Pdflrf equivalent for Image Magick

[baliantas]

Hello,

My first post here. Hope I'm doing everything "according to the book"!

I've been using pdflrf for a while now, and although the results are not bad at all, I feel I need to have some more control. I'm hoping I can use Image Magick to do what pdflrf does, with the added advantage that I can adjust settings more easily.

Specifically, I want to be able to:
- review/modify the generated png files before compiling them into an ebook
- convert to epub instead of lrf

My workflow would be something like this:
- convert the pdf file to png
- get rid of the white margins (Image Magick does a wonderful job here)
- rotate 90° and resize the height so it's equal to the width of my ereader screen (Sony PRS-350)
- make chunks with dimensions that fit my ereader
- if text lines are broken, find a line that's completely white and split there instead (cf. "smart cut" in pdflrf)
- thicken the edges, to increase readability
- join pages (so I don't have pages with just a couple of lines on them, cf. pdflrf)
- compile to epub

This would be (mostly) the same workflow as in pdflrf...


Can this be done with Image Magick? (or another tool, perhaps)

Thanks,
Baliantas

[DSpider]

Yes, Scan Tailor: http://scantailor.sourceforge.net

To edit PNG images you can use GIMP: http://www.gimp.org


They are both very powerful open source tools.


Edit: Scan Tailor outputs .tif images, but it's all the same (lossless just like PNG). If you're keen on PNG you could do a batch conversion using IrfanView or similar. But honestly, PNG/TIFF, it's all the same to me - though it may not be the all the same with the e-reader, so you need to experiment.

AFAIK, ePUB splits graphics automatically. So you could, perhaps, set the top margin in Scan Tailor at 1 (maybe except for the chapter headings) for the pages to transition more smoothly. Then you would basically have an ePUB full of graphics, specifically "tailored" for the PRS-350, which I believe is what you're trying to accomplish. But you won't be able to search or highlight, they're images. Then again, same with a book borrowed from a library... You may be onto something.

[baliantas]

Originally Posted by DSpider

Yes, Scan Tailor: http://scantailor.sourceforge.net

To edit PNG images you can use GIMP: http://www.gimp.org
They are both very powerful open source tools.

Scan Taylor does everything I want, EXCEPT:

- resize the height so it's equal to the width of my ereader screen (Sony PRS-350) --> so the images should be split into 800x600 pixels chunks
- equivalent for "smart cut" in pdflrf
- join pages (so I don't have pages with just a couple of lines on them, cf. pdflrf)

The Gimp is a wonderful piece of software, but I'm not willing to manually edit hundreds of images...

Thanks for the quick reply!
B.

[eksor]

Originally Posted by baliantas

Hello,

My first post here. Hope I'm doing everything "according to the book"!

I've been using pdflrf for a while now, and although the results are not bad at all, I feel I need to have some more control. I'm hoping I can use Image Magick to do what pdflrf does, with the added advantage that I can adjust settings more easily.

Specifically, I want to be able to:
- review/modify the generated png files before compiling them into an ebook
- convert to epub instead of lrf

My workflow would be something like this:
- convert the pdf file to png
- get rid of the white margins (Image Magick does a wonderful job here)
- rotate 90° and resize the height so it's equal to the width of my ereader screen (Sony PRS-350)
- make chunks with dimensions that fit my ereader
- if text lines are broken, find a line that's completely white and split there instead (cf. "smart cut" in pdflrf)
- thicken the edges, to increase readability
- join pages (so I don't have pages with just a couple of lines on them, cf. pdflrf)
- compile to epub

This would be (mostly) the same workflow as in pdflrf...


Can this be done with Image Magick? (or another tool, perhaps)

Thanks,
Baliantas

Hi:

Try gscan2pdf (gscan2pdf.sourceforge.net), splits, joins, crops and rotates. Also, interfaces with unpaper (http://unpaper.berlios.de/) for blacl areas removal and deskewing.

Regards