EPUB 3.0 Workflow

Today I attended part two of the Understanding EPUB 3 webcast series, called Working with EPUB 3 : How EPUB 3.0 Will Affect Your Workflow. Read about the first webcast here.

Ron Severdia, the CTO of Metrodigi, was the guest speaker. Metrodigi is a conversion house that has been around for two years, located in the San Francisco Bay Area. The company focuses on highly-designed and very interactive ebooks. Clients include Time, Salesforce.com, Apple, and Triumph.

Metrodigi uses the tool Chaucer, an ebook conversion software that converts InDesign, Quark, PDF, and other source files to epub, KF8 and B&N ebooks. At some point the company plans to license this tool to other businesses.

Some background on ePub:

ePub began in September 2007, from Open eBook. When the Kindle was first released in November 2007, eInk was predominant. In mid 2010, ePub 2 came around, and just a little over a year later we have ePub 3.

So, according to Ron, the big question is, is ePub 3 the holy grail?

“It’s not really the holy grail, as some have said, but it is the dawn of a new era,” he said. “And that’s why it’s so exciting.”

Ron said that timing is everything and there is a harmonic convergence for ePub 3. First, HTML5 is developing and flash is going away for mobile devices. Second, CSS3 has a good amount of browser support. And third, Javascript frameworks are hitting their stride. Although the libraries and frameworks tend to be overly powerful for ePub, they do make it easy to add a lot of interactivity.

There are a few new features in ePub 3:

  • XHTML5 and CSS3
  • Audio and video, both local and remote–meaning large videos files can be stored externally. Also, media overlays (read aloud capabilities)
  • Javascript
  • MathML –only a subset of the full spec so far
  • Global language support. This means there’s left-to-right (LTR), right-to-left (RTL), and vertical writing capabilities, which is good for languages such as Arabic and Japanese.
  • Embedded fonts–OTF and WOFF
Some of these features were already being used, but ePub 3 makes them official, such as embedded fonts, and audio and video.
What went away with ePub 3
  • DTbook
  • Tours
  • Guide–although Apple still uses this, but the NCX file is better. The guide basically duplicates the table of contents

New terminologies to know:

  • Open Publication Structure 2.0.1 becomes ePub Content Documents 3.0
  • Open Package Format 2.0.1 becomes ePub Publications 3.0
  • Open Container Format 2.0.1 updated to 3.0
  • New spec for NCX file: ePub Publications 3.0

From a structural and coding standpoint, these docs are much more unified, according to Ron.

In ePub 3, CSS3 is not entirely CSS3 yet. Currently it’s CSS 2.1 support and certain CSS3 attributes, one being the ability to animate. CSS3 also allows for multi-column layouts and directional printing. Additionally, vendor prefixes such as -epub-text-align-last is good for fixed layouts.

Although audio and video is not new, it is now an official standard. At a minimum you have to use MP3, but ePub 3 also allows for MP4 and AAC. Also there is the ability to have media overlays and TTS, via SMIL, PLS, or SSML. However, there is no specific video format required, which means it all depends on the device.

Other things to consider:

  • Accessibility improvements (for users with disabilities)
  • Content switching (determine the size of screen being used, and which device)
  • SVG (now in the spine too)

Still, Ron said that some areas could be improved. For many items in the specs, reader system support is not mandatory–there is a balance between being strict and being flexible, and Ron thinks that the spec should have been more strict in audio and video support. He also thinks video should be allowed on eInk devices and audio, javascript, and media overlays should be universally mandatory. Additionally, Ron said he thinks the canonical fragment identifier is not quite ready–it’s difficult to use and not intuitive.

In terms of devices, both tablet and ereader support are “all over the map.” Most support what Ron calls “ePub 2.5”, which basically means devices are built on the ePub 2 standard, and are not quite ePub 3. But “backwards compatibility with ePub 2.0.1 still remains to be seen,” he said. Things are likely to change drastically this year though, he said.

Right now, there are four main tablet devices on the market:

  • iPad
  • Kindle Fire
  • Nook Color/Tablet
  • Kobo Vox
Metrodigi has a four-step workflow that they recommend publishers should also use.
  1. Planning (the most time-consuming stage)
  2. Design
  3. Preflight
  4. Production
The idea, Ron said, is to deal with problems early in the process so when you go to production, there are no extra issues compounded due to poor planning.


Publishers should make ePub part of the print publishing workflow, and not as an afterthought, Ron said. Publishers should also think about font and media licensing, because it’s not cheap. Many font vendors see licensing digital fonts as an opportunity to make up for lost revenue, which unfortunately makes it expensive. However, there are open source font websites to try, such as Font Squirrel, Google Webfonts, and dafont. Metrodigi also has a free font directory of 25,000+ fonts for clients. Ron recommends that publishers create their own in-house directory of fonts for print and digital publications.

In terms of media production, Ron said publishers should plan their enhancements. Don’t go overboard; it’s better to serve the story. Think about what your enhancements are. Users and ebook readers are starting to expect more and be more savvy. Having the same digital copy as the one in print is not enough anymore. What extra materials can you give that the print book doesn’t have? Utilize the media to the best of your aptitude. The marketing plan for digital content also needs to be incorporated into the regular marketing plan. Digital content can “piggyback” the efforts on the print release. The format of the book, whether it’s fixed-layout or reflowable, also needs to be considered. How will this affect the process? Vertical layout is usually preferred for fiction books, because people want to hold the device like a book. Horizontal layouts are more appropriate for journalistic books and books with photos.

It’s also important to gather metadata. Determine the most appropriate fields–examples include title, author, and publisher. Metrodigi has a metadata information sheet everyone can fill out for the most commonly used metadata fields. Metadata is important for the searchability of the ebook. Publishers also need to decide whether or not to use DRM. Read more about the benefits of DRM-free in my DBW12 post. Lastly, publishers are now questioning whether or not to use iBooks Author. It looks great if you do it right and feels like an app, but not all books are appropriate for iBooks Author. Read more on the pros and cons about iBooks Author here.

There are some reality checks all publishers should remember. Currently the cross-device user experience can’t be consistent. For example, Kindle doesn’t support audio, and iPad and Nook work differently, so it’s not possible to replicate the read aloud experience on all three devices. Additionally, Amazon only half embraces ePub and will probably not change course. Lastly, reader system support varies and ebooks will not look and behave like their print counterparts (which is a good thing).

But, according to Ron, the future is probably bright for ePub3, in terms of device support:


What’s the best way to present the material? There are many choices to make: 7″ v. 10″ screens, vertical v. horizontal, and single page v. spreads. The final aspect ratios are critical, Ron said, since the iPad is different from the Kindle. Publishers should also consider relative font sizes, especially for 7″ screens.

Metrodigi creates templates for individual devices and “shared” templates for multiple devices. “Hot zones” are important to consider (where the reader taps the screen to turn the page–you don’t want to include any interactive elements in the hot zones).

Ron also suggested publishers create stock interactive page elements to reuse in layouts as placeholders, such as buttons and videos. Metrodigi has different layouts for difference screen sizes, but they all have the same look and feel.

The orientation and format is also important to consider, as is white space. Although many designers like a lot of white space, it’s not appropriate on small screens. Publishers should also use InDesign object, paragraph, and character styles, even if they’re not exporting from InDesign to ePub (which is not recommended because it’s very messy). This makes it easier to tag and identify everything later.

Designers also need to consider how to break up large pieces of content into bite-sized bits. Although sometimes it’s nice to replicate the print version, it’s not always necessary. One example to consider is how to break up a large table.

Lastly, publishing designers should use pop-ups (which is possible in iOS and KF8) to show larger items. This enables you to add video and other interactive elements without distorting your layout.


Always check the following:

  • Fonts are in the right format and are not missing (getting open-type fonts is key)
  • There are no low-res or missing images
  • That the InDesign package exports and that you use one InDesign file, not separate chapter files (it’s much easier if you are exporting InDesign)


Prototyping in PDF is much faster, more flexible, and allows multiple iterations in a short time, Ron said. The PDF can test that font sizes look good, margins are good, and everything is clear on particular screen sizes.

When hiring a conversion house for the first time, always ask for a sample (you should feel comfortable before committing). Publishers should also build in extra QA time. Ron recommended physically testing ePub files on all the devices. Additionally, publishers should look into courses/books to train their in-house team. “Get ahead of the curve,” he said. The more in-house knowledge about the process, the better.

Ron had a few other tips, starting with tools for ePub 2 “guerrilla bookmaking”:

  • Sigil
  • Calibre (not recommended for use with creation, since it doesn’t always produce valid ePubs. Much better for reading than converting)
There are also validation tools:
Publishers can also use readers to test ePub 3:

Publishers can also play with sample ePub3s at this site.

Lastly, Ron suggested some rules publishers should follow. First, there’s the golden rule of Keep it Simple (KISS). Second, both design and code are grid-based, so for best optimization stick to the grid. Third, limit the tricks used in average web markup in ePub code structure. Fourth, don’t overdo enhancements. And, make sure the ebook is intuitive and not too difficult to use enhancements.

Other advice that Ron offered was to ease into ePub3, and wait for the reader support. Don’t dive in, since the support is not there yet.

“Build towards what devices can do today, not too much for what they can do tomorrow,” he said. “Publishers get too hung up on ‘what’s going to happen in 6 months?'” But things keep changing quickly. Publishers should go with the best look and the best ebook they can make today. Additionally, the ePub 2 spec is part of the ePub3 spec, so publishers are in effect, future-proofing.

Also, publishers can use Jquery in ebooks, but they are a bit heavy, so Ron recommends only taking out the parts that are needed. And lastly, PDF is a great format, but not quite suitable for ebooks. For example, readers can’t resize text, change font-sizes, or have full-bleed images adapt to the screen.


7 Replies to “EPUB 3.0 Workflow”

  1. This was a fantastic article. Thank you so much for taking the time to write this out. I’ve been a major advocate of ePub 3.0, but I still am having so much trouble with video renderings in iOS. Normal jQuery functions work beautifully, but I’m unable to (a) render videos to react properly to the page dimensions and (b) use jQuery video players. Any ideas?

    1. Thank you! I don’t think you need JQuery to embed videos, since ePub 3 uses HTML5. It should work if you use the video tag. Something like (video id=”name” controls=”controls” src=”name.m4v” poster=”image.jpg” width=”x” height=”y” title=””)(/video)
      **I replaced the less than and greater than signs with ( and ) so WordPress would let me post the code

  2. Right, I guess you can just use the native player – that’s actually much easier. Sorry to bug with another question – but do you know an optimal iPad video resolution? I’ve tried 480×270, 640×360, and many others that are H.264 MP4, but there’s are black borders on the sides of the video when you render in iOS (does not happen with desktop eReaders). My goal is to have the width maximized and have the height adjust accordingly, which is a normal CSS strategy, but iBooks does not read this right or something?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s