k2pdfopt: optimizes PDFs for viewing on e-readers

[willus]

Originally Posted by killout

Hello, Willus.

So this is what I executed:
./k2pdfopt -mode fitwidth -ls- -w 2048 -h 1536 -dpi 264 ansible-for-devops-sample.pdf

And in the attach is what I got. They output does has proportions,what I want, width: 2048, height: 1536, so I can read on ipad in landscape mode.The thing is large white margins, which I would like to get rid of, by enlarging content to the size of ipad screen, so desired output pdf is 2048*1534, but with no margins. Is it possible to do?

Yes, you can eliminate the margins. Just make the -dpi value larger, e.g. -dpi 400. At -dpi 264, you are telling k2pdfopt that your device screen is 2048/264 = 7.75 inches wide, whereas your source PDF is only 7 inches wide, so k2pdfopt does not magnify it. It displays it on your iPad so that the page width will physically be 7 inches if you measure it with a ruler--to match the source document size. By specifying -dpi 400, you essentially tell k2pdfopt to magnify the source file. I.e. you are telling k2pdfopt that your device is 2048/400 = about 5 inches wide (even though in reality it is 7.75 inches wide). Now that the "width" of your device (5 inches) is less than the source text width, k2pdfopt will fit the text to your screen, with no margins (assuming you specify -mode fitwidth).

[killout]

Originally Posted by willus

Yes, you can eliminate the margins. Just make the -dpi value larger, e.g. -dpi 400. At -dpi 264, you are telling k2pdfopt that your device screen is 2048/264 = 7.75 inches wide, whereas your source PDF is only 7 inches wide, so k2pdfopt does not magnify it. It displays it on your iPad so that the page width will physically be 7 inches if you measure it with a ruler--to match the source document size. By specifying -dpi 400, you essentially tell k2pdfopt to magnify the source file. I.e. you are telling k2pdfopt that your device is 2048/400 = about 5 inches wide (even though in reality it is 7.75 inches wide). Now that the "width" of your device (5 inches) is less than the source text width, k2pdfopt will fit the text to your screen, with no margins (assuming you specify -mode fitwidth).

Oh, thank you very much for the explanation!
Now I will be slice all my pdf perfectly to read on ipad, just need to adjust dpi depending on input pdf size.
Thanks again for such a great tool

[willus]

Originally Posted by killout

Oh, thank you very much for the explanation!
Now I will be slice all my pdf perfectly to read on ipad, just need to adjust dpi depending on input pdf size.
Thanks again for such a great tool

You can generally err on the side of a larger dpi. If the dpi is too large, the "-mode fitwidth" will still fit the source contents to the device width (effectively reducing the dpi).

[willus]

K2pdfopt v2.40 is released. Tesseract OCR processing is now multi-threaded and so runs significantly faster on modern, multi-threaded CPUs like Intel's Core-i7. See details at the web site.

[Chen Megatroon]

Hi, I meet some problem with the command -col. In the converted file, a paragraph is dived into 2 column, which I don't want. But It seems I can't set -col 1 because there are some 2-column pages in the PDF. So I wonder if I can set diffierent value of max column to diffierent pages (I remember -cbox has this function) or is there any other solution to my problem? Can anyone help me? Thanks a lot.

[willus]

Originally Posted by Chen Megatroon

Hi, I meet some problem with the command -col. In the converted file, a paragraph is dived into 2 column, which I don't want. But It seems I can't set -col 1 because there are some 2-column pages in the PDF. So I wonder if I can set diffierent value of max column to diffierent pages (I remember -cbox has this function) or is there any other solution to my problem? Can anyone help me? Thanks a lot.

Can you post an example from the source file--both the page that has the paragraph that you don't want divided and a page or two with legitimate 2-column text? There is probably a setting you can adjust. Maybe look through the -c... options, e.g. -cg, -cgmax, -ch.

[Chen Megatroon]

Originally Posted by willus

Can you post an example from the source file--both the page that has the paragraph that you don't want divided and a page or two with legitimate 2-column text? There is probably a setting you can adjust. Maybe look through the -c... options, e.g. -cg, -cgmax, -ch.

Thanks for your help! I have fixed it using -cg.

[fortea]

Hi all!
I would like to use k2pdfopt to get a pdf with one slide per page from a document with 6 slides per page.

I tried to use the option "-comax -1" but it doesn't seems to work. I'm using k2pdfopt version 2.34a on Manjaro (Linux Arch based).

I attach a sample pdf.

Thanks in advance!

-------------------------------------------------------------------------------------

I managed to crop correctly the document using "-grid" option, so take this post as an issue segnalation and not as a help request

[willus]

Originally Posted by fortea

Hi all!
I would like to use k2pdfopt to get a pdf with one slide per page from a document with 6 slides per page.

I tried to use the option "-comax -1" but it doesn't seems to work. I'm using k2pdfopt version 2.34a on Manjaro (Linux Arch based).

I attach a sample pdf.

Thanks in advance!

-------------------------------------------------------------------------------------

I managed to crop correctly the document using "-grid" option, so take this post as an issue segnalation and not as a help request

You have to use the -grid or -cbox options, as you correctly figured out. The "-comax -1" option only works when the column divider changes positions, which it does not in your case. See the help page on it.

[Kuang-che Wu]

First, thanks for your great tool. I have a feature suggestion for landscape mode.

Currently, "fitwidth" mode is implemented by rotating the output content 90 degree. While the user can rotate his device to landscape orientation to read the book, the device still think it is in portrait orientation. This posed UI problems. (see fitwidth.png)
1. The page count, menu, and any UI elements are still in portrait orientation.
2. The page-flip virtual buttons are on top and bottom instead of left and right.

My feature suggestion: add a new landscape mode, which rotating page size instead of rotating content. In this mode, I can use the "landscape mode" feature of my device.
Conceptually, this feature could be done by swapping the value of -w and -h.
For example, the original size of my device, paperwhite2, is -w 718 -h 965, so use -w 965 -h 718 instead.

However, "-w 965 -h 718" doesn't work well for me. Because the visible area size in device's landscape mode is different to the size in device's portrait mode, the result page looks distortion due to scaling.

So if you want to add this mode, you have to collect again the size for device's landscape mode like https://www.mobileread.com/forums/sh....php?p=3013298

From my test, "-w 931 -h 690" is good value for paperwhite2.
My test pdf and result are attached.
test_native_landscape.pdf is my source pdf.
k2pdfopt command is -dev kpw -wrap- -bp -h 690 -w 931 -ws 0.2
test_native_landscape_k2opt.png is the result screenshot.

p.s. there are small dark pixels on the right side. I guess they are due to the size of my source pdf is not perfect aligned with k2pdfopt's expectation. (they already exist in test_native_landscape_k2opt.pdf)

[G400S]

I want to optimize a PDF book to use on my Pocketbook Touch Lux 2.

Native mode seems to be the way to go, but whatever I do, it only resizes the titles and the rest of the text remains unchanged (see screenshots). I want to use a 0.5" bottom margin because reasons. I also want to set fixed output font size because the font is too small without it. Is there anything I can do?

Non-native mode works fine and after some experimentation I was able to make the output look decent.


If the full size image doesn't show at first, just refresh the page.

[willus]

Originally Posted by Kuang-che Wu

First, thanks for your great tool. I have a feature suggestion for landscape mode...
...
p.s. there are small dark pixels on the right side. I guess they are due to the size of my source pdf is not perfect aligned with k2pdfopt's expectation. (they already exist in test_native_landscape_k2opt.pdf)

Thanks for the suggestion. The small dots are there to force the kindle to fit them in it's view area, otherwise, at least on my old kindle 2, it would try to auto-scale the visible part of the PDF page to its screen. You can try turning them off with -mc-.

[willus]

Originally Posted by G400S

I want to optimize a PDF book to use on my Pocketbook Touch Lux 2.

Native mode seems to be the way to go, but whatever I do, it only resizes the titles and the rest of the text remains unchanged (see screenshots). I want to use a 0.5" bottom margin because reasons. I also want to set fixed output font size because the font is too small without it. Is there anything I can do?

Non-native mode works fine and after some experimentation I was able to make the output look decent.

Text cannot be re-flowed in native output, so that is why it can't be displayed any larger than it is. You can either rotate the display to landscape mode so that you have more display width (which you have probably already tried), or do what you did and not use native mode.

[G400S]

Originally Posted by willus

Text cannot be re-flowed in native output, so that is why it can't be displayed any larger than it is. You can either rotate the display to landscape mode so that you have more display width (which you have probably already tried), or do what you did and not use native mode.

I see, but why does JUST the title get enlarged? Could it be a different kind of... object or whatever it's called?

[willus]

Originally Posted by G400S

I see, but why does JUST the title get enlarged? Could it be a different kind of... object or whatever it's called?

k2pdfopt is trying to accomodate your request to have a uniform font size within the constraint of native mode output, which means it cannot re-flow text. That means it can only size up the main body of text until it fits the width of your device, and no larger. The titles, because they aren't as wide as the main body of text, can be sized larger and more closely to your desired font size, and still fit completely within the width of your device.

You might try clicking the "Generate marked up source" check box and clicking the preview button. It will show you how k2pdfopt is parsing your source document.