k2pdfopt: optimizes PDFs for viewing on e-readers

[drizh]

Hi,
Thank you very much for this great application. It is the best tool for converting that I have ever used before.
It made it possible to read PDF books on my Kindle.
But I still have some minor issues. I tried to find answers in that thread but was unsuccessful.
I use the next command for converting:
k2pdfopt %f -w 1236 -h 1648 -c -wrap+ -hy- -vb -2 -col 1 -o %%s_p.pdf
Could someone help me with that, please?

1. Output file size was significantly increased 5 Mb -> 117.8 Mb

2. Part of the letters moved to the next page. (see 2.png)


3. The gap between words is different if that word is from the end of the line in the original document.See 3_1.png (original document) 3_2.png (output document after wrapping)



The most important thing for me is decreasing the file size, but I can do nothing.
I appreciate any help you can provide.

[drizh]

Hello again.
I played a bit with arguments and results.
So I have got answers to some questions and will put them here:
1. The output file is the image. So I can use greyscale images with 4 bits (the maximum that my reader supports). Use the exact device resolution and DPI. And probably compression.
2. -om 0.03 fixed the issue.
3. Still can do nothing with that.
Have a good day.

[VladimirS]

https://www.willus.com/k2pdfopt/help/options.shtml
Did you allready check this ? Cheers !

[drizh]

Originally Posted by VladimirS

https://www.willus.com/k2pdfopt/help/options.shtml
Did you allready check this ? Cheers !

Hello.
First of all, thanks for your answer.
And yes, I have read the whole forum, wiki, and manual but did not find an answer to my third point.
Could you please point me to the post or parameter I should use to fix issue #3. Many thanks for considering my request.

[willus]

Originally Posted by drizh

1. Output file size was significantly increased 5 Mb -> 117.8 Mb

As you figured out, the output size for bitmapped text re-flow is dominated by the resolution that you choose for the output page along with how many gray levels. If you can avoid text re-flow, you can likely reduce the size much further.

Originally Posted by drizh

2. Part of the letters moved to the next page. (see 2.png)

Where k2pdfopt breaks lines is controlled by the -gtr option. Read up on that option and you can try adjusting it.

Originally Posted by drizh

3. The gap between words is different if that word is from the end of the line in the original document.See 3_1.png (original document) 3_2.png (output document after wrapping)

Word spacing is probably most controlled by the output justification -- the -j option.

[drizh]

Hello.
Thanks for your patience with newcomers.
For file size, I prefer a bigger size to landscape mode. So I set the optimal resolution, dpi, and color depth to reach the optimal file size.

Word spacing is probably most controlled by the output justification -- the -j option.

I have played a lot with that option but without any results. I prefer the full justification, but sometimes I get some results I can't understand.
In general, the text looks great, but some lines have a lot of space at the end. At the same time, other pages look pretty good. And about the gap between words on the same line that can be quite different. Looks like it was calculated before re-flowing the text.
[IMG]1.png[/IMG]
But finally, it's not a big deal. It's awesome that I got the possibility to read PDF files with my reader. I have been searching for that many years.

Thanks again.

[willus]

You could also try adjusting the -ws (word spacing) value.

[MarjaE]

A lot of game books, and comics, have a mix of black text on light backgrounds and light text on black backgrounds.

k2pdfopt tends to make the black text bolder, but the light text thinner, and sometimes breaks it up. I've experimented with gamma and the erosion filter; is there another option to make both dark and light text reasonably bold?

[willus]

I think generally the gamma option (-g) is the best at this, maybe mixed with contrast adjustment (-cmax). There may not be an ideal option for you.

[MarjaE]

Thanks. It looks like a recurring problem with the source pdfs, then.

[tetraxi33]

Hello. I liked this software very much and tried different command line settings for a whole day. Everything is good, but whenever an image is inserted into the text, I could not make the software detect this image as a separate block. Do you think there is a way to do this? I added screenshots. I tried -cg -ch etc. and also -rsf and -to command line options with no success.

[tetraxi33]

Originally Posted by tetraxi33

Hello. I liked this software very much and tried different command line settings for a whole day. Everything is good, but whenever an image is inserted into the text, I could not make the software detect this image as a separate block. Do you think there is a way to do this? I added screenshots. I tried -cg -ch etc. and also -rsf and -to command line options with no success.

I mean I couldn't make the software detect both the image and the text next to it as seperate blocks. So the text next to the image is always in small font size. I beg your help. Thanks beforehand.

[tetraxi33]

Originally Posted by drizh

Hello.
Thanks for your patience with newcomers.
For file size, I prefer a bigger size to landscape mode. So I set the optimal resolution, dpi, and color depth to reach the optimal file size.


I have played a lot with that option but without any results. I prefer the full justification, but sometimes I get some results I can't understand.
In general, the text looks great, but some lines have a lot of space at the end. At the same time, other pages look pretty good. And about the gap between words on the same line that can be quite different. Looks like it was calculated before re-flowing the text.
[IMG]1.png[/IMG]
But finally, it's not a big deal. It's awesome that I got the possibility to read PDF files with my reader. I have been searching for that many years.

Thanks again.

The solution is "-j 0+". If you write this as an additional code it will justify the text without gaps.

[willus]

Originally Posted by tetraxi33

I mean I couldn't make the software detect both the image and the text next to it as seperate blocks. So the text next to the image is always in small font size. I beg your help. Thanks beforehand.

Okay--I made a PDF of your screen shot (at 96 dpi) and was able to get k2pdfopt to parse it the way you want with this command:

k2pdfopt -sm -crgh .01 -gtr .02 -f2p -1 sample.pdf

The -sm option just creates a PDF showing how k2pdfopt parsed the page--this is a useful debugging tool.

The -crgh option default is .013 (1/72) inch. I lowered it. This makes it easier to detect a 2-column sub-section of the document. See this page for how it is used.

The -gtr option makes it easier for k2pdfopt to detect the row breaks since your text lines are kind of crowded together. The default value is .006. The higher value means there can be more "black" in the gaps between rows--they don't have to have pristinely all-white gaps between them.

Finally, -ftp -1 shrinks the figures onto a single page if they are extra tall.

[tetraxi33]

Originally Posted by willus

Okay--I made a PDF of your screen shot (at 96 dpi) and was able to get k2pdfopt to parse it the way you want with this command:

k2pdfopt -sm -crgh .01 -gtr .02 -f2p -1 sample.pdf

The -sm option just creates a PDF showing how k2pdfopt parsed the page--this is a useful debugging tool.

The -crgh option default is .013 (1/72) inch. I lowered it. This makes it easier to detect a 2-column sub-section of the document. See this page for how it is used.

The -gtr option makes it easier for k2pdfopt to detect the row breaks since your text lines are kind of crowded together. The default value is .006. The higher value means there can be more "black" in the gaps between rows--they don't have to have pristinely all-white gaps between them.

Finally, -ftp -1 shrinks the figures onto a single page if they are extra tall.

Thank you sir. It seems that carefully tweaking the settings can make almost any pdf readable.