Re: [Fontforge-users] First steps in scanning glyph images?

[<mps...@hi...>]

Hi,

Following was my procedure to make Yi font by FontForge.

https://www.codeblog.org/blog/mpsuzuki/20060710.html

1. I printed a 16x16 matrix onto A4 paper by thin blue ink,
   and draw glyph chart with black ink by hand.

2. scan the images per page at 300dpi.

3. split the image into each glyphs in BMP format,
   by ImageMagick's "convert".

4. make EPS outline data by "potrace" (potrace -c -b eps),
   and name each EPS file by Unicode codepoint.

5. collect the EPS files by FontForge script like this:

------------------------------------------------------------
#!/usr/bin/env fontforge

New()
# Reencode( "iso10646-1" )
Reencode( "unicode" )
ucs2_cp = 0xA000
while ( ucs2_cp < 0xA4CF )
  Select( ucs2_cp )
  # Select( ucs2_cp - 0xA000 )
  eps_filename  = "U" + ucs2_cp + $1 + ".eps"
  Print( eps_filename )
  Import( eps_filename )
  # Transform( 400, 0, 0, 400, -1000, -1100 )
  Transform( 300, 0, 0, 300, -40000, -30000 )

  bbox = GlyphInfo( "BBox" )
  ll_x = bbox[ 0 ]
  ll_y = bbox[ 1 ]
  ru_x = bbox[ 2 ]
  ru_y = bbox[ 3 ]
  c_x = ( ll_x + ru_x ) / 2
  c_y = ( ll_y + ru_y ) / 2
  # Transform( 100, 0, 0, 100, (50000 - c_x) , (50000 - c_y) )
  Transform( 100, 0, 0, 100, (500 - c_x) * 100, (500 - c_y) * 100 )
  SetVWidth( 1000 )
  SetWidth( 1000 )
  ucs2_cp = ucs2_cp + 1
endloop
Save( "Yi-syllables.sfd" )
------------------------------------------------------------

On Tue, 19 Dec 2006 16:36:43 -0700
"Kenneth Reid Beesley" <krb...@gm...> wrote:

>Newbie questions:
>
>I plan to recreate a non-roman font from 1869 by scanning an original book,
>and then importing the glyph images and tracing them, probably manually, in
>FontForge.
>
>Are there any instructions/guidelines/suggestions about how to go about the
>scanning steps? e.g.
>
>1. What is the recommended resolution of the original scans?

It depends on the quality of the glyph. In my case, the
quality of hand-writing glyph is poor, and 300dpi was
too much. I recommend you to check one or two glyph for
trial.

>2.  Is one image file format easier to deal with than another?

If you want to check both of potrace and autotrace, I
recommend PPM format.

>3. How do you best go about separating out the individual images from
>the page scans?

This is strongly dependent with the script you want to
deal...

>4.  How do you best divide the work of positioning and standardizing
>the glyphs? before or after they are imported into FontForge?

If the glyph collection is small (less than 50?), tuning
in FontForge (after import) is worthful to consider, because
you can tune hinting info. If the glyph collection is great,
I recommend to lookup your favorite vector draw editor, to
tune tons of graphic data before import into FontForge.

>5.  Anything else to think about?

I don't know good light-weight vector draw editor which is
recommended for font production. If you have some favorite,
please let me know.

Regards,
mpsuzuki