Preview "save as" creates bigger file than input. Why?

I just looked over all of the versions I have. The OS 9, Office 2008 and Office 2011 versions have no restrictions at all. However, the Type 1 PS version of Lucida Sans has Only printing and previewing of the document is allowed (read-only) on all four styles. The following are the four choices you can place on a font:

What you can't see in this screen shot is the checkbox behind the drop down menu choices. That's to Allow subsetting, which is tuned on.

So if the Type 1 PS version of Lucida Sans is what you used, that's what may be causing the issue. Scribus is seeing the copyright restriction of only printing and previewing being allowed, and not looking past that. Since subset embedding is allowed, it should still be copying the font into the new document as is instead of turning it into Type 3 embedded fonts. Preview is compounding the issue by turning all of those conversions into artwork.

This post in the Scribus forum seems to describe the problem: http://forums.scribus.net/index.php?topic=159.0

I still think it is a bug in Preview. The output from Preview can't be selected as text anymore. It is obviously still vector, but acts like an image. I'm starting to dislike PDF more and more.

marty39 wrote:

There are bugs, misfeatures, and missing features in Scribus. If the output from Scribus is buggy, then, strictly speaking, Preview is not required to deal with it at all. But being able to handle buggy output is a market advantage, so it would be better if Preview could handle it.

For Scribus, that's true. If it were Acrobat, it wouldn't be true. Adobe really pulled a fast one with PDF. Release a "standard" and then keep updating it every year. But I digress. From my point of view, I can select text coming in to Preview and I lose the selectable text going out. I call that a bug.

I'm just asking why Preview handles the output from Scribus the way it does. That might involve asking what's special about the output from Scribus. There are some clues in the font listing that Adobe Reader shows, but the full answer might have to come from somebody who can understand what's in the PDF file itself.

Unfortunately, we don't have anyone with that level of knowledge. (If anyone wants to prove me wrong, now is the time to chime in.) I don't think Scribus is doing anything wrong. I have seen image PDFs that have been re-text-ified (for want of a real word) so that you can select the text again. So, there must be some facility in PDF to associate graphics with text strings. I don't like it that Preview loses that. The file size expansion may be a separate issue. I consider the loss of the text more important than the gain of 7 MB.

I tried to create a booklet in SintraWorks PDF Nomad and the same thing happened. Whatever is going on is deep in the OS. You might try telling Scribus to export to a postscript file instead. You might also try to track down some of the older command-line booklet tools like ps2up.

marty39 wrote:

Hmmm. You tried SintraWorks PDF Nomad. I went to their website and they say it requires Lion. This is a Snow Leopard forum. I chose not to upgrade to the Lion.

The older PDF Clerk Pro works in Snow Leopard. I hadn't tried PDF Nomad and though that, if it worked, it might be easier.

I tried telling Scribus to export to PostScript (actually, to print PostScript to a file). Preview and Ghostscript both complained about PostScript errors. I told you Scribus is buggy.

Have you considered Apple Pages? I haven't pushed it for complex layout work, but it handled a 1300 page document pretty easily.

Just for some quick testing results. I opened apr12.pdf in Preview under Snow Leopard and simply did a Save As to a new name. As you noted, the same document is suddenly 7.1 MB.

Opening the same PDF in the Adobe Acrobat Reader (version 10.1.3) and doing a Save As resulted in a slightly smaller PDF.

Preview is doing something very wrong. Giving them a quick look over, etresoft's thought that it may be rasterizing the fonts appears to be correct. The original and copy made from the Reader both have selectable text. The copy from Preview can't select anything. Which means the pages have been turned completely into raster images with no fonts being used at all.

Edit: Nope, wrong. I can't select the text in the 7.1 MB copy made by Preview, but I can if I open the PDF in Adobe Illustrator. So the text still is font outlines. The difference in size is being caused by the loss of the embedded fonts. The Reader maintains the embedded fonts in the new copy, so it remains small. When I open the copy from Preview in Illustrator, it notes that the expected fonts are missing.

So what it looks like Preview has done is turn every bit of text into individual outlines. Which means instead of having all text being displayed from the font's outlines (one outline set for everything), the text instead is now all outlines, as if you had drawn each individual letter by hand in Illustrator, Freehand or other vector editing program.

Is the text still selectable?

Nevermind. I guess we can both edit

Whatever Preview is doing is even funkier than that. As marty said, you zoom way, way in, the file is still obviously a vector file. Obviously Preview is creating a raster, but at least it is kind enough to continue displaying the vector data.

I thought it was something related to the copyright too, but when Scribus actually embeds the fonts, it works as expected. It is Scribus that is converting the fonts into pure outlines, perhaps to stay legal. Then it is Preview that scrambles that.

I filed a bug report on it. If Apple choses to fix it, I sincerely doubt any Snow Leopard users will ever see it.

I sincerely doubt any Snow Leopard users will ever see it.

That's just it. Snow Leopard is what caused the PDF to engorge itself to 7.1 MB in my test. And that simply by opening the PDF in Preview and doing a Save As to a new name. Doing the same thing with the Adobe Reader didn't cause that.

It is Scribus that is converting the fonts into pure outlines, perhaps to stay legal. Then it is Preview that scrambles that.

Nope, I found the real problem. I had only been looking at the documents in Illustrator (not the best way to examine it), Preview and the Acrobat Reader. Silly me, I should have been looking at it with Acrobat Pro.

There are no restrictions on the document at all, and no copyrights on the fonts. It uses 5 TrueType fonts. And the problem, 55 Type 3 PostScript fonts.

Acrobat Pro and the Reader have no trouble moving them into a new document intact. Not surprising, Type 3 PS was an Adobe creation and they can handle them any way they want.

Preview, however, cannot seem to handle Type 3 PS fonts. So every glyph using them is turned into a vector object. Examining the PDF copied by Preview in Acrobat Pro, the only embedded fonts left are the five TrueType fonts. All of the Type 3 PS fonts have been stripped and turned into vector art.

Kurt Lang wrote:

I sincerely doubt any Snow Leopard users will ever see it.

That's just it. Snow Leopard is what caused the PDF to engorge itself to 7.1 MB in my test.

I meant Snow Leopard as opposed to Lion. The original poster doesn't want to upgrade to Lion. If Apple fixes this, I doubt it would show up until 10.8.3 at the earliest.

Nope, I found the real problem. I had only been looking at the documents in Illustrator (not the best way to examine it), Preview and the Acrobat Reader. Silly me, I should have been looking at it with Acrobat Pro.

There are no restrictions on the document at all, and no copyrights on the fonts. It uses 5 TrueType fonts. And the problem, 55 Type 3 PostScript fonts.

Acrobat Pro and the Reader have no trouble moving them into a new document intact. Not surprising, Type 3 PS was an Adobe creation and they can handle them any way they want.

You are just feeding my Adobe conspiracy theories now. They publish the specs and standards for Postscript and PDF and the world rejoices. However, they still retain their market dominance so they immediately start flooding the market with new versions that no other tools can interpret.

Then people show up here and complain about Apple. Adobe is effectively making more work for Apple to do. Sun did the same thing with Java and we all know how that turned out. To add insult to injury, Apple codes to said standards with Adobe doesn't actually support themselves. Then Apple has to fix it again to be compliant with Adobe's de-facto standards. Tell me again why Apple is getting out of the print market? PDF is dead. I can't stand it anymore.