[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#424404: PDF files are much larger with lmodern than with CM fonts



On 2007-05-16 13:31:44 +0200, Norbert Preining wrote:
> OTOH I would say that it really doesn't hurt too badly as soon as we
> consider real world problems. It is true that for one char documents
> this is a bad thing, but as soon as there are a lot of characters
> included, the difference should get less severe.

Less severe, but still significant, as in practice, many fonts may be
included (e.g., because of the different levels of heading, I assume).
For instance, on a typical computer-science article:

name                                 type              emb sub uni object ID
------------------------------------ ----------------- --- --- --- ---------
FCHZZX+NimbusSanL-Regu               Type 1            yes yes no       6  0
BIQIWQ+NimbusRomNo9L-Regu            Type 1            yes yes no       7  0
TLLMPC+NimbusRomNo9L-MediItal        Type 1            yes yes no       8  0
NZKSYR+NimbusRomNo9L-Medi            Type 1            yes yes no       9  0
HOFHNU+Times-BoldItalic              Type 1C           yes yes no      14  0
IJUGPW+LMRoman12-Bold                Type 1            yes yes no      28  0
JEXNLM+LMRoman12-Regular             Type 1            yes yes no      29  0
YLCJDP+LMRoman10-Regular             Type 1            yes yes no      30  0
YLCJDP+LMRoman10-Regular             Type 1            yes yes no      31  0
BLFSLL+LMRoman10-Bold                Type 1            yes yes no      32  0
KFJEKH+LMRoman10-Italic              Type 1            yes yes no      33  0
ZJMJPW+LMRoman8-Regular              Type 1            yes yes no      42  0
YLCJDP+LMRoman10-Regular             Type 1            yes yes no      43  0
XGNKOW+LMRoman6-Regular              Type 1            yes yes no      44  0
XMPDUW+LMRoman9-Regular              Type 1            yes yes no      45  0
XMPDUW+LMRoman9-Regular              Type 1            yes yes no      46  0
LBHWLJ+LMMathItalic10-Italic         Type 1            yes yes no      50  0
DPOLVU+LMMathSymbols10-Italic        Type 1            yes yes no      51  0
ZJMJPW+LMRoman8-Regular              Type 1            yes yes no      52  0
IMCDFK+LMTypewriter10-Regular        Type 1            yes yes no      53  0
YJCHEJ+LMMathItalic8-Italic          Type 1            yes yes no      54  0
KBUCPP+LMTypewriter9-Regular         Type 1            yes yes no      55  0
XMPDUW+LMRoman9-Regular              Type 1            yes yes no      56  0
QSWYSG+LMRoman9-Italic               Type 1            yes yes no      57  0
QLYZBA+LMMathItalic9-Italic          Type 1            yes yes no      58  0
XGNKOW+LMRoman6-Regular              Type 1            yes yes no      59  0
HHUADA+LMMathSymbols8-Italic         Type 1            yes yes no      63  0
XAAFEP+LMMathExtension10-Regular     Type 1            yes yes no      71  0
BLFSLL+LMRoman10-Bold                Type 1            yes yes no      72  0
RPZDSR+LMRoman10-BoldItalic          Type 1            yes yes no      82  0
TFLFXD+LMSans10-Regular              Type 1            yes yes no     103  0

On this article, the conversion from CM (using aeguill) to LM leads
to a 80% size increase (more than 200 KB). So, this is still quite
bad in real world (both concerning network transmissions, which are
not always very fast, and concerning data storage on mobile devices,
where space is rather limited).

-- 
Vincent Lefèvre <vincent@vinc17.org> - Web: <http://www.vinc17.org/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.org/blog/>
Work: CR INRIA - computer arithmetic / Arenaire project (LIP, ENS-Lyon)



Reply to: