Re: [DejaVu-fonts] Instructions of U+0434, U+0446, U+0449, U+045f in Sans

[Andrey V. P. <pa...@ca...>]

On Thursday 21 December 2006 01:26, Eugeniy Meshcheryakov wrote:
> Hello,
>
> 4 =D0=B3=D1=80=D1=83=D0=B4=D0=BD=D1=8F 2006 =D0=BE 13:03 +1000 Andrey V=
. Panov =D0=BD=D0=B0=D0=BF=D0=B8=D1=81=D0=B0=D0=B2(-=D0=BB=D0=B0):
> > On Tuesday 21 November 2006 11:25, Andrey V. Panov wrote:
> > > At some sizes (17-20px) descenders of U+0434, U+0446, U+0449, U+045=
f in
> > > Sans looks too narrow.
> >
> > This issue also reveals on printing these glyphs from Windows (IMHO i=
t
> > utilizes tt instructions on printing).
>
> Width ratio for descenders vs strokes for lowercase letters is 79%. On =
your
> screenshot it is 3/4=3D75% - not too different, isn't it? At print it s=
hould

It occures also with another sizes, e.g. 34 px (see attachement). Both=20
horizontal stem and the descender should have the same width.

> be more like 79%, if you print it on modern printer (with large enough
> dpi, at dpi=3D300 size 9 means 37 PPEM, even if instructions are used w=
ith
> such sizes [I do not think so], this makes no difference).

Windows renders instructions in different from freetype way so it may rev=
eal=20
issues at different circumstances. In theory the instructions should not =
be=20
utilized on printing at all.

>
> > Another issue with instructions: at
> > some sizes "breve" in U+439 (=D0=B9) and U+45e (=D1=9E) is shifted to=
 the right,
> > this bug also appears on printing from Windows. Really "breve" in U+4=
39
> > (=D0=B9) and U+45e (=D1=9E) is positioned preciously in glyph`s middl=
e.
>
> I cant see this with large PPEM values. What printer do you use?

I use Samsung ML1250 printer (PCL5/6 compatible) with 600 dpi. I see also=
 the=20
issue with another accents (at least dieresis) and I think this is relate=
d to=20
https://bugs.freedesktop.org/show_bug.cgi?id=3D7224 . These instructions =
were=20
imported from Vera fonts.

--=20
Andrey V. Panov
panov /@/ canopus.iacp.dvo.ru