How flexible are the Macintosh outline fonts?
Dan Sahlin
several things are still unclear about their capabilities.
- Is there any support for rotating the fonts? How will the "hints" work
for rotated fonts?
- Is it possible to make variants of the fonts in the way it is done
in Laser FX or the Letraset fonts? Laser FX is using PostScript to reach
those effects, and is uncapable of showing the final result on the screen.
The software from Letraset uses special fonts, which are only possible
to use with their software.
- Acorn DTP uses knowledge about the form of the letters to make the
kerning "perfect" and automatic. For instance, the pair "AV" can kern
very much due to the shape of the letters, whereas "VV" cannot kern much
at all. This method for kerning should also be possible for the outline
fonts for the Mac as the shape is available.
- "Fuzzy fonts" are fonts where greyscale is used to make the edges
smoother. I've heard that fuzzy fonts are standard for the NeXT. Tests
have shown that fuzzy fonts are much more legible than fonts with sharp
edges.
The Mac already has support for color and grayscale fonts, but can the
outline fonts generate "fuzzy fonts"?
- In close relation to fuzzy fonts is the concept "subpixel addressing".
By varying the intensity of gray of the edges, the character may appear to
move smoothly between the pixels. This method will effectively increase
the resolution of the screen several times.
The Acorn DTP already supports both "fuzzy fonts" and "subpixel addressing".
What about the forthcoming Macintosh "outline" fonts?
- The Mac already has support for color fonts. Will the outline fonts
also support color?
- Many fonts come in families that do not fit the model "Bold/Italic/
Underline/...". For instance some fonts have variants such as "light",
"extra heavy" but lack some of the standard variants. Will the new
font system support such font families better?
/Dan Sahlin
email: d...@sics.se
John Bowler
>
> - Acorn DTP uses knowledge about the form of the letters to make the
> kerning "perfect" and automatic. For instance, the pair "AV" can kern
> very much due to the shape of the letters, whereas "VV" cannot kern much
> at all. This method for kerning should also be possible for the outline
> fonts for the Mac as the shape is available.
>
for character pair kerning information, but this is not used. Manual
kerning is certainly possible, but automatic kerning seems too much to
hope for at present. For example how close should the letter pair OO
be placed? If kerning is done on the basis of character shape the two
letters would almost certainly be placed too far apart. Perhaps some
form of weighted average using a series of character widths entered by
the font designer would give reasonable results (ie better than no kerning).
The idea is certainly interesting.
The point about the Acorn DTP font format (which may have given rise to
the above confusion) is that the hints combined with the support for
anti-aliased fonts within the operating system allow the production of
quite reasonable screen fonts at small point sizes. Of course, the other
advantage is that the software is available now :-) (In Europe)
John Bowler, Acorn Computers Ltd
[The opinions are my own - not necessarily Acorn's]
Dan Sahlin
> - Acorn DTP uses knowledge about the form of the letters to make the
> kerning "perfect" and automatic.
And John Bowler at Acorn replied:
>Unfortunately this is not the case - currently our font format has space
>for character pair kerning information, but this is not used.
I'm sorry about my mistake. I really meant that "Calamus 1.09.2" on an
Atari has this functionality. I quote and translate from the Swedish
magazine \CAP, June 1989, page 30:
"The excellent kerning function in Calamus is also using the
vectorised fonts and using them automatically computes the most
appropriate distance between the characters".
So there does not seem to exist a system which combines fuzzy fonts and
automatic kerning, or does it?
/Dan Sahlin
email: d...@sics.se
James 'Kibo' Parry
> "The excellent kerning function in Calamus is also using the
>vectorised fonts and using them automatically computes the most
>appropriate distance between the characters".
I believe the mechanism Calamus uses is that in the font file, for each
character, there are 16 numbers -- 8 describing the widths of the character
at eight different "latitudes" on the left, and 8 for the right. (I think
eight is the number.) This way it can fit letter together by saying
to itself, "This letter is wide at the top and narrow at the bottom, but
this one's perfectly straight... hmm, how close can I make them..."
Of course, not being privy to the secrets of the designers of Calamus, etc.,
I may be off the mark somewhat. I've deduced the above by dumping some
Calamus demo fonts (like CGLOGO.CFN, the CompuGraphic insignia) in hex
and spending a few hours examining them; I'm trying to write a program
to interchange font formats.
Whatever mechanism Calamus uses, ultimately, the characters wind up
very nicely spaced... the output is beautiful. (I like how it uses
the outline fotns on both the screen and printer; there are no seperate
bitmapped screen fonts. WYSIWYG down to the pixel level...)
james "kibo" parry, 138 birch lane, scotia, ny 12302 usa / All colors are
kibo%pawl.r...@itsgw.rpi.edu _________________________/ arbitrary.
ki...@mts.rpi.edu / Kibology / Anything I say is the opinion
user...@rpitsmts.bitnet / is better! / of myself, and not of Xibo.
Eric S. Tiedemann
>I've heard that fuzzy fonts are standard for the NeXT.
Where did you here this? As far as I know this isn't true..which is a
shame, since the NeXT screen is an ideal delivery vehicle for fuzzy fonts.
Eric
e...@cs.nyu.edu