Quote:
Originally Posted by mkgtu
FThe jpg trick doesn't seem to work if the book in question was originally an Amazon book, ie purchased from Amazon or borrowed as a Kindle book through public library via Overdrive.
...
So, bottom line, if the book was originally in Amazon's clutches, it needs a double laundering. Thank you, Jeff Bezos!
|
Quote:
Originally Posted by JSWolf
You may have to edit the OPF to remove any traces of Amazon. Do that and try again.
|
The issue is not caused by Amazon recognizing that the book originally came from them. It is a consequence of the conversion from AZW3 to EPUB done using calibre.
When you email a file with an image extension it is processed by Amazon using kindlegen. That program recognizes that the format is actually EPUB and processes it accordingly.
However kindlegen is designed to convert books as part of the publishing process. It often intentionally rejects books with content that fails to meet Amazon's publishing guidelines. That makes sense in a publishing environment where the publisher is expected to deliver a properly formatted book and needs to know if that is not the case.
When calibre converts AZW3 to EPUB it tends to create links to HTML ids that are placed on body elements in the XHTML files. Those links, while correct according to the EPUB specification, are rejected by kindlegen if they occur in the table of contents.
One workaround for this is to use the
Kindleunpack plugin to convert AZW3 to EPUB instead of using calibre's Convert Books function. That produces an EPUB that better reflects the original content of the AZW3 file and is more likely to be acceptable to kindlegen.