Re: Convert old formats to OpenType?

In article <11c01tuof23o438@xxxxxxxxxxxxxxxxxx>,
*** Margulis <margulisd@xxxxxxxxxxx> wrote:

> Character wrote:
>
> > M G Berberich wrote:
> >
> >> On Fri, 24 Jun 2005 23:51:56 GMT, RSD99 <rsdwla.NOSPAM@xxxxxxx> wrote:
> >>
> >>
> >>> "...
> >>> From cubics to quadratics
> >>>
> >>> It is possible that the outlines you obtain will be described as cubic
> >>> curves. If so, it is not possible to perform an exact conversion between
> >>> the two formats. It is, however, possible to convert the curve
> >>> descriptions
> >>> to the degree of accuracy required.
> >>> ..."
> >>
> >>
> >>
> >> That's worth nothing. You can convert it to a polygon "to the degree
> >> of accuracy required" too -- that's simply a question of number of
> >> segments.
> >>
> >> MfG
> >> bmg
> >>
> >
> > To be even more pedantic, it should read "to the degree of *precision*
> > required".
> >
> > - Character
>
> This came up in a somewhat different context on a job I was doing
> several years ago. The word for the procedure of converting a curve to a
> series of straight segments is "faceting." The PRO-E CAD software that
> was being used at the time (running on HP UNIX workstations) faceted all
> curves that were destined for output to a plotter, on the reasonable
> assumption that plotters only know how to render straight lines in any
> case. However, it also faceted all curves that were being exported in
> any of its available file formats. The only such format that I had an
> import filter for in Interleaf was IGES. So I would get these e-NOR-mous
> files that would take forever to render, whenever the item in question
> had a lot of curves--which was all the time.
>
> I think we ended up figuring out a way for them to output PostScript to
> a file, and IIRC they were able to turn off the faceting function. Or
> maybe I don't recall correctly and I just suffered with the problem
> until the end of the project.
>
> The moral of the story is that when you find some EPS that someone sends
> you seems to be way too big, it may be because the engine that created
> it faceted all the curves at some minute scale, creating a gazillion
> nodes and a gazillion line segments, requiring an inordinate amount of
> time to render. Not all EPS is created equal, in other words. But you
> knew that already ;-)

Sounds like a real nightmare. It also reminds me of PostScript's
"flatness" which determines something similar to the faceting you speak
of.

Ever since I got a bootleg copy of Illustrator .901b (beta), I've been
in "love" with the Bezier curve. Having over the past week or so looked
at several hundred faces in both Beziers (Type 1/PostScript) and
"quadratic" (TTF) curves, my earliest feelings about the math have been
validated. I was way off-course when I wondered if Type-1 was a "lower
form of life" because of something Character had said about the number
of M space segments available for placement of nodes in each respective
method -- something I still haven't confirmed, although I have searched
for information about this ever since. (I've found nothing.) It would
seem that the math behind Type 1 is superior to True Type, but True
Type's robust self-contained hinting is potentially superior to Type
1's, depending on the PostScript Interpreter, which these days is not
limited by such issues. I would guess that the only time you "might" be
in trouble with PostScript's Type 1 hinting would be when printing on a
low-resolution non-PostScript printer. But that's in theory only; my
experience shows a very acceptable look to Type 1 fonts printed at low
resolution on a non-PostScript printer. Which has been a surprise to me!

So there you go. Much ado about nothing, except conversion. :-)

Michael
.