Problem converting to azw to epub

bobcdy · 05-03-2015, 10:12 PM

Recently I purchased a azw book from Amazon and converted it to epub with the most recent calibre version. It converted ok, but there were abundant spelling errors throughout the book. It looked like the results of a ocr that had converted text and couldn't distinguished many letters and letter combos. I had never encountered such with calibre before.

In addition, after the conversion the results were an epub and an htmlz book - usually calibre (2013 version) gives the original azw file rather than an htmlz file.

Does anyone have any ideas what happened and how to correct it?

eschwartz · 05-03-2015, 10:24 PM

HTMLZ? I regret to inform you that you have stumbled upon one of the rare Topaz format experiments, and discovered exactly why it is so horrible.

Topaz used, IIRC, some form of embedded image, backed by OCR. The DeDRM plugin and calibre as well can only extract the OCR layer.

Basically it is the PDF of ebook formats.

cybmole · 05-04-2015, 12:59 AM

you have 7 days to return any e-book to amazon if you are not happy. they will give a full refund. Just don't make a habit of it or your account could get suspended for excessive returns

bobcdy · 05-04-2015, 06:08 PM

Thanks to both of you for the information about my kindle book conversion. The Kindle book has almost no errors but there are tons of spelling errors in the epub conversion. I'm in the process of correcting them with Sigil spellcheck.

eschwartz · 05-04-2015, 07:13 PM

The Topaz format looks proper when you first read it, because it uses images of the proper text. But then you try DeDRMing and converting, or highlighting, and you hit the invisible OCR layer.

The Kindle book has lots of errors, just not in the image overlay.

AnemicOak · 05-05-2015, 10:43 AM

Quote:

Originally Posted by bobcdy

Thanks to both of you for the information about my kindle book conversion. The Kindle book has almost no errors but there are tons of spelling errors in the epub conversion. I'm in the process of correcting them with Sigil spellcheck.

Italics will also be missing so if you want those you'll need to put them in manually.

HarryT · 05-05-2015, 12:04 PM

Quote:

Originally Posted by eschwartz

HTMLZ? I regret to inform you that you have stumbled upon one of the rare Topaz format experiments, and discovered exactly why it is so horrible.

Topaz used, IIRC, some form of embedded image, backed by OCR. The DeDRM plugin and calibre as well can only extract the OCR layer.

Basically it is the PDF of ebook formats.

Topaz is absolutely fine for the purpose it's intended for - it's just not a good format for conversion. Its intended use is the situation when the source material is not available in digital form, and you either can't afford to do a digital conversion, or the material isn't amenable to digital conversion.

I have a number of old Egyptology books in Topaz format that really can't practically be converted to Mob/ePub format because they contain a lot of non-standard glyphs (Egyptian hieroglyphs, etc). Topaz is an ideal format for such material, because it's a fraction the size that a page-scan would be, and it's reflowable.

eschwartz · 05-05-2015, 02:37 PM

Hmm, I suppose it *might* have certain niche use. I tend not to read books with those character sets, and generally don't think of them.

One Topaz book that I remember getting was an innocuous copy of The Princess Bride.

Some lazy publisher didn't want to bother converting, I guess.

Question: Isn't that what the Private Use Areas of unicode are for, and shouldn't those books use a custom font for that purpose, as part of an otherwise-bog-standard EPUB/AZW3?

HarryT · 05-05-2015, 02:51 PM

Sure, but these are 100+ year-old books that have been out of print for decades, and there's just not the market that would make it economic to create ePub versions of them at a sensible price. I'm happy to have them available, and Topaz works well if you don't have a requirement to convert them.

bobcdy · 05-08-2015, 12:03 PM

I first concluded I could correct the spelling mistakes but found that there were no italics and even worse there were no paragraphs - each chapter was one paragraph. I could have inserted them with Sigil (I have the paperback version, but the small print of the 600p book is getting too difficult for my eyes for extended reading) but was too much work so I gave up on correcting the conversion. Lack of paragraphs was too much, so I finally read the topaz version on my largest android tablet. Larger text in the newer Kindle for android is possible, although the margins are comparatively wide with larger text sizes. In the topaz often the letters were poorly formed, esp. the H in which the first upright looked like an I next to the crossbar and right upright together. I won't send the topaz back because I read it, but probably will if I ever buy another one.

05-03-2015, 10:12 PM	#1
bobcdy Fanatic Posts: 527 Karma: 1048576 Join Date: May 2009 Device: bebook; prs-950; nook simple touch; HTC Jetstream tablet	Problem converting to azw to epub Recently I purchased a azw book from Amazon and converted it to epub with the most recent calibre version. It converted ok, but there were abundant spelling errors throughout the book. It looked like the results of a ocr that had converted text and couldn't distinguished many letters and letter combos. I had never encountered such with calibre before. In addition, after the conversion the results were an epub and an htmlz book - usually calibre (2013 version) gives the original azw file rather than an htmlz file. Does anyone have any ideas what happened and how to correct it? Last edited by bobcdy; 05-03-2015 at 10:17 PM.

05-08-2015, 12:03 PM	#10
bobcdy Fanatic Posts: 527 Karma: 1048576 Join Date: May 2009 Device: bebook; prs-950; nook simple touch; HTC Jetstream tablet	I first concluded I could correct the spelling mistakes but found that there were no italics and even worse there were no paragraphs - each chapter was one paragraph. I could have inserted them with Sigil (I have the paperback version, but the small print of the 600p book is getting too difficult for my eyes for extended reading) but was too much work so I gave up on correcting the conversion. Lack of paragraphs was too much, so I finally read the topaz version on my largest android tablet. Larger text in the newer Kindle for android is possible, although the margins are comparatively wide with larger text sizes. In the topaz often the letters were poorly formed, esp. the H in which the first upright looked like an I next to the crossbar and right upright together. I won't send the topaz back because I read it, but probably will if I ever buy another one. Last edited by bobcdy; 05-08-2015 at 12:14 PM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[Closed] AZW not converting to ePub	Tatezi	Conversion	9	03-22-2015 11:59 PM
Problem converting AZW to epub	Alleyoop	Sony Reader	6	03-17-2012 06:32 AM
Maintaining highlighted text converting AZW to EPUB?	Mkoenig	Conversion	2	06-04-2011 09:48 PM
Converting AZW to ePub	narfit	Introduce Yourself	3	02-28-2011 09:18 AM
Problem converting Google Scholar HTML to AZW	JerryCC	Amazon Kindle	5	02-08-2010 08:38 PM

05-03-2015, 10:24 PM	#2
eschwartz Ex-Helpdesk Junkie Posts: 19,421 Karma: 85397180 Join Date: Nov 2012 Location: The Beaten Path, USA, Roundworld, This Side of Infinity Device: Kindle Touch fw5.3.7 (Wifi only)	HTMLZ? I regret to inform you that you have stumbled upon one of the rare Topaz format experiments, and discovered exactly why it is so horrible. Topaz used, IIRC, some form of embedded image, backed by OCR. The DeDRM plugin and calibre as well can only extract the OCR layer. Basically it is the PDF of ebook formats.

05-04-2015, 12:59 AM	#3
cybmole Wizard Posts: 3,720 Karma: 1759970 Join Date: Sep 2010 Device: none	you have 7 days to return any e-book to amazon if you are not happy. they will give a full refund. Just don't make a habit of it or your account could get suspended for excessive returns

05-04-2015, 06:08 PM	#4
bobcdy Fanatic Posts: 527 Karma: 1048576 Join Date: May 2009 Device: bebook; prs-950; nook simple touch; HTC Jetstream tablet	Thanks to both of you for the information about my kindle book conversion. The Kindle book has almost no errors but there are tons of spelling errors in the epub conversion. I'm in the process of correcting them with Sigil spellcheck.

05-04-2015, 07:13 PM	#5
eschwartz Ex-Helpdesk Junkie Posts: 19,421 Karma: 85397180 Join Date: Nov 2012 Location: The Beaten Path, USA, Roundworld, This Side of Infinity Device: Kindle Touch fw5.3.7 (Wifi only)	The Topaz format looks proper when you first read it, because it uses images of the proper text. But then you try DeDRMing and converting, or highlighting, and you hit the invisible OCR layer. The Kindle book has lots of errors, just not in the image overlay.

05-05-2015, 02:37 PM	#8
eschwartz Ex-Helpdesk Junkie Posts: 19,421 Karma: 85397180 Join Date: Nov 2012 Location: The Beaten Path, USA, Roundworld, This Side of Infinity Device: Kindle Touch fw5.3.7 (Wifi only)	Hmm, I suppose it might have certain niche use. I tend not to read books with those character sets, and generally don't think of them. One Topaz book that I remember getting was an innocuous copy of The Princess Bride. Some lazy publisher didn't want to bother converting, I guess. Question: Isn't that what the Private Use Areas of unicode are for, and shouldn't those books use a custom font for that purpose, as part of an otherwise-bog-standard EPUB/AZW3?

05-05-2015, 02:51 PM	#9
HarryT eBook Enthusiast Posts: 85,544 Karma: 93383043 Join Date: Nov 2006 Location: UK Device: Kindle Oasis 2, iPad Pro 10.5", iPhone 6	Sure, but these are 100+ year-old books that have been out of print for decades, and there's just not the market that would make it economic to create ePub versions of them at a sensible price. I'm happy to have them available, and Topaz works well if you don't have a requirement to convert them.

Advert

Advert