I feel like it would be negligible degradation for this purpose. Still might not anonymize whomever shares it though, could be watermarked with the same Metadata (https://en.m.wikipedia.org/wiki/Machine_Identification_Code) without being noticeable to the naked eye
You’re pushing it through one system that converts a PDF file into printer instructions, and then through another system that converts printer instructions into a PDF file. Each step probably has to make adjustments with the data it’s pushing through.
Without looking deeply into the systems involved, I have to assume it’s not a lossless process.
You should maybe look a bit more into it. How do you think commercial printers or even hobbyists maintain fidelity in their images? Most images pass through multiple programs during the printing process and still maintain the quality. It’s not just copy/paste.
As a trivial example, if you use the wrong paper size (like Letter instead of A4) then it might crop parts of the page or add borders or resize everything. Again I’ll admit, in 99% of cases it doesn’t matter, but it might matter if, say, an embedded picture was meant to be exactly to scale.
My friend, I worked in commercial printing for 2 decades. You’re still making assumptions that are wrong. There are ways to transfer files that are lossless and even ways to improve and upscale artwork. Why do you care so much about this?
You’re still wrong. the only place where it could cause quality loss if embedded bitmap images are compressed with lower quality settings (which you can adjust). PDF is a vector format, i.e. a mathematical description of what is to be rendered on screen. It was explicitly designed to be scalable, transmittable and rendered on a wide variety of devices without quality loss.
No point discussing this if neither of us is going to prove it one way or the other.
Bitmaps are actually a key part of what I was thinking about, so you agree with me there it seems. There’s also the issue of using the wrong paper size. .IIRC Windows usually defaults to Letter for printing even in places where A4 is the only common size and no one has heard of Letter, and most people don’t realise their prints are cropped/resized. This would still apply when printing to PDF.
My point is that all these things can be controlled in the settings of your PDF printer driver. So it’s not completely straightforward but definitely doable.
I don’t understand the “that’s no how PDFs work” criticism.
Removing data from the original file is the whole point of the exercise! Of course unique tokens can be hidden in plain sight in images, letter spacing, etc. If we want to make sure to remove that we need to degrade the quality of the PDF so that this information is lost in said lossy conversion.
I feel like this will cause quality degradation, like repeatedly re-compressing a jpeg. Relevant xkcd
Edit: though obviously for most use cases it shouldn’t matter
I feel like it would be negligible degradation for this purpose. Still might not anonymize whomever shares it though, could be watermarked with the same Metadata (https://en.m.wikipedia.org/wiki/Machine_Identification_Code) without being noticeable to the naked eye
Why would it cause degradation? You’re not recompressing anything, you’re taking the visible content and writing it to a new PDF file.
You’re pushing it through one system that converts a PDF file into printer instructions, and then through another system that converts printer instructions into a PDF file. Each step probably has to make adjustments with the data it’s pushing through.
Without looking deeply into the systems involved, I have to assume it’s not a lossless process.
You should maybe look a bit more into it. How do you think commercial printers or even hobbyists maintain fidelity in their images? Most images pass through multiple programs during the printing process and still maintain the quality. It’s not just copy/paste.
Removed by mod
They maintain a high quality but not lossless.
As a trivial example, if you use the wrong paper size (like Letter instead of A4) then it might crop parts of the page or add borders or resize everything. Again I’ll admit, in 99% of cases it doesn’t matter, but it might matter if, say, an embedded picture was meant to be exactly to scale.
Lossless is the default for print output.
My friend, I worked in commercial printing for 2 decades. You’re still making assumptions that are wrong. There are ways to transfer files that are lossless and even ways to improve and upscale artwork. Why do you care so much about this?
“There are ways” ≠ this is what happens by default when done by the average user
Those printer instructions are called Postscript and they’re the basis of PDF.
You are thinking that the printing process will rasterize the PDF and then essentially OCR/vector map it back. It’s (usually) not that complicated.
Unless of course you print everything and then scan it again, like this guy probably does.
That’s not how PDF works at all.
See my reply to another comment
You’re still wrong. the only place where it could cause quality loss if embedded bitmap images are compressed with lower quality settings (which you can adjust). PDF is a vector format, i.e. a mathematical description of what is to be rendered on screen. It was explicitly designed to be scalable, transmittable and rendered on a wide variety of devices without quality loss.
No point discussing this if neither of us is going to prove it one way or the other.
Bitmaps are actually a key part of what I was thinking about, so you agree with me there it seems. There’s also the issue of using the wrong paper size. .IIRC Windows usually defaults to Letter for printing even in places where A4 is the only common size and no one has heard of Letter, and most people don’t realise their prints are cropped/resized. This would still apply when printing to PDF.
My point is that all these things can be controlled in the settings of your PDF printer driver. So it’s not completely straightforward but definitely doable.
You can ask ChatGPT to spit out the latex code
What
I don’t understand the “that’s no how PDFs work” criticism.
Removing data from the original file is the whole point of the exercise! Of course unique tokens can be hidden in plain sight in images, letter spacing, etc. If we want to make sure to remove that we need to degrade the quality of the PDF so that this information is lost in said lossy conversion.