Best way to convert JPEG page images to PDF

[HarryT]

The Bodleian Library in Oxford have recently digitised and made available online a number of their old books. The only way to download them, however, is as a ZIP file of individual JPEG images. I want to convert these to PDF to read on my iPad (they're page scans, so PDF makes sense).

What I've done so far is to rename all the images in the ZIP file to reflect the correct page order, renamed the ZIP file with a ".cbz" extension, and then used Calibre to convert from CBZ to PDF. This works pretty well, but I was wondering if there were any settings I should be using to optimize the process? What, for example, does the "Disable comic processing" setting do?

I've never done this type of conversion before, so all advice would be welcome!

Thanks,

[eschwartz]

I've never done CBR/CBZ conversions either, but I would assume "Disable processing" means "don't use all the options at the bottom, like despeckle, normalize, sharpening, trimming, etc."

If the JPEGs are of good quality, I think you should probably avoid messing with them. If they need clearing up, you might want to let calibre do it, or you might want to do it by hand.



Make sure your output profile (Common Options ==> Page Setup) is set to "tablet" or the JPEGs may be resized to fit the resolution of your currently-selected default output device.

[cybmole]

The other route is to use a pdf program which can import jpg and can save as a single pdf. I am not aware of any good free ones. Adobe is v expensive, pdf. exchange pro is more reasonably priced. The pdf creator programs can also crop, add footers etc

[cybmole]

https://www.tracker-software.com/pro...df-xchange-pro
I see it is shareware, so free to try.

[theducks]

I was thinking along the same lines as eschwartz:

use a Comic format

The real reason, is to keep them in order

Windows has a Slideshow mode to look at photos.

Get a PC

[HarryT]

Originally Posted by eschwartz

I've never done CBR/CBZ conversions either, but I would assume "Disable processing" means "don't use all the options at the bottom, like despeckle, normalize, sharpening, trimming, etc."

If the JPEGs are of good quality, I think you should probably avoid messing with them. If they need clearing up, you might want to let calibre do it, or you might want to do it by hand.

They're not the world's greatest quality scans. I've attached one so you can see what they're like (they're reproductions of carvings on Egyptian monuments). I'll try the various processing options in Calibre and see what the results are; my aim is to make the hieroglyphic inscriptions are clear as possible, so I can work on translating them. There are really too many to do by hand.

Make sure your output profile (Common Options ==> Page Setup) is set to "tablet" or the JPEGs may be resized to fit the resolution of your currently-selected default output device.

Oh yes, I always have the "Tablet" output option selected!

[eschwartz]

Those images want fixing up!

Hopefully calibre can do some good with them -- although calibre is really outsourcing the work to imagemagick, so if you know what you are doing you can probably do better than off-the-shelf settings in calibre.
imagemagick and a variety of other tools can do batch processing just like calibre, so it's not like your choices are "open each file in Photoshop/GIMP and fix it" or "use those calibre settings".


Perhaps you should hunt down the people in charge and ask them to rescan using a better-quality scanner and output to PNG. They could be doing a much better job digitizing.

[Billi]

Originally Posted by HarryT

I've never done this type of conversion before, so all advice would be welcome!

If advice outside of Calibre is welcome: you might want to have a look at imagemagick and its conversion tools
(http://www.imagemagick.org/script/convert.php)
which would have been my first idea for such a task.

[HarryT]

Thanks for the suggestions! I'll do some experiments and see what works best.

Even with the defaults, the PDF looks a lot better than the original page scans, so I'm happy .

[evianluw]

I recommend PDF shaper. It's free and it contains more useful tools for PDF. http://www.glorylogic.com/pdf-shaper.html

[HarryT]

To follow up on this, the CBZ to PDF conversion worked beautifully for several books that were around 200 pages long (a CBZ file about 70MB in size). I then tried it for a book that was 450 pages long, and Calibre (32-bit Calibre running on a Windows 7 laptop with 4GB RAM) ran out of memory in the final PDF output stage of the conversion process.

Luckily my work laptop has 16GB RAM, so I installed 64-bit Calibre on that, and it converts the books with no problems . The largest book I've so far tried it on has been 953 pages (a 168MB CBZ) and 64-bit Calibre converts it fine, creating a 322MB PDF. Pretty big, but "GoodReader" on my iPad displays it fine.

Does anyone have any tips for ways to shrink the size of the PDF? They are page scans, and completely unsuited for OCR. It's a hand-written book - basically the diary of an expedition to Egypt in the 1820s - and every page has little sketches of buildings, hieroglyphic inscriptions, etc, on it. I've attached one of the page scans to shows what a typical page looks like. As you can see, there's absolutely no way it could sensibly be OCR'd!. Can anyone think of a way to reduce the size of the PDF?

I should add that it's not particularly a problem that the PDF is 322MB in size; I have a 128GB iPad so I'm not short of storage space, but obviously smaller is better .


Many thanks,

[eschwartz]

There are various utilities out there that promise to recompress a PDF with better compression schemes, or apparently converting to PostScript and back to PDF improves file size a lot (says half the internet, so it must be true ).
Depending on the solution, though, it may be lossy.


Another thing that might help is to reduce the size of the original images. Usually that means stripping out pointless metadata, as well as converting the color depth to grayscale if there is no color anyway. But again I am not sure how to effectively do that for lossy formats like JPEG.