[PDF] Text to image: pdftocairo or pdftoppm?

[Shohreh]

Hello,

Calling the Poppler experts

As an easier way than trying to turn it into EPUB, a work-around to read a PDF on an e-reader is to simply turn each page that contains complicated layout into an image. I know it's not as nice as real text, it won't be flowable, but it could be good enough to read.

I see that Poppler offers two applications for this purpose: pdftocairo and pdftoppm.

I have a few questions:

1. Which application do you recommend?
2. Should pages be saved as JPG, PNG, or SVG?
3. What settings should I use for an e-reader whose screen is 768x1024 pixels?
4. How to crop pages and get rid of the circles on the sides (used by printers for adjustement?)?

Thank you.

[kovidgoyal]

There's already software that does this, splitting up pdf into reader screen sized images. k2pdfopt IIRC

[Shohreh]

Thanks for the tip.

Using the default settings, some parts were a lot better than Poppler, but unsurprisingly and just like Poppler, it has a hard time with tables and insets.

I'll go through the settings and see if it can do better.

[Shohreh]

Since, unlike Poppler, the Windows version of MuPDF is up to date, I'm giving mutool as try.

Unless I missed it, the only way to turn pages of a PDF into pictures is to a) split the file into individual PDF files (one page = one PDF), b) turn the PDF into a PNG, c) turn the PNG into a PDF, and then d) merge all the PDF files back into a single file.

1. For some reason, the 40KB PNG turns into a 2MB PDF:

#212 dpi, 768x1024
mutool draw -r 212 -w 768 -h 1024 -o 30.png input.pdf 30

#From 40KB to 2MB!
mutool draw -o 30.pdf 30.png

2. Is there a way to crop the pictures, as way to get rid of the borders?

Thank you.

--
Edit: Found it.

mutool convert -O compress -o 30.pdf 30.png