k2pdfopt: optimizes PDFs for viewing on e-readers

[tigerjack89]

Also, I have a scanned PDF with some underlines and some gray stains. What I obtained from the conversion is something like the attached image (and this is yet a great result btw ).
There is some way to have better results?

[willus]

Originally Posted by tigerjack89

I tried to open the file with different readers, but I can't see this grey line on the PDF. Am I going blind or am I missing something?

The line is at the very right edge of the page, so it is difficult to see. I had to really zoom in with my reader to convince myself.

Originally Posted by tigerjack89

How can I see the dpi of the images?

I zoomed in to a very high magnification and manually did an approximate pixel count over a small region. But it's not that important to know. You can just try a few different values for -odpi and you should be able to tell when the output quality stops improving.

Originally Posted by tigerjack89

Also, I have a scanned PDF with some underlines and some gray stains. What I obtained from the conversion is something like the attached image (and this is yet a great result btw ).
There is some way to have better results?

You might try adjusting the white threshold value (-wt option -- see the list of command line options). Start with -wt 175 and if that doesn't work, try lowering that value. As you lower the value, the processing should become more tolerant to the gray stains, but it also might have trouble picking up the actual text and other features if you go too low. You could also try a different contrast value (-cmax option).

[tigerjack89]

ty ty
I had never thought it was possible

I fix the problem using -wt 100 with -odpi 150 or -wt 80 with -odpi 200.
Btw, what is the range span of -cmax? Is there any good value to start from?

[willus]

Originally Posted by tigerjack89

ty ty
I had never thought it was possible

I fix the problem using -wt 100 with -odpi 150 or -wt 80 with -odpi 200.
Btw, what is the range span of -cmax? Is there any good value to start from?

There is no limit to cmax, though realistically you shouldn't need to go beyond 10 or so. The value, if positive, only specifies a limit which k2pdfopt is not allowed to exceed. If k2pdfopt thinks a lower value works better, it will use that. Use a negative value to force the value to be used. The default value is 2.

[Jorram]

Hi Willus, I just stumbled across your amazing tool and thought I should repeat thousands others in thanking you for your efforts and persistence to improve it!

I was wondering if anyone has tried using Calibre to convert the resulting PDFs into MOBIs. I wanted to search for answers to this question but didn't find any meaningful results.

The reason i ask is because I'm dying to have PDF covers in cover view and from what I've read so far it seems that no matter how good the PDF output is, it's most likely that covers don't show for PDFs on the kindle (i have a kindle paperwhite).

For this reason only I'm considering converting to MOBI. For bitmap-based PDFs i guess the conversion is painless but i'd rather have the native output PDFs and so i was wondering if Calibre's conversion to MOBI would screw up your brilliant results. Has anybody tried that?

[willus]

Originally Posted by Jorram

Hi Willus, I just stumbled across your amazing tool and thought I should repeat thousands others in thanking you for your efforts and persistence to improve it!

I was wondering if anyone has tried using Calibre to convert the resulting PDFs into MOBIs. I wanted to search for answers to this question but didn't find any meaningful results.

The reason i ask is because I'm dying to have PDF covers in cover view and from what I've read so far it seems that no matter how good the PDF output is, it's most likely that covers don't show for PDFs on the kindle (i have a kindle paperwhite).

For this reason only I'm considering converting to MOBI. For bitmap-based PDFs i guess the conversion is painless but i'd rather have the native output PDFs and so i was wondering if Calibre's conversion to MOBI would screw up your brilliant results. Has anybody tried that?

@Jorram -- thank you for the post and nice comments. Unfortunately I have little first hand experience with calibre or MOBI files. But there should be a number of experts here at MR. I wonder if your post wouldn't be better off in the MOBI sub-forum? You could request that a moderator move it if you think so. Hopefully some others will chime in.

[tigerjack89]

Originally Posted by willus

There is no limit to cmax, though realistically you shouldn't need to go beyond 10 or so. The value, if positive, only specifies a limit which k2pdfopt is not allowed to exceed. If k2pdfopt thinks a lower value works better, it will use that. Use a negative value to force the value to be used. The default value is 2.

thanks again for your help

Just a little bug to signal.
If there is yet a file with the same name of the output file to be created, it tells something like "Do you want to replace it?"; so far so good.
But if the previous file is currently opened (and then locked) by another program, k2pdfopt just closes automatically without any message.

[sounonyma]

Thanks for this program, it really helps.
But I also read in japanese and I was ondering if it was possible to make k2pdfopt recognize a text in japanese (to read from the top to the bottom like in this text :
http://dl.free.fr/cpUrN4Tmk) ?

[willus]

Originally Posted by sounonyma

Thanks for this program, it really helps.
But I also read in japanese and I was ondering if it was possible to make k2pdfopt recognize a text in japanese (to read from the top to the bottom like in this text :
http://dl.free.fr/cpUrN4Tmk) ?

@sounonyma -- Welcome to MR! I cannot see the link to your file. I do not speak French and don't see an obvious download link. Can you attach it to a post?

[sounonyma]

Sorry, here is the photo.

[willus]

Originally Posted by sounonyma

Sorry, here is the photo.

It is meant to be read starting with the right-most column of characters, correct?

[sounonyma]

yes, from right to left and top to bottom.

[willus]

Originally Posted by sounonyma

yes, from right to left and top to bottom.

k2pdfopt will not "wrap" columns of text the same way it can wrap rows of text. The best I can offer is to use the settings shown in the attached screen shot. I've also attached your sample as a PDF.

[mike2003]

where is all profile store?
why this so simple app dont use "portable concept"???
i loose all my profile after reinstalling win7

[Smith-Corona]

hi Willus,

first, thanks for a very useful piece of software, it opens up many reading possibilities. I can do some basic things with k2pdfopt, but I am having trouble understanding how margin settings work.

First of all, I am not sure what "Crop margins" means. To me that suggests, how much would you like to crop the margins by, ie how much would you like to chop off. But the default setting of 0.00 removes the margin altogether, so I guess "Crop margins" refers to the actual margin width? I've tried various combinations of right and left "crop margins" and it never seems to do what I want it to do, which is to *set* the margins to a certain width. Increasing the margin values just seems to "eat into" the text width.

What I would like to do is set margins relative, not to the text, but to the page edge. I have some PDFs with widely varying margins on each page, because of sidebars etc. Because k2pdfopt sets margins relative to the text, I end up with different font sizes, depending on whether k2pdfopt is working with a wide margin page or a narrow margin page. This is with default settings, ie margin = 0. I have also tried "trim" and for the most part this seems to keep font sizes the same, but then the pages end up as very different sizes...

So basically I want all the pages cropped as much as they can be with reference to the maximum text width found in the document (with column set to 1). Then to make all the pages that size, and have the font the same size throughout.

I am using Native PDF Output to keep the file size down.

cheers

Smith-Corona