Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.postscript > #3042 > unrolled thread

How to get glyph metric info of CID-keyed font?

Back to article view | Back to comp.lang.postscript

Contents

  How to get glyph metric info of CID-keyed font? Jinsong Zhao <jszhao@yeah.net> - 2017-04-09 15:56 +0800
    Re: How to get glyph metric info of CID-keyed font? ken <ken@spamcop.net> - 2017-04-10 08:13 +0100
      Re: How to get glyph metric info of CID-keyed font? Jinsong Zhao <jszhao@yeah.net> - 2017-04-18 11:32 +0800

#3042 — How to get glyph metric info of CID-keyed font?

Hi there,

I am using R (www.r-project.org) to plot statistical figure. I prefer to 
use postscript device. Instead of Type1 fonts, I hope to use Truetype or 
Opentype fonts in my figure, that means I have to use CIDFont function. 
I read the help page of CIDFont function which state that the CID-keyed 
font are intended only for use for the glyphs of East Asian languages 
that are all treated as filling the same bounding box. In the source 
code of ``devPS.c'', the ascent, descent and width of glyph is assumed, 
and not calculated based on actual font. In fact, I think, R can't get 
the exact metric info for the glyph in CID-keyed font, and can't typeset 
correctly or beautifully. It seems that R can't find or access to the 
actual font. After output to a postscript file, Ghostscript with 
correctly configuration can render the figure correctly.

My question is how to get glyph metric info of CID-keyed font in C code 
directly or by call Ghostscript?

Any help will be really appreciated, and thanks in advance.

Best,
Jinsong

[toc] | [next] | [standalone]

#3044

In article <occp90$tl6$1@dont-email.me>, jszhao@yeah.net says...
> 
> Hi there,
> 
> I am using R (www.r-project.org) to plot statistical figure. I prefer to 
> use postscript device. Instead of Type1 fonts, I hope to use Truetype or 
> Opentype fonts in my figure,

You cannot directly use TrueType fonts or OpenType fonts with a 
PostScript interpreter, because PostScript deos not support either 
TrueType or OpenType.

Language Level 3 PostScript interpreters can use both type 42 (which are 
TrueType fonts 'bundled up' in a way which PostScript interpreters can 
understand) and CFF font outlines. Earlier language levels did not 
support CFF and may not support type, support for CFF was added with 
level 3, and support for type 42 was introduced as a supplement to the 
original level 2 sepcification. CIDFonts were also added (in version 
2015 as I recall) during the lifetime of language level 2.

Note (1) that for OpenType fonts, you would need to check the outline 
type of the font and then supply either a Type 42 (TrueType) font or a 
CFF font, in each case extracting the relevant data from the font.

Note (2) Ghostscript *can* make use directly of both TrueType and 
OpenType fonts as substitutes for missing fonts. However this involves 
the interpreter 'guessing' at certain information which would normally 
be presented unambiguously in the tytpe 42 font. So sometimes it will 
get it wrong.


> that means I have to use CIDFont function. 

Nope. Type 42 fonts are not CIDFonts. A CIDFont may have a number of 
different kinds of outlines, including type 42 outlines, but its 
entirely possible to have a type 42 font which is not a CIDFont.

The same is also true for CFF outlines (OpenType fonts may have CFF 
outline data, instead of TrueType.


> I read the help page of CIDFont function which state that the CID-keyed 
> font are intended only for use for the glyphs of East Asian languages 
> that are all treated as filling the same bounding box.

No, sorry, that's not really true either. It is the case that the 
original impetus for development of CIDFonts was the complexity of the 
written forms of Asian languages, but nowadays CIDFonts are used 
extensively for all languages.


> In the source 
> code of ``devPS.c'', the ascent, descent and width of glyph is assumed, 
> and not calculated based on actual font. In fact, I think, R can't get 
> the exact metric info for the glyph in CID-keyed font, and can't typeset 
> correctly or beautifully. It seems that R can't find or access to the 
> actual font. After output to a postscript file, Ghostscript with 
> correctly configuration can render the figure correctly.
> 
> My question is how to get glyph metric info of CID-keyed font in C code 
> directly or by call Ghostscript?

You can interpret the font. There is no equivalent of the '.pfm' files 
that were once supplied with Type 1 fonts and contained the metrics of 
the glyphs.

Ghostscript, like any font rendering engine, *must* interpret the font 
in order to render it. The font (whether TrueType or CFF outlines) will 
contain its own set of metrics which cover all the glyphs in the font, 
each glyph will additionally contain information such as the BoundingBox 
(or hsbw, Horizontal Side Bearing and Width) information in the glyph 
description.

Ghostscript (mostly) uses FreeType for its font rendering, I would think 
that the simplest approach for you would be to do the same.

I'm afraid I can't offer you any help with R, I know nothing about it.

				Ken

[toc] | [prev] | [next] | [standalone]

#3048

On 2017/4/10 15:13, ken wrote:
> In article <occp90$tl6$1@dont-email.me>, jszhao@yeah.net says...
>>
>> Hi there,
>>
>> I am using R (www.r-project.org) to plot statistical figure. I prefer to
>> use postscript device. Instead of Type1 fonts, I hope to use Truetype or
>> Opentype fonts in my figure,
>
> You cannot directly use TrueType fonts or OpenType fonts with a
> PostScript interpreter, because PostScript deos not support either
> TrueType or OpenType.
>
> Language Level 3 PostScript interpreters can use both type 42 (which are
> TrueType fonts 'bundled up' in a way which PostScript interpreters can
> understand) and CFF font outlines. Earlier language levels did not
> support CFF and may not support type, support for CFF was added with
> level 3, and support for type 42 was introduced as a supplement to the
> original level 2 sepcification. CIDFonts were also added (in version
> 2015 as I recall) during the lifetime of language level 2.
>
> Note (1) that for OpenType fonts, you would need to check the outline
> type of the font and then supply either a Type 42 (TrueType) font or a
> CFF font, in each case extracting the relevant data from the font.
>
> Note (2) Ghostscript *can* make use directly of both TrueType and
> OpenType fonts as substitutes for missing fonts. However this involves
> the interpreter 'guessing' at certain information which would normally
> be presented unambiguously in the tytpe 42 font. So sometimes it will
> get it wrong.
>
>
>> that means I have to use CIDFont function.
>
> Nope. Type 42 fonts are not CIDFonts. A CIDFont may have a number of
> different kinds of outlines, including type 42 outlines, but its
> entirely possible to have a type 42 font which is not a CIDFont.
>
> The same is also true for CFF outlines (OpenType fonts may have CFF
> outline data, instead of TrueType.
>
>
>> I read the help page of CIDFont function which state that the CID-keyed
>> font are intended only for use for the glyphs of East Asian languages
>> that are all treated as filling the same bounding box.
>
> No, sorry, that's not really true either. It is the case that the
> original impetus for development of CIDFonts was the complexity of the
> written forms of Asian languages, but nowadays CIDFonts are used
> extensively for all languages.
>
>
>> In the source
>> code of ``devPS.c'', the ascent, descent and width of glyph is assumed,
>> and not calculated based on actual font. In fact, I think, R can't get
>> the exact metric info for the glyph in CID-keyed font, and can't typeset
>> correctly or beautifully. It seems that R can't find or access to the
>> actual font. After output to a postscript file, Ghostscript with
>> correctly configuration can render the figure correctly.
>>
>> My question is how to get glyph metric info of CID-keyed font in C code
>> directly or by call Ghostscript?
>
> You can interpret the font. There is no equivalent of the '.pfm' files
> that were once supplied with Type 1 fonts and contained the metrics of
> the glyphs.
>
> Ghostscript, like any font rendering engine, *must* interpret the font
> in order to render it. The font (whether TrueType or CFF outlines) will
> contain its own set of metrics which cover all the glyphs in the font,
> each glyph will additionally contain information such as the BoundingBox
> (or hsbw, Horizontal Side Bearing and Width) information in the glyph
> description.
>
> Ghostscript (mostly) uses FreeType for its font rendering, I would think
> that the simplest approach for you would be to do the same.
>
> I'm afraid I can't offer you any help with R, I know nothing about it.
>
> 				Ken
>

Thank you very much for the reply and explanation. I now use cairo_ps 
function in R to plot, which use cairo graphics API and can use CFF 
fonts directly.

Best,
Jinsong

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.postscript

csiph-web