Ultimate PDF to Epub/Mobi conversion tips - Page 2

JSWolf · 01-10-2013, 10:40 AM

Quote:

Originally Posted by BlackVoid

Yes, that is true if you want a perfect conversion. But a readable conversion can be had with much less work. I used to use BookDesigner 5 to convert from PDF to LRF with acceptable results and minimal editing. Unfortunately BookDesigner does not support ePub, so an extra step would be needed.

Calibre is also one of the worst converters, it totally ruins paragraph and page layout with 1 line and empty pages, etc.

I will try the methods described by the thread starter.

Calibre is not as bad as you say. But given that I like my formatting a certain way, I do take the Calibre converted ePub and tweak it so it displays as I want. I also make sure the underlying code is fairly clean.

Pranananda · 02-17-2013, 10:49 PM

saoir, on the Mac you might try the free application TextWrangler. It has regular expression support, so you can find and replace tags. It also can find text across line boundaries. And you can specific a range using the text selection for where a "Replace All" operation will take place. It also allows you to write perl or shell scripts and use those scripts to modify the selected text.

gekkoh · 08-19-2014, 04:43 AM

This is a great topic! Lots of great tips.

I tried converting a PDF to epub, but having some issues — the book has lots of tables and equations, and they all just end up like regular text thrown around the page and totally messed up.

Does anyone have any tips to handle tables and equations? Can they be automatically converted to images to make sure they look exactly as in the PDF?

Thanks!

Toxaris · 08-19-2014, 07:20 AM

For the best PDF to ePUB result you really must use the OCR method. That is still the best method and after the OCR you have to check the results with the PDF to ensure it is correct.
For example ABBYY has support for tables, but I would not recommend exporting to ePUB directly. It will generate a mess.
Equations are troublesome, since MathML is not supported in ePUB2 (it should be in ePUB3, but most readers or applications cannot read ePUB3). I usually recreate the equation at the Codecogs site and create a SVG from it. That one can be easily imported into an ePUB.

mrmikel · 08-19-2014, 07:51 AM

Assuming you can convert tables or formulas, there remains the issue of size. A table with over 3-4 columns can get folded, spindled and mutilated by anyone with a finger on the increase text size button.

Sometimes it is better to break them apart, make a rotated image or figure out another way to portray the information than trying to put a 16 column table in a 5" screen epub.

gekkoh · 08-19-2014, 08:10 AM

I was hoping there is some automated way to convert the tables and equations, but I guess I will have to do it all manually.

DaleDe · 08-19-2014, 01:07 PM

Quote:

Originally Posted by gekkoh

I was hoping there is some automated way to convert the tables and equations, but I guess I will have to do it all manually.

There is some tools that can convert equations to SVG. See our wiki for some of them. SVG is supported in ePub 2.

Dale

crich70 · 08-19-2014, 10:31 PM

Quote:

If you are going to use Calibre:
7. Drop your pdf file into Calibre.

8. Convert you pdf into htmlz.

Actually, Htmlz is nothing to do with html. It is a special zip file, just like epub which has all the necessary files inside to produce an ebook.

You can use Calibre's conversion settings to remove fonts, font size, margin etc.

9. Once conversion is complete, open htmlz file with a zip software, you will see a file called index.html along with css and opf files. You can edit index file and repackage.

10. During final conversion, use htmlz as your source.

Could you post some suggestions on what those conversion settings should be to get a good looking result?

JSWolf · 08-19-2014, 10:40 PM

Quote:

Originally Posted by crich70

Could you post some suggestions on what those conversion settings should be to get a good looking result?

There are NO conversion settings that can get an error free conversion. The only way to do that is to 100% A/B compare the converted ePub with the PDF until you have every single character compared.

crich70 · 08-19-2014, 11:04 PM

Quote:

Originally Posted by JSWolf

There are NO conversion settings that can get an error free conversion. The only way to do that is to 100% A/B compare the converted ePub with the PDF until you have every single character compared.

Actually I was referring to the line in the OP:

Quote:

You can use Calibre's conversion settings to remove fonts, font size, margin etc.

When I posted the question I was talking about that. I am aware that there is no magic setting to get a completely 100% clean conversion from PDF to Epub. If someone did manage to develop such a program they could make a fortune I imagine.

Hitch · 08-28-2014, 04:02 AM

Quote:

Originally Posted by crich70

Actually I was referring to the line in the OP:

When I posted the question I was talking about that. I am aware that there is no magic setting to get a completely 100% clean conversion from PDF to Epub. If someone did manage to develop such a program they could make a fortune I imagine.

Hey, Crich:

Given that not even ADOBE can do it, I suspect we can all agree that it's not in the immediate pipeline! ;-)

Hitch

DaleDe · 08-28-2014, 02:05 PM

Quote:

Originally Posted by crich70

Actually I was referring to the line in the OP:

When I posted the question I was talking about that. I am aware that there is no magic setting to get a completely 100% clean conversion from PDF to Epub. If someone did manage to develop such a program they could make a fortune I imagine.

One trick is to print the PDF to a PDF file in lieu of the printer. While printing will lose all intelligence like book marks and live links it will rearrange the text to be in order and make the file more uniform. This can be a major advantage in converting PDF successfully. Many of the PDF files are edited, particularly those made by Adobe. The editing process does not produce a clean document but rather patches it.

Dale

willus · 08-30-2014, 12:46 PM

Quote:

Originally Posted by gekkoh

This is a great topic! Lots of great tips.

I tried converting a PDF to epub, but having some issues [...]

Does anyone have any tips to handle tables and equations? Can they be automatically converted to images to make sure they look exactly as in the PDF?

Thanks for resurrecting this thread. I know some people don't always like reviving old threads, but this was a thread I wouldn't have noticed otherwise. If you want to preserve the look of the PDF file, particularly equations, you might look at the tools in the MR PDF forum. The drawback is that they mostly convert PDF only to (re-formatted) PDF, not PDF to ePub. But most mobile devices can read PDF files just fine, particularly if the PDF pages are sized for the device. You just get a different mix of features compared to the ePub format, as you no doubt are aware, and part of that depends on the device itself (some devices have better PDF viewing software than others). It depends on what your end goal is (e.g. do you just want to get a document to where you can read it on your device, or do you truly want to convert it, as accurately as possible, to a new format).

shevirsy · 09-12-2014, 05:20 AM

Anybody knows a way with *libre* software?

Alix78 · 10-07-2014, 12:11 PM

I had to register on here just to say a BIG THANK YOU for this post! Since a few days I have my Kindle Paperwhite, but to my big frustration my PDF books looked awful on it. I was searching the net for hours to find a solution. I almost gave up because I read that it's not possible to make a PDF look neat on a Kindle. Then I stumbled upon your post and you saved me!
Big kudos to guys out there doing these sort of things and making our expensive gears be more user friendly!

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
The Ultimate Kobo Tips & Tricks Thread	ficbot	Kobo Reader	449	01-25-2023 10:00 AM
Conversion Issues - PDF/Word to Mobi/ePub	MajC	Conversion	1	02-14-2011 08:27 AM
Any chance for chm to epub/mobi/pdf conversion support	joblack	Calibre	4	11-02-2010 01:06 AM
PDF to Mobi Conversion	rayh	Calibre	2	09-24-2010 02:33 AM
Epub/Mobi TO pdf conversion problem	Hitch	Calibre	4	06-15-2010 05:28 PM

02-17-2013, 10:49 PM	#17
Pranananda Connoisseur Posts: 98 Karma: 122982 Join Date: Apr 2010 Location: Humboldt County, California Device: ipad, iPod touch, JetBook Lite	saoir, on the Mac you might try the free application TextWrangler. It has regular expression support, so you can find and replace tags. It also can find text across line boundaries. And you can specific a range using the text selection for where a "Replace All" operation will take place. It also allows you to write perl or shell scripts and use those scripts to modify the selected text.

08-19-2014, 04:43 AM	#18
gekkoh Junior Member Posts: 2 Karma: 10 Join Date: Aug 2014 Device: iPad	This is a great topic! Lots of great tips. I tried converting a PDF to epub, but having some issues — the book has lots of tables and equations, and they all just end up like regular text thrown around the page and totally messed up. Does anyone have any tips to handle tables and equations? Can they be automatically converted to images to make sure they look exactly as in the PDF? Thanks!

08-19-2014, 07:20 AM	#19
Toxaris Wizard Posts: 4,520 Karma: 121692313 Join Date: Oct 2009 Location: Heemskerk, NL Device: PRS-T1, Kobo Touch, Kobo Aura	For the best PDF to ePUB result you really must use the OCR method. That is still the best method and after the OCR you have to check the results with the PDF to ensure it is correct. For example ABBYY has support for tables, but I would not recommend exporting to ePUB directly. It will generate a mess. Equations are troublesome, since MathML is not supported in ePUB2 (it should be in ePUB3, but most readers or applications cannot read ePUB3). I usually recreate the equation at the Codecogs site and create a SVG from it. That one can be easily imported into an ePUB.

08-19-2014, 07:51 AM	#20
mrmikel Color me gone Posts: 2,089 Karma: 1445295 Join Date: Apr 2008 Location: Central Oregon Coast Device: PRS-300	Assuming you can convert tables or formulas, there remains the issue of size. A table with over 3-4 columns can get folded, spindled and mutilated by anyone with a finger on the increase text size button. Sometimes it is better to break them apart, make a rotated image or figure out another way to portray the information than trying to put a 16 column table in a 5" screen epub.

08-19-2014, 08:10 AM	#21
gekkoh Junior Member Posts: 2 Karma: 10 Join Date: Aug 2014 Device: iPad	I was hoping there is some automated way to convert the tables and equations, but I guess I will have to do it all manually.

09-12-2014, 05:20 AM	#29
shevirsy Banned Posts: 28 Karma: 31454 Join Date: Sep 2014 Location: France Device: Kindle 3	Anybody knows a way with libre software?

10-07-2014, 12:11 PM	#30
Alix78 Junior Member Posts: 1 Karma: 12696 Join Date: Oct 2014 Device: Kindle Paperwhite	I had to register on here just to say a BIG THANK YOU for this post! Since a few days I have my Kindle Paperwhite, but to my big frustration my PDF books looked awful on it. I was searching the net for hours to find a solution. I almost gave up because I read that it's not possible to make a PDF look neat on a Kindle. Then I stumbled upon your post and you saved me! Big kudos to guys out there doing these sort of things and making our expensive gears be more user friendly!

Advert

Advert