Link This | Email this | Blog This | Comments (0)

Halftone patterns and diffusion dithers

September 15, 2008 Halftones, developed by darkroom technician Frederic Ives in New York City in 1878, started their lives as straight lines that would get thicker and thinner as the tonality of the original changed. Ives did this by exposing a printing plate through a sheet of glass etched with straight lines. To improve the process, he made a second glass sheet with lines etched at 90-degree angles. As light passes through the the etched glass (the halftone screen) the light is partly interrupted by the etched part, but not affected by the clear glass. More intense light can pass-through the etched part more easily than less-intense light.

The light that reaches the plate (or in more modern terms, film) is modulated by the screen, and it creates tiny spots that are larger with more intense exposure, and smaller with less exposure. The little spots of light are called “circles of confusion” and they are the source of halftone dots.

Ask the average person how many colors are printed on a black and white newspaper page, and there will be a long pause. Why, it’s just black ink on paper, right? Well, then how do you get the gray tones in a photograph? Is it done with smoke and mirrors?

A halftone pattern made up of dots produces the illusion of tonality by printing the darker areas of photos with bigger dots, while printing the lighter areas of photos with smaller dots. Conventional dots are always the same distance apart, but their size varies in proportion to the tonality of the original.

The making of an analog halftone was laborious and moderately complex. In recent times, the etched glass screen was replaced by polyester screens, and the exposures were calculated on expensive exposure calculation devices. The film, when processed in a machine, was adequately consistent. If processed by hand (immersing the film in trays of developer, stop-bath and fixer under red or amber light) there was usually inconsistency due to agitation, timing, and chemical strength. (Just writing about this makes me thank Adobe for Photoshop. Thank you Adobe!)

The electronic version of halftones was first done by Dr. Rudolph Hell whose drum scanners revolutionized the imaging industry. Various iterations of drum-scanned halftone machines helped to bring us into the digital world. I remember a particularly odious machine from Fairchild called the Scan-a-graver. It used a heated needle to burn holes in a sheet of plastic in response to a continuous-tone photo mounted on a drum. It was stinky, slow, and the results were – at best – awful.

In about 1953, RCA developed a photographic method for setting type with a cathode ray tube. They were able to include photos in the image, but the cathode ray tube didn’t have adequate resolution to make a good halftone. So, they developed the diffusion-dither method for imparting “tonality” to a grayscale image.


The difference between a conventional halftone and a diffusion-dither halftone is that conventional halftones have dots at an even spacing (this is called Amplitude-Modulated), where the diffusion-dither divided the tonality of a grayscale image into patterns without a fixed spacing (called Frequency-Modulated). The little clumps of pixels were arranged randomly according to the tonality of the original.

Diffusion-dither halftones work delightfully well. The greatest problem is that there is a characteristic tonal jump that occurs at 50 percent. This is caused by spots that don’t quite touch (approximately 48%) appearing to touch, and dots that do touch increasing in value due to dot-gain on press and paper. The tonal jump is quite noticeable, and it makes some photos look really odd. What happens is that the 50-percent areas of an image appear to have contour lines drawn around them.

Adobe Photoshop has an excellent diffusion-dither tool, which will take any grayscale image and do to it what RCA did back in 1953. The results are mathematically the same.

Another problem with diffusion-dither on offset presses was the ability of hand-processed plates being able to hold the microscopic dot-clumps generated by the process. It was nearly impossible to hold them on film, and even harder to keep them on the plates. Modern technologies have solved that problem.

The printing industry had to wait another generation for the stochastic pattern to emerge. And, that, I will discuss in my next blog.

Posted by Brian Lawler on September 15, 2008 | Comments (0)