Why are Mac generated pdfs so huge?

I'm happy for my switch to Mac last year, but one thing has disappointed me: pdfs generated on the Mac are huge, generally 10 times the size of corresponding ones generated on my previous PC! I just bought Acrobat Pro, thinking it was the built in OS X pdf generator that was inefficient. Surprise: the pdfs generated by Acrobat Pro are about the same size (sometimes bigger) than those generated by OS X. Notice that I'm talking about pure text documents, with no bitmap graphics. Some examples

text file X: PC .doc 500 kB -> pdf 164kb.
Mac (same file): .docx 1 Mb -> pdf 3 Mb.
Keynote: 7.5 Mb -> pdf 9.7 Mb
Powerpoint (PC) 884 -> pdf 512 kb.

So, on my PC files usually got smaller when converting to pdf, but on Mac they get larger. Making things worse, when pdf files are combined, the size of the combined file is larger than the sum of the sizes (example, two pure text files, 110kB and 150kB becomes a 560 kB file when combined!). Previously I used pdf as a compact format for online distribution, but no more.

My questions are: why are Mac pdfs so huge? And can I do anything about it?

MacBook Air 1.6/HDD, Mac OS X (10.5.6), MacBook (10.4.11) iPhone 2G Dell Windows Vista

Posted on Feb 3, 2009 6:47 AM

Reply
11 replies

Feb 3, 2009 7:20 AM in response to Malcolm Rayfield

At least when using Acrobat pro, the size of combined files is <= the sum of the file sizes. I tried to optimize sizes in Acrobat Pro, but it only changes little (e.g., the 3Mb file I mentioned above can be reduced to 2Mb, still larger than the original word document, and much larger than the 160kB PC generated document). I wonder if it has something to do with fonts? Without knowing too much about it, I understand it so that Apple uses its own fonts, which are incompatible with Adobe and Microsoft. Could this be a reason why Mac generated pdfs are huge?

Feb 3, 2009 8:01 AM in response to ahostmadsen

It may be your settings. Since there are no graphics, then there's no reason not to squeeze the file down as much as possible. Just to make sure we were comparing similar items, I took a text only Word document and kept pasting all 20 pages of text back into it until the file was just a bit over 500 KB.

Saving the file as a PDF using OS X's built in PDF function produced an 864 KB file. Through Acrobat Pro 9's Distiller, using the option for "Smallest File Size", it produced a PDF file of 432 KB. That without embedding any fonts with Acrobat. Which is most likely the main difference in the sizes. OS X automatically embeds fonts, Acrobat must be told to do so.

What I don't know is exactly how OS X embeds fonts. That is, whether they are a subset of the fonts used, or entire sets. The drawback though of leaving fonts out of the document is the likelihood that the PDF file will not display correctly on any system that does not have the same fonts active you used when creating your file.

Feb 3, 2009 10:26 AM in response to Kurt Lang

Kurt Lang wrote:
Saving the file as a PDF using OS X's built in PDF function produced an 864 KB file. Through Acrobat Pro 9's Distiller, using the option for "Smallest File Size", it produced a PDF file of 432 KB. That without embedding any fonts with Acrobat. Which is most likely the main difference in the sizes. OS X automatically embeds fonts, Acrobat must be told to do so.


This mirrors my experience. I can see that the pdf I generated on my former PC has a few fonts embedded. However, the one generated on the Mac from the same document has a million (slight exaggeration) fonts embedded. If I strip all fonts the file size does get considerably reduced, but then it doesn't look good on a PC. Still, it doesn't quite make sense to me. If it's just the embedded fonts that make the file huge, then it should be an additive increase (i.e., a small file gets large, but when you continue to add pages, the file should increase only slightly), as the fonts are reused on new pages. But the file size increase seems to be multiplicative.

Feb 3, 2009 11:55 AM in response to ahostmadsen

I made a new document and saved it as a straight text document to remove all formatting. Then opened it again and assigned Comic Sans MS to the whole thing. Made one paragraph Arial Bold and another Trebuchet MS Bold Italic. Then copy/pasted the entire body text until a .doc file was 500 KB.

OS X made a 548 KB PDF file. Acrobat with no fonts embedded came out to 540 KB. I then made one PDF file with full font sets embedded and another using subsets with Acrobat. Both came out to 2.8 MB.

This was a very simple document using nothing more than letters A-Z and a-z with the usual punctuation marks. Digging a little in Acrobat Pro, it shows that the fonts were not embedded just once, but for every instance of a break in style. Since I copy/pasted the same text about 40 times, it embedded the same fonts over and over.

Hard to believe that OS X would be smarter than Acrobat Pro's Distiller and embed used fonts only once, but that sure seems to be the case.

Feb 3, 2009 5:26 PM in response to Kurt Lang

This is interesting. I repeated your experiment. A word file with two fonts alternating hundreds of times. The word file is 43kB. The OS X generated pdf file 76 kB. It includes two fonts, as expected. The Acrobat Pro generated file 1Mb!!! Hundreds of repeated fonts included. Hey, this makes Acrobat Pro essentially useless. I just wasted my money! I only got Acrobat Pro in the hope it would reduce my file sizes. Perhaps I can find a way to uninstall it (it's a site license with a fee per copy, maybe I can delete it and say I don't have a copy and not pay). The only good thing about Acrobat Pro for me is that at least it doesn't increase the file size when combining files, as OS X does.

What remains is that the pdf files I generated on my PC were much smaller. But I can see now that perhaps it is not a systematic problem in OS X. Some of it could have to do with embedded equations. It seems that Mac Word includes a copy of fonts for each equation when generating pdf, while PC Word has only one copy of math fonts for all equations. Similar, in Keynote I have embedded equations as PDF from LaTeX, and again it appears it includes the font for every equation when generating pdf. Not sure what I can do about that. Does a program exist that can consolidate fonts in pdf files? It doesn't seem Acrobat Pro can do it.

Feb 3, 2009 7:27 PM in response to ahostmadsen

Does a program exist that can consolidate fonts in pdf files? It doesn't seem Acrobat Pro can do it.


Not sure, but then, I haven't looked into it. This also may simply be an ugly bug in version 9. I don't recall version 8 having this issue. It certainly makes no sense to embed the same font repeatedly.

I too looked through Acrobat's settings (somewhat cursively) and couldn't find a quick solution. Guess I'll mosey over to Adobe's site and see if others have reported the issue on their forums.

Edit: Found this thread at Adobe's forums.

http://www.adobeforums.com/webx/.59b6b855/0

Sounds backwards to set the embedding option to 0%, but can't hurt anything to try. It's late here, I'm going to give it a try tomorrow.

Hey, this makes Acrobat Pro essentially useless. I just wasted my money!


Heh! Well, that does depend on what you're making a PDF of. For documents that are nothing but text, Acrobat is certainly overkill. But if you need to create print quality PDF files with high resolution images, then it's a necessity. OS X's built in PDF generator always knocks embedded images down to highly compressed, low resolution JPEGs. You have no control over it.

Feb 4, 2009 6:48 AM in response to ahostmadsen

Aha! And the note on Adobe's forum about TrueType fonts is exactly the issue. Did some more experimenting this morning to prove what happens when.

I once again started with some simple TrueType fonts with the text repeated about ten times in a Word document so each type face was also repeated. Note that this happens with any TrueType font. Both old 8 bit fonts and OpenType TrueType fonts.

As it did yesterday, the same fonts were embedded over and over no matter how I set things in the Distiller. From a 208 KB Word document, OS X would create 232 KB PSD file and Acrobat a 1 MB file.

So I started over using only PostScript fonts. Two Type 1 PostScript fonts and one PostScript OpenType font. From a 172 KB Word document, OS X produced a 180 KB file, but Acrobat's was only 100 KB. Success! Checking in Acrobat Pro, each font was indeed only embedded once. With the TrueType fonts, they were embedded, as before, for as many instances of a change in type face.

Now, you can clean the TrueType versions up, but it takes an extra step. A real nuisance if you have to do it to many documents. You can open the resulting PDF from the Distiller in Acrobat Pro and do a Save As. Choose "Adobe PDF Files, Optimized" and click on the Settings button. Click on the Fonts heading at the left and unembed all fonts. Then click one of each font face and click the retain button. Click OK and save your file. That took what was a 1 MB PDF file down to 140 KB.

Or save yourself a lot of monkeying around and use only PostScript fonts. 😉

Feb 4, 2009 2:34 PM in response to Kurt Lang

Thanks for all your research on this. It makes things a little more clear. But there are still things I don't understand. For example, I have a keynote document, 7.5 Mb, containing text, vector graphics (eps) (but no bitmaps) and equations (imported as pdf). With OS X pdf generation, I get a 9.7Mb document, with Acrobat Pro 2.8 Mb. Both have a million fonts embedded, but the Acrobat more so. Still, the Acrobat document is smaller. So, it's not just fonts.

This thread has been closed by the system or the community team. You may vote for any posts you find helpful, or search the Community for additional answers.

Why are Mac generated pdfs so huge?

Welcome to Apple Support Community
A forum where Apple customers help each other with their products. Get started with your Apple Account.