[scribus] scribus Digest, Vol 91, Issue 30
williambader at hotmail.com
Sun Nov 1 01:35:31 UTC 2015
> To: scribus at lists.scribus.net
> From: gpittman at iglou.com
> Date: Sat, 31 Oct 2015 19:11:03 -0400
> Subject: Re: [scribus] scribus Digest, Vol 91, Issue 30
> On 10/31/2015 06:38 PM, Great Dane wrote:
> > On 1 November 2015 at 02:16, Johannes Deutch <zelbstzz at gmail.com> wrote:
> >> For Windows, the following thread, "How to convert PDF File into smaller
> >> size" will give you a couple of leads:
> >> http://www.techsupportalert.com/freeware-forum/home-office-and-productivity/2442-how-to-convert-pdf-file-into-smaller-size.html#post16978
> > This prompted me to try something to see if it worked. (BTW my first post,
> > I am no expert, but hopefully may be of some use to the group)
> > I have Windows 7 Professional, and use Acrobat Pro 10. I opened a 13 page
> > document authored in Scribus and exported as a pdf, which was just over 3mb
> > in size in Acrobat. I then clicked Print, chose Adobe pdf, clicked on
> > Properties, and then the Adobe pdf settings, changed the Default Settings
> > to smallest file size, then OK and Print, saving it with an altered file
> > name. The resulting file was down to just under 700kb.
> > Looking at both files at 100%, I can hardly see any change of quality of
> > the images, and none for text. Certainly if your “clientele” have problems
> > handling larger file sizes in emails, this seems to work for me.
> > Trying the Save As>Reduced File Size pdf, hardly made any significant
> > difference in file size, whereas this little procedure certainly did.
> I can understand that this would work, but it's a black box solution.
> What does Acrobat do to accomplish this?
> We know that the size of images will increase PDF size, also that
> embedding fonts will. Are there other tricks?
Most applications that claim dramatic decreases in the size of PDF files use examples where the original PDF has a large amount of embedded images that are either uncompressed or have an excessively high resolution. Solutions using ghostscript usually help with both of those issues using the preset values of ps2pdf -dPDFSETTINGS=/prepress as listed in http://ghostscript.com/doc/current/Ps2pdf.htm
After that, the next decreases come with subsetting fonts and with combining font references (for example, if the original PDF embeds a new copy of the font once for each object that uses it or once for each page that uses it instead of embedding only a single copy for the entire document).
After that, you can sometimes gain a little by combining duplicate PDF objects.
The pdfsizeopt project https://github.com/pts/pdfsizeopt has a python script that attempts a number of transformations including converting Type1 fonts to Type1C, unifying subsets of the same Type1C font, converting inline images to XObjects (which makes them easier to unify), optimizing images with pngout and jbig2, removing unused objects, removing duplicate objects, removing duplicate trees of objects, removing gaps in object numbers, and reordering most-referenced objects to come early. The author has charts showing the effectiveness of the various steps from a 2009 talk at https://github.com/pts/pdfsizeopt/blob/master/pts_pdfsizeopt2009_talk/pdfsizeopt_charts.pdf He also published an article at http://pdfsizeopt.googlecode.com/files/pts_pdfsizeopt2009.psom.pdf
I had issues with pdfsizeopt failing, producing incorrect output, and taking very long to run, but it is interesting to experiment with.
Regards, William Bader, Director of Research and Development at SCS, http://www.newspapersystems.com
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the scribus