Preserving foreign characters from epub to mobi in Calibre

markpearl · 07-30-2011, 04:01 PM

I bought an epub book and have been trying to convert it to mobi for my wife's Kindle. It's translated from Japanese to English, but some of the principle characters named "Sōseki" and "Sanshirō" (I hope the symbols can be seen with everyone's browser), for example, are still accented.

With no "look & feel" options checked, the accented characters appear as question marks in mobi. And when I convert with the "unicode to ASCII" option, the accented nouns are replaced with "vanilla" English equivalents -- as the the option says it does.

Is there any way to preserve the identical accent marks and symbols in mobi with Calibre's conversion process? Any settings I may have overlooked? Or is this the best I can expect?

Thank you in advance for some assistance with this issue!

itimpi · 07-30-2011, 04:51 PM

A question mark appearing in the mobi file is a symptom of the Kindle not having font/code page combination for the character. To handle this sort of characters you need to specify the code page to be used. The reader also needs to have fonts corresponding to the specified code page.

In theory the ePub file should already be specifying the code page, but it is possible that it is being handled incorrectly. You could alwys use the Calibre 'teak epub' feature to explode the epub file and peek at the html contained in it to see if that helps - often the code page is specified at the start of the html file. Unfortunately working out the correct code page can be a bit trial-and -error. You might find that it is something like utf-8 if you are lucky.

ldolse · 07-30-2011, 10:53 PM

Valid ePub already has to be UTF8 or UTF16, no other code page is allowed. Granted I agree that I have seen badly encoded ePubs as well, but usually the ePub won't display the characters correctly in that case.

You might want to check this page:
http://freekindlebooks.org/Unicode/unicode.html

If the Kindle can support that character or a similar glyph it would show up in one of those test files. You can use search and replace to search for the any variant it does support.

I suspect 'transliterate unicode to ascii' might be your only option that doesn't require hacking the Kindle's fonts, etc.

markpearl · 08-02-2011, 10:47 AM

Thanks for your replies!

Oddly enough, my issues with the distorted unicodes disappeared when I transferred the converted mobi to my wife's Kindle! The problem was with the way Kindle for PC displayed it, not with the actual Kindle. Next time I'll know better before suspecting an "issue" with Calibre :0)

07-30-2011, 04:01 PM	#1
markpearl Enthusiast Posts: 43 Karma: 576732 Join Date: Jun 2011 Device: Kindle	Preserving foreign characters from epub to mobi in Calibre I bought an epub book and have been trying to convert it to mobi for my wife's Kindle. It's translated from Japanese to English, but some of the principle characters named "Sōseki" and "Sanshirō" (I hope the symbols can be seen with everyone's browser), for example, are still accented. With no "look & feel" options checked, the accented characters appear as question marks in mobi. And when I convert with the "unicode to ASCII" option, the accented nouns are replaced with "vanilla" English equivalents -- as the the option says it does. Is there any way to preserve the identical accent marks and symbols in mobi with Calibre's conversion process? Any settings I may have overlooked? Or is this the best I can expect? Thank you in advance for some assistance with this issue! Last edited by markpearl; 07-30-2011 at 04:05 PM.

07-30-2011, 04:51 PM	#2
itimpi Wizard Posts: 4,553 Karma: 950151 Join Date: Nov 2008 Device: Sony PRS-950, iphone/ipad (Marvin/iBooks/QuickReader)	A question mark appearing in the mobi file is a symptom of the Kindle not having font/code page combination for the character. To handle this sort of characters you need to specify the code page to be used. The reader also needs to have fonts corresponding to the specified code page. In theory the ePub file should already be specifying the code page, but it is possible that it is being handled incorrectly. You could alwys use the Calibre 'teak epub' feature to explode the epub file and peek at the html contained in it to see if that helps - often the code page is specified at the start of the html file. Unfortunately working out the correct code page can be a bit trial-and -error. You might find that it is something like utf-8 if you are lucky. Last edited by itimpi; 07-30-2011 at 04:54 PM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
PRS-505 cannot show foreign characters?	paulpeer	Sony Reader	19	05-22-2013 02:25 AM
foreign characters not showing up?	sovre	ePub	18	07-29-2011 12:10 PM
Preserving images from DRM'd mobi	Paulinafrica	Amazon Kindle	15	05-18-2011 01:16 PM
Preserving format from ePub (Sigil) to Mobi	jeff47	Calibre	9	10-22-2010 10:17 PM
Foreign accented characters and libprs500	Stingo	Calibre	6	02-24-2008 07:51 PM

07-30-2011, 10:53 PM	#3
ldolse Wizard Posts: 1,337 Karma: 123455 Join Date: Apr 2009 Location: Malaysia Device: PRS-650, iPhone	Valid ePub already has to be UTF8 or UTF16, no other code page is allowed. Granted I agree that I have seen badly encoded ePubs as well, but usually the ePub won't display the characters correctly in that case. You might want to check this page: http://freekindlebooks.org/Unicode/unicode.html If the Kindle can support that character or a similar glyph it would show up in one of those test files. You can use search and replace to search for the any variant it does support. I suspect 'transliterate unicode to ascii' might be your only option that doesn't require hacking the Kindle's fonts, etc.

08-02-2011, 10:47 AM	#4
markpearl Enthusiast Posts: 43 Karma: 576732 Join Date: Jun 2011 Device: Kindle	Thanks for your replies! Oddly enough, my issues with the distorted unicodes disappeared when I transferred the converted mobi to my wife's Kindle! The problem was with the way Kindle for PC displayed it, not with the actual Kindle. Next time I'll know better before suspecting an "issue" with Calibre :0)

Advert