Hello!

A PDF file at heart is just a bunch of objects referenced by a cross reference table written to disk.
There are two ways for a PDF Writer to save a cross reference table:

Save cross reference sections made up of sub sections containing the objects
Save cross reference streams utilizing object streams since PDF 1.5

The main job of pdfcpu optimize is getting rid of redundant resource objects.
What that means is eg. if your file is embedding the same font multiple times perhaps for each page then pdfcpu makes sure the font file is embedded only once (and fixing all references to old objects) or for instance if the same image appears multiple times then pdfcpu also can get rid of unneeded copies of this image.

Once these (and other) optimization steps are done pdfcpu writes back the cross reference table using 2) utilizing the DEFLATE stream compression filter which is zlib/deflate under the hood.

To answer you question:
Unless your PDF files are really old chances are they are already saved using cross reference streams and in this case you are not really gaining much out of pdfcpu optimize so
your mileage may vary. It really depends on how your file was written and to what extent it uses embedded fonts and images.

In my case I had to deal with a PDF with a filesize of a couple of gigabytes written by a popular Java PDF library and after optimization and removing embedded fonts it was a lightweight of a couple of hundred megabytes. This was actually my motivation for starting pdfcpu in the first place. So for me it worked out well.

Unfortunately pdfcpu does not provide any magic switches but in pkg/pdfcpu/configuration.go there are some settings but they are already preconfigured for achieving the best compact results in the most efficient way.

Let me know if you have further questions.
Thank you for using pdfcpu 💚

Uh oh!

Benefits of pdfcpu optimize #135

Description

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions