Fonts-be-gone (from my PDF)!

August 28, 2008 3:07 PM

Why are there 20 Mb of embedded fonts in my PDF and how can I make them go away?

I've run into a pretty obscure Word/OSX/PDF problem that I'm hoping to get some insight on. I just finished working on a relatively substantial Word document. The Word document itself is around 11 meg, and I've pretty diligently compressed figures to keep the file from getting too large. When I PDF this file (using the PDF button in the standard OS X print dialog box), it comes out around 28 Mb. When I run "Assess Space Usage" on the file using Acrobat Pro 8, it tells me that the vast majority of the file is embedded fonts.

The "Optimize PDF..." tool will show me the embedded fonts, and indeed there are a million of them. I'm using two main fonts (Meta and Hoefler Text) with a bunch of varieties for, e.g., footers, different header levels, figure captions, etc. In the embedded font list I get mostly entries for those two fonts (a standard entry would be "MetaBoldLF-Roman (Subset)", repeated many tens of times), with a handful of entries for fonts I'm not intentionally using but might have somehow snuck in anyway (Times and Arial). Any attempt to use the Optimize PDF tool to remove fonts will crash Acrobat.

I've printed a single page PDF and tried removing fonts from that. It seems to work okay, although I can't really tell if removing fonts causes problems in the output. I suspect since I have the fonts that are embedded, it will display the same for me regardless of whether or not it's embedded. I've also tried using the "Reduce File Size" option. That reliably crashes Acrobat at the very end of the process. This happens to other people I've had try it, too. I've tried using Acrobat Distiller from Word, but pressing the "convert to PDF" button on their toolbar does absolutely nothing. No idea why not.

So, is there any hope for me? My main goal is to get the file size lower without changing the appearance. I feel like there should be some way to remove most of the embedded fonts and just embed each font once and have it work out. Is this wishful thinking? What's causing this proliferation of embedded fonts? Are they each mapped to Word styles, or something?

I'm happy to give a link to the PDF or Word file if you want to play with them yourself, but since they're loaded with personally identifying information I'd rather not publicly link to them. MeMail me for links.

Thanks!

posted by heresiarch to computers & internet (6 comments total)

On Windows there are all kinds of free "virtual printer" applications that when installed create a fake printer that generates a PDF file when you print to it. Sometimes I've been able to resolve problems with PDF files by simply printing from my PDF viewer app to one of these virtual printers. Is there anything like that for OSX you could try?
posted by XMLicious at 4:25 PM on August 28, 2008

I've MeMailed you for the link, but a thought occurs to me; Rather than using OS X's Print-To-PDF function, I'd be more inclined to save as a PostScript file and let Acrobat Distiller do the work of converting the .ps file into a PDF.

I don't know much about Word's inner workings, but you may be able to "save as" a .ps file for distilling instead. I know that Publisher can do this, but that's on a Windows machine.
posted by lekvar at 5:00 PM on August 28, 2008

Assuming that you used styles to manage your large document, I would imagine that it is due to leftover markup in the document's "code" to indicate styles. See if you can delete some styles from within Word. The multiples of current fonts could be stemming from using different styles of each font – italic, bold, heavy, etc. – because recent versions of Word appear to do use font styles properly.

You are correct thinking that you won't be able to tell when the fonts are removed from the pdf, because you system will just use it's installed fonts to render the page.

In a worst case, the long document may be corrupt, as often happens when they start getting long.

Some options:

• What about opening Acrobat and choosing something like Create from File, and see if Acrobat will do the conversion there.
• Try an online .pdf converter.
• Try a different version of Word (are you using 2008?)
• Save to .rtf and then back to Word

posted by PixelatorOfTime at 5:19 PM on August 28, 2008

Tried:

• Saving to .rtf and back again - no dice, blew the file size up again.
• Tried an online PDF convertor which produced a great slim file, but it doesn't have the fonts in it so it lays out all wrong. Spent a while trying to embed fonts in a file that's already been created but can't for the life of me figure it out.
• Create from File wants a postscript document as far as I can tell.
• Printing direct to a postscript file and then loading that into disiller fails like this "%%[ Error: syntaxerror; OffendingCommand: true,head%%[ Flushing: rest of job (to end-of-file) will be ignored ]%%"
• Don't really want to change word versions - 2008 has ligatures and I loves me some ligatures and don't want to give it up.

Got some other stuff I'm fussing with, but nothing's really working out that well here. I have no idea what I did in this document to make PDFing it so hellish.
posted by heresiarch at 6:52 PM on August 28, 2008

I'm out of ideas at the moment, especially since I don't have access to a Mac right now.

Try posting here. It is a great site made up of print industry members, and they are usually pretty quick with detailed replies. Good luck, and yay for ligatures!
posted by PixelatorOfTime at 7:25 PM on August 28, 2008

Try opening the document in Open Office 3.0, export to pdf and see if that solves your problems.
posted by tesseract420 at 2:58 PM on August 29, 2008