850 kB is not too much, that could be some document level metadata. Are those 4 pages coming from a larger document by chance, possibly with a table of contents or something like that?

Without looking at the file it will be hard to dig deeper here.

There may be other structures on the file level which would end up in each output file. (And I think we merged a PR to clear one of those, but I'm not entirely sure and it may have gotten into a release yet.)

Found it, it is in 1.12, please update to that (or use the flatpak) and try with the box unchecked, see #1156

I'd expect the multi page and single page documents to be roughly the same size in this case

I agree that duplication might not be good to be used to reproduce the problem. However, it still shows that while the original has 2 MB, the "duplicated" one with 6 pages has 3.5 MB, and every extracted single page also has 3.5 MB. I'd expect it to have a filesize of around 2 MB instead.

Are those 4 pages coming from a larger document by chance,
possibly with a table of contents or something like that?

It is a print@home ticket for 4 people from a local theater. No TOC, just 4 pages with mostly the same content, but different ticket IDs, barcodes, etc.. Same graphics, address, etc.

Found it, it is in 1.12, please update to that (or use the flatpak)
and try with the box unchecked, see
#1156

Tried with 1.12.1 and the setting unchecked: No difference in file sizes for both my original PDF and the base64image PDF mentioned above.

Without looking at the file it will be hard to dig deeper here.

The base64image PDF won't help?

Thanks for the input. Though honestly I don't see this as a relevant issue and it is unlikely that I will spend more time looking into it. (But maybe someone else will, in this case I'd recommend checking with the qpdf cli to figure out if that is a PDF Arranger or a qpdf issue.)

As you are saying the tickets contain mostly the same graphics, it is likely that they get compressed or deduplicated in the PDF; saving it as single PDFs prevents that and therefore the sum of the parts must be larger than the original file. We are - in my view - also not talking about a massive amount of data here which further limits my interest.

I just loaded the base64image-1.pdf in PDF Arranger and saved it without changes. Came out with 3.5 MB. Qpdf always rewrites the whole file and due to compression settings or other details the size can vary.

That explanation makes sense. Thanks for taking time to look into this, much appreciated =)