Conveting huge PDF to EPUB, what should I expect?

[Xzahn]

Hi folks,
I'm currently running a converting job on 1400 page high-quality PDF with lots of pictures. The file is 260MBs.
I ran the file before and it went for about 800 minutes before I stopped it because I thought it was bugged or something. Then I read that large files can take hours to convert and as long as I do not get errors the program is working. So I ran it again and it has been 530 minutes so far. Calibre currently takes 15% of my CPU and 2GB of RAM.
What should I expect? Anyone converted something like that? How long did it take?
I've been using the program for 2 days now and converted a few 50-80MBs books. All took less than a minute to complete except a 30MB one that took 40 mins. File size doesn't seem to have any direct effect.
By the way, can I somehow direct CPU power towards Calibre to speed up the process? I'm already using the 64-bit version.
Thanks!

[Divingduck]

With complex documents it will take some time. I had conversions in the past with a duration about 36 hours (...but not for a pdf). As long as the little job symbol in bottom right is animated the conversion is still running.
In Preferences ->Behavior you can tweak the Job priority, but keep in mind your system will be not responsive during long tasks.
A last tweak, depending on your system is in Preferences ->Miscellaneous for the usage of CPU cores (but I do not know if this have really an impact for a conversion of one file).


Calibre maybe isn't the best solution for this kind of project. If you have an actual Acrobat and MS Word, the better way around is to open the PDF via Acrobat and save the file as docx MS-Word file or, alternative open the PDF directly in MS-Word (you need minimum version 2013, better is version 2016) and save it as docx file. This file you can edit to your needs and then convert with calibre with a much better result.

[PeterT]

You could also use one of the PDF splitting tools to get multiple smaller PDFs and then do multiple conversions in parallel.

Sent from my Nexus 7 using Tapatalk

[Xzahn]

Originally Posted by Divingduck

In Preferences ->Behavior you can tweak the Job priority...

...in Preferences ->Miscellaneous for the usage of CPU cores

...the better way around is to open the PDF via Acrobat and save the file as docx... This file you can edit to your needs and then convert with calibre with a much better result.

Very useful information!

The file converted for about 650 minutes. It ended up half the size of the original but it was corrupted and unusable.
A few other files converted successfully but the text was mangled and unreadable. It seemed to work fine for a few others though. So the type of the PDF seems to affect the outcome greatly.

Before I used K2pdfopt and it would just rearrange the text and zoom in for a better read in eBook readers. It seemed to be working alright and it's fast but definitely not the cleanest solution.

Originally Posted by PeterT

You could also use one of the PDF splitting tools to get multiple smaller PDFs and then do multiple conversions in parallel.

Sent from my Nexus 7 using Tapatalk

Also very useful information especially for large books. An off-topic question: have you used your Nexus 7 for reading? How do you find it?

And thank you, both for the info!

[PeterT]

I don't read much on the Nexus; I do most of my reading on a Kobo Glo HD. I do have Libby, Moon+ Reader and a few others on the tablet, and it's ok; I just prefer the eReader.

On a tablet I think it's more a question of the app being used as opposed to the device.

Sent from my Nexus 7 using Tapatalk

[DoctorOhh]

Just an FYI, there are sticky posts at the top of each forum that often have answers to questions. The sticky post at the top of the Conversion forum titled, "Read this before Posting PDF Questions" doesn't talk about time, but it does cover a bunch of other pertinent info in regards to converting PDF files. And in the second sentence links to the calibre manual titled, Convert PDF documents.

The section of the calibre manual concludes with the following statement.

To re-iterate PDF is a really, really bad format to use as input. If you absolutely must use PDF, then be prepared for an output ranging anywhere from decent to unusable, depending on the input PDF.