I Don’t Hate Calibre, But…

duck coverJudging by the heated storm I roused the last time I criticized Calibre it’s probably a mistake to do so again. What the hell. I have to say it:


Yesterday I did a troubleshoot and repair on a writer’s ebook. The EPUB she had converted from Word via Calibre was perfectly fine. It was a valid EPUB, and it displayed the way it should on both Calibre and Adobe Digital Editions. Problems arose when she converted the EPUB into AZW3 and MOBI. The ebook worked when she loaded it onto her Paperwhite Kindle, though it had some disturbing issues. Amazon rejected the file outright.

I took her EPUB and ran it through the Kindle Previewer to see what the problem was. It converted with WARNINGS (never want to see that). The ebook opened, but the button under Devices for Kindle for iOS was greyed out. And the cover didn’t display. When I loaded it on my Paperwhite, some of the user controls were locked. I then went back and looked inside the EPUB. The Calibre conversion had declared font-families–Georgia and Times New Roman–neither of which display on Kindles, but they aren’t ignored either, and hence cause all sorts of interesting little problems. Plus, it had built a cover page.

I’m not a techie person, so I don’t know if I can explain it adequately, but I will try. When a file is uploaded to Amazon it converts the source file into a 3-in-1 ebook. If you’ll compare the size of an EPUB to a Kindle, you’ll notice that the MOBI file is about 3 times bigger than the EPUB. The ebook you buy from Amazon will work across several types of devices: e-ink readers, Fire tablets, and a variety of apps, including iOS for Apple. This is tricky stuff. While Calibre can and does convert your files into the MOBI format or AZW3 format, and the files are suitable for personal use, they aren’t suitable for commercial use because the probability is about 97.3% that an ebook converted through Calibre and then uploaded to Amazon will be broken.

This is all about Amazon. They have proprietary platforms and they are constantly updating and improving and tinkering and it’s difficult for outsiders to keep up. Calibre is a library management tool, not a commercial conversion tool. They can’t be expected to stay ahead of the Amazon updates and bugs. That’s not their purpose.

The fix for the writer’s problem ebook was fairly simple. I removed the font-family declarations, removed the cover page, added a line of code to the metadata in the content.opf so Amazon knew the cover was in the file, resized the cover and ran it through the Kindle Previewer. Voila! Everything worked the way it should. The writer was able to upload the file to Amazon with no problems.

My recommendation to the writer for future projects was this: continue using Calibre to convert her Word files to EPUB files since there appear to be no problems and her ebooks are working fine. But, then to refresh her knowledge about using stylesheets in Word via Mark Coker’s Smashwords Style Guide, and direct load her Word file to Amazon. It’s a bit of a pain since you can’t examine and adjust the file before you start the uploading process, but you can take full advantage of the previewer at Amazon and make sure everything works before you hit the Publish button. That way there won’t be any extraneous items like cover pages and “embedded” fonts to muck up the works.

So seriously, folks, don’t kill the messenger or throw rotten tomatoes. This is just the reality. Calibre is not the right tool for converting files to sell on Amazon. Use the program for what it’s intended–managing your ebook library–and find other means to deal with Amazon.


17 thoughts on “I Don’t Hate Calibre, But…

  1. Hi Jaye,

    I would absolutely agree with you that using Calibre to convert Word files to EPUB or MOBI is a very bad idea and can introduce all sorts of formatting gremlins.

    However, Calibre is perfectly fine for converting HTML files to EPUB or MOBI (which is how I format, without issues – Guido Henkel’s method, basically).

    In short, after doing a bunch of search-and-replaces to put in my named entities etc., I dump the Word file into something like Notepad to remove any hidden HTML (I call this bit The Sheep Dip), and then use a HTML editor like Notepad++ to clean it all up and lay out the backmatter etc. Then I convert with Caliber into MOBI, run the conversion once more for EPUB after changing the backmatter links etc., and then I test, upload, then test some more.

    Seems to work fine,


    • Hey David. Good to know about html. But, since you’re comfortable with it anyway, the next time you have a book to do email me and I’ll give you a template for a Kindle format. You won’t have to convert your files through Calibre. You can use the Kindle Previewer which won’t add any extraneous junk to the file.

      And a tip, if you encode your Notepad++ files in UTF-8, you can skip the step of having to use all those named entities for ASCI characters. You’ll need them for reserved characters (ampersands, straight quotes and straight apostrophes) but just about everything else will work fine.

      • Oh, interesting! I’ll definitely take a look at that template when the time comes. I have a HTML shell I drop my text into, but it’s quite basic.

    • Dave, I started that way, too. Then Jaye pointed out some things in the resulting file that were non-optimal for Kindle, and I switched over to the Sigil method where it’s much more under my direct control. Way better.

  2. Oh, my brain shut off when I read the word ‘code’. I have no clue how you do what you do which is why you do it for me.

      • Best kind – they know their limitations.

        Real geeks aren’t supposed to HAVE limitations, much less ADMIT them. It is kind of fun to watch them do the ‘explaining dance’ with hand motions.

        I can say this: I have three of ’em for children. I understand them, most of the time.

        Limitations? We don’t have no stinkin’ limitations.

    • Not hard. Trouble-shooting is fun (but only when I get results). But yes, got a blog post out of it. 😉 It also gets me riled up that the device makers and platform developers focus on proprietary stuff instead of making everything device-neutral. If those noodleheads would do that, producers wouldn’t have to spend so much time trying to figure out the quirks and foibles of every distributor’s device and could spend more time on the important stuff, like writing.

  3. Meant to say: the photo really took me back – to the days when I was in first grade in S. California (before I moved to Mexico), and we had to do those drills where we pushed our desks against the inner walls and then got under them, row by row. We expected to be blown up by an atomic bomb any day. But of course I have no photos, just memories. 1950s.

  4. I keep hearing about Jutoh. Do you know anything about that? Worth trying out?

    Also, I always heard converting straight from Word was a VERY BAD THING.

    Seems like every time I feel like I have the hang of this formatting stuff, the ground shifts under my feet. Oy!

    • The only person I know personally who uses Jutoh is PG over at the Passive Voice. He gets excellent results and swears by the program. Since I buy and read the books he produces, I can say first hand that they are quality formats. As for learning curve, difficulty, or cost, I can’t even guess about any of that.

      The thing with Word. IF the format is very simple (straight fiction, minimal paragraph styles, no graphics, ‘standard’ layout); and IF the producer is meticulous about cleaning up the source file; and IF they use ‘accepted’ styling according to the Smashwords style guide, they can convert it into a serviceable ebook that works for Amazon. The converter onsite at Amazon is set up to recognize and handle certain fonts and paragraph styles in Word. The major problem with using Word is that while ebooks are pretty simple, Word (and other word processors) are very complicated. (If you were examine a straight html format side by side with the html produced in a Word doc, you would see immediately what I mean) Word allows users to do amazing things, but when it comes time to convert those ‘amazing’ things, strange and awful things can happen. Even something as innocuous (on the surface) as justifying text can break the ebook.

      A MOBI/KF8 file (for Kindle) is a converted EPUB file. When I format ebooks, I create EPUB files that are then converted for Kindle. BUT, I don’t build them the say way I build them for B&N or Kobo or iPads. I have specific styling to work on Kindle devices. The differences are subtle, but they are there and they do make a difference.

      That’s what causes problems with Calibre. It converts to EPUB based on specs that make it compatible with EPUB readers. But then it takes the same file, with the same specs, and converts it MOBI or AZW3 and while the file will work on Kindle devices, the styling is not quite right and it’s not going to work properly across the many types of Kindles. The simple turns complicated and therein lies the problem.

      Rob, if you’re looking for a serviceable program that is relatively user-friendly and you’re willing to dig around in the ‘engine” so to speak, Sigil can be used to create ebooks for sale on Amazon. it requires modification to the stylesheet and content.opf and toc.ncx (and if you understand what that means, then you’re more than qualified to do a good job of it). But if you don’t understand, you can learn. Sigil will allow you to do some very sophisticated ebooks.

      • Thanks for the detailed reply. Yeah, I have no clue what you meant about content.opf and toc.ncx. LOL. But I’ll still take a look at Sigil. Do you know if there’s a good guide somewhere on how to use it to make ebooks?

        Thanks again!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s