karl@umb.umb.edu (Karl Berry) (10/26/89)
In one sense, grayscale characters *are* higher resolution. Why? Because the human eye can trade intensity resolution for spatial resolution. The basic problem with low-resolution characters is that there are not enough pixels to go around, i.e., there is not enough spatial resolution. In signal processing terms, a sufficiently high frequency cannot be represented. (What is ``sufficiently high''? Well, the human eye (most human eyes, I suppose I should say) typically cannot distinguish lines closer together than one minute of arc, which comes to about 60 cycles per degree of visual angle, i.e., 60 alternations from black to white. (This is called ``grating acuity'' in the vision research.)) But jaggies are really a question of being able to distinguish an offset of one line from another, not just distinguish two lines as being separate. This is called ``vernier acuity'', and it is about five times higher than grating acuity, i.e., 12 seconds of arc. Furthermore, with practice, it can be reduced to 6 seconds of arc. Some trigonometry will show you that 6 seconds corresponds to 1800dpi at 18" (which is normal reading distance) and 12 seconds to 1200dpi -- and most typesetters are manufactured in this range. (Incidentally, halftoning sometimes requires more resolution than text; hence, it is not pointless to manufacture machines with higher resolution.) Anyway, the edge information that is lost by using grayscale fonts is recovered by the human eye. If it didn't, grayscale characters would appear to be amorphous blobs, but in fact, they are more readable. By increasing our intensity resolution, we also increase our spatial resolution, because of the eye. There are lots of references on all this, if anyone is interested. (And if anyone reading this is doing active research on letterforms/image processing/signal processing/human vision, I'd love to hear from you.) karl@umb.edu