<html> <head> <meta content="text/html; charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body bgcolor="#FFFFFF" text="#000000"> Am 08.02.2012 15:29, schrieb Stefano Emilio Campanini: <blockquote cite="mid:CAOTGqbYkzPmKjPViKHzXwTgL_VBtuVM5P1xcv=pmebVjYGdncg@mail.gmail.com" type="cite"> <pre wrap="">On 8 February 2012 15:09, Stefano Emilio Campanini <a class="moz-txt-link-rfc2396E" href="mailto:stefano.campanini@tinvention.net"><stefano.campanini@tinvention.net></a> wrote: </pre> <blockquote type="cite"> <pre wrap="">2012/2/8 suzuki toshiya <a class="moz-txt-link-rfc2396E" href="mailto:mpsuzuki@hiroshima-u.ac.jp"><mpsuzuki@hiroshima-u.ac.jp></a>: </pre> <blockquote type="cite"> <pre wrap="">Hi, Could you post a sample PDF (or upload it to somewhere) showing the warning? It seems that the object reference tree in the PDF is looking like as if the PDF includes embedded fonts for ArialNarrow, CourierNewPS, etc, but poppler could not load them (and substituted them by external font resource). </pre> </blockquote> <pre wrap=""> They are eBooks not free, how can give you they without break copyrights ? There is another way ? Can I extract a single page, without modify what you need for investigate the problem ? Suggest me the way , please. </pre> </blockquote> <pre wrap=""> Hi mpsuzuki, I have trained on linux and the error disappear, any ideas to solve it on Windows ?</pre> </blockquote> Font searching behaves completely different on Windows and Unix: on Unix fontconfig is used to find the font file, where as under Windows the font must be installed with THIS name, otherwise it takes ALWAYS helvetica. Now I'm wondering a little bit over Your error messages under Windows:  UMVSUI+ArialNarrow-Bold hints to the fact, that the font data should be embedded in the PDF, if the font is NOT emebedded in the PDF font names look more readable, in this case I would expect something like ArialNarrow-Bold without the prefix. So I see two possibilities for Your problem: the PDF is coruupted so that pdftoppm is not able to find the font data in the PDF file or the PDF is ok, the font data is embedded but pdftoppm has a problem with this PDF. Because I can't look into the PDF, You must do it by Your own: Open the PDF with a text editor like textpad and search for UMVSUI+ArialNarrow-Bold. You should find something like (be careful, just a sample) 28 0 obj<</StemV 88/FontName/XCFNXW+MyriadPro-Regular/FontFile3 27 0 R/Flags 4/Descent -250/FontBBox[-157 -250 1126 952]/Ascent 952/CapHeight 674/XHeight 484/Type/FontDescriptor/ItalicAngle 0>> The interesting is the FontFileX, it says, that the font is embedded. and You should then have, in this case, an object 27 0 obj like: 27 0 obj<</Subtype/CIDFontType0C/Length 3506/Filter/FlateDecode>>stream The interesting pattern is here the stream keyword. If You find something like this we probably have bug in pdftoppm, but we couldn't solve that without having the PDF file. Another idea could be to do a "better" substitute, recognize that we search for Arial Narrow Bold (this is the name for it under windows) and take that. This still would give You the error messages but probably gives You a better output! Regards, Thomas <blockquote cite="mid:CAOTGqbYkzPmKjPViKHzXwTgL_VBtuVM5P1xcv=pmebVjYGdncg@mail.gmail.com" type="cite"> <pre wrap=""> I'm using noarch pre compiled version of poppler , downloaded , as suggested by this mailing list, from here: <a class="moz-txt-link-freetext" href="http://download.opensuse.org/repositories/windows:/mingw:/win64/openSUSE_12.1/noarch/">http://download.opensuse.org/repositories/windows:/mingw:/win64/openSUSE_12.1/noarch/</a> Thanks Stefano </pre> </blockquote> </body> </html>