PDF export settings for InDesign: ZIP vs JPG (Maximum), PDF/X or none

[Vince at Teldon]

Hey, folks. Apologies in advance if I'm in the wrong forum, but I figured I'd give it a shot. I work for a nationally distributed magazine and one of my tasks is to generate high-resolution PDFs for sending to print from the layout files. We've been using the same PDF settings for years without too many problems. To summarize:

• Standard: PDF/X-1a:2001, Compatibility: Acrobat 4 (PDF 1.3)
• Bicubic Downsampling to 300 PPI for images above 450 PPI. (colour/grayscale images)
  Compression: Automatic (JPEG) , Image quality: Maximum.
  Crop image data to frames [CHECKED]

After noticing some preflight errors on some of files we've been submitting, and realizing that some of them might be introduced during the encoding process (specifically: black overprinting objects set to OPM 1 are being switched over to OPM 0), I asked the printer if they had any ideas, which they responded by sending me their PDF settings (Job.options file), which are...

• Standard: None, Compatibility: Acrobat 4 (PDF 1.3)
• Bicubic Downsampling to 300 PPI for images above 450 PPI. (colour/grayscale images)
  Compression: ZIP, Image quality: 8 bit [greyed out, no other options available]
  Crop image data to frames [UNCHECKED]

When trying out their settings on our own files and comparing them to what we sent them earlier, overall, the resulting PDFs are massive, with some being upwards of 10x larger than the ones we submitted. On top of that, there are objects at OPM1 that get set to OPM0.

So, what I'd like to know:

• In earlier tests of our settings, I did notice that including PDF/X compliance would generally increase the file size, but not too noticeably. And since PDF/X compliance allows for more uniform files and fewer errors with fonts and colours, is there any reason why I would NOT want to include it in my files?
• Is there any actual visually apparent benefit to using ZIP compression over JPG(Maximum) or even ZIP2000 Lossless? I do realize that JPG compression can cause some severe hits in image quality, but will anyone ever notice the difference once it's in print, like when it's at Maximum quality?
• Is there any particular reason why InDesign won't allow me to select any of the PDF/X settings if I change the compatibility to Acrobat 5-7? Based on what I read, I should be able to select PDF/X-1a:2003 and PDF/X-3:2003 with Acrobat 5 (PDF 1.4).

[Dov Isaacs · Accepted solution]

Some good questions here.

Adobe most strongly recommends the  PDF/X-4 settings for print publishing. Every RIP/DFE using Adobe (or even Global Graphics Harlequin) technology released over the last eight years or more will yield much better results with PDF/X-4 than PDF/X-1a. PDF/X-1a is not color-managed and transparency is flattened in the process of creating PDF. This ties your PDF file to a particular device and resolution. Bad, contrary to what many Luddite and poorly informed print service providers / printers will tell you (unfortunately, many of them “learned” their trade from their elders but never bothered to update their skills based on the major technological advancements of the last decade!

In terms of ZIP versus the Automatic (JPEG) / maximum quality image compression, Adobe's PDF creation differentiates between images that are “vector-like” in content versus those that are photographic. With the Automatic (JPEG) setting, each image is analyzed and based on its characteristics, either ZIP or JPEG compression is used. Unless you are printing exceptionally critical material, we strongly recommend using Automatic (JPEG) which provides an excellent balance between compression and quality. (Many old-timers shutter at the concept of “lossy compression,” which JPEG provides, but most often can't tell the difference between a ZIP-compressed and a JPEG maximum quality-compressed photographic image!)

Not checking the Crop image data to frames buys you absolutely nothing other than a bloated PDF file unless you are using a particular image many places within the same PDF file with different croppings - highly unusual though.

PDF/X-4 files are a bit bigger than the equivalent PDF 1.6 file without PDF/X-4 due to inclusion of a CMYK output intent ICC profile. PDF/X-4 files are typically smaller than PDF/X-1a files of comparable content due to transparency flattening bloating the file size.

JPEG2000 lossless compression will typically give you better compression than ZIP compression, but JPEG2000 compression is not available with PDF/X-1a or PDF/X-3 (neither of which you want to use anyway). The only reason not to use JPEG2000 compression is that some non-Adobe PDF viewers croak on JPEG2000-compressed images.

We've heard of this symptom of overprint OPM=1 being switched to OPM=0. I can assure that this is not Adobe software doing this. Unfortunately, many print service providers have various third-party programs that try to “inspect and fix” PDF files. They often ruin them. PDF files exported from InDesign are typically perfectly fine to directly print. In other words, we believe that it is your printer who is messing up your files, not something from InDesign or any Adobe software. (Quite frankly, if you could find out what the printer is using in their workflow between your submission of PDF/X-4 and a messed-up file with the OPM value changed, we'd be glad to hear about it and follow up on same!)

          - Dov

[Vince at Teldon]

Thanks! Much better information than I got from the printer, which was just to send me their PDF settings with any explanation whatsoever. I went back and did a bit more tweaking of settings, trying out PDF/X-4. One thing I did like about using PDF/X-1 is that it auto-converted RGB objects to CMYK, which is ideal since a lot of our ad clients don't even bother converting the images before submitting, and I'm used to printers only wanting CMYK (with PDF/X-4, it has to be manually selected). But then, our print settings haven't changed in years and I'm doing more reading indicating that submitting in RGB might not be a bad thing (and in a lot of cases, is preferred).

Given the resistance I'm getting from my current printer, I might stay with our current settings (even though it's PDF/X-1a), as we haven't been getting that many problems with it, although it's something I definitely want to revisit in the near future.

[rob day]

One thing I did like about using PDF/X-1 is that it auto-converted RGB objects to CMYK, which is ideal since a lot of our ad clients don't even bother converting

There's no reason why you cant do that with the PDF/X-4 standard even when a printer is requesting all CMYK. This would convert all RGB and CMYK with conflicting profiles into the document's assigned CMYK profile and maintain the X-4 standard. If there were no transparency in the document it would produce the same result as X-1a

The downside of this approach is you have to know the press profile before you make the conversion on export.

[Test Screen Name]

To add to Dov's excellent reply, an absolute rule of PDF/X-1a and PDF/X-3 is PDF 1.3 or older. Among other things this forces transparency flattening which can have huge file size implications. Later PDF standards allow transparency and should be safe if ALL parts of the workflow (including satellite sites and outsourcing) are up to date. You've standardised on very old technology indeed.

[rob day]

I do realize that JPG compression can cause some severe hits in image quality, but will anyone ever notice the difference once it's in print, like when it's at Maximum quality?

You can use Photoshop's Difference blending mode to tell you exactly what the effect of compression is on an image and the Max quality is far from severe and it is doubtful that any offset press is capable of printing the change in pixel structure.

So if you duplicate an image to a new layer and set the mode to Difference the histogram shows that there's no deviation in the pixel because the layers are identical.

If I save out the image as a low quality JPEG, reopen and add it as a layer, there's a significant 5% deviation and the artifacts are visible and printable:

With maximum quality there are no visible artifacts and the deviation is less than a half a percent. The visual noise from a typical halftone screen would certainly hide any compression artifact when the quality is maximum.