k2pdfopt: optimizes PDFs for viewing on e-readers

[willus]

Originally Posted by kos25k

thanks,just compared and its same!Btw,do you have any idea on how to make reading on my kindle as much similar to prototype? (left) I would like to remove all blank areas from everywhere and smaller gaps between words..

Try unclicking "Re-flow text". That will at least crop out the excess margins on the left and right. It will not put words closer together. k2pdfopt does not have an option to do that.

[kos25k]

Originally Posted by willus

Try unclicking "Re-flow text". That will at least crop out the excess margins on the left and right. It will not put words closer together. k2pdfopt does not have an option to do that.

Thank you.So,also no conversion to other format than pdf can do this? (gaps between words) Btw,if you compare carefully,i think that the gaps between words are smaller on prototype pdf than the converted.

[Quoth]

I was baffled for a while by the Linux command line version and the menu, but eventually realised from Help (h) that I needed to do a final "enter" alone to execute, rather than q to quit, after picking options!

[Quoth]

Is there a recommended command line setting for coloured scan PDFs from archive.org books that are often too dark and slow on eink?

[willus]

Originally Posted by Quoth

I was baffled for a while by the Linux command line version and the menu, but eventually realised from Help (h) that I needed to do a final "enter" alone to execute, rather than q to quit, after picking options!

Is there a recommended command line setting for coloured scan PDFs from archive.org books that are often too dark and slow on eink?

Yeah, sorry about that. I haven't put much energy into the linux menu options in some time. It is a poor man's gui, and not a very good one.

As for converting archive.org books, do you have a particular example in mind that you could post the link to? Archive.org typically displays slowly on e-inks because of the compression used (JPX/JPEG 2000). K2pdfopt can certainly be used to convert to "flate" compession (z-lib), which displays faster, but will result in a larger file size.

[willus]

I recommended this for Quoth's archive.org download (with substantially yellowed pages):

k2pdfopt -mode copy -dpi 400 -c- -g 1 -bpc 4 -cmax -5 sourcefile.pdf

-dpi 400 for 400 dpi. The source file has a dpi of 600.

-c- to turn off color (make it grayscale)

-g 1 to turn off gamma correction (default of -g 0.5 makes things darker)

-bpc 4 to get 4-bit grey scale (the default)

-cmax -5 to force the contrast level to 5 to get the page background mostly white.

[0jay]

Hey Willus, thanks for the tool and all the support you provide here.

I’ve noodled through this thread and haven’t spotted a general guide and am hoping you are aware of one posted somewhere.

I’m working with a plethora of various types of PDFs, hence general guide needed. I had a play and was unable to get things to adjust as expected from some pretty tidy academic journal PDFs (most puzzlingly margins didn’t seem to adjust when I thought I was inputing adjusted margin allowances). I’m not particularly comfortable in Terminal so started with what I could get from making Elipsa screen size settings in the Python GUI.

Can you direct me to something that might give me a bit of a better approach?

Thanks for your time.
Oscar

[Quoth]

The terminal / text mode that has a menu to build option string works far better than the python GUI on Linux. The options do what they say, even though initially I was a bit confused.

Willus is very helpful.

Play with the command line/console version. The Python version is slow and seems to behave unexpectedly. The python GUI is only generating the command line anyway and the preview in it is too slow.

[0jay]

Originally Posted by Quoth

The options do what they say

Play with the command line/console version.

I started with a string spat out by the GUI after configuring screen size there (copy/paste from the GUI into terminal).

Aware of the help menu but found adjusting the margins as per help resulted in the same margins I’d started with.

I general guide’d be good. If there isn’t one it might be an idea to write one. I’d volunteer but have at thesis to write this year, might be time enough after that.

[willus]

Originally Posted by 0jay

Can you direct me to something that might give me a bit of a better approach?

I've long been wanting to make a web page that is exactly what you describe--sort of a decision tree for when to use what options depending on what you want to do with your PDF. I've just never gotten around to it.

I think the main decision to start with is if you want to re-flow the text. If you want to re-flow the text, it sort of sends you down one set of options, whereas if you don't, and you just want to crop out various areas or trim margins, that sends you down another set of options.

Honestly I think the best thing you can do is post a few basic examples of the kinds of documents you want to convert for small screen reading and I can give you advice if you like. You can PM me docs as well.

The video on my web page, though dated, does walk through three types of conversions. It focuses on the MS Windows GUI, but if you keep your eye on the "command-line version of options" box, you can see how the GUI selections translate into command-line options you can use in Linux.

[willus]

Originally Posted by 0jay

I general guide’d be good. If there isn’t one it might be an idea to write one. I’d volunteer but have at thesis to write this year, might be time enough after that.

I had started a guide a while ago and never made it available. I've added it to my help pages, though it is still a work in progress.

[Lukusaukko]

I've been trying to use k2pdfopt to convert a multi-column pdf to single-column format so that I can then pass it on to calibre's conversion tool to convert it to epub with bold and italics (pdftotext which otherwise is ok with multicolumn pdf's loses italics and bolding), which seems to be working fine. But there's one issue I've discovered - setting the margins in the menu doesn't work: if I choose the "m" option and input a comma-separated list, it seems to ignore it, and only gives "-m 0,0,0,0" regardless of what you type in, and if I type in a single number, it sets "-m 1,0,0,0". I can circumvent it by setting up the margins on the command line, though, so it's just the UI.
This is on the Linux 64-bit 2.54 version.

[Lukusaukko]

Originally Posted by Lukusaukko

I've been trying to use k2pdfopt to convert a multi-column pdf to single-column format so that I can then pass it on to calibre's conversion tool to convert it to epub with bold and italics (pdftotext which otherwise is ok with multicolumn pdf's loses italics and bolding), which seems to be working fine.

Actually, scratch that - it's only worked with two fairly simplistic pdf's, anything else just hangs calibre's conversion... still looking for a tool that could convert multi-column pdf to html or epub with character-level formatting intact.

[willus]

Originally Posted by Lukusaukko

... I've discovered - setting the margins in the menu doesn't work: if I choose the "m" option and input a comma-separated list, it seems to ignore it, and only gives "-m 0,0,0,0" regardless of what you type in, and if I type in a single number, it sets "-m 1,0,0,0". I can circumvent it by setting up the margins on the command line, though, so it's just the UI...

Unfortunately I haven't touched the Linux UI in a while, so this is not a surprise. Like you said, you can override with a direct command-line option.

[0jay]

Originally Posted by willus

I had started a guide a while ago and never made it available. I've added it to my help pages, though it is still a work in progress.

Originally Posted by willus

the best thing you can do is post a few basic examples of the kinds of documents you want to convert for small screen reading and I can give you advice if you like. You can PM me docs as well.

The video on my web page, though dated, does walk through three types of conversions.

Thanks willus, I’ll snuffle around and see if I can’t get a better idea from these resources.
Oscar


Sent from my iPhone using Tapatalk