Type 1 Specs
31 messages in this thread
Use the RM Function to read
these messages.
I thought Adobe's reaction to the announcement was very honorable, as well.
The press is being rather blind about it all – Microsoft and Apple have
demonstrated little to nothing about their proposed technology, while Adobe has
such a great track record for maintaining the PostScript standard, plus the
existence of so many PS applications. It boggles the mind.
Hi, David –
To paraphrase a classic office saying:
How is the Apple/Microsoft PDL agreement like elephants mating?
1. It is done on a high level between great beasts.
2. There is much bellowing and trumpeting.
3. It takes a year to get results.
Frankly, regardless of the number of RIP vendors announcing that they will
support the new standard, it is still vaporware as far as the user community is
concerned. That's great if Hyphen is already interpreting Royal fonts, but
there's more to a PDL than fonts, and Microsoft's contribution here is still
hard to assess either in terms of results, current progress, or remaining
development effort.
The recentness of the coalition announcement revealed – at least to me – that
both MS and Apple didn't have complete solutions on their own. Also, it's going
to take time for the two firms' development teams to synchronize efforts.
Another related favorite classic is, "It is not possible to hire 9 women to
produce a baby in one month." Unless there are some Jokers hiding in this
particular deck, the critical path is at least 8 months long due to
staff/vendor/developer communications alone.
Now I'm _fairly_ sure that I didn't sleep through all three years of calculus
classes (despite what my professors may claim…), but quadratics are 10-20%
more compute-intensive then Beziers. Also, I programmed quadratic curve
routines for NASA some 17 years ago in machine language, so there is a little
rationale for the delusion that I might know something about this mess.
Now this may sound too, too simple, but – what problem does this proposed joint
PDL solve? Performance? Naw – they'll have the same hurdles to overcome, and
the clone shops are supposed to have 120 page/minute capability now, anyway.
Screen display quality? Uh uh – ATM fixes that, and Display PostScript
certainly works well on other platforms. Enriched features? Nope – I wouldn't
be surprised if it had fewer features. Adobe's large market share? Bingo! –
this PDL venture is purely a predatory move which would do justice to IBM's
gorilla tactics of the
[continued in reply]
[continued]
'60's.
I'm going to boot-up my crystal ball here:
– Apple's rumored announcement of QD-based printers in March '90
will likely not incorporate Microsoft's end of the PDL.
– System 7.0 release may be delayed as much as 4 months due to
glitches in doing the "hooks" to MS's code.
– Adobe will launch an anti-trust suit against the alliance shortly
before roll-out of any "AMs" PDL-based product. Capturing a
market with proprietary technology is one thing, but it's another
thing altogether to form alliances with competitors in order to
take market from a third party.
– The "AMs" PDL will find rough going in the professional, high-res
arena due to the market saturation by PostScript (and clone)
engines that will likely occur during the "AMs" gestation, debug,
and acceptance (and lawsuit!) period.
– "AMs" will likely become another OS/2 – "established standard"
on a proprietary platform with few supporting applications for an
indeterminate period of time.
– The former "simple & user-friendly" Macintosh will drown
in a cesspool of competing standards in which only the "techies"
will have a full grasp (job security?) – shades of MS-DOS!
High-end users will embrace the NeXT box, if only out of
simplicity.
The wheel has been invented, and it is round, rolls well, and just plain works.
The Apple/Microsoft deal will only confuse everybody and turn the user
community against Apple. (Who really _likes_ Microsoft?).
…mm
Well, for whatever its worth, I would agree with Mike A bird in the hand and so
forth The whole PDL scene is so crazy now I appreciate a few simple standards
'nuff said Blaise
I've got some replies to your questions about what the MS-Apple deal will
solve.
Performance? You bet. No matter how much we may all love Adobe's PS, there's no
denying that it's incredibly, and sometimes miraculously, SLOW. Much of the
reason for that is probably its "less than admirable" (I'm being nice) memory
management system. Much more of the reason is that they never had much
incentive to speed it up, because they HAD NO COMPETITORS!
Screen display quality? ATM is not a good general solution for Mac owners,
because it only images Adobe fonts!! (You're going to say that everyone will
have Type 1 fonts after Adobe releases the info, but that's another point for
me – they never would have released that information without the MS-Apple
agreement). As for Display PS working well on other platforms, D-PS was to my
understanding the main reason drawing programs work so insufferably slow on the
33-MHz 68030 NeXT machine.
Enriched features? Certainly. The MS_Apple solution will be a superset of Adobe
PS features. They will probably also improve the existing features, such as the
moronic (still being nice) 1500 turnpoint limit that every PS user knows of, or
PS screens that are inaccurate, etc.
I don't think anti-trust really is a threat; not after 8 years of Reagon
court-appointees. Or with the Big 3 car companies teaming up with Asian
competitors (presumably to take market share away from others).
Finally, I'm betting that if it's done right (and I think it can), most of the
changes will be transparent to the user, and he won't have to care either way.
Particularily after Royal fonts become the new standard (presumably after all
the PS fonts have been converted to it).
Adobe's really got no complaints. None of this would have come about without
their eexec nonsense, and font encryptions, and hidden PS utilities that only
Adobe products could use, and 1500 turnpoints (we've put up with that one for
four years now!), and high prices, and, and, and . . .
— Pete
Earl-
You make this MS-Apple font/PDL scenario sound like a good thing. I
suspect Mike's analysis is more accurate: it's a political move, almost purely
a market-share grab on Apple's part. Keep in mind that Adobe (and ALTSYS, I
imagine) is still run by visionary types. But Apple is a slave to
shareholders, and thereby to quarterly reports. That's a different universe,
with a different set of rules.
In a corporate strategy session it takes about 5 minutes for the bright
boys to realise that brewing a PR storm gets a hell of a lot more bang for the
buck than actually DELIVERING a superior product.
I see little for the users of this technology to cheer about.
David Lynn
David:
Where did you hear that Royal fonts "have provisions for defining size-specific
features"?
I remember hearing something similar a while back, but when I asked several
people, including two of the guys coordinating the digitalization of the URW
font library as well as the president of Monotype, everyone said that b-spline
quadratics are just a different way of producing digital type -inherently no
better than the bezier curves than Adobe's PostScript fonts use.
As an ad typographer, optical scaling of fonts is an important concern when
looking at the PostScript typefaces. Most faces I've seen are digitized to look
best at text sizes. Several of the faces I've used just don't look right in
large point sizes. An electronic way to optically scale the type would be a big
plus in our minds or, failing that, font companies should offer two printer
outlines (one for text sizes, one for display).
I keep looking back through various magazines to find the reference I saw
concerning the Royal fonts, but can't seem to locate it. Any ideas?
–jon
Jon,
When I was at Seybold, I attended the meeting of the SPS Association
(PostScript standards group). Mike Parker from the The Company, which is the US
company which handles URW fonts and font technology, spoke on font developments
in PostScript. When referring to the Royal fonts, he said that they were
constructed with "three levels" of quality. Level 1 provided hinting analogous
to that provided by Adobe with its Type 1 fonts. Level 2 provided "outline
manipulation" to make sure the display of fonts "looked nice." He said this is
analagous to what Adobe does with its ATM fill routines. Level 3 is something
not provided by Adobe in its PostScript fonts. It is a complete font
manipulation language. Apparently this could be used by a font vendor to
produce (among other things) size-specific features.
Again, quoting from my notes from Parker's talk, font vendors don't have to use
all three levels of Royal font technology. They can choose to implement the
ones they want to. And output devices don't have to use all three levels. Level
3 would only take effect if: (1) the font vendor put such information into
their fonts and (2) the output device (screen, printer) was set up to use it.
Otherwise the output would fall back to the simpler levels.
I hope haven't mangled his ideas too much. I'm doing this from notes I took two
weeks ago.
^Steve^
Steve:
Thanks a lot for the info on Parker's speech. I wonder if Adobe's bezier fonts
could also be manipulated in a manner similar to that envision in a Royal
"Level 3" font, given PostScript language extensions to handle such font
manipulation, or if it is something that only the Royal quadratics can do. The
obvious question is whether any of the big type foundries will bother to take
advantage of this feature or just continue to produce fonts that do linear
scales from one master.
I suppose the market that would appreciate such a move is limited. Additional
font manipulation in PostScript would be nice in other areas beside optical
scaling. In small caps, in addition to horizontally scaling them a little more
than the parent font (which is fairly easily accomplished in most DTP programs)
it would be nice to be able to heavy-up the small caps so they match the weight
of the parent cap.
Adobe reps that I recently spoke to said a fairly extensive upgrade in
PostScript is almost ready that will provide quite a few additional type
manipulation operators. I'm curious to see what we get.
–jon
Jon,
We'll probably know a lot more about the current font capabilities of Type 1
fonts after the specs are published. Until then, what we're talking about is
pretty speculative.
The PostScript upgrade would be welcome. The SPS Association has quite an
extensive list of extensions they would like to see.
^Steve^
Adobe type 1 fonts can contain two or more versions, so they could provide
separate masters for different size ranges (with the obvious restriction that
the widths would have to be the same–the computer doesn't know about size
ranges or alternate font metrics). I was surprised to learn about this when I
encountered a flaw in the "S" character in Optima Bold on a high-res
imagesetter that didn't appear on a LaserWriter (set at 600 point on both
machines). Adobe tech support found that there are two versions of Optima in
the basic font, selected according to the printer's resolution. I feel certain
that separate versions could be chosen for different size ranges, or even based
on some combination of size and resolution.
Furthermore, the sort of continuous optical scaling promised (and shown on
paper) by Monotype holds great promise for creating a continuously scaled type
face that would look good at all sizes. I agree that present Adobe PostScript
fonts don't look their best at extreme sizes (too large or too small), and look
forward to seeing this scaling approach implemented by Adobe as well as its
competitors.
Stephen:
That's real interesting about the Type 1 fonts. So, beyond just straight hints,
there may often be two masters. No wonder the Adobe fonts look so much better
on low resolution machines (if the hints can't correct a problem face, they
just load up a version optimized for low-rez devices).
Can't wait to see the Type 1 specs when they get released. Wonder what else is
in there that isn't generally known about Adobe fonts.
–jon
I asked how often this two-master trick was used, and the people I was in
contact with thought it was very infrequent. Optima is a very subtle face, and
they recommend against its use on low-res devices; for those who insist, they
provide a "simplified" version of the outline font to do as well as possible.
They told me that the switch-over from one master to the other was based on the
resolution of the printer, which doesn't seem the most sensible method to me.
I'd rather see large type on a low-res printer created from the "high-res"
master, but apparently a 300dpi printer always gets the "low-res" master
whatever the size. If they use several masters to provide optimized appearance
for text and display sizes, the switch-over should be based on both size and
resolution, with more emphasis given to size.
What would be ideal, using the two master theory, is for the font description
to be device smart. It would pull the default matrix and use it to test what
size/resolution ratio it's printing at. At that point it would grab two
versions of the same character, each specified for a particular size/resolution
ratio. Then it would perform an in-betweening or ex-tweening if the size is out
of range, to create a new face unlike the two given, but a mixture of them
depending on it's ratio's position in comparison to the other two. Make sense?
Monotype has shown type 1 PostScript fonts that adjust stroke weights according
to the size of the finished character, which seems like a good idea. Whether
their technique is clever enough to account for global scaling surrounding the
text isn't clear, and I haven't gotten any of their new faces to experiment
with. But it's clear that some nifty stuff could be done with the type 1
character spec to make us "fussy typographers" happier. It will be very
interesting to see what creative solutions come out of the opening up of the
type 1 spec.
One of the current "bit players" (Was that a pun? I'm not sure…) in the
high-quality PS font world is Kingsley/American Type Founders, who used to be
one of the dominant names in foundry (lead) type. While I haven't personally
acquired any of their faces to prove it, they are supposed to be doing some
manner of dynamic stroke adjustment with their line of _Type_3_ fonts.
…mm
Oops! I think I may have been thinking of the ATF library when I mentioned
size-related stroke width scaling, rather than Monotype. I expect we'll see
more of this from several vendors in the near future.
Stephen, Jon, and Steve,
Very interesting discussion re Adobe's font mechanisms. I was wondering if one
of you could comment about the new PostScript interpreter (if that's what was
meant re new font handling routines). How would such an interpreter be added to
a LaserWriter or other PostScript device?
Thanks,
Rich
Chances are that updates would be impossible in existing LW's unless you'd
agree to an expensive board swap (might be just as easy to buy the latest LW
with the upgrade in it).
I'm afraid I haven't heard anything about a new PostScript interpreter, so I
don't have any good rumors to pass on. In the past, Adobe has maintained that
version upgrades to printers in the field were the responsibility of the
printer manufacturer, so they were unwilling or unable to offer any assistance
to LaserWriter owners except through Apple. Perhaps Apple's impending "divorce"
from Adobe will make Adobe more willing to consider direct support for the
LaserWriter community.
The only thing I've heard about improvements to PostScript's font support is
that the new algorithms in Adobe Type Manager are better at interpreting hints
and faster overall than those in the printer PostScript interpreters, and that
future printers will incorporate the ATM scheme for font handling. But there
aren't any new features or capabilities in ATM relative to printer interpreters
now, as far as I know.
Richard,
In most cases a PostScript printer would require a board swap. It'll probably
be up to each printer manufacturer to incorporate a newer version of a
PostScript interpreter and to offer it to their customer base. We've had
Linotronic RIPs since 1985, and we've gone from PostScript version 38 to 42 to
47.1 to 49.3. Sometimes the board was swapped, sometimes it required the
purchase of a new RIP.
^Steve^
47.1 to 49.3 was a PROM change for us, Steve. I still have the old PROMs, too.
…mm
Steve, Nathan, Stephen,
Thanks. When I bought my LWII NT I really only wanted a proof press before
going to Lino. I didn't worry too much at the time about upgrades but thought I
could rely on the machine as conforming to an industry-wide standard PDL. I
couldn't afford a II NTX at the time, but figured that I would probably be able
to upgrade the 68000 micro processor to a 68020 or maybe even an 030 if that
became necessary. I thought for sure that future changes in the PostScript
interpreter would be nothing more than a ROM change. I also (wrongly) assumed
that there would be a third party market for upgrades (speedups, etc). After
all, the machine is 'just' a 680×0 computer with a laser driven xerographic
device rather than a CRT.
Now with changes in the PostScript standard evolving, and the Apple/Adobe
estrangement, these machines may become orphans. It may not be possible to see
output on a LW that will closely enough approximate the output on a PostScipt
image setter which is using a version of PostScript that has been updated. I do
not think that Apple is going to be interested in updating the machines'
innards to keep pace with Adobe-PostScript. I sure hope that Adobe itself or
some third party will perceive a market need in this area.
Rich
Important to remember, though, that changes to PostScript from Adobe are merely
enhancements for speed improvement and device specific things such as
halftoning accuracy/speed. Any increment to the PostScript version will never
make the RED book obsolete, nor any printer adhering to at least version 23.0
obsolete (the version of the RED book). The RED book commands will always be
supported. It is the underlying standard. Any enhancements above that are
merely extensions.
Don't worry unnecessarily. There has been no promise–or even implied
promise–of any changes to PostScript that would affect your ability to produce
an image on your LaserWriter IINT that will correspond to the output of any
other PostScript imagesetter. Indeed, Adobe has been quite insistent that
PostScript will remain a true standard, and I have no reason to doubt them in
this. The only enhancement to PostScript that requires a different or newer
version of the interpreter is Kanji PostScript, which is available in Apple
LaserWriters in Japan; Adobe has mentioned (but not yet delivered) a
downloadable version of the Kanji PostScript code to be used on conventional
LaserWriters.
At Altsys, we're fairly certain that different filling and stroking algorithms
are used in Type 1. We believe that when a PS outline is rasterized, pixels
which the path goes over are turned on only if more than half the pixel would
be covered by the fill (in Type 1). Type 3 fonts have to use the regular PS
filling routines, which turn on pixels under the outline no matter how little
of the pixel is actually covered. This is probably why Type 3 fonts print
significantly darker than Type 1 fonts. This is just conjecture, but we're not
the only ones who hold this opinion~
— Pete
This anti-aliasing, if it could be called that with 1 bit pixels, seems logical
but would take an astronimcal amount of calculations comparatively.
You're right, I've been absolutely selfish about this. I haven't heard a word
about where *I* can find out about Royal fonts. However, I've been able to buy
PostScript books at any computer bookstore I've entered. Must I be a
million-dollar RIP maker to find out about the Microsoft/Royal fonts? If
you've got some tips on tools or documentation on the new format, please help
me alleviate my ignorance.
Use the RM Function to read
these messages.