A Start
This book is about making information disappear. For some people, this topic is a parlor trick, an amazing intellectual exercise that rattles around about the foundations of knowledge. For others, the topic has immense practical importance. An enemy can only control your message if they can find it. If you hide data, you can protect your thoughts from censorship and discovery.
The book describes a number of different techniques that people can use to hide information. The sound files and images that float about the network today are great locations filled with possibilities. Large messages can be hidden in the noise of these images or sound files where no one can expect to find them. About one eighth of an image file can be used to hide information without any significant change in the quality of the image.
Information can also be converted into something innocuous. You can use the algorithms from Chapter 7 to turn data into something entirely innocent like the voice-over to a baseball game. Bad poetry is even easier to create.
If you want to broadcast information without revealing your location, the algorithms from Chapter 11 show how a group of people can communicate without revealing who is talking. Completely anonymous conversations can let people speak their mind without endangering their lives.
The early chapters of the book are devoted to material that forms the basic bag of tricks like private-key encryption, secret sharing, and error-correcting codes. The later chapters describe how to apply these techniques in various ways to hide information. Each of them is designed to give you an introduction and enough information to use the data if you want.
The information in each chapter is roughly arranged in order of importance and difficulty. Each begins with a high-level summary for those who want to understand the concepts without wading through technical details, and a introductory set of details, for those who want to create their own programs from the information. People who are not interested in the deepest, most mathematical details can skip the last part of each chapter without missing any of the highlights. Programmers who are inspired to implement some algorithms will want to dig into the last pages.
Many of the chapters also come with allegorical narratives that may illustrate some of the ideas in the chapters. You may find them funny, you may find them stupid, but I hope you'll find some better insight into the game afoot.
For the most part, this book is about having fun with information. But knowledge is power and people in power want to increase their control. So the final chapter is an essay devoted to some of the political questions that lie just below the surface of all of these morphing bits.
0.1. Notes On the Third Edition
When I first wrote this book in 1994 and 1995, no one seemed to know what the word “steganography” meant. I wanted to call the book Being and Nothingness on the Net. The publisher sidesteped that suggestion by calling it Disappearing Cryptography and putting the part about Being and Nothingness in the subtitle. He didn't want to put the the word “steganography” in the title because it might frighten someone.
When it came time for the second edition, everything changed. The publisher insisted we get terms like steganography in the title and added terms like Information Hiding for good measure. Everyone knew the words now and he wanted to make sure that the book would show up on a search of Amazon or Google.
This time, there will be no change to the title. The field is much bigger now and everyone has settled on some of the major terms. That simplified a bit of the reworking of the book, but it did nothing to reduce the sheer amount of work in the field. There are a number of good academic conferences, several excellent journals and a growing devotion to building solid tools at least in the areas of digital rights management.
The problem is that the book is now even farther from comprehensive. What began as an exploration in hiding information in plain sight is now just an introduction to a field with growing economic importance.
Watermarking information is an important tool that may allow content creators to unleash their products in the anarchy of the web. Steganography is used in many different places in the infrastructure of the web. It is now impossible to do a good job squeezing all of the good techniques for hiding information into a single book.
0.2. Notes On the Second Edition
The world of steganography and hidden information changed dramatically during the five years since the first edition appeared. The interest from the scientific community grew and separate conferences devoted to the topic flourished. A number of new ideas, approaches, and techniques appeared and many are included in the book.
The burgeoning interest was not confined to labs. The business community embraced the field in the hope that the hidden information would give creators of music and images a chance to control their progeny. The hidden information is usually called a watermark. This hidden payload might include information about the creator, the copyright holder, the purchaser or even special instructions about who could consume the information and how often they could push the button.
Many of the private companies have also helped the art of information hiding, but sometimes the drive for scientific advancement clashed with the desires of some in the business community. The scientists want the news of the strengths and weaknesses of steganographic algorithms to flow freely. Some businessmen fear that this information will be used to attack their systems and so they push to keep the knowledge hidden.
This struggle errupted into an open battle when the recording industry began focusing on the work of Scott A. Craver, John P McGregor, Min Wu, Bede Liu, Adam Stubblefield, Ben Swartzlander, Dan S. Wallach, Drew Dean, and Edward W. Felten. The group attacked a number of techniques distributed by the Secure Digital Music Initiative, an organization devoted to creating a watermark system and sponsored by the members of the music industry. The attacks were invited by SDMI in a public contest intended to test the strengths of the algorithms. Unfortunately, the leaders of the SDMI also tried to hamstring the people who entered the contest by forcing them to sign a pledge of secrecy to collect their prize. In essence, the group was trying to gain all of the political advantages of public scrutiny while trying to silence anyone who attempted to spread the results of their scrutiny to the public. When the group tried to present their work at the Information Hiding Workshop in April in Pittsburgh, the Recording Industry Association of America (RIAA) sent them a letter suggesting that public discussion would be punished by a lawsuit. The group withdrew the paper and filed their own suit claiming that the RIAA and the music industry was attempting to stiffle their First Amendment Rights. The group later presented their work at the USENIX conference in Washington, DC, but it is clear that the battle lines still exist. On one side are the people who believe in open sharing of information, even if it produces an unpleasant effect, and on the other are those who believe that censorship and control will keep the world right.
This conflict seems to come from the perception that the algorithms for hiding information are fragile. If someone knows the mechanism in play, they can destroy the message by writing over the messages or scrambling the noise. The recording industry is worried that someone might use the knowledge of how to break the SDMI algorithms to destroy the watermarking information– something that is not difficult to do. The only solution, in some eyes, is to add security by prohibiting knowledge.
This attitude is quite different from the approach taken with the close cousin, cryptography. Most of the industry agrees that public scrutiny is the best way to create secure algorithms. Security through obscurity is not as successful as a well-designed algorithm. As a result, public scrutiny has identified many weaknesses in cryptographic algorithms and helped researchers develop sophisticated solutions.
Some companies trying to create watermarking tools may feel that they have no choice but to push for secrecy. The watermarking tools aren't secure enough to withstand assault so the companies hope that some additional secrecy will make them more secure.
Unfortunately, the additional secrecy buys little extra. Hidden information is easy to remove by compressing, reformatting, and rerecording the camouflaging information. Most common tools used in recording studios, video shops, and print shops are also good enough to remove watermarks. There's nothing you can do about it. Bits are bits and information is information. There is not a solid link between the two.
At this writing the battle between the copyright holders and the scientists is just beginning. Secret algorithms never worked for long before and there's no reason why it will work now. In the meantime, enjoy the information in the book while you can. There's no way to tell how long it will be legal to read this book.