Why are my PDFs becoming heavyweights?

October 23, 2009 9:55 AM

Why do minor changes to a PDF in Photoshop CS3 radically increase the file size (from 200k to 6.5MB)?

I'm no graphic artist (as evidenced by this question), but I'm an intermediate user of Photoshop and frequently asked to make changes to existing documents. Right now, I'm trying to change the address on an old PDF to our new address. The old PDF is roughly 200kb before editing, but 6.5MB after editing. What am I doing wrong?

Here's my step-by-step:

1) Open PDF in Photoshop CS3
2) Create new layer
3) Use text tool to type out the new address (basic Arial font) over top of the old address
4) In background layer (the original PDF), use the paintbrush tool to cover up the old address
5) Save as PSD
6) Flatten layers
7) Save as PDF, without Photoshop editing capabilities


The doc is a full sheet, 8.5x11; It's a photo-heavy step-by-step guide, but the pictures aren't high-res or great quality. Photoshop CS3, Adobe Acrobat, Win XP SP3. If you need any other info or want the files themselves, I'll be glad to provide.

posted by steeb2er to computers & internet (14 answers total)

When you open the PDF, what settings do you choose in the Rasterize dialog box?
posted by odinsdream at 10:04 AM on October 23, 2009

When you open the PDF in photoshop it rasterizes the content (turns it into pixels), pixels take up more disk space than vectors (which basically boil down to mathematical formulas.

The better way to manipulate a PDF is to open it in Illustrator, though this can create different problems if you're missing the documents original fonts.
posted by wabbittwax at 10:05 AM on October 23, 2009

If you have Adobe Acrobat Pro, try the TouchUp Text Tool, under Tools, Advanced Editing.
posted by hydrophonic at 10:22 AM on October 23, 2009

Another problem to watch out for is that photoshop uses rich blacks by default. This may or may not be a problem depending if your docs are being printed and by whom.
posted by travis08 at 11:08 AM on October 23, 2009

odinsdream - There is no rasterize dialog(or I'm blind?). There's an import box, which I leave at the defaults. Crop to bounding box, Anti-Aliased checked, 8.5"x11", 300dpi, sRGB color mode, 8-bit. Now that you've all got me thinking about rasterization, it totally makes sense that a 300dpi full-page couldn't be 200kb.

wabbittwax - Tried this, but Illustrator doesn't seem to acknowledge that there's text. I don't know the PDF's history, but my guess is that it was scanned in or imported from a Word doc. Maybe Illustrator is viewing it as one giant graphic ... ?

hydrophonic - I tried, and the mouse cursor doesn't change. It's just the regular pointing arrow. I tried it with another PDF and was able to make it work. This supports my thought that this isn't a "native" PDF.
posted by steeb2er at 11:34 AM on October 23, 2009

I no longer have Adobe Acrobat on my computer, but I recall there that you could once use Document -> Reduce File Size. Don't know if that option still exists, but it would successfully pare down PDFs that Photoshop had engorged.
posted by coolguymichael at 11:36 AM on October 23, 2009

The old PDF is roughly 200kb before editing, but 6.5MB after editing. What am I doing wrong?

Are you doing anything wrong? Hard drive space is cheap, so while the change in size is odd, is it really a problem?

There is no rasterize dialog(or I'm blind?). There's an import box, which I leave at the defaults. Crop to bounding box, Anti-Aliased checked, 8.5"x11", 300dpi, sRGB color mode, 8-bit. Now that you've all got me thinking about rasterization, it totally makes sense that a 300dpi full-page couldn't be 200kb.

That's a rasterization dialog, not an import dialog. I know it says "Import PDF", but since Photoshops is a pixel based program, it's going to rasterize anything it imports.
posted by Brandon Blatcher at 11:55 AM on October 23, 2009

If the PDF is graphics-heavy, my guess is the smaller original was saved with pretty crappy graphics compression quality values. Then, when you open it up in PS and later re-save it as a standard or high quality PDF, you're re-saving those crap graphics in a much higher quality, but you wouldn't be able to tell that by appearance because they were already degraded. Nonetheless the new file takes a big hit in terms of size. Like re-saving a poor-quality video as HD-quality video file -- you're not going to improve how it looks, but you're going to end up with a huge file.

Obviously the best solution would be for you to have access to the original file the PDF was created from, but I already know (from having been asked to do similar things by my coworkers) the answer to that question, and that those asking for changes might not even know there *was* an original file. The real answer to people who want you to "edit PDFs" is: "You're using PDFs incorrectly." But people sure don't like to hear that.
posted by aught at 12:03 PM on October 23, 2009

Seconding aught: the PDF format is a delivery format, not a working format. Try your best to get access to the original document file, and try to impress upon your co-workers that editing PDF files is a bad idea and won't result in anything like the quality you'd otherwise be capable of.

If you absolutely must get your final PDF down in size, try decreasing the dpi of your working image somewhat. You can also switch modes around (grayscale? indexed color?) and save some major filespace that way too.
posted by Aquaman at 12:06 PM on October 23, 2009

but Illustrator doesn't seem to acknowledge that there's text.

It's possible that it is all just one big graphic, but that doesn't really square with the original small file size. It's also still possible that your text is buried under a number of clipping masks and compound paths in the file (as most PDFs are in Illustrator). If you open the layer menu in Illustrator, is there just one layer with no sub-layers or anything else? Another possibility is that the text was outlined before it was output as a PDF, which would mean there's certainly no editable text there. But that wouldn't stop you from using the same process you described doing in Photoshop. Create a white box, cover up the text you want to edit, and type new text in its place.

Are you doing anything wrong? Hard drive space is cheap, so while the change in size is odd, is it really a problem?

If he started out with a good quality PDF and he's rasterizing it, then I'd venture that he is doing something wrong, because rasterizing it inevitably degrades its quality.

By all means, if you can get the original files, do so.
posted by wabbittwax at 12:10 PM on October 23, 2009

coolguymichael - Reduce file size helped alot (down to 600k, good enough for now). Thanks!

Everyone else ... I'll go get me some classes or books on these programs. I've got lots to learn.
posted by steeb2er at 12:25 PM on October 23, 2009

Yes, if I had the original files, I'd obviously be working with those. But this has been a big 'game of telephone' as is, so I'm just trying to solve the problem and move on.

Hard drive space is cheap, but I wanted to understand what I was doing to suddenly jack up the file size. MeFites to the rescue!
posted by steeb2er at 12:35 PM on October 23, 2009

The old PDF is roughly 200kb before editing, but 6.5MB after editing. What am I doing wrong?

Are you doing anything wrong? Hard drive space is cheap, so while the change in size is odd, is it really a problem?

One good reason is that it's annoying to e-mail a 6.5MB file. Some e-mail systems would reject it outright.
posted by odinsdream at 12:39 PM on October 23, 2009

I think the illustrator suggestion is probably the best answer though, since as everyone has pointed out, it's the vector vs. bitmap issue that's getting you. But if you don't have illustrator handy, try the following using photoshop/image-ready.

After you flatten the layers in photoshop, send it to image ready (there should be a button at the bottom of the toolbar), then in image-ready, you can use the optimization tools to shrink the filesize, then save the optimized version. I've gotten very good filesize reduction with very minimal, if any, obvious visual degradation.
posted by johnstein at 3:14 PM on October 23, 2009