[ghmerrill]

This is goofy. I'll appreciate any insight into it.

I'm in the final stages of publishing a book on Amazon and my Sigil-made Epub3 file passes the Epub validator with no errors, uploads to Kindle, converts perfectly and displays on all the devices and viewers I have available to me. Now I need to generate the PDF for the print version. I suck it into Calibre, set what seem to be reasonable parameters, convert it to PDF. I look at it in Acrobat Reader and it has frequent rendering errors -- which appear to be characterers overlaid on one another. Large chunks of it are fine, but it looks like someone just scattered random wrong characters all over. This Epub3 book contains 20 XHTML files and is about 140K words. I've spent all day tracking this down, and whittled it down to a test case consisting of only two short XHTML files (my Copyright and my Preface) that exhibits the same problem.

The first file is rendered perfectly in PDF. The second file is mostly rendered fine, but the first header (e.g., <h1> or <h2>) is render incorrectly, and then seemingly random words are rendered incorrectly. It doesn't matter which order the files are in, it's always the second (and subsequent) file in the order that exhibits the problem. For example, in the "<h1>Preface</h1>" tag, what appears is an upper case P followed by a space, then an "e", an "f" overlaid with an "a", and the final "e". The first letter of the following sentence begins with "This", but has the "T" and "h" overlaid. The rest of that paragraph is fine, but half-way through the following paragraph there is another overlay of "t" and "h" in what should be "the".

Multiple header tags in the first file are rendered without any problems.

I have not yet tried converting this to Epub2 to see if that would make a difference. I'm just trying to convert my book to PDF so I can submit that format to Amazon for the print version.

Any ideas?