is the sun shining through a window.

The right sphere also reflects the scene around the ball, including the white paper background and the dark room behind the camera. In the right ball, you can even see a reflection of the middle ball, with a highlight in the middle of that reflection.

That same pattern of reflections of the paper, room, and neighboring ball is subtly visible in the middle ball as well.

In the middle ball, there is a second, smaller highlight just to the right of the primary highlight. This secondary highlight is the sunlight is reflected three times. The light bounces off the middle ball, bounces back off the right ball, and bounces again off that little highlight on middle ball back to your eye.

So we've arrived at a definition: A highlight is a specular reflection (Latin "speculum"=mirror) of the light source on a shiny surface. The shinier and smoother the surface, the brighter and clearer the highlight.

These three diagrams show what's happening at the surface level. On the matte surface, light arrives from the top left and hits the rough surface. Some light gets absorbed and the rest scatters away in all directions. This is what happens when light hits a matte surface like a sand dune or a sweater.

The glossy surface of the middle ball bounces a a portion of the light at the same relative angle as the incoming light, but some of the light rays hit uneven spots and bounce in random directions. This is like bouncing a golf ball on a country road. It will probably bounce the way you want it to unless the golf ball hits a crack or a pebble.

The mirror-like surface is so smooth that all the light bounces off the surface predictably at the same relative angle, like bouncing a ball off a basketball court. (Edit: a true mirror-like reflection requires a metallic surface, and cannot be achieved by smoothness alone. Thanks, David!)

Since highlights belong to the world of specular reflection, they should be thought of as somewhat separate and distinct from the normal modeling factors (light, halftone, and shadow) of diffuse reflection. Artists in the world of 3D computer graphics can control the form-modeling and the specularity as separate components.

I wrote an article on this subject of "highlights and specularity" for the current issue of International Artist magazine, which should be on the newsstands now. This blog post is just a little piece of it.

The six-page article contains a lot more examples and explanation, and it includes artwork that has never been reproduced in print before. I'll talk more on the blog about highlights in real-world examples tomorrow.
International Artist magazine, Issue 90
This subject is also covered in my book Color and Light: A Guide for the Realist Painter (Amazon), also available signed from my website store.


Marvin said...

David Briggs has written something a little different about reflections and highlights at
According to Briggs, it actually has nothing to do with surface "roughness". I am talking about this page:

This is my first post by the way. I like your blog posts, videos and books very much! Thank you for all your hard work!

James Gurney said...

Marvin, thanks for mentioning that point and sharing that link. It is indeed an important qualification. David is generally correct to raise the point that subsurface scattering can play a role in diffuse reflection, especially in the examples I gave of the sweater and the sand dune, the fibers and grains of which are quite translucent. On a microscopic level many, if not most matte surfaces diffusely reflect light that has penetrated the surface.

But it's not true to say that all diffuse reflection includes subsurface scattering. For example there's virtually no light transported below the surface of brushed or sanded metal, nor below the surface of the opaque spray primer that I used on those spheres.

So subsurface scattering is another variable feature of surfaces, just as specularity is a variable feature. For the purposes of this brief post and article, I simplified the discussion to keep it from getting overly technical.

By the way, I highly recommend David Briggs' website huevaluechroma.

David Still said...

There is some possibility for confusion, since some sources use the word 'highlight' in reference to the lightest part of the form light, and not the specular highlight. And it is important to remember that the specular highlight is not necessarily in the same place as this lightest part of the light.

James Gurney said...

David, you're right. I call that the "center light," the lightest part of the form modeling, where the light strikes the object at a 90 degree angle. It's in a different position from the highlight unless the light is coming from directly behind the viewer. The highlight is a reflection from the light source, so it's usually closer to the viewer than the center light.

David Briggs said...

It's easy to talk at cross-purposes because of semantics here, but the crucial point is that the difference between a glossy ball and a mirror-like ball is not that "some of the light rays hit uneven spots and bounce in random directions", although this myth is still widely repeated. No matter how smooth you make that glossy ball, it's not going to look like a mirror! There may well be some rays that "hit uneven spots and bounce in random directions" but such rays will still have the colour of the light source, like the highlight, and not the colour of the object. Polishing concentrates this specular or interface reflection, but does not reduce the main diffuse or body reflection, which results from light that penetrates into the so-called "opaque" (i.e. highly scattering) paint layer, bounces around between the grains, and emerges almost equally in all directions, coloured by selective absorption by the pigment grains. A mirror-like finish is possible only on certain substances such as metals that have no diffuse or body reflection, and so can reflect nearly 100% of the incident light as a specular or interface reflection.

James Gurney said...

David, thanks for explaining. So am I hearing you right that specularity is a matter of degree, but that a true mirror-like finish is a quality reserved to metallic materials? I guess that makes sense, and I defer to you.

I was trying to think of examples of surfaces that, without changing the composition of the material, can become more specular simply by smoothing the surface. One example is plasticine clay smoothed by a car designer into a slick, shiny surface, and another is what happens to a rough stone when you put it in a rock tumbler. My question is why does the chroma of the stone intensify when it comes out of the polisher? Is it because it's not so swamped with diffusely scattered light off the surface?

David Briggs said...

Yes it is, so there are two kinds of "diffuse" light coming off the rough rock: the light-source coloured "fuzzy" interface reflection and the rock-coloured body reflection. Polishing the rock concentrates (rather than increases) the interface reflection to a highlight, revealing the underlying rock-coloured body reflection more clearly, but not diminishing it. The polishing will only result in a truly mirror-like finish on any minerals present that have a metallic lustre, such as native metals or pyrite, galena, etc.; the rest of the rock will only ever look glossy, though it's just as smooth.