Digital filters are awesome for post image manipulation if you have enough bits to throw away. Glass filters, though, work at the highest resolution possible, in the camera head itself, and you'll never have a better image to tweak than that.
A year ago I had the pleasure to interview Ira Tiffen, formerly of Tiffen, Inc., and Bob Zupka, currently of Schneider Optics, for a magazine article. Here, with a few tweaks and updates, is that article.
The class of filter most often spoken of, and yet the least understood, is that of diffusion. Twenty years ago, Ira Tiffen sought to formulate a methodology of classifying diffusion filters to make it easier for cinematographers to understand what a particular diffusion filter was doing, and to make it possible to guesstimate what another filter would do based on knowledge of a particular filter's effects. He proposed numbering the three components that a diffusion filter affected, on a scale from zero to ten, and using those numbers as part of the filter name. The three components were contrast, resolution and flare.
Filters that reduce resolution are said to affect "high frequency detail." To those of us who got A's in English and D's in Physics, that means areas of the image that contain fine detail are affected more than areas that contain coarse detail. Wrinkles are a fine detail, lips and eyes are a coarse detail. How much fine detail is removed depends on the strength of the filter used.
Flare can also be thought of as halation, or "halos." Flare causes highlights to scatter, and the quality of that scatter determines a lot about how the filter is used. For example, a Tiffen Pro-Mist may only scatter highlights a short distance across the filter so the highlight appears to glow. A Schneider Fog filter would create highlights that spread a great distance, effectively distributing the halation across the entire image and creating an overall "fog" effect.
Contrast filters affect--wait for it!--contrast. These work off flare as well, but instead of creating halation in the scene they scatter the light into the shadows, revealing details in the darkest tones that might not otherwise be visible.
Tiffen's system would have designated each filter with a three digit sequence that told how much of each effect was present in the filter. If, for example, a 1/2 White Pro-Mist had a score of 2 for reducing high frequency detail, 3 for contrast, and 5 for halation, it would be relatively simple to find a filter that provided roughly the same effect but without the halation (a Soft F/X filter, for example). While the system itself never caught on, the thought behind it did.
On page 2: the three aspects of diffusion...