Editing and creating EPUB files

[holyschist]

I recently picked up a used Sony PRS-600 to try out ereaders properly, and so far I like it. However, I have noticed that EPUB files converted from PDF annoy the hell out of me, because they have hard page breaks, which do not coincide with ereader pages on my preferred font size. So I have two questions:

1) What's the easiest way to edit EPUB files created by someone else to add a table of contents, fix formatting issues and typos, etc.?

2) What's the easiest way to create basic EPUB files from scratch, if one felt inclined to do so? (No images, no fancy formatting.)

I could do it entirely by hand, but that seems less than efficient--I'd rather generate the file automatically and clean up by hand if necessary. I'm using a Mac, current operating system, and I do have Adobe InDesign (CS3), but that sounds like it might be overkill. Free is good.

Has anyone tried Sigil or eCub? Have another favorite EPUB editor/creator?

Also--does anyone happen to know how the PRS-600 handles epub ebooks that were originally created from HTML that had non-Roman unicode characters (e.g. Chinese)? I know I am going to run into this on occasion. Do ebooks with non-Roman characters have to be PDFs to display correctly?

Thanks!

Comments

[elf]

Non-Roman characters shouldn't have to be PDFs but may need the fonts embedded in the epubs. I know nothing of how to do this; I just know it's more-or-less possible. PDFs manage by embedded part or all of the font in the PDF.

I haven't gotten to creating epubs by hand, but the easiest way to learn how, if you're code-minded, is to take one apart and look at the pieces:

1) Take an epub (AO3's are simple; Calibre's output less simple)
2) Rename from "epub" to "zip"
3) Unzip. Examine contents at will.
4) Tinker with contents; rezip; change extension; view on reader.

I gather that the build-an-epub part is easy; the reason people have programs is to do the .css and other formatting parts. A basic .epub consists of an html file and a tiny group of very simple metadata files.

Converting from PDF is always going to be problematic. I'm comfortable with Word, so I convert to doc or rtf and edit to remove the useless page breaks & such; other people prefer to convert them to HTML and edit that instead.

[holyschist]

Hmmm. Interesting. So far I have not yet found anything about embedding fonts in epubs, but I will have to see how they tick, first.

Do you know offhand how to get Calibre NOT to automatically turn the first page into a cover? It's really annoying when it makes the first page of text a cover. Do I have to have a cover image to avoid that?

[elf]

I barely use Calibre; when I make epubs, I always add a cover, even if it's Calibre's auto-generated one. I didn't know it would auto-convert some other part of the book into the cover.

Embedding fonts: there's apparently two options; putting it inside the zip (which is considered wasteful) and making the .css attach to an external font.

Mobileread thread from a couple of years ago, which means Calibre may have changed so much it's completely invalid now.

Embed fonts in epub instructions, which loses me at Step 2.

Person who claims embedding doesn't work, which I don't think is quite true, with a comment including link to epub with embedded font/s and an invitation to take it apart and see how it ticks. Link is dead, but commenter's site has sample ebooks which presumably could be played with.

[holyschist]

Hmmm. The EPUB files show the fonts fine on my computer, but my Sony obviously does not have a default set of non-Roman fonts, so I think they would have to be embedded somehow (as images, even, but that would be a giant PITA) or referenced externally. Those links are really helpful, thanks! I think I will figure out the basics of making an EPUB before I try to figure out the font thing.

[valiha]

I can attest that font embedding wasn't easy to solve. I had major problems trying to embed fonts so I could display Central European characters in my epubs in a textbook I was using to prepare for an exam. I had a 6" Astak reader at the time, which now belongs to my sister.

I posted threads in both the Astak reader forum and the Sigil forum. The Astak thread was about fixing embedding in the textbook, and the Sigil about doing the same in my lecture notes.

The Sigil posts contain links to some of the files used, and a detailed and simple description of what needs to be done to get embedding to work. I was close to pulling all my hair out.

Sigil creator Valloric noted that "the style rule that uses a new font family name has to come after the @font-face rules that declare it."

[holyschist]

Oh dear, there are problems with diacriticaled Roman letters, too? That seems rather unfortunate.

Well, I will attempt to figure this out later.

[jumpuphigh]

Do you know offhand how to get Calibre NOT to automatically turn the first page into a cover?

I think you have to have a cover image (or at least use Calibre's cover format) but I'm not certain. There are a number of ways to play with cover formation - metadata, the convert pop-up window, preferences. Are you looking for no cover or just not a page of text as the cover?

[holyschist]

Either one, although no image-cover would be the easiest. I don't know--it seems to do it sometimes and not others, so I guess I need to play with settings more.