k2pdfopt: optimizes PDFs for viewing on e-readers

[greg3210]

Thank you. I got it working now. However why not just let -n default to turning off anything that conflicts with it? Isn't that what a user reading the command name would reasonably expect to happen?

[willus]

Originally Posted by greg3210

Thank you. I got it working now. However why not just let -n default to turning off anything that conflicts with it? Isn't that what a user reading the command name would reasonably expect to happen?

You are right, that -n should do that. The problem is that it is already on by default, but it is trumped by text wrapping, so if you turn off text wrapping, you will get native mode in the current version. That's just confusing, though, so I'll fix it in the next version (probably by defaulting native mode to off in the next version and having -n turn it on and disable anything that conflicts with it).

[zola]

I've a kindle 3 and Mac on OSX. I've changed the pdf document to k2opt.pdf , but after putting it on my kindle it says 'This pdf cannot be opened due to embedded features not yet supported by Kindle'.

Do you know how to fix that?`

Thanks

[willus]

Originally Posted by zola

I've a kindle 3 and Mac on OSX. I've changed the pdf document to k2opt.pdf , but after putting it on my kindle it says 'This pdf cannot be opened due to embedded features not yet supported by Kindle'.

Do you know how to fix that?

Hard to say what it might be. If you could attach the source PDF file to a post or send it to me in a private message, that would be most helpful. Did you convert using native PDF output, or using the default settings (or what settings did you use)? Are you able to view the source PDF document on your kindle? How did you transfer the PDF file to your kindle?

[davr]

Is it theoretically possible to do both text wrapping and native pdf output? That seems like it would be the best of both worlds -- smaller, higher quality files, while still getting the nifty reformatting. You can do something similar with the desktop adobe pdf reader (view->zoom->reflow), though it's not perfect, some formatting is lost.

[willus]

Originally Posted by davr

Is it theoretically possible to do both text wrapping and native pdf output? That seems like it would be the best of both worlds -- smaller, higher quality files, while still getting the nifty reformatting. You can do something similar with the desktop adobe pdf reader (view->zoom->reflow), though it's not perfect, some formatting is lost.

I've learned never to say anything is impossible. I was not aware Adobe had that option in their PDF reader--I'll have to check it out. The problem with the way I do native PDF output is that for each rectangular region I place on an output page, I render the entire source page (cropping out everything outside the crop box). This is a generally reliable method since I'm not trying to go inside each page stream and tinker with it. But if you used text re-flow, you'd then have hundreds of crop boxes (one for each word), and it would take a very long time to render each page. Plus, selecting and searching text would be a nightmare because even though the content outside the crop box is visibly eliminated, most readers/viewers still act as if the text is there anyway, so it would be like trying to select text on hundreds of overlaid pages. Trust me that you would not prefer this output. Bitmapped pages with OCR are a much better alternative. This is why I try to encourage the use of the "-mode" options which try to minimize the number of crop boxes used on each page. SoPdf limits itself to one per page, for example, which is ideal. There is more than one way to skin a cat, though, and I may not be using an optimal method, but for now it's what I've got. To go further I will have to understand more about PDF files.

[willus]

K2pdfopt v1.62 is released. I have significantly re-organized the k2pdfopt source code to be more easily integrated with the KindlePDFViewer application (K3 thread here). The bulk of the k2pdfopt-specific source code now resides in a k2pdfopt library, and I have eliminated all global variables. There are also (optional) conditional compile directives so that third-party library dependencies can be eliminated. There are some minor new features and bug fixes as well. See the web site for details.

[willus]

Originally Posted by saeed.geek

I want the output regardless of input be white text on black background.
smartphones have backlight and in dark areas(like in bed) this backlight hurts the eyes. I have a smartphones then I want night mode( white text on black background) for these situations.
best regards.

"Night mode" has been implemented in v1.62. Use -neg command line option (or under "c" option in menu).

[rostamiani]

Hi
I cannot reflow one of my PDFs
after execution this error is shown :

Detecting document orientation ... warning: not building glyph bbox table for fo
nt 'OMLTVR++Cambria' with 4722 glyphs

[willus]

Originally Posted by rostamiani

Hi
I cannot reflow one of my PDFs
after execution this error is shown :

Detecting document orientation ... warning: not building glyph bbox table for fo
nt 'OMLTVR++Cambria' with 4722 glyphs

If you'd like to send me the PDF in a private message or post it here, I'll have a look at it on my system. That's an error message printed by the MuPDF library, but the lack of re-flow may have nothing to do with that message. What system are you running on? Are you using any special options? Does k2pdfopt exit prematurely, or does it finish but leave an un-reflowed or incomplete output file?

[kut]

THis is genius! thanks a lot!!

I've a kindle 3 and Mac on OSX. I've changed the pdf document to k2opt.pdf , but after putting it on my kindle it says 'This pdf cannot be opened due to embedded features not yet supported by Kindle'.

I got the same error with an imcomplete pdf file. However after the conversion was finished the pdf worked just great!

[rostamiani]

Originally Posted by rostamiani

Hi
I cannot reflow one of my PDFs
after execution this error is shown :

Detecting document orientation ... warning: not building glyph bbox table for fo
nt 'OMLTVR++Cambria' with 4722 glyphs

oh ... the file has reflowed correctly
I didn't noticed !

This is the file :
http://www.4shared.com/office/0eJCbF2b/R_online.html

But the watermark conflicts the process !
Can you add an option for removing watermark ?

I did a simple job using Matlab .Just subtracted average of all pages from every page

. This is the original page:
http://www.iranimg.ir/?di=213549606095

And this one is with partly removed watermark :
http://www.iranimg.ir/?di=15135493865414

[willus]

Originally Posted by rostamiani

But the watermark conflicts the process !
Can you add an option for removing watermark ?

I did a simple job using Matlab .Just subtracted average of all pages from every page

. This is the original page:
http://www.iranimg.ir/?di=213549606095

And this one is with partly removed watermark :
http://www.iranimg.ir/?di=15135493865414

If the watermark is light enough, you can just set the white threshold level so that it will be ignored (option -wt). Beyond that, I don't anticipate implementing an algorithm to remove the watermark at this time.

[valex]

I have been a long time k2pdfopt user via its integration with KindlePDFViewer. Now I have a pdf file with very wide margins which I want to print on a standard letter-size paper (8.5in by 11in) but without the margins and in "2 pages per sheet" mode. I need the reflow feature since the font comes out a little small even after the margins are cropped. Which options should I be using? Since I am printing on paper I need high dpi. I tried to use the dimensions of half of the letter-size page with the highest dpi k2pdfopt allows (-p 10-20 -w 5.5in -h 8.5in -odpi 538) but the output pdf has way too small font size and no reflow.

Thanks.

[willus]

Originally Posted by valex

I have been a long time k2pdfopt user via its integration with KindlePDFViewer. Now I have a pdf file with very wide margins which I want to print on a standard letter-size paper (8.5in by 11in) but without the margins and in "2 pages per sheet" mode. I need the reflow feature since the font comes out a little small even after the margins are cropped. Which options should I be using? Since I am printing on paper I need high dpi. I tried to use the dimensions of half of the letter-size page with the highest dpi k2pdfopt allows (-p 10-20 -w 5.5in -h 8.5in -odpi 538) but the output pdf has way too small font size and no reflow.

Because you are specifying the width and height in inches rather than pixels, you cannot use higher values of -odpi to magnify the text. That only works when you specify the width and height in pixels. Use the -ds (document scale) option instead, to trick k2pdfopt into thinking your source document is larger than it is, e.g. -ds 2.0 (with your other settings above) will magnify the text by a factor of 2, but if you use a value like -ds 2, you can probably reduce your output dpi to 300 for perfectly fine printable resolution and also set -idpi 300 so that the input dpi will match your output dpi (normally the input dpi is set to twice the output dpi, but that's overkill for what you're doing here). Higher values for the dpi will slow down the processing and make your output file unnecessarily large.