Value Separations

Update in July, 2019

After several years, I wandered back around to my value separation program. (See original article, below.) I had neglected it for two reasons. First of all, it was very tedious to operate, making it not much fun to work with. Second, the only way to find out if a given photo was a good starting point for a value separation was to spend some time with it. In a program that was hard to use. Only to find out it was not going to work out.

So I did some more programming. The first task was to create a program that would batch process a bunch of photos and output a good first cut value separation from each one. Using this program, I could produce value separations from my 120 or so "art" shots from the previous year, import them into Lightroom, and make a "contact sheet". All in 30 minutes. This let me select the photos most likely to be worth more attention.

Then I did some work on the user interface of the original program to make it more straightforward to use. I also increased the program's speed so that any changes you make are now visible instantly. Much better.

My test photo during this programming process was one of my favorite pictures of my wife. I'll walk you through the stages of the transformation from original to value separation. First, here's the original color photo.

The program detects that you are loading a color photo and converts it to grayscale for internal use.

It then creates five separation zones based on the values in the photo. The idea is to create a five-zone separation that looks pretty much like the original photo.

It is obvious that the "drawing" in the face here is not very good (as my graphics professor would have said), but this is a good starting point for evaluation purposes. This five zone image is what the batch program would produce; and it's the initial display in the interactive program. From here, you're on your own. Here's one variation that I like with three zones and the colors shifted to sepia.

If you want a simpler and more graphic look, you can drop down to just two zones.

More imaging possibilities through the miracle of software.

Original Article from January, 2016

Value separations have been around for a long time. I was first aware of them when I was taking pictures for the school newspaper and annual at Clemson in the late 1960s. Here's one that was in the 1967 Clemson annual. (Not my photo.)

We considered this pretty far out at the time. The printing company did it from a straight black-and-white photo; so, to us photographers, it was kind of a magic process.

Later, at the University of Georgia in the mid 1970s, we learned how to do value separations for ourselves. We started out simply, using only photographic paper in the darkroom. Here are a couple of examples from that time:

We weren't allowed to stop after doing a "normal" looking separation; we had to do variations that put each of the three gray levels in each of the separation zones. Just to show that we could, you know. We also did value separations using Color Key material. These let you put a color or a texture in each separation zone. Here's a portrait of my photography professor:

I would call this portrait and the football image above "false color" separations, because the separation zones contain colors that aren't derived from the photos' original colors. (Although they do adhere to the originals' light/dark values.)

I like value separations. They remind me of some of the ink-based print processes like screen printing and lithography. So, being a programmer, I decided to write a program that would let me convert photos into value separations with some degree of control over the zones and values. (Although they are all "natural color" separations, because I'm currently too lazy to program a color picker.)

Here's the program applied to a grayscale original. It contains five separation zones, each with a different shade of gray.

Here's a similar treatment applied to a color image. In this case, the hues are all retained, but the colors in each separation zone are all the same brightness.

PS--After writing that my value separation program didn't have "false color" capability, I felt a little embarrassed. So I spent the afternoon upgrading it to include false color. Here's a not-so-special example, just to show what one looks like. There are five separation zones, each containing a single color.