No double LL's-PDF to EPUB

[Pancho Harrera]

Everytime i convert a PDF to Epub , the epub will skip the first L and you will only see the last L. For example actua ly. What do i do? I have checked the "keep ligatures" box , but still doesen't work.

[AprilHare]

What application are you using to generate EPUB files from PDFs?
I assume you're testing with an iPhone - what app?
Details, details..

[tomsem]

I gather you are using calibre for the conversion.

For reasons I don't understand, some double 'l' sequences in some PDFs are converted successfully, and some are not. In each case where I've investigated the PDF source, they are not encoded as ligatures, and nevertheless many 'll' sequences are converted to 'l ' in the HTML. So it is not about ligatures, I'm not surprised that option has no effect.

I haven't seen a problem with other letter combinations, but it's entirely possible that such problems exist.

I can only think that it's a bug in the (open source?) PDF converter code that calibre uses: if I export a 'problem' PDF to HTML from Acrobat Professional, the HTML has the correct sequence of characters, but it is deficient in other respects.

My attempts to reproduce a problem by creating my own PDFs and converting with calibre have been unsuccessful.

It is weird, and frustrating, and I'm not sure there are better one click PDF conversion options available at any reasonable price. You can always clean up the results, of course, but that often involves more time than I'm willing to invest.

[ardeegee]

I was just about to ask a similar question but I searched for posts with "ligatures" first. I'm having the same problem with a PDF. Text looks fine in the PDF, I export it to a text file and each of the "bad" characters is exported as a period. When Calibre is used to convert it (and when one on-line converter I found is used) the following problems happen (there might be more combinations, but this is what I've found):

If "ll", second one missing. If "ff", both missing. If "tt", second one missing. If "fl", both missing.

[ardeegee]

I think I know why my PDF, at least, has conversion problems. It has a font called "Charis SIL" embedded, and is something called a "smart font"

Charis SIL is a TrueType font with “smart font” capabilities added using the Graphite, OpenType®, and AAT font technologies. This means that complex typographic issues such as the placement of diacritics or the formation of ligatures are handled by the font, provided you are running an application that provides an adequate level of support for one of these smart font technologies. With the old font (and its derivatives), diacritic placement was handled using non-standard character encodings that incorporated multiple versions of a diacritic as distinctly-encoded characters.

http://scripts.sil.org/cms/scripts/p...=charissilfont

[tomsem]

Originally Posted by ardeegee

I think I know why my PDF, at least, has conversion problems. It has a font called "Charis SIL" embedded, and is something called a "smart font"

Charis SIL is a TrueType font with “smart font” capabilities added using the Graphite, OpenType®, and AAT font technologies. This means that complex typographic issues such as the placement of diacritics or the formation of ligatures are handled by the font, provided you are running an application that provides an adequate level of support for one of these smart font technologies. With the old font (and its derivatives), diacritic placement was handled using non-standard character encodings that incorporated multiple versions of a diacritic as distinctly-encoded characters.

http://scripts.sil.org/cms/scripts/p...=charissilfont

That is interesting. I checked my 'problem' PDF to see if it matched the pattern you've identified. Sure enough, my conversion problem involves Adobe Garamond Pro, an OpenType font and therefore (I assume) one with smart font capabilities. The same PDF has a small sample of text with some of the same letter combinations, using a different font (TimesNewRoman Italic, I believe, which is a TrueType font and therefore without smart font capabilities), and this does not show a conversion problem.

But when I tried to create a new PDF using the problematic letter combinations and using these two fonts, the conversion went ok. So I'm still only able to see a problem with this one PDF file, and never with anything I create myself.

I still think it points to a problem with the (open source?) PDF library calibre is using, because it cannot convert this particular PDF file to any other format without corruption, whereas Acrobat, and Adobe's free PDF2TXT and PDF2HTML services, don't have this issue.

[Freeshadow]

that's why PDF is for printing

complex typographic issues

[JSWolf]

Originally Posted by ardeegee

I think I know why my PDF, at least, has conversion problems. It has a font called "Charis SIL" embedded, and is something called a "smart font"

Charis SIL is a TrueType font with “smart font” capabilities added using the Graphite, OpenType®, and AAT font technologies. This means that complex typographic issues such as the placement of diacritics or the formation of ligatures are handled by the font, provided you are running an application that provides an adequate level of support for one of these smart font technologies. With the old font (and its derivatives), diacritic placement was handled using non-standard character encodings that incorporated multiple versions of a diacritic as distinctly-encoded characters.

http://scripts.sil.org/cms/scripts/p...=charissilfont

That font family is used in a lot of commercial ePub since it does look better then the default ADE serif font.