Breaking Barriers to Popularize Internet Streaming Broadcast

TDI RCAC Project


Internet streaming broadcast has become a strong empowerment tool for individuals and small groups to distribute their opinions in various audio and video formats. The audio broadcast requires less bandwidth and has become more popular than the video. In this paper, the author focuses on the audio distribution issues.

This paper first analyzes the restrictions imposed on Internet audio-streaming broadcast, focusing on the content handling issues. The paper argues that the following current practices undermine the opportunities for individuals and small groups with low budgets to broadcast on the Internet:

The paper proposes an alternative model to give more opportunities for Internet streaming broadcast to individuals and small groups, while preserving intellectual property rights. The model includes the following schemes:


Introduction: the status of Internet audio-streaming technology

Audio streaming has become a production-level application on the Internet. The bandwidth available to a dial-up Internet user is sufficient to receive and send radio-quality audio streams. Most dial-up Internet users work on 28.8kbps/33.6kbps voice-grade phone links or 56kbps/64kbps Integrated Services Digital Network (ISDN) links. The bandwidth required to exchange an audio stream is less than these limits. For example, a RealAudio [1] stream of 8.5kbps conveys adequate information for human-voice content; a 16kbps stream is sufficient for an AM-radio-quality stream, with voice and music mixed. Also, an MP3 [2] stream of 24kbps is commonly used for near-FM-radio quality monaural audio distribution. The dial-up links can also be used to provide an upstream link of streaming broadcasts. Software tools to listen to these audio streams, such as RealPlayer [1] or WinAmp [3], are free.

Music distribution in higher fidelity than the real-time streaming has also become a popular application on the Internet. MP3 encoding in 128kbps joint-stereo format has been a de-facto standard for disseminating "CD-quality" music on a non-real-time (on-demand) basis for dial-up Internet users, or even by real-time streaming for those who have access to higher Mbps-class bandwidth, such as cable TV Internet. MP3 format is also widely used for transferring content to portable music players and CD-ROM distribution of music. While the encoding algorithm of MP3 is patented, the encoder software packages are affordable.

The streaming server and encoder programs have also been widely available as free software packages. RealServer Basic, the minimal 25-stream server, and the encoder RealProducer are available online free of charge from RealNetworks [1]. For MP3 streaming, a proprietary server system SHOUTcast [3] and the compatible open-source system called icecast [4] have been widely deployed since their emergence in January 1999. A hypertext transfer protocol (HTTP) server such as Apache is also sufficient for on-demand MP3 audio distribution, combined with the HTTP-streaming-capable client software such as WinAmp. Many companies such as RBN [1] and Yahoo! Broadcast [5] provide distribution service for Internet audio and video streams by connecting their streaming servers to Internet service providers with high-bandwidth links.

The wide availability of Internet audio-streaming tools resulted in the phenomenon that many individuals and small groups, as well as the major broadcasters including BBC, CBC/Radio-Canada, and World Radio Network, have popped up on the Net instead on the air. For example, the author of this paper has performed live DJ sessions called Neuro Net Radio [6] twice from home with independent music provided by Neuro Net Recordings [6], using electronic mail and IRC (Internet Relay Chat) as feedback tools for the listeners. The sessions gained 25 simultaneous listeners at maximum from various regions in Japan. This shows that the potential of Internet audio streaming is enormous.

Current barriers

Internet audio-streaming technology has already become a popular method of distributing audio content. Technical, political, and legal issues regarding the making of the audio content, however, are still thick barriers to public access and awareness, especially for low-budget Internet broadcasters. The existing media industry such as music distributors enforce financial, legal, and technological restrictions on audio recording and distribution activities, often through lobbying government bodies. In this section the core issues of the barriers are explained.

SCMS: a double standard of content copy protection

The methodology of creating audio content rapidly migrates from the analog domain to the digital one. Most of the editing is performed on computers or on dedicated digital devices, for the efficiency of the operation and for critical preprocessing of the content, such as strong noise reduction and multi-band audio-level compression. The overall quality of professional-type and consumer-type is almost the same and the difference is getting smaller, as even the professional broadcasters widely use the consumer-type devices for content production. The current copy-protection scheme imposed on the consumer-type devices, however, significantly reduces their usability.

The Digital Audio Interface (DAI), standardized as IEC60958 [7], has been the de facto standard for exchanging digital audio information among Digital Audio Tape (DAT), MiniDisc (MD), and many other devices. This standard has a built-in copy-protection system called SCMS (Serial Copy Management System). SCMS is mandatory for the consumer-type device, and was originally intended to prevent illegal duplication. But so far SCMS has not been mandated on computer interface devices for digital audio processing, and this is a fatal loophole for the entire scheme. The fact that computers are exempted from SCMS protection creates a controversial double standard, which means the copy protection just irritates the rule-abiding users and that SCMS is not effective at all against piracy.

SCMS has three copy-protection states as follows:

Currently consumer-type digital audio devices have a mandatory SCMS implementation that treats all materials as copyright-protected. The user is supposed to be able to make only single-generation digital copies of the original.

The single-copy restriction of SCMS, however, has already been compromised and become useless. Here are some of the applicable methods to circumvent the restriction:

SCMS has another fundamental loophole: it does not protect against duplication in the analog domain. Connecting two MD recorders with an analog stereo audio cable will make a nearly perfect copy of the original, adding very little amount of noise. Then why is copying disabled in the digital domain?

SCMS was originally intended to protect against mass piracy of commercially produced content, such as a music compilation. But the reasoning behind SCMS is of little importance today, since the market for prepackaged content for DAT and MD has been virtually non-existent compared to the de facto markets of compact disc (CD) and digital video/versatile disc (DVD). In fact, DAT decks were available in 1987, but before the full implementation of SCMS the music industry refused to provide a commercial compilation, so the industry killed a market of DAT prepackaged content by itself.

The fact that DAT and MD are mainly used for non-packaged recording suggests that the SCMS of DAI should be reconsidered for their main application--content editing. SCMS is just an annoyance for content editing and does not really protect anything, so it should not be mandated in any device with DAI.

Taxation of private copying: another double standard

Making broadcast content requires a lot of editing. Editing audio content means making quite a few private copies. The existing media industry forces governments to restrict all private copying by lobbying to change the applicable legislation, such as the copyright laws. For example, Japanese Copyright Law prohibits use of copy-protection-breaking devices in Article 30. The article also mandates the manufacturers of recording media to pay the compensation fee for private copying to the copyright owners [KA1998]. This article has two goals: to reduce copyright infringements and to compensate for the loss caused by making private copies. The implementation of this article, however, far from satisfies the goals.

The Japanese system of paying a compensation fee for private recording that has been imposed on the recording media has many loopholes. For example, CD recordable (CD-R) and CD recordable/writeable (CD-RW) media for computer use are not required to pay the fee, while the consumer-music-type CD-R and CD-RW media are under the coverage of the system. This creates another controversial double standard and a significant loophole in the whole system. Using a computer with an attached CD-R device, anyone can make a copy of a music CD without paying the compensation fee. On the other hand, if a user buys a consumer-type music CD-R recorder, the user has to buy the consumer-music-type CD-R media since the recorder rejects the non-compliant ones. A smart user will choose to buy a computer rather than a restricted music-only CD-R recorder.

Another example of the controversy in the Japanese compensation fee system is found by comparing MD and MP3 devices. MD is under the coverage of the system, while MP3 format and devices are not. The overall quality of sound of the two media is almost the same. And many personal computers have the capability to record the content of a CD or external audio streams in digital format. A smart user will choose to buy an MP3 portable player and attach it to a computer, rather than connect a CD player and an MD recorder, especially if the user has a large collection of CD music. The running cost of computer storage devices such as hard disks is much less expensive than that of MD media. And the user does not have to pay the compensation fee. This sort of double-standard taxation is far from fair and should be corrected.

Prohibition of recording streamed content: double standard again

There is a principle proven by the history of the battles between the cryptographers and the code-breakers [BS1999]. Following is the author's interpretation of the principle for audio data:

When a listener wants to listen to audio content, the content must be available in the listening device in a plain unencrypted format. If the content is encrypted, it has to be decrypted on the device before being listened to. If the encryption is software-based, a clever person may crack into the software code and capture the plain-format data. Even if the encryption is hardware-based and encapsulated in a so-called tamperproof device, the plain-format device would soon be captured once the encryption key is retrieved from the device. And it is almost impossible to make hardware completely tamperproof.

To make this short, the plain-format audio content must be available to the listener when he or she listens to it. This leads to a conclusion that nobody can effectively copy-protect a plain-format content.

Popular streaming media players have the ability to record the streams, but in most cases the sender of the stream tells the player to prohibit the recording. While the author does not intend to deny the right of the sender to prohibit the recording, this sort of a voluntary copy-protection mechanism has already been technically compromised. Here are some of the applicable methods:

Those who use the methods mentioned in the previous list can record the content of a streaming broadcast, whether the recording is prohibited by the sender or not. This is another double standard of the existing copy-protection schemes, which are ineffective and only irritating the legitimate users. In many countries including Japan, private recording is legal as long as the recorded item is not transferred to the third party. Disabling this ability on streaming broadcast players significantly hampers the reusability of the streamed content.

SDMI: securing the industry while making listeners insecure

The media industry, including the music distributors, has continuously failed to implement an effective copy-protection method. SCMS, mentioned previously in this paper, is a good example. In November 1999, a DVD copy-protection scheme called Content Scrambling System (CSS) was broken by the leakage of a key. It is reported that the persons who invented the code-breaking program called DeCSS had no intention of piracy; they wanted to play the legal DVDs on their Linux computers. Since there was no Linux player for DVD, the attempt to make DeCSS does not necessarily make a piracy case [ES2000].

The CSS has another flaw; it was not designed in an open process and apparently was not thoroughly tested against possible attacks. There is no security in obscurity. On the other hand, the U.S. government is in the selection phase of many open candidate algorithms to choose the Advanced Encryption Standard (AES), their next-generation standard of common-key cryptography, through an open-discussion process [11]. History proves that the open, published, and publicly trusted encryption algorithms are always better than the proprietary obfuscation methods. The Secure Digital Music Initiative (SDMI), however, which is going to be a de facto standard among digital music players and content distribution, has yet been pursuing the obfuscation method based on their proprietary schemes instead.

The usage of the word secure in the name of SDMI misleads the listeners; using SDMI will not secure the listeners at all, and may even threaten their privacy. It is mandatory for SDMI-compliant digital music players to use storage media with unique identification numbers; otherwise the player itself should be uniquely identified, which is an SDMI requirement [12]. This means each SDMI music player and the listener are uniquely identified during the transaction between the music content provider. This is a violation of the right to listen to the music anonymously. From this point of view, the listeners are less secured by SDMI.

SDMI is also controversial regarding whom the standard secures. Currently the American music industry has been dominated by only five major distributors: Time-Warner, EMI, Universal Music, Sony Music, and BMG. On January 24, 2000, Time-Warner and EMI agreed to merge and form a new company, Warner EMI Music, so the number of major distributors is only four. SDMI had 159 members as of January 14, 2000 [13], but most of the members are from the existing industry giants including electronic device makers, copyright control organizations, and the music industry companies. Independent distributors such as have been excluded from membership, and the listener's perspective is little reflected in the proposed SDMI specifications. The fairness of mandating a technical specification just to protect a small number of the industry giant companies is questionable.

An alternative model

This paper has been concerned with the technical, political, and legal impediments against the popularization of Internet audio content distribution. This section presents a set of alternative ways to resolve the double standards and enable public access to the audio-streaming technology, instead of letting the technology dominated by the corporate giants.

The basic principle: no copy protection

The idea of copy protection for public content distribution is self-contradictory. The only effect of copy protection is to raise the barrier of access for the users, whether honest or dishonest. The users must obtain a plain unencrypted copy of the content to use it [BS1998].

The author proposes to stop imposing mandatory copy protection for mass distribution on DAI, Internet streaming media, and all the methods of transferring content. Removing the copy-protection requirement may reduce the cost of digital audio devices and ease the process of the content editing. Public dissemination of audio content is not a private communication; the content creators want their pieces to be listened to by as many people as possible.

Regarding the no-copy-protection principle, SDMI is a very controversial movement and may backfire on the move to popularize music distribution over the Internet. Why do music content distributors have to secure themselves from the listeners? When demand for music is higher than supply, the distributor should consider this as a chance, rather than a risk [ES2000]. If distributors want to protect their intellectual property rights, they should do so by other means.

On the other hand, when exchanging an unpublished piece of music under ongoing production over Internet, the content must be carefully protected from being tapped, as well as the data streams of virtual private networks or secure Web transactions. The distributor must use well-trusted-and-proven encryption and authentication algorithms to protect the content. This method, however, does not scale well for a public music distribution system, due to the difficulties of maintaining a public-key infrastructure. And even with this method, the distributor cannot prevent the content listeners from obtaining the plain format content.

Termination of blanket taxation to recording media

The blanket coverage principle of taxing private recording for preventing illegal music copying is doubtful from the point of allowing fair-use of media, and should immediately be terminated. Blanket taxation to recording media results to the higher costs unfairly imposed to the law-abiding users and contributes little to reduce piracy.

Here is an example of a failure to force the blanket taxation. Canadian government has tried to impose the compensation fee scheme in 1999, but the movement is legally suspended [14]. While in Japan the taxation is focused onto digital recording media, in Canada the legislation proposed to impose tax also onto the analog cassette tape. This leads to a nation-wide controversy because a CDN$ (Canadian Dollar) 0.50 cassette tape would be raised to CDN$2.25. A Canadian Reform Party MP Inky Mark points this out as follows [15]:

One manufacturer alone sells more than one million voice-quality audiotapes every year to churches who record their worship services for shut-in's, hospital patients, and others. "CD-R" technology is used for data storage. Educational institutions likewise use hundreds of thousands of audiotapes for purposes quite apart from dubbing any copyrighted performances.

Whether the media is digital or analog, the same conclusion applies: taxing the recording media without regarding the usage is an unfair way of collecting money. Taking money from the tapes from the worship services for protecting music industry is illogical, since the Holy Bible is not a music product.

Watermarking to ensure traceability of content

A content creator and a distributor have the right to claim the copyright on their products. The implementation is effectively performed by standardizing watermarking of the audio content. The existing industry should focus on advancing this technology than wasting time on developing ineffective copy-protection mechanisms.

Watermarking is a method of embedding an independent stream of information while minimizing the degradation of the original audio content. The technology is already robust enough against practical data encoding methods such as MP3 encoding, resampling, filtering, and cropping into the pieces of shorter length [MASK1999]. This means the technology can be used not only for the protection of an audio content file, but for continuous audio streams and the derived content.

A combination of watermarking with the search engines is a powerful systematic tool to detect possible copyright infringement. Music distributors who want to protect their copyrights can make an alliance with companies running search-engine servers and collect the necessary information to prove the unauthorized duplication of their products.

Service of broadcasting assistance providers

Purchasing a license for rebroadcasting music is expensive for low-budget broadcasters. The author proposes a business model of a company that obtains the necessary licenses for rebroadcasting and leases them to the low-budget broadcasters. The company can be called a broadcasting assistance provider (BAP). The important point is that the low-budget client does not have to purchase the license; the client can let the BAP broadcast the content partly or solely chosen or controlled by the client. BAP acts as a clearinghouse for rebroadcast licenses.

A primitive form of BAP is already used. has started a new service called Instant Listening Service to members [16], on January 13, 2000. claims this as a virtual CD player service, which provides audio content confirmed to be previously purchased and owned by the customer as on-demand MP3 audio streams, played back from's servers. This service realizes an intriguing transaction protocol as follows:

The roles of BAP are briefly defined as follows:

The current Internet audio live broadcast model is as follows:

Content producer -(content)-> encoder+server --> listeners

The BAP acts as a intermediate body to solve the license issues:

Content producer -(instructions and content)-> BAP --> listeners

BAP can be an effective solution not only for clearing license restriction, but for providing a virtual warehouse of Internet broadcasting facilities. Concentrating expensive resources such as high-bandwidth Internet links and high-performance servers, BAP can also be a practical solution for reducing bandwidth consumption and increasing reliability of Internet streaming broadcast.


In this paper, the copy-protection issue has been analyzed and the unfair double standard imposed by the existing industries is explained. The author explains how current practices such as SCMS, taxation to recording media, prohibition of recording Internet streaming broadcasts, and SDMI have controversial adverse effects on the popularization of Internet audio content distribution, including streaming broadcast.

This paper also shows that the main way to popularize Internet audio streaming is to break the existing political, social, and legal barriers set by the existing industries, and that the technological foundation for sending and receiving streaming and on-demand audio has already been widely deployed. The potential of streaming may further evolve by increased availability of global multicasting technology.

In this paper the author proposes a new model of copyright control by phasing out copyright protection and shifting to the watermarking technology, since the deployment of plain unencrypted data cannot be technically restricted. The author also proposes a model of broadcasting assistance provider, which serves as a clearinghouse of licensing issues and technical restrictions for low-budget individuals and small groups to perform Internet streaming and on-demand broadcast.

The intention of the author has not been to suggest completely denying existing scheme of copyright control, nor imposing a strict regulation, but that depending solely on copy protection is no longer practical to protect the copyright of audio content, and that watermarking is a viable alternative to solve this issue.


[BS1998] Bruce Schneier, Software Copy Protection, Crypto-Gram, 15 November 1998, available at

[BS1999] Bruce Schneier, DVD Encryption Broken, Crypto-Gram, 15 November 1999, available at

[CM2000] Courtney Macavinta,'s move to copy CDs stirs debate, CNET, 28 January 2000, available at

[ES2000] Eric Scheirer, Three Lessons From DVD for SDMI, news article, 19 January 2000, available at

[KA1998] Kazuhiro Ando, Yoku Wakaru Ongaku Chosakuken Business 1: Kiso-Hen (Practical Music Copyright Business, Part 1: The Basics), Rittor-music, 1998 (in Japanese).

[MASK1999] Michael Arnold, Sebastian Kanka, MP3 robust Audio Watermarking, from the Proceedings of DFG VIIIDII Watermarking Workshop 1999, 5-6 October 1999, Erlangen, Germany, available at


[1] Technical details of the RealNetworks' products, such as RealSystem, RealAudio, RealServer, and RBN (Real Broadcasting Network) are available at

[2] MP3 stands for MPEG Audio Layer 3, a method of audio data compression format. The technical details are available at

[3] SHOUTcast and WinAmp are products of Nullsoft, an America Online company. See and for the details.

[4] The details of icecast are available at

[5] The details of Yahoo! Broadcast are available at

[6] The details of Neuro Net Recordings and Neuro Net Radio are available at

[7] The overview of Digital Audio Interface is available at

[8] The author has successfully disabled the SCMS protection between two DAT decks using the Gina digital audio card of Echo Audio on a PC. See for the details of the Gina card.

[9] A program called Total Recorder by High Criteria Inc. The details are available at

[10] Information about Streambox Ripper! is available at

[11] Information about AES is available at

[12] SDMI Portable Device - Part 1, Version 1.0 (Document Number PDWG99070802) mandates that an SDMI-compliant Portable Media has to be bound to an identifier in the SDMI environment. For further details, see Section 5.4 of the document, available at

[13] SDMI's member list as of 14 January 2000 is available at

[14] Information about the Canadian attempt to tax recording media for private recording is available at

[15] The details of the Canadian Reform Party's claim on the Tape Tax issue are available at

[16] Instant Listening Service of is available at


The author would like to acknowledge the support and cooperation of TDI RCAC Project, the Executive Director Kazuo Hirono, and the staff members. The author would also like to acknowledge the generous support of Neuro Net Recordings, an independent organization of electronic music creators, for whom the author serves as the Internet Archivist.


The views expressed herein are solely of the author. They do not necessarily represent those of the TDI RCAC Project or any members thereof.