Adding one pixel can increase more or less the real size of a halftone dot, depending on the current shape of the dot, and the position of the newly added pixel. A pixel, which should add 1% to a halftone dot, can sometimes add only 0.3% on the film, at other times the same pixel added to another halftone dot, in a different position, can add 2-3%. This phenomenon can be amplified by the dot gain on the paper.

Many RIPs do not offer a solution to this problem. Thus, even if they do provide 4096 shades of gray, the vignettes will have visible tonal jumps (rather than steps, as indeed there are 4096 shades).

StudioRIP uses a low-pass filtering technique to avoid the tonal jumps, by using dots of slightly different sizes together, dispersed stochastically (e.g. achieving 50% by also using 48%, 49%, 50%, 51% and 52% halftone dots). On the picture below, you can see the imagesetter/CTP response curve after a linearization:

· with an insufficient number of shades: steps are visible
· with 4096 shades, but without a filtering technique: the curve is smooth, but has tonal jumps (non-linearities),
· with 4096 shades and after the filtering is done: the curve is smooth, and the tonal jumps filtered out, giving an almost perfectly linear result.