Embedding fonts - strange issue

[Barty]

Originally Posted by Katsunami

I'd love to see those two options work together: completely remove any font that may be in the file, and THEN embed the one selected by the user. Now it seems to work the other way around: Calibre removing the font it just embedded itself. Maybe it'd be a good feature request.

Yes. It seems when you filter out font-family calibre will filter out the font it is itself embedding! Which means when the book you are converting already has an embedded font, you need two conversions, one to strip out any font, then embed your own.

[Katsunami]

Originally Posted by Barty

Yes. It seems when you filter out font-family calibre will filter out the font it is itself embedding! Which means when the book you are converting already has an embedded font, you need two conversions, one to strip out any font, then embed your own.

That's one of the reasons I use Modify EPUB first; to strip all embedded fonts (except on books where I know they will be needed). I would like it if Modify EPUB would also contain an option to actually remove the font-family; it leaves that behind, referring to fonts that are actually not in the EPUB, but if they are in/on the reader, it could be that they actually supersede the font you set yourself; this depends on the device, of course.

[7hir7een]

Originally Posted by Katsunami

I'd love to see those two options work together: completely remove any font that may be in the file, and THEN embed the one selected by the user. Now it seems to work the other way around: Calibre removing the font it just embedded itself. Maybe it'd be a good feature request.

You can always convert once with remove fonts checked, use the remove books of a specific format option to remove the original file, then convert again to embed. That's what I do for the tricky ones that already have embedded fonts.

A bit kludgy, but it works.

[dordale]

Originally Posted by 7hir7een

You can always convert once with remove fonts checked, use the remove books of a specific format option to remove the original file, then convert again to embed. That's what I do for the tricky ones that already have embedded fonts.

A bit kludgy, but it works.

This is what I'm doing as well--actually I convert ePub to ePub--removing embedded fonts. Then I convert ePub to azw3 and embed my font of choice. That way, I can load the ePubs on my Sonys (350/650 with PRS+) and use the fonts of my choice. But to use these same fonts on my KPW2, I have to embed them into the AZW3 file. The bummer about this is depending on my mood/lighting/etc I might want to change to a different font--but on the KPW2, I'm stuck with one embedded font (plus the default options.)

[Katsunami]

Originally Posted by 7hir7een

You can always convert once with remove fonts checked, use the remove books of a specific format option to remove the original file, then convert again to embed. That's what I do for the tricky ones that already have embedded fonts.

A bit kludgy, but it works.

Yeah, LOL. I buy the huge books from DelphiClassics.com. I could buy Kindle versions, but I don't want to; I prefer EPUB.

Calibre cannot convert them to AZW3; somehting needs to be done to create "multiple indices", and Calibre does not yet know how. However, it *can* convert the books from EPUB to EPUB, and I've found that KindleGen can convert them to a MOBI/KF8 combination format. KindleUnpack (the script) can split that up to AZW3.

So, after fixing the metadata, I do an EPUB->EPUB conversion to embed the fonts, save the book to disk, and then use KindleGen and KindleUnpack to create the AZW3's.

Kludgy... but it works. And the AZW3's display fine. I keep my original files safe; what happens to the files during conversion for my current reader doesn't matter a thing to me as long as they display fine after the conversion.

[sparklemotion]

Thanks so much for the replies and helpful tips everyone. I really appreciate it. I'm sure it will be helpful to others too.

So far I cannot seem to embed the font I want into the book. However, I made the realization it's only one particular book that it won't take on. Or it least it seems that way so far. I tried another book with that font and it worked fine. Hmm.

Originally Posted by sparklemotion

Thanks so much for the replies and helpful tips everyone. I really appreciate it. I'm sure it will be helpful to others too.

So far I cannot seem to embed the font I want into the book. However, I made the realization it's only one particular book that it won't take on. Or it least it seems that way so far. I tried another book with that font and it worked fine. Hmm.

It could be as simple as sloppy HTML/CSS. I've modified several eBooks in the past that looked like they were created by MS Word then incorrectly converted to HTML. In these books, rather than use simple and effective CSS class and heading tags to assign formatting, they use span tags on nearly every line of HTML. These kind of sloppy files are a mess and a PITA to work with. I wouldn't be surprised if calibre just choked on them.

[sparklemotion]

Originally Posted by jswinden

It could be as simple as sloppy HTML/CSS. I've modified several eBooks in the past that looked like they were created by MS Word then incorrectly converted to HTML. In these books, rather than use simple and effective CSS class and heading tags to assign formatting, they use span tags on nearly every line of HTML. These kind of sloppy files are a mess and a PITA to work with. I wouldn't be surprised if calibre just choked on them.

I think you may be right. This book does have some formatting issues too.

[Katsunami]

Originally Posted by jswinden

It could be as simple as sloppy HTML/CSS. I've modified several eBooks in the past that looked like they were created by MS Word then incorrectly converted to HTML. In these books, rather than use simple and effective CSS class and heading tags to assign formatting, they use span tags on nearly every line of HTML. These kind of sloppy files are a mess and a PITA to work with. I wouldn't be surprised if calibre just choked on them.

If a book sets the font on every paragraph, Calibre will be unable to embed a font. It will embed it allright, but the CSS on each paragraph will override it.

The only way to fix that is to open the book in Sigil (if it's not an EPUB you must convert to it first, further messing up the code), and then use a regular expression to rip all the font-family properties out of the <p> tags.

Maybe the book is even using <font> tags...

Coding like that is really old style. If you do it on a website nowadays, you'll probably get ridiculed, fired, stoned, or shot.

[Atunah]

I tried that embedding thingy too in calibre, but I am calibre and "hack" challenged, so of course I wasn't going to figure that one out.
I have the PW1 and couldn't even figure out how to get fonts on there and then I realized, I didn't really want to to that in case something goes wrong.

So I tried the embedding to see what another font would look like, but none of the books I tried it on show it on my PW.

Oh well, it is what it is.

[sparklemotion]

Originally Posted by Atunah

I tried that embedding thingy too in calibre, but I am calibre and "hack" challenged, so of course I wasn't going to figure that one out.
I have the PW1 and couldn't even figure out how to get fonts on there and then I realized, I didn't really want to to that in case something goes wrong.

So I tried the embedding to see what another font would look like, but none of the books I tried it on show it on my PW.

Oh well, it is what it is.

Did you see dordale's earlier post? It might be the problem.

Originally Posted by dordale

I had that problem (publisher font showing as helvetica) when I first started trying to embed my own fonts with Calibre. Turns out I under the "Filter Style Information" tab on the "Look and Feel" section of the conversion screen I also had the box checked to "completely remove fonts". When I unchecked that and tried the conversion again, it worked perfectly. (Thanks to booklover6 for pointing me in the right direction on this!)

[Atunah]

Originally Posted by sparklemotion

Did you see dordale's earlier post? It might be the problem.

I tried so many variations of converting back and forth to this and that so I just gave up.
I have no clue about some of the other stuff that is being talked about here like sigil or CSS. No clue what that is. So for me, unless amazon gives me the option to unlock or put properly more fonts on my PW, its just way above my skill.

Maybe one day they give us more fonts on the kindles. That is all I can hope for really.

Originally Posted by Katsunami

If a book sets the font on every paragraph, Calibre will be unable to embed a font. It will embed it allright, but the CSS on each paragraph will override it.

The only way to fix that is to open the book in Sigil (if it's not an EPUB you must convert to it first, further messing up the code), and then use a regular expression to rip all the font-family properties out of the <p> tags.

Maybe the book is even using <font> tags...

Coding like that is really old style. If you do it on a website nowadays, you'll probably get ridiculed, fired, stoned, or shot.

I agree it is old school and sloppy! You have to get your hands dirty to clean it up in an HTML editor or Sigil. MS Word is notorious for saving files converted into HTML like that, though Word does have alternate save options some of the eBook writers/editors/publishers still pick the wrong save as options.

Basically, if they use the STYLE tag with any formatting within the HTML tags, that is a no no. I also hate it when idiots use all P tags even for headings, and just change the style of the P tags' class(es) to match an H1, H2,...Hx. Bad form if you ask me.

[Hitch]

Originally Posted by jswinden

It could be as simple as sloppy HTML/CSS. I've modified several eBooks in the past that looked like they were created by MS Word then incorrectly converted to HTML. In these books, rather than use simple and effective CSS class and heading tags to assign formatting, they use span tags on nearly every line of HTML. These kind of sloppy files are a mess and a PITA to work with. I wouldn't be surprised if calibre just choked on them.

Generally, that's not a Word file output, even from the biggest idiots. THAT, boys and girls, is a PDF that's been "saved as Word." Or output from Adobe Acrobat Pro X or XI. Sometimes, depending on the designer, output as RTF from within InDesign, "with formatting." OR, it's been run through one of those "Convert PDF to Word!" freebie websites (or worse, some that charge).

You mostly won't find all those spans in a straight-up Word-->HTML file. Those spans tend to be exclusively from Adobe. (Of course, you will see them for things like "text-transform" or "smallcaps," and any special font features that an author may use, even directly from Word.)

Hope that helps. Use regex to clean up valueless spans--you can do it in one-two passes. Then you'll have fundamentally clean HTML to work with. In the alternative, I highly recommend Toxaris' Word eBook Tools program--works a treat.

P.S.: on the font-embedding front, I'll tell you guys what I tell my clients. The biggest "bang" for the buck is simply embedding a title page and chapter heads. Leave the body text, so that you can change it depending on time of day, lighting and mood. I do.

Hitch

I created a simple test book in Sigil to illustrate how a few different fonts look on the Paperwhite 2nd generation (see attached). I included a short paragraph for each font as follows:

• Caecilia Kindle font
• Caecilia Heavy
• DejaVuSerifEInk (by Katsunami)
• LexiaDaMa
• Amasis

I just wanted to see what various fonts would look like on my PW2.

I deleted the ePub and mobi attachments because I'm not sure about the distribution of all of these fonts. I attached a screen capture from PW2 in their place.