Scribusoid

Author Topic: PDFtoEpub converter disaster  (Read 6816 times)

founder

  • Full Member
  • ***
  • Posts: 104
  • Karma: 0
PDFtoEpub converter disaster
« on: February 11, 2012, 10:22:49 pm »
Just checking Kindle and how to come there.

So I downloaded from http://www.pdftokindle.com/

The conversion was a total disaster, the result are some fragments of my text.

It's not small failures, it's nearly nothing to see on Kindle what was origingal in the PDF.

Who knows a more successful way to convert a PDF produced by Scribus to Kindle?

BTW, I make heavy usage of shapes and errors in Scribus.

founder

  • Full Member
  • ***
  • Posts: 104
  • Karma: 0
Re: PDFtoEpub converter disaster
« Reply #1 on: February 12, 2012, 09:44:37 am »
Just tested http://calibre-ebook.com/

It looks a liitle bit better than PEFtoEpub converter, but it's still garbage.

Thought maybe it's a problem how Scribus makes PDF files.

So I converted with both programs a PDF created by OpenOffice. It's the same garbage what comes out.

Meho R.

  • Full Member
  • ***
  • Posts: 198
  • Karma: 0
Re: PDFtoEpub converter disaster
« Reply #2 on: February 12, 2012, 12:45:35 pm »
You should be aware that PDF is intended to be the final output of a document creation process, not a source for further processing/converting, which means that any PDF conversion is unreliable and very likely to be a "garbage", no matter which converter you use, and no matter which app you use to create PDF in the first place. The exception is a PDF file containing only plain text, in which case conversion might go well.

In my experience, Calibre converter is one of the best, but as you noticed, its output still isn't satisfactory in most cases. So, if you need an ebook, the best way (and I dare to say, the proper way) for now is to create it from scratch, either by using "pure" approach, or by using an app like Sigil.
« Last Edit: February 12, 2012, 12:49:40 pm by Meho R. »

Nermander

  • Hero Member
  • *****
  • Posts: 1236
  • Karma: 16
Re: PDFtoEpub converter disaster
« Reply #3 on: February 12, 2012, 12:54:26 pm »
Why go from PDF to ePUB?

It's a bit like going from JPEG to HTML.

As far as I understand it ePUB is more of a markup language than a page description language, so going from postscript or PDF to ePUB does not make sense. The other way around makes more sense.

That was for example what formats like Docbook were made for, to be able to create several different outputs from the same source.

From the wikipedia article on Docbook:
Quote
As a semantic language, DocBook enables its users to create document content in a presentation-neutral form that captures the logical structure of the content; that content can then be published in a variety of formats, including HTML, XHTML, EPUB, PDF, man pages, Web help and HTML Help, without requiring users to make any changes to the source.

founder

  • Full Member
  • ***
  • Posts: 104
  • Karma: 0
Re: PDFtoEpub converter disaster
« Reply #4 on: February 12, 2012, 02:42:50 pm »
Why go from PDF to ePUB?

Only because the Amazon offer for Kindle.
70% royalities for authors publishing with Amazon Kindle.

But seems it's not possible to go with all my shapes and lines used in Scribus to Kindle

Nermander

  • Hero Member
  • *****
  • Posts: 1236
  • Karma: 16
Re: PDFtoEpub converter disaster
« Reply #5 on: February 13, 2012, 08:36:03 am »
Well, you fried your egg and now want it boiled instead. You probably need some magic. Or use a new egg.

a.l.e

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1727
  • Karma: 26
    • the graphics lab
Re: PDFtoEpub converter disaster
« Reply #6 on: February 13, 2012, 03:16:26 pm »
pdf to kindle/epub will something that will be hard to support for scribut...

... but there are serious plans to have a .sla to .epub working (and at least two persons have done some work in that direction!)

providing your work as a pdf / .sla with your wishes how it should look as .epub could help getting the project make some steps further! :-)

ciao
a.l.e

founder

  • Full Member
  • ***
  • Posts: 104
  • Karma: 0
Re: PDFtoEpub converter disaster
« Reply #7 on: February 13, 2012, 05:36:26 pm »
pdf to kindle/epub will something that will be hard to support for scribut...

... but there are serious plans to have a .sla to .epub working (and at least two persons have done some work in that direction!)

providing your work as a pdf / .sla with your wishes how it should look as .epub could help getting the project make some steps further! :-)

ciao
a.l.e

That's great!

Where should I sent the it to?

It shows the biggest problems at the conversion.
I write yellow on the dark part of a photo
It's converted to the photo and after the photo, text in yellow.
This would have been even more funny in white.

I assume it could be neceassary to convert parts like this into PNG graphics
to show right on Kindle.

Nermander

  • Hero Member
  • *****
  • Posts: 1236
  • Karma: 16
Re: PDFtoEpub converter disaster
« Reply #8 on: February 14, 2012, 09:29:25 am »
And that again shows the problem, how should an automatic converter know which text should be made into an image and which text can still be text?

(And I won't go into the problems people using a braille display or text to speech to read an ePUB would get if parts of the text has been converted to images.)

founder

  • Full Member
  • ***
  • Posts: 104
  • Karma: 0
Re: PDFtoEpub converter disaster
« Reply #9 on: February 14, 2012, 09:40:20 am »
And that again shows the problem, how should an automatic converter know which text should be made into an image and which text can still be text?

(And I won't go into the problems people using a braille display or text to speech to read an ePUB would get if parts of the text has been converted to images.)

Maybe we use a new frame types

"Convert to PNG if converted to eBook"
"Convert to JPG if converted to eBook"

All the area covered by this frame types will be converted to PNG of JPG.

I once had to to this manual as I converted my first book to a web site.
Each page of the book one page of the web site.

Nermander

  • Hero Member
  • *****
  • Posts: 1236
  • Karma: 16
Re: PDFtoEpub converter disaster
« Reply #10 on: February 14, 2012, 12:40:42 pm »
Well, that is well outside the scope of Scribus.

As I wrote earlier, there are ways to write a book so you can get both ePUB, PDF, HTML and like a dozen of other output formats without having to do any changes to the source. And those ways have been available for at least 10 years.

I think maybe writing an ePUB import filter for Scribus would be a better thing to look at. Write your book as ePUB, import into Scribus and do some fine tuning of the layout to get a PDF.

a.l.e

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1727
  • Karma: 26
    • the graphics lab
Re: PDFtoEpub converter disaster
« Reply #11 on: February 16, 2012, 05:48:24 pm »
hi founder,

you can attach some sample documents to this thread.

however, from the discussions we had, there won't be any 1:1 conversion of the layout from scribus to epub... it will be more a semantic thing.

ciao
a.l.e

founder

  • Full Member
  • ***
  • Posts: 104
  • Karma: 0
Re: PDFtoEpub converter disaster
« Reply #12 on: February 19, 2012, 09:49:07 pm »
hi founder,

you can attach some sample documents to this thread.

however, from the discussions we had, there won't be any 1:1 conversion of the layout from scribus to epub... it will be more a semantic thing.

ciao
a.l.e

This chapter contains all the ebook format problems

* Pictures where I write white or yellow on it
* Charts created with shapes

I made for this the proposal with special frame types for the ebook conversion

* Convert what is below this frame to a PNG
* Convert what is below this frame to a JPG

a.l.e

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1727
  • Karma: 26
    • the graphics lab
Re: PDFtoEpub converter disaster
« Reply #13 on: February 21, 2012, 02:04:21 pm »
hi founder

i had a look at your documents and there are things in there which are hard to support while converting a document from scribus to epub.

if you want to create good epubs, i think that creating PNGs and JPGs is a no go. the result must be scalable!

now, the question is indeed if there is a way to export parts of the document as artwork: maybe...

i will submit your sample to the person who is thinking about the user experience, usability and features and we will see what she says :-)

ciao
a.l.e

founder

  • Full Member
  • ***
  • Posts: 104
  • Karma: 0
Re: PDFtoEpub converter disaster
« Reply #14 on: April 19, 2012, 06:35:42 pm »
hi founder,

you can attach some sample documents to this thread.

however, from the discussions we had, there won't be any 1:1 conversion of the layout from scribus to epub... it will be more a semantic thing.

ciao
a.l.e

This chapter contains all the ebook format problems

* Pictures where I write white or yellow on it
* Charts created with shapes

I made for this the proposal with special frame types for the ebook conversion

* Convert what is below this frame to a PNG
* Convert what is below this frame to a JPG

I use most times Scribus direct as a vector drawing program.
So maybe a frame

* Convert what is below into a vector graphic