IF they're using it properly, i.e. creating a stochastic screen function and then just telling the imagesetter to output a box so many points high by so many points wide, filled 80%, then the resolution is device dependent, and is determined by the RIP, accounting for the matrix size and rotation. For a 256 gray scale at 0 degrees (typically used), it's a 4x4 matrix, though the rotation, if used affects the matrix size. (2450/4=ppi)
There's a good explanation of Postscript halftones in the Postscript reference books, I think the blue book, though it's been years (~20) since I've written raw Postscript, so I may be mistaken. The reality is likely that they're doing it from within a graphics program, and unless they're intimately familiar with Postscript (and I've only met 3 or 4 people that care enough to be) the likelihood is the graphics software is outputting a 600 ppi image (Illustrator, in particular is pretty bad about this, since version 7, as I recall), as opposed to the code to create a stochastic screen function at the raw imaging unit resolution.
(yes, at one time I did that sort of stuff, and had 2 Lino 330s of my own)
Oops I meant to say more in the 300-400 ppi range. Mark In a message dated 5/5/07 6:46:26 PM, firstname.lastname@example.org writes:Hi Mark,