Audio Watermark

Abstract

The recording industry is losing money. Piration should be blamed. But are there any way to prevent the trend? Yes, the answer lies in the watermark technology. In my paper, first I will talk about the original and definition of audio watermark. Then the main idea, that is to say, the mechanism behind watermark will be discussed. Thirdly, properties would be elaborated. Then different applications are discussed. The last but not least, a vivid case in real life is presented.

Introduction

According to figures published by the IFPI, the recording industry lost 4.2 billion dollars almost every year. The illegal manufacturer burn CDs and sell them to the customer, who can buy the same CD the tenth price as in some music store. What is more, some website provide download or online listening service to music fans. However, an audio watermark breakthrough the recording industry needs. Actually the audio watermark is like a signature, different to every copy and piece of music. It could be used to track the owner of illegal distribution and could be also used as a copyright for the owner. Naturely, it has to make sure that it could not be heard and interfere with the original sound quality and it could survive the audio signal distortion. In my paper, first I give the origin and definition to audio watermark, so to say, what is watermark, audio watermark, where does it come from. Secondly I will discuss the main idea behind the technology and then one of them-bitstream watermark. Third, properties of audio watermark would be elaborated. That is what requirement of an audio watermark should be met. The forth, I will talk about the applications of audio water, how it could be used? The last but not least, a special case of music trace company is presented. I will concentrate on several business module based on the audio watermark technology.

Origin and definition 

Actually, the word watermark comes from paper industry first. A watermark is a design embossed into a piece of paper during its production and used for identification of the paper and papermaker. The watermark can be seen when the paper is held up to light. The design could be some pattern or images which indicate quality of trademark of the paper. Then the word watermark in introduced to digital field. It is an identification code carrying information about the copyright owner, the creator of the work, authorized consumers and so on(1). The watermark is invisible and permanently embedded into digital data for copyright protection and for checking if the data has been corrupted. The electronic documents which is watermarked can be digital image, audio or even video. But I will only concentrate on the audio watermark, which means a watermark embedded in an audio stream to identify its origin. 

The main idea

In the watermark embedding block, a copyright owner uses a private key to create an inaudible watermark on the audio file. So the private key is used to encode the digital watermark into the music. Later on, the owner can examine whether a given audio file, which us suspected for being illegally copied, contains his or her own watermark and use it as legal prove. In the watermark detector block,  a public key is used  to decode the watermark from the music. The basic scheme contains a lot of possibilities both for compressed file and for uncompressed files. Below is how the audio watermark is embedded into the compressed files such as MP3.

The mechanism is called bitstream watermarking.

Properties of audio watermark

Some basic requirements should be met by an audio watermark. These are low computational cost, no distortion of original signal and change of format, robustness and imperceptibility.

Low computational cost means it should not take long for a watermark to be embedded and detected. It is a crucial factor in the watermark system. The watermark imbedded in the audio file should in no way reduce the quality of the audio file and change the format. If a watermark is embedded in an MP3 file, the customer can also play it on some MP3 supported player, they do not need to buy a new player again. Robustness refer to the ability to detect the watermark after common signal processing operations and hostile attacks. Example of common operations performed on audio files include noise, reduction, volume adjustment or normalization, digital to analog conversion and so forth. As far as attacks concerned, there is different kinds of attack on the audio watermark. They are collusion attack, signal diminishment attacks and inversibility attacks. A collusion attack is having copies of audio files with the same watermark, the attacker estimate the value of watermark by averaging the value of the audio signal. Only about 10 samples are enough for such estimation. As a coutermeasurement to this kind of attack, it is better to embed more than one watermark in an audio file or embed different watermark for different files. The signal deminishment attack refer to different signal operation such as lossy compression, modulation, conversion, etc. Now it is possible to embed more than one watermark in a audio file. If Bob steals the watermarked one from Alice and watermark it with his private key. It is sure to find Alice’s watermark in Bob’s file, but vice versa not possible. However, an attack to this procedure is created. So it is hard to decide who is the real owner of the audio file(2). The other property of audio watermark is imperceptibility.  There should be almost no difference between the watermarked one and the unwatermarked one. Here is a sound wave graph of the same music with and without watermark. As we can see, there is only small changes which could not be noticed at the first sight and could also be ignored.

Applications

The audio watermark is used in many ways. First, it is used in proof of ownership. Information during production, copyright information in the form of a watermark can be anchored directly in the recording. Specific ownership has different information imbedded. It could also used for access control, the watermark becomes a trigger to give allowance for playing the music. Tracing illegal copies is also an very important application. Personalized information is embedded in the music. The personalized information could be the customer number which moniter the trace of the music. If such music is found in internet, the watermark will be distracted and the customer number could be known for later law actions.

Case study

Watermark is used in real life business. The fraunhofer institution in Germany invented the watermark embedding technology and sold it to the company music trace. Company music trace unfold two main business module based on the technology. They are watermark embedding for audio signals and internet search for illegal use. The company use specific software to listen on the internet all day and night and search the matched music. Of course, the customer should provide several criteria for selecting the music first. For example, the music with specific title should be caught. Then the watermarks would be distracted as evidence. But only transaction numbers or customer numbers could be extracted. The identity of original purchasers can only be determined by the customers self.

Another business module is watermark embedding for audio signals. Copy right information, recipient information and title information could be imbedded. The imbedded content is different. The Product family contentmark permits embedding a digital watermark in audio signals. Depending on the product, this can involve either uncompressed or compressed audio signal(3). ContentmarkPCm is a technology to embed watermark in uncompressed files and the watermarked files are still uncompressed files. The result is unchanged. The file could be PCM files or Wave files. The contentmarkCD technology make it possible to create different watermarks on the same copy files. ContentmarkMP3 is used in compressed files.

Conclusion

As the conception watermark goes into digital area, recording industry has some hope not to lose money. But it needs time. The watermark is used as a special identification for every audio file. Different watermark algorithm and scheme have been developed. Nevertheless, it should be expanded and amended to cope with malicious attacks.