k2pdfopt: optimizes PDFs for viewing on e-readers

[willus]

Originally Posted by abe1

I use that settings because that's how converting my bangla books works properly. Check the differance between 0.200 vs 0.001

The other settings are pointless however. I just checked. I don't know which settings does what. Just checked/unchecked settings as I went along till I found a good one. It's not perfect however. But it's good enough.

Look closely at your example--see attached. By using -ws -0.001, you've encouraged k2pdfopt to break a "double-row" of text in the middle of that double row--this causes the words to be out of order in the converted PDF. What gives a correct conversion is to increase the value of -gtr, e.g. -gtr 0.1 instead of -gtr 0.05.

Also--the -as option isn't necessary on your particular source document since it is a computer-generated PDF and is perfectly straight. It doesn't hurt except that it burns extra cpu. And why are you using -r to convert right-to-left? Isn't Bengali read left to right?

I don't mind at all that you shotgun your solution until you find something that works, or that you are trying to post helpful answers in this thread, but I do think you should fully understand your settings before recommending them to other users in this forum.

[abe1]

Originally Posted by willus

Look closely at your example--see attached. By using -ws -0.001, you've encouraged k2pdfopt to break a "double-row" of text in the middle of that double row--this causes the words to be out of order in the converted PDF. What gives a correct conversion is to increase the value of -gtr, e.g. -gtr 0.1 instead of -gtr 0.05.

Also--the -as option isn't necessary on your particular source document since it is a computer-generated PDF and is perfectly straight. It doesn't hurt except that it burns extra cpu. And why are you using -r to convert right-to-left? Isn't Bengali read left to right?

I don't mind at all that you shotgun your solution until you find something that works, or that you are trying to post helpful answers in this thread, but I do think you should fully understand your settings before recommending them to other users in this forum.

I tried with -gtr 0.1 but it doesn't work. It makes these weird output..
This only happens when using the "gtr" option.. And if I don't use "gtr" some lines becomes smaller as I've shown before.




This is how it should look for example. The last line is shown.



This happen in my settings too but to a much lesser extent. Please help me fix it if you can. I understand this is a foreign language. What's happening is the software is separating the top/bottom 'signs' from them.

I'm attaching the file so you can take a look.

Manto.pdf

Thanks.

P.S. Yes Bangla isn't right-to-left. No idea how that got turned on.

[willus]

Originally Posted by abe1

I tried with -gtr 0.1 but it doesn't work. It makes these weird output..
This only happens when using the "gtr" option.. And if I don't use "gtr" some lines becomes smaller as I've shown before.


This is how it should look for example. The last line is shown.

The source PDF is problematic because there are lines of text that don't have a clear gap between them (see attached), causing k2pdfopt to guess wrong at where it should cut them. Higher -gtr leads to more aggressive cutting, but, as you can see, sometimes it is overzealous. I can't find a set of options that work without some mistakes. I'd have to go back and see if the algorithm could be tweaked to work slightly better for this particular case.

[abe1]

Originally Posted by willus

The source PDF is problematic because there are lines of text that don't have a clear gap between them (see attached), causing k2pdfopt to guess wrong at where it should cut them. Higher -gtr leads to more aggressive cutting, but, as you can see, sometimes it is overzealous. I can't find a set of options that work without some mistakes. I'd have to go back and see if the algorithm could be tweaked to work slightly better for this particular case.

Thank for your hard work.

[Loyd]

Originally Posted by willus

Default options (i.e. no special options) seem to work pretty well. Not identical to abe1's conversion, but close.

Hello willus,

I need your help for the same file. In deed, after conversion, I loose resolution in the pdf (so the quality is not good while reading under my PW3). See attached files please.



Also, I need to know how can I specify the number of words per line wanted and the numbers of lines in each page after conversion ?

Thank your for your time !

Edit : I join the PDF again in this post

[willus]

Originally Posted by Loyd

I need your help for the same file. In deed, after conversion, I loose resolution in the pdf (so the quality is not good while reading under my PW3). See attached files please.
Also, I need to know how can I specify the number of words per line wanted and the numbers of lines in each page after conversion ?

Are you using the latest version of k2pdfopt (v2.51a)? Your output looks about right for a PW3. The document itself is computer-generated type, so it will scale well to any resolution, as your first image shows. The k2pdfopt output is bitmapped and therefore at a fixed resolution (300 dpi for the PW3). You can increase the "Document Resolution Factor" in the gui or use the -dr command-line option to improve the resolution, but there's not much point to going above 300 dpi since that is the resolution of your PW3. It will also cause the size of the output file (in bytes) to increase. Attached is a screenshot from using -dr 2 (which doubles the resolution to 600 dpi).

You cannot directly specify words per line or lines per page, but you can specify the size of the font in points the output file checking the "Fixed output font size" box and putting in the value you want, and this will indirectly determine the words per line and lines per page. You can also separately specify the vertical line spacing in the output file with the -vls option. E.g. putting -vls 2 in the "Additional options" box will double-space the output text. See attached screen shot for example.

[The Free Reader]

Hi
I'm using k2pdfoopt
I don't know how to use it
Sometime I convert the Arabic book then I lose some words or I convert and find words are bigger than the others
What is the solution ?
I think here I can find the developer of this software so
The best help I can get it from here
Best wishes 😊

[Loyd]

Originally Posted by willus

Are you using the latest version of k2pdfopt (v2.51a)? Your output looks about right for a PW3. The document itself is computer-generated type, so it will scale well to any resolution, as your first image shows. The k2pdfopt output is bitmapped and therefore at a fixed resolution (300 dpi for the PW3). You can increase the "Document Resolution Factor" in the gui or use the -dr command-line option to improve the resolution, but there's not much point to going above 300 dpi since that is the resolution of your PW3. It will also cause the size of the output file (in bytes) to increase. Attached is a screenshot from using -dr 2 (which doubles the resolution to 600 dpi).

You cannot directly specify words per line or lines per page, but you can specify the size of the font in points the output file checking the "Fixed output font size" box and putting in the value you want, and this will indirectly determine the words per line and lines per page. You can also separately specify the vertical line spacing in the output file with the -vls option. E.g. putting -vls 2 in the "Additional options" box will double-space the output text. See attached screen shot for example.

Thank you for your quick reply and for valuable information.

I compiled the latest version of k2pdfopt (v2.51a) after your message. My precedent results have been obtained with v2.42.

Since I am using the mac version of k2pdfopt I need to ask what is the equivalent of "Fixed output font size" box in the command line ?

Based on your suggestion, here is what I use currently :

-gs -gtr 0.18 -col 1 -mag 1.65 -dr 2.75 -idpi 600 -vls 1.6 kp3 -x

Maybe some commands are not necessary…

My purpose is to get closer to what I have when reading an AZW3 (see attached Kindle' screenshots), and it is almost achieved. Thank you.

What I try to get right now is :

if it is possible :

1. Get the text adjusted (no space in the end of lines). Just like AZW3.

2. Specify the same font in the output file that I use when reading under Kindle PW3 (I am using a personalized one called CharInk).

3. Overcome the artefact encountered on page 16, line 5 (see pdf attached).

Great thanks in advance for your time.

[willus]

Originally Posted by The Free Reader

Hi
I'm using k2pdfoopt
I don't know how to use it
Sometime I convert the Arabic book then I lose some words or I convert and find words are bigger than the others
What is the solution ?
I think here I can find the developer of this software so
The best help I can get it from here
Best wishes 😊

Try reading back through the previous few posts. Abe1 had a similar problem that has been discussed just recently on this thread. Have you tried reading the FAQ page as well, or the help pages in general? Also, it would help if you could post a sample of the source document and the problems you are encountering. No two source PDFs are the same, so it really helps.

[willus]

Originally Posted by Loyd

I compiled the latest version of k2pdfopt (v2.51a) after your message. My precedent results have been obtained with v2.42.

Since I am using the mac version of k2pdfopt I need to ask what is the equivalent of "Fixed output font size" box in the command line ?

Based on your suggestion, here is what I use currently :

-gs -gtr 0.18 -col 1 -mag 1.65 -dr 2.75 -idpi 600 -vls 1.6 kp3 -x

Maybe some commands are not necessary…

My purpose is to get closer to what I have when reading an AZW3 (see attached Kindle' screenshots), and it is almost achieved. Thank you.

What I try to get right now is :

if it is possible :

1. Get the text adjusted (no space in the end of lines). Just like AZW3.

2. Specify the same font in the output file that I use when reading under Kindle PW3 (I am using a personalized one called CharInk).

3. Overcome the artefact encountered on page 16, line 5 (see pdf attached).

Great thanks in advance for your time.

You should not have had to compile 2.51 yourself--there is an OSX binary download on my site. The command-line font size option is -fs, e.g. -fs 12 for 12-point font.

You don't need the -gs option unless you did not compile with MuPDF and are thus forced to use ghostscript. The Mac OSX binaries from my site are compiled with the MuPDF library.

To fully justify the text (no spaces at the end of lines in the converted file) use: -j 0+

To eliminate the artifact on p. 16, remove your -gtr option. It's too high and causing overly aggressive line cutting. The default works fine for your document since there is already plenty of gap between the lines in the source file. You should only need to increase -gtr when the lines in the source file are so close to each other that they slightly overlap in places.

There is no way to change the font with k2pdfopt. That's not part of its functionality. To do that you need to fully convert the PDF to a different type, such as mobi, which often requires a lot of manual editing, or you need the source document that created the PDF.

Also, I think you're over-killing the -dr option. You shouldn't need nearly 3x the resolution of your reading device. All that does is to make the conversion unnecessarily slow and the output file unnecessarily large (in bytes).

So my final recommendation is:
k2pdfopt -dev kp3 -j 0+ -col 1 -mag 1.2 -dr 2 -vls 1.6 -x m.pdf

Converted file attached.

[ahmd5800]

hi i want to convert scaned pdf book to kindle but if i used reflow text the font is not in same size and big output size
and in fitpage, copy and even 2 column the size is small
is there any way to convert each input page to 2 output pages???

[ahmd5800]

can i increase font size??
i used 2 column in single column input book because it make the page some how little lines than the origion but i hope to find a way to make the size more biger

[willus]

Originally Posted by ahmd5800

can i increase font size??
i used 2 column in single column input book because it make the page some how little lines than the origion but i hope to find a way to make the size more biger

Please spend some time reading my help pages and also the last several posts to this thread--several of them cover these same issues. It would also help if you could post your source PDF file.

[ahmd5800]

Originally Posted by willus

Please spend some time reading my help pages and also the last several posts to this thread--several of them cover these same issues. It would also help if you could post your source PDF file.

i spend all time trying to solve the problem in scaned pdf file font size

[ahmd5800]

Originally Posted by willus

Please spend some time reading my help pages and also the last several posts to this thread--several of them cover these same issues. It would also help if you could post your source PDF file.

i found the best result come when i chose 2 column conversion mod just i ask if their any way to make the font size biger or in other word to make the page contain like only 12 lines per page?becase the font size order work just in reflow text not in other conversion mod, and if i chose reflow check the wards become not in that good shape of the source
thank you very very much