Watermarking & Social DRM for ebooks

[Len]

The vast majority of major publishers in North America use very intrusive DRM by Adobe that prevents users from easily backing up their online purchases and prevents them from loading them on many smartphones and e-readers. If you purchase a Kindle book, it is impossible to transfer it to the Nook or Kobo, due to the incompatibility of DRM. Social DRM has been around for a number of years and is growing in popularity in Europe. This technology employs digital watermarks and makes it easy to sideload the ebooks on all of your devices and if the book appears on a pirate site, it is easily tracked down to the user who initially bought it. Is there a future in social DRM?

Digital watermarking is a term that first appeared in the mid-1990s. The term “watermarking” is meant as an analogy to a watermark in paper: a hidden mark that only becomes visible when the paper is subjected to a special detection process, such as immersion in certain liquids. Watermarks were originally meant to convey the identity of the paper manufacturer or vouch for the authenticity of documents such as postage stamps, paper currency, passports, and other government documents.

Analogously, digital watermarking involves embedding information – called a payload – in digital files that indicates the identity of the content or its owner; and as we’ll see, watermarks can also be used to embed other information such as the identities of e-book retailers and users who purchase e-books. Unlike paper watermarks, digital watermarks may or may not be hidden (imperceptible to users); nevertheless, academics and technologists often categorize watermarking as a type of steganography or “hidden writing.”

Digital watermarks were first applied to audio and visual information. The original schemes were algorithms for embedding data into digital files in ways that were imperceptible to humans. Watermarking algorithms for those types of content typically start by analyzing the content to determine areas of the content, such as portions of an image, where embedding data is least likely to change the user’s perception of it. Then they apply an embedding algorithm to those areas.

E-book watermarking works somewhat differently. Because e-book file formats such as EPUB and PDF are based on text characters, it’s not possible to apply an algorithm to alter the text itself so that the changes are imperceptible; the text of the e-book must remain as the author and publisher intended.

Instead, e-book watermarking algorithms affect aspects other than the text itself. There are three general ways to insert payloads into e-books:

Readable: the payload is directly readable and meaningful to the public, such as a name or email address.

Obfuscated: the payload is human readable but not meaningful, such as an alphanumeric identifier, user ID, or encrypted email address.

Imperceptible: the payload is embedded in a way that is not readily visible to human readers.

Watermarking prevents piracy

The most common application of watermarking for e-books is in reducing unauthorized distribution such as piracy. E-book watermarking is sometimes known as “social DRM” or “soft DRM” because it can serve as a deterrent to distribution beyond a user’s close associates; the user may not be comfortable sharing a document with a user identifier in it. These terms are used in the industry even though watermarking is not DRM, as we will discuss below.

A high-quality watermarking scheme will provide a range of options for where the watermark is placed in the e-book (e.g., with or without Ex Libris pages; with or without page or chapter footers) as well as the type of watermark (readable, obfuscated, or imperceptible). The frequency and type of watermarks in a given title reflects tradeoffs among several factors, including:
 

Effectiveness in deterring piracy

Utility in forensic piracy detection (see below)

Robustness (see above)

Impact on the look of the finished e-book

For example, an email address at the bottom of every page (the technique O’Reilly and Springer use) serves as a constant reminder not to “overshare” the e-book, while their value in forensically tracing the time and device of purchase is limited. On the other hand, the single obfuscated (yet visible) transaction identifiers embedded in Pottermore Harry Potter e-books are more for forensic use in piracy detection than as “social DRM” reminders to users.

Watermarking is especially effective as a forensic anti-piracy enforcement measure when combined with a copyright monitoring service. Copyright monitoring services continuously search websites, file storage services, and other online locations that could harbor unauthorized copies of your e-books. They record instances of such copies and report them to publishers or other customers. Digimarc Guardian and Link-Busters are examples of copyright monitoring services that work with e-books.

A copyright monitoring service can use the watermarking vendor’s tools to extract watermark payloads and report them back to the publisher. This can complement other anti-piracy techniques that these services use, such as:
 

Fingerprinting: searching for files or web pages containing the actual text of an e-book, or at least a significant part of the text of the e-book. This technique is often accomplished with a variant of the hash technique for integrity checking discussed above.

Metadata: searching for filenames, titles, and other descriptive metadata on web pages, file storage services, and so on that indicate the title in question.

Watermarking vs. DRM

Watermarking is often confused with Digital Rights Management (DRM). DRM is a technique that involves encrypting files and requiring special software/hardware and credentials, such as user or device IDs and encryption keys, to view the content. Many major e-book retailers, such as Amazon, Apple, and Kobo, use their own DRM technologies, which are not interoperable with one another.

DRM has become well-known over the past two decades because of its controversial nature; there has been much press coverage of DRM, including in mainstream media, whereas coverage of watermarking is practically nonexistent and limited to industry trade and academic publications.

The fact that DRM is a widely-recognized term has led some writers to use it in describing watermarking; in fact the term “social DRM” was originally used – by Bill McCoy, then the head of Adobe’s e-book software business – in 2007 to describe an alternative to DRM that involved inserting the user’s name in e-books.

Nevertheless, watermarking is not DRM; the technologies are fundamentally distinct. The biggest practical difference between watermarking and DRM is that DRM-protected files can only be viewed on devices or apps intended for it (e.g., Amazon Kindle e-books on Amazon Kindle devices and apps), whereas watermarked e-books can be read on any device or app that renders the format (e.g., any e-reader device or app that reads EPUB).

In fact, interoperability is one reason why most retailers (not affiliated with publishers) are reluctant to support watermarking as an alternative to DRM: DRM enables retailer “lockin,” while watermarking does not.

Otherwise, here are some of the practical differences between watermarking and DRM for e-books:
 

Interoperability: DRM-protected files can only be viewed on devices or apps intended for it. For example, Amazon Kindle e-books with DRM can only be viewed on Kindle devices and apps. Watermarked e-books are fully interoperable and can be used with any e-reader that uses the file format; for example any e-reader device or app that reads EPUB can read a watermarked EPUB file. This includes Apple iBooks, Nook, and Kobo, as well as independent e-reader apps such as Calibre, Bluefire Reader, Icecream, and FBReader.

Usage restrictions: DRM imposes restrictions on the usage of e-books. Only those with proper credentials can view DRM-protected files. Depending on the DRM and its configuration, DRM-protected files may have restrictions on functions such as transferring to a user’s other devices, copying text to the clipboard for pasting elsewhere, printing, and text-to-speech synthesis. Watermarking does not restrict any of these functions.

Robustness: The resistance of DRM technologies to hacking varies; but once a DRM technology has been hacked, it is difficult to plug the security hole, as it can involve updating software or hardware already in the field. The robustness of watermarking schemes also varies. If a watermark stripping technique is discovered, new files can always be watermarked with a different scheme to avoid the hack, and no e-readers will be affected. In contrast, with DRM it is not necessarily possible to use a different encryption scheme for files without changing the DRM functionality in users’ e-readers.

Effects of hacking: once a hack to a DRM scheme is available, it is possible to use the hack to create DRM-free e-books. You know when you have succeeded in stripping DRM from a file, and there is unlikely to be any evidence that the resulting file was created by stripping DRM or where it came from, other than (for example) a retailer that uses a specific e-book format. In contrast, with watermark stripping, the hacker can’t be sure that a watermark with built-in redundancies is completely gone.

Evidentiary value: DRM-stripped files do not contain any evidence of where they came from or who obtained them in the first place. Watermarks may have value as evidence in copyright infringement litigation. If a hacker strips a watermark, the ambiguity over whether a watermark still remains creates the possibility that the file contains evidence that can lead to the original owner of the file and can be produced in litigation.

Accessibility: most DRM schemes for e-books control or inhibit e-reader features that make text accessible to the print disabled, such as text-to-speech synthesis. While some DRM schemes make it possible for publishers to enable these features, others do not support them at all. With watermarks, authors and publishers can be sure that all accessibility features supported in an e-reader will work properly.

What companies are responsible for social DRM?

Digimarc is one of the industry leaders in making readers lives simpler and has signed up a number of big name clients over the years. Digimarc has been at the forefront of digital watermarking technology for almost two decades. After the acquisition in late 2012 of what is now the Digimarc Guardian platform, the leading anti-piracy solution for the publishing industry, we turned our focus to developing a service for watermarking e-books.

A few years ago HarperCollins Publishers, one of the largest publishers of consumer books in the world, and LibreDigital, a leading provider of e-book distribution and fulfillment services, signed on as partners.

BooXtream is a cloud based service to watermark and personalize EPUB e-books. In 2010 BooXtream was the first dedicated API based watermarking and personalization web service for EPUB e-books. The BooXtream web service provides an advanced watermarking engine, a master e-book repository, and a download link fulfilment module. The API has been specifically designed for an easy e-commerce integration. The web based BooXtream Dashboard offers management and reporting tools as well as a manual mode to watermark single e-book (review) copies.

BooXtream has a worldwide customer base and was winner of the International Book Industry Technology Supplier Award at the London Book Fair International Excellence Awards 2015. BooXtream is created by Icontact, a Netherlands-based software company specialized in solutions for the publishing and library industry.

Conclusion

Watermarking for e-books has existed since at least the early 2000s, but it hasn’t gotten anywhere near the attention that DRM has. Yet it is a complement and worthwhile alternative to DRM as well as other techniques for curbing piracy.

Despite its lack of publicity, e-book watermarking is becoming more and more widely used – especially in Europe. Here are some indications of watermarking increasing popularity in European markets:
 

In Germany, several publishers including Random House, Holtzbrinck, and Bonnier abandoned DRM in favor of watermarking.

Over half (55%) of e-books sold in Italy use watermarking instead of DRM.

Almost all of the e-books sold in the Netherlands are watermarked.

In Sweden, 98% of the e-books distributed by eLib, the country’s biggest e-book distributor, are watermarked.

Watermarking is used in the vast majority of e-books in Central and Eastern Europe, including Poland, Hungary, Estonia, Slovakia, and Austria.

Watermarking is versatile: it gives publishers and retailers control over the level of personal information they can embed in e-book files, and it offers opportunities for adding value to e-books through personalization. It is time for the publishing industry to recognize the importance and potential of e-book watermarking. Every book publisher should consider adopting watermarking and advocating its use throughout the e-book publishing value chain.