About ePub > PDF conversion

roger64 · 05-23-2018, 12:38 PM

Hi

I have used for a long time the Prince PDF plugin. I have been very favorably impressed by the quality of Calibre ePub-to-PDF conversion.

I have some questions.

1- the PDF produced by Calibre does not hyphen the text (and the hyphenation is not explicity forbidden or limited in the html). Though it does not seem to be important for English texts, it can give sometimes an irregular aspect to the display of text when it's written in languages which use lots of long words. Can hyphenation be added (even as an option)?

2- the size of the output PDF is amazingly compact. I can only match it with Prince if I use cpdf for "squeezing" further the PDF. How do you compress the output PDF?

3- Once the exact settings have been found (which requires some tries), it would be nice if we could save the output profile somewhere. One can imagine that we may use several kinds of PDF, one for each ebook-reader size. Maybe I am misled, but I did not find such an option.
Even repeating the same conversion is no small feat. As it seems that Calibre fails to keep faithfully the last settings, the user has to remember several values scattered among several screens.

4- About conversion itself, I noticed two minor points:
a- drop-caps over two lines are not converted correctly (Calibre displays them on the ePub OK) Drop-caps over three lines are OK everywhere.
b- some full page images are converted with the wrong margins (Calibre displays them on the ePub OK). The latter case may be due to the fact that I did not find the right setting (?). The cover image uses exactly the same code and is perfectly displayed (with margin 0). Other images inside the text, alone in their own file, with exactly the same code, fail to display with margin 0.

I can provide you with the source ePub.

kovidgoyal · 05-23-2018, 11:01 PM

1. Sure it is easy to do, open a bug for it

2. It does not do anythong to compress, it simply has a very efficiently designed PDF backend that I wrote from scratch.

3. All calibre conversion settings are saved. https://manual.calibre-ebook.com/conversion.html (see the section on how settings are saved)

4. THere is no way to gruarantee "correct" conversion of a dropcaps, just as there is no way to guarantee correct display of one. It depends on the font size/dpi/other css settings, you will need to play with them.

5. The best way to display images with margins zero is to use the standard svg cover markup that calibre generates for its epub cover pages. That is guaranteed to work.

roger64 · 05-24-2018, 03:46 AM

Quote:

Originally Posted by kovidgoyal

1. Sure it is easy to do, open a bug for it

2. It does not do anythong to compress, it simply has a very efficiently designed PDF backend that I wrote from scratch.

3. All calibre conversion settings are saved. https://manual.calibre-ebook.com/conversion.html (see the section on how settings are saved)

4. THere is no way to gruarantee "correct" conversion of a dropcaps, just as there is no way to guarantee correct display of one. It depends on the font size/dpi/other css settings, you will need to play with them.

5. The best way to display images with margins zero is to use the standard svg cover markup that calibre generates for its epub cover pages. That is guaranteed to work.

1. bug report opened.

2. thank you

3. Apologies for this. I saved indeed easily my PDF conversion preferences.

4. I'll come back on this one later.

5. About full-page images

I'd like to come back on this point, because this puzzles me: when the right conversion options are ticked, Calibre converts precisely the cover image (margin 0). However, it adds margins to other images using the exact same code, even when they are isolated in their own file.

Here is the very plain code (without even a reference to a stylesheet) I use. I fail to understand why it should not work for all images making use of it.

Spoiler:

This code is well displayed with the ePub (with Calibre viewer too). As for PDF, I think that to allow a different treatment between the cover image and others images using exactly the same code does not seem to be right, the more so since this relative value seem to be well suited for the PDF fixed display.

DoctorOhh · 05-24-2018, 04:06 AM

Quote:

Originally Posted by roger64

Quote:

Originally Posted by kovidgoyal

5. The best way to display images with margins zero is to use the standard svg cover markup that calibre generates for its epub cover pages. That is guaranteed to work.

5. About full-page images

I'd like to come back on this point, because this puzzles me: when the right conversion options are ticked, Calibre converts precisely the cover image (margin 0). However, it adds margins to other images using the exact same code, even when they are isolated in their own file.

I believe this is an example of the "standard svg cover markup" that Kovid is referring to. I grabbed it from the titlepage of one of my books that calibre created.

Code:

  <div>

    <svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" version="1.1" width="100%" height="100%" viewBox="0 0 1200 1600" preserveAspectRatio="none">
      <image width="1200" height="1600" xlink:href="cover.jpeg"/>
    </svg>

  </div>

Adjust as required.

Update: I'm curious how any image with a width of 88.0% wouldn't have left and right margins?

roger64 · 05-24-2018, 04:25 AM

Quote:

Originally Posted by DoctorOhh

Update: I'm curious how any image with a width of 88.0% wouldn't have left and right margins?

Thanks for your example.

To answer your question:
The image is centered.
The display is fixed (this will be a PDF with 9×12 format, fit for my PW3)
No margin is allowed for full-page image display (see code above).

The percentage is calculated according to the shape of the image relative to a 3/4 display. I use for this a -yet- unpublished Sigil plugin.

The end result is that it displays very exactly on this kind of display without any vertical margin. Obviously, due to the shape of the image, a whitespace is automatically left out on both sides (this is the result of the "auto" job). In this particular case, a 6% whitespace will appear on each side. Again, it's the consequence of the undistorted image shape, and is not related to the margins of the document.

DoctorOhh · 05-24-2018, 07:02 AM

Quote:

Originally Posted by roger64

The end result is that it displays very exactly on this kind of display without any vertical margin. Obviously, due to the shape of the image, a whitespace is automatically left out on both sides (this is the result of the "auto" job). In this particular case, a 6% whitespace will appear on each side. Again, it's the consequence of the undistorted image shape, and is not related to the margins of the document.

Thanks for the explanation. In that case this might be the better example for you. It preserves the aspect ratio.

Code:

  <div>

    <svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" version="1.1" width="100%" height="100%" viewBox="0 0 1252 2024" preserveAspectRatio="xMidYMid meet">
      <image width="1252" height="2024" xlink:href="cover1.jpeg"/>
    </svg>

  </div>

Full Titlepage code

Spoiler:

roger64 · 05-24-2018, 11:58 AM

Hi

I did some tries I used the same source ePub containing an image 606×912, size 49.5KB format jpg.

I converted to PDF with Calibre using the code given above
I converted to PDF with Calibre using the code for SVG images provided by the InsertImageSVG v 1.4.0 Sigil plugin.

Spoiler:

These two kinds of code gave exactly the same results*. (see left screenhot1). There are document margins around the image (including vertical margins).

If I convert to PDF with Prince using the code given above, (see right screenshot2). There are NO document margins around the image (only the 6% left and right due to the shape of the image).

So, with Calibre, even with a svg wrapper, the display is not the one expected. The only exception with Calibre, concerns the cover image which can be displayed without document margins.

Why, what can be done for the cover image could not be extended to other full page images using the same code?

* Note that I do not pretend that these two codes are equivalent. If the results are the same for this particular test, it is because the first code is calculated for screens with a 3/4 ratio. There would be differences for others shapes of screen.

DoctorOhh · 05-24-2018, 08:41 PM

Quote:

Originally Posted by roger64

I converted to PDF with Calibre using the code for SVG images provided by the InsertImageSVG v 1.4.0 Sigil plugin.

Spoiler:

These two kinds of code gave exactly the same results*. (see left screenhot1). There are document margins around the image (including vertical margins).

Did they have the margin in the epub when viewed by calibre? If so they should continue to have it during conversion to PDF.

Quote:

Originally Posted by roger64

If I convert to PDF with Prince using the code given above, (see right screenshot2). There are NO document margins around the image (only the 6% left and right due to the shape of the image).

If Prince converts the epub to your liking I guess this answers my question and the epub must look as you expect it to look in calibre's viewer.

The other option may be to print the PDF from the calibre viewer and see how that output looks.

Quote:

Originally Posted by roger64

So, with Calibre, even with a svg wrapper, the display is not the one expected. The only exception with Calibre, concerns the cover image which can be displayed without document margins.

We'll see what Kovid says.

kovidgoyal · 05-24-2018, 10:11 PM

The PDF Output in calibre does not use @page you have to set the margins to zero in the conversion settings. I've never heard of anyone creating PDF files with zero margins, they would not print very well. But, it is a trivial change to add an option to get the plugin to respect @page margins if needed.

kovidgoyal · 05-24-2018, 10:23 PM

Actually, on second thoughts it is probably not a good idea to respect margins in the HTML files, that way would lead to lots of hard to debug-inconsistent rendering.

Instead probably better to just special case full page image markup, just like the viewer does. Open a bug report for it and I will get to it when I have a moment.

roger64 · 05-25-2018, 02:15 AM

When one wish to display a PDF on a small screen (say the widely used six-inch screen), the possibility to display an image really full screen makes a difference, for example for maps, drawings of all kinds. This way we extend the quality of display given to the cover image to other images of the same kind within the book.

@Kovid Goyal

Thanks you very much for your interest. Bug report opened. If need be, I can provide a full ePub for testing purposes by PM.

roger64 · 06-02-2018, 04:48 PM

A short word but a big thank you to Kovid Goyal for enabling so quickly and so efficiently the two points I submitted to him about ePub-to-PDF conversion:

- hyphenation is working quite well (at least for French language) provided you tick the ad hoc check box in the Preferences page (PDF output).

- Calibre can display perfectly all your full page images without document margins. To get it working, you must tick a check box in the Preferences Page PDF output (Use page margins ...) and use a svg wrapper (a compliant code -including @page 0- has been published above).
To ensure that the cover image is not displayed twice (...) declare this image to be the cover_image (right click in the relevant html file).

Again.

kovidgoyal · 06-03-2018, 12:15 AM

You're welcome

05-23-2018, 12:38 PM	#1
roger64 Wizard Posts: 2,608 Karma: 3000161 Join Date: Jan 2009 Device: Kindle PW3 (wifi)	About ePub > PDF conversion Hi I have used for a long time the Prince PDF plugin. I have been very favorably impressed by the quality of Calibre ePub-to-PDF conversion. I have some questions. 1- the PDF produced by Calibre does not hyphen the text (and the hyphenation is not explicity forbidden or limited in the html). Though it does not seem to be important for English texts, it can give sometimes an irregular aspect to the display of text when it's written in languages which use lots of long words. Can hyphenation be added (even as an option)? 2- the size of the output PDF is amazingly compact. I can only match it with Prince if I use cpdf for "squeezing" further the PDF. How do you compress the output PDF? 3- Once the exact settings have been found (which requires some tries), it would be nice if we could save the output profile somewhere. One can imagine that we may use several kinds of PDF, one for each ebook-reader size. Maybe I am misled, but I did not find such an option. Even repeating the same conversion is no small feat. As it seems that Calibre fails to keep faithfully the last settings, the user has to remember several values scattered among several screens. 4- About conversion itself, I noticed two minor points: a- drop-caps over two lines are not converted correctly (Calibre displays them on the ePub OK) Drop-caps over three lines are OK everywhere. b- some full page images are converted with the wrong margins (Calibre displays them on the ePub OK). The latter case may be due to the fact that I did not find the right setting (?). The cover image uses exactly the same code and is perfectly displayed (with margin 0). Other images inside the text, alone in their own file, with exactly the same code, fail to display with margin 0. I can provide you with the source ePub. Last edited by roger64; 05-23-2018 at 04:56 PM.

05-24-2018, 11:58 AM	#7
roger64 Wizard Posts: 2,608 Karma: 3000161 Join Date: Jan 2009 Device: Kindle PW3 (wifi)	Hi I did some tries I used the same source ePub containing an image 606×912, size 49.5KB format jpg. I converted to PDF with Calibre using the code given above I converted to PDF with Calibre using the code for SVG images provided by the InsertImageSVG v 1.4.0 Sigil plugin. Spoiler: Code: <?xml version='1.0' encoding='utf-8'?> <html xmlns="http://www.w3.org/1999/xhtml"> <head> <title></title> <style type="text/css"> @page { padding: 0; margin: 0; } body { text-align: center; padding: 0; margin: 0; } </style> </head> <body> <div> <svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" height="100%" width="100%" preserveAspectRatio="xMidYMid meet" version="1.1" viewBox="0 0 264 400"> <image width="264" height="400" xlink:href="../Images/xxx.jpg"/> </svg> </div> </body> </html> These two kinds of code gave exactly the same results. (see left screenhot1). There are document margins around the image (including vertical margins). If I convert to PDF with Prince using the code given above, (see right screenshot2). There are NO document margins around the image (only the 6% left and right due to the shape of the image). So, with Calibre, even with a svg wrapper, the display is not the one expected. The only exception with Calibre, concerns the cover image which can be displayed without document margins. Why, what can be done for the cover image could not be extended to other full page images using the same code? Note that I do not pretend that these two codes are equivalent. If the results are the same for this particular test, it is because the first code is calculated for screens with a 3/4 ratio. There would be differences for others shapes of screen. Attached Thumbnails Last edited by roger64; 05-24-2018 at 12:43 PM.

05-25-2018, 02:15 AM	#11
roger64 Wizard Posts: 2,608 Karma: 3000161 Join Date: Jan 2009 Device: Kindle PW3 (wifi)	When one wish to display a PDF on a small screen (say the widely used six-inch screen), the possibility to display an image really full screen makes a difference, for example for maps, drawings of all kinds. This way we extend the quality of display given to the cover image to other images of the same kind within the book. @Kovid Goyal Thanks you very much for your interest. Bug report opened. If need be, I can provide a full ePub for testing purposes by PM. Last edited by roger64; 05-25-2018 at 03:53 AM. Reason: opened

06-02-2018, 04:48 PM	#12
roger64 Wizard Posts: 2,608 Karma: 3000161 Join Date: Jan 2009 Device: Kindle PW3 (wifi)	A short word but a big thank you to Kovid Goyal for enabling so quickly and so efficiently the two points I submitted to him about ePub-to-PDF conversion: - hyphenation is working quite well (at least for French language) provided you tick the ad hoc check box in the Preferences page (PDF output). - Calibre can display perfectly all your full page images without document margins. To get it working, you must tick a check box in the Preferences Page PDF output (Use page margins ...) and use a svg wrapper (a compliant code -including @page 0- has been published above). To ensure that the cover image is not displayed twice (...) declare this image to be the cover_image (right click in the relevant html file). Again. Last edited by roger64; 06-02-2018 at 05:08 PM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Conversion PDF -> EPUB ?	JPS	Assistance	6	05-19-2016 03:34 PM
Conversion PDF -> Epub	rudmueller	Conversion	0	09-11-2011 10:44 AM
conversion from pdf to epub help	slushbilly	Workshop	1	01-31-2011 08:07 AM
PDF to EPUB Conversion	LuchoResto	General Discussions	1	11-19-2010 04:54 PM
Help with conversion from PDF to EPUB	Fizz	Calibre	5	10-25-2009 11:48 AM

05-23-2018, 11:01 PM	#2
kovidgoyal creator of calibre Posts: 43,856 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	1. Sure it is easy to do, open a bug for it 2. It does not do anythong to compress, it simply has a very efficiently designed PDF backend that I wrote from scratch. 3. All calibre conversion settings are saved. https://manual.calibre-ebook.com/conversion.html (see the section on how settings are saved) 4. THere is no way to gruarantee "correct" conversion of a dropcaps, just as there is no way to guarantee correct display of one. It depends on the font size/dpi/other css settings, you will need to play with them. 5. The best way to display images with margins zero is to use the standard svg cover markup that calibre generates for its epub cover pages. That is guaranteed to work.

05-24-2018, 10:11 PM	#9
kovidgoyal creator of calibre Posts: 43,856 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	The PDF Output in calibre does not use @page you have to set the margins to zero in the conversion settings. I've never heard of anyone creating PDF files with zero margins, they would not print very well. But, it is a trivial change to add an option to get the plugin to respect @page margins if needed.

05-24-2018, 10:23 PM	#10
kovidgoyal creator of calibre Posts: 43,856 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	Actually, on second thoughts it is probably not a good idea to respect margins in the HTML files, that way would lead to lots of hard to debug-inconsistent rendering. Instead probably better to just special case full page image markup, just like the viewer does. Open a bug report for it and I will get to it when I have a moment.

06-03-2018, 12:15 AM	#13
kovidgoyal creator of calibre Posts: 43,856 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	You're welcome

Advert

Advert