Converting non-ascii/non-unicode text - pictures the way to go?

politicorific · 03-17-2009, 05:38 PM

I recently purchased an ebook reader and want to convert a book to an ebook. The issue is that the book uses a lot of arcs, lines, arrows, and boxes incorporated into the text which have no ascii or unicode equivalents. Ocring this stuff is useless, although it does work for 60% of the regular print. Thankfully most of the material follows a set pattern and is purely black and white. However using word's drawing tool looks like a poor option.

So I'm considering using photoshop to create a template for most of these items, but I see several potential drawbacks: loss of zoom functionality, mismatched dpi/sizing issues, file size, rastered graphics issues, and overall trouble with placement in word.

Is there a guide for graphics/images in ebooks?

Should I just bite the bullet and go the photoshop route or can I somehow use vectered graphics tool like illustrator so the image is resizeable?

Valloric · 03-17-2009, 07:22 PM

SVG sounds perfect for what you need. Create all those shapes as SVG images (Inkscape is free, and Adobe Illustrator is good if you can afford it) and embed them in an epub file with the rest of the text. You can even inline the SVG code in the XHTML.

Illustrator should even do a decent job of converting the line drawings to SVG if the scans are good enough.

igorsk · 03-17-2009, 08:40 PM

An embedded font with custom characters would work too.

darkmonk · 03-18-2009, 07:30 PM

If you learn enough about svg, you can even hand code your drawing. And yes that is definitely the preferred format for all non-picture images.

politicorific · 03-24-2009, 03:36 AM

Thanks for the suggestions. I looked at some free font software, but nothing stuck out as being particularly easy to work with or useful for my applications. I ended up experimenting with hand coded SVG files to produce Bezier curves which nearly matched what was created in this document. As a bonus, I learned a new skill which will help me convert the concepts in this document to a website. Just a little concerned about placement during conversion, which program to set this up in since SVG isn't supported by my copy of word, and what final format to pick, but I guess I'll experiment.

politicorific · 04-02-2009, 05:59 AM

Progress report:

It's a week later and I've made difficult progress.
After giving up on Word I turned to writing my own xhtml files with svgs written in-line.

I attempted to preserve the layout of the book - trying to mimic every indent, italicized, bolded, word - this was dumb since the page width of the scanned book was fixed, while my browser wasn't. The first 25 pages without mathematical symbols went fine.

Then the horrors began. The svgs caused numerous problems. They looked different from browser to browser - in firefox they'd end up having an 80 pixel bottom margin and in safari about 600. The exact same SVG repeated twice in a row would look different - the lines appeared aliased, even if was repeated just 100 pixels down the page. This is troubling because I'd like to use svgs for a later project involving the same book. I also attempted to convert the first page using calibre to all three of the formats and none of them displayed the svg. Maybe making them external files would work.

So tonight I've downloaded a LaTeX editor Postext and hope to have better luck drawing the symbols with some of the addons.

Word to xhtml to LaTeX. Never expected that progression.

03-17-2009, 05:38 PM	#1
politicorific Junior Member Posts: 7 Karma: 10 Join Date: Mar 2009 Device: prs-505	Converting non-ascii/non-unicode text - pictures the way to go? I recently purchased an ebook reader and want to convert a book to an ebook. The issue is that the book uses a lot of arcs, lines, arrows, and boxes incorporated into the text which have no ascii or unicode equivalents. Ocring this stuff is useless, although it does work for 60% of the regular print. Thankfully most of the material follows a set pattern and is purely black and white. However using word's drawing tool looks like a poor option. So I'm considering using photoshop to create a template for most of these items, but I see several potential drawbacks: loss of zoom functionality, mismatched dpi/sizing issues, file size, rastered graphics issues, and overall trouble with placement in word. Is there a guide for graphics/images in ebooks? Should I just bite the bullet and go the photoshop route or can I somehow use vectered graphics tool like illustrator so the image is resizeable? Last edited by politicorific; 03-17-2009 at 06:10 PM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Classic Pictures overwrite the text!	JDługosz	Barnes & Noble NOOK	0	09-09-2010 11:54 PM
Help a beginner:Python/Recipe Unicode and ASCII	Starson17	Calibre	2	02-15-2010 11:10 AM
Unicode characters OK in text but wrong in TOC	paulpeer	ePub	8	01-15-2010 06:17 PM
Using Unicode Fonts for darker text	Damætas	Kindle Developer's Corner	11	04-19-2009 03:44 PM
Can I preserve entities when converting from html? (To avoid unicode on kindle)	krunkster	Calibre	1	04-07-2009 05:11 PM

03-17-2009, 07:22 PM	#2
Valloric Created Sigil, FlightCrew Posts: 1,982 Karma: 350515 Join Date: Feb 2008 Device: Kobo Clara HD	SVG sounds perfect for what you need. Create all those shapes as SVG images (Inkscape is free, and Adobe Illustrator is good if you can afford it) and embed them in an epub file with the rest of the text. You can even inline the SVG code in the XHTML. Illustrator should even do a decent job of converting the line drawings to SVG if the scans are good enough.

03-17-2009, 08:40 PM	#3
igorsk Wizard Posts: 3,442 Karma: 300001 Join Date: Sep 2006 Location: Belgium Device: PRS-500/505/700, Kindle, Cybook Gen3, Words Gear	An embedded font with custom characters would work too.

03-18-2009, 07:30 PM	#4
darkmonk Connoisseur Posts: 58 Karma: 12 Join Date: Jan 2009 Device: none	If you learn enough about svg, you can even hand code your drawing. And yes that is definitely the preferred format for all non-picture images.

03-24-2009, 03:36 AM	#5
politicorific Junior Member Posts: 7 Karma: 10 Join Date: Mar 2009 Device: prs-505	Thanks for the suggestions. I looked at some free font software, but nothing stuck out as being particularly easy to work with or useful for my applications. I ended up experimenting with hand coded SVG files to produce Bezier curves which nearly matched what was created in this document. As a bonus, I learned a new skill which will help me convert the concepts in this document to a website. Just a little concerned about placement during conversion, which program to set this up in since SVG isn't supported by my copy of word, and what final format to pick, but I guess I'll experiment.

04-02-2009, 05:59 AM	#6
politicorific Junior Member Posts: 7 Karma: 10 Join Date: Mar 2009 Device: prs-505	Progress report: It's a week later and I've made difficult progress. After giving up on Word I turned to writing my own xhtml files with svgs written in-line. I attempted to preserve the layout of the book - trying to mimic every indent, italicized, bolded, word - this was dumb since the page width of the scanned book was fixed, while my browser wasn't. The first 25 pages without mathematical symbols went fine. Then the horrors began. The svgs caused numerous problems. They looked different from browser to browser - in firefox they'd end up having an 80 pixel bottom margin and in safari about 600. The exact same SVG repeated twice in a row would look different - the lines appeared aliased, even if was repeated just 100 pixels down the page. This is troubling because I'd like to use svgs for a later project involving the same book. I also attempted to convert the first page using calibre to all three of the formats and none of them displayed the svg. Maybe making them external files would work. So tonight I've downloaded a LaTeX editor Postext and hope to have better luck drawing the symbols with some of the addons. Word to xhtml to LaTeX. Never expected that progression.

Advert

Advert