PDF File Size

[taks]

A cursory search seems to indicate nobody mentioned this, so forgive me if I'm wrong, but why are the remastered books so much larger than their original counterparts despite maintaining the same (or similar) page counts? For example, the new Dark Archive is 246MB, but the original is 34MB, while both have the same page count.

Not complaining, just curious what the additional content is.

[moosher12]

If I was to hazard a guess, it could be they worked the new book off of a raw file, and forgot to compress the images going into the Remastered version. I'd imagine if anything is taking up room on the file, it'd be the multitude of images.

[thenobledrake]

I thought frequent complaints of the crunchy images in PDF products had led to Paizo doing less compression on files, but I could be misremembering.

It certainly would explain the difference, though.

[Ascalaphus]

My suspicion is that they're missing some optimizations for repeated images, like page border decorations. When I run ghostscript to optimize with medium settings I can reduce file size by around 80% consistently without it looking any different to me.

[Tridus]

Yeah that is absurdly huge for a PDF like this. They really need to run optimization on it.

[AFigureOfBlue]

This isn't a brand new thing - the first volume of Shades of Blood and second volume of Triumph of the Tusk are each around 200MB despite having less than half the page count of Dark Archive.

I think it started shortly after Rage of Elements, which was famously over-compressed to the point that most of the illustrations had a very visibly grainy/pixelated look to them. It feels like perhaps Paizo went too far in the other direction after that. Either that or it is something like repeated images being repeated in the data when they shouldn't be.

But it is an issue for sure - on the weaker of my two computers, navigating some of these larger PDFs is actively laggy compared to the smaller PDFs that we used to get.

[Karys]

I recall this being somewhat common with Starfinder 1e PDFs as well, particularly APs. Several ranging between 100-300 MB. Catches me off guard every time I see it happen, and mildly annoying to make space for on some devices sometimes.

[Moth Mariner]

yeah my phone's pdf reader has been struggling with the latest ones for sure.

[taks]

Huh, ok. Well, I just happened to be looking at that one in particular, didn't realize it's been a thing for a while. Thanks all.

[Conscious Meat]

Necroing this to note that

* this is ongoing and perhaps getting crazier (SF2e Player Core is ~370 MB; even the chapter-by-chapter files have files exceeding 50+ MB ).

* because of this I did a bit of poking and found an open-source tool that can use Ghostscript to recompress PDFs with configurable DPI for the art ( https://nikkhokkho.sourceforge.io/?page=FileOptimizer )

Downsampling to 72 DPI using that reduced SF2e Player Core to... ~12 MB. The art is obviously significantly reduced in quality, but the text is fine.

[Ascalaphus]

You can optimize PDFs with ghostscript without sacrificing visible quality. I suspect there's something wrong with the way background elements (borders etc.) get duplicated for every page they're used on.

[JiCi]

Consider you guys lucky that it's only the file size that got higher... and not the price...

Yeah, with all what's going on about digital media, I wouldn't be surprised if books of all kinds become as expensive as their physical versions...

[Moth Mariner]

Probably true, inflation makes everything go up in price. Paizo staff need to pay rent and buy food. But that's not relevant to this conversation :)

Hopefully we can get some better compression on the PDFs from Paizo sooner rather than later.

[Conscious Meat]

Yeah, would be nice if that's doable w/o compatibility issues wrt. commonly-used readers.

FWIW, GS-based optimization *without* downsampling dropped SF2e Player Core to ~150 MB. That's still absolutely massive, but less crazily, and suggests that the single-chapter files wouldn't necessarily be ludicrous.

Guns and Gears, Treasure Vault and Dark Archive (all remastered) still all exceeded 200 MB with a no-downsampling optimization. Downsampling G&G with the 150dpi (e-book) preset resulted in it dropping from ~230 MB to ~24 MB. That seems to be... manageable performance-wise for a Kindle Paperwhite, assuming that one is willing to pinch and zoom a lot in order to be able to read the text given that it's in a two-column format.