Streaming audio

from Wikipedia, the free encyclopedia

Streaming audio is a variant of streaming media in which audio data is continuously transmitted over a computer network. In its simplest form, streaming audio can be thought of as radio broadcasting on the Internet; the process then ranks among the potential new mass media .

However, streaming over the public Internet is only a - albeit significant - special form. In the form of local streaming, on the other hand , streaming audio is - like the stereo system - a component of private " entertainment electronics ".

Streaming audio enables a multitude of personalized "broadcast" programs and formats, access to sound and music archives as well as sophisticated combinations with other multimedia technologies for web presentations and web conferences . In practice, however, such visions face a number of legal, economic and technical hurdles.



Streaming audio , especially in its special form of internet radio , promises to solve a classic problem of broadcasting : the scarcity of frequencies . Only certain sections of the frequency spectrum are suitable for radio broadcasts , which is why the number of terrestrial channels is limited in principle.

Cable distribution systems, satellite communication and the digitization of transmission technology reduce the frequency shortage , but they can never completely eliminate it. In contrast, an unlimited number of “channels” can be transmitted over the Internet. The only physical limit is the gross bandwidth of the data lines between the Internet “sender” and Internet radio “receiver”. In addition, the available bandwidth is allocated dynamically and is therefore only used when required, so there is no need for a fixed allocation of transmission frequencies for permanent use.

Digital data and special formats

While traditional radio broadcasts analog signals, streaming audio operates with digitized data; these signals are converted into a special format that can be broken down into data packets and transmitted over a network; Such a sequence of related data packets is called a stream .

The conversion into a special streaming format is carried out by the streaming provider using a so-called encoder ; this is primarily responsible for reducing the data rate : The digital audio data stream of an audio CD ( PCM ) has a data rate of 176 kByte / s (= 1,408 kbit / s); a single ISDN - channel provides only 64 kbit / s, and even a DSL port with 1024 kbit / s is not enough, an uncompressed sound in full CD quality to receive.

Data compression

Special psychoacoustic data reduction processes and data compression processes drastically reduce the amount of data in audio files and streams. This compression required for audio streaming goes well beyond that of MP3 . Even with MP3 files, disturbing distortions can still be heard with certain pieces of music at data rates of 128 kbit / s, which is why encoding with 192 kbit / s or higher is recommended. Since significantly lower data rates are required for audio streaming over the Internet - so that the broadest possible target group can be reached with a narrow-band Internet connection - a comparatively poor sound quality is to be expected.

Sound quality

The unsatisfactory quality of streaming was also criticized by a study by the US company Keynote Systems from October 2000, which showed that the Internet is too slow for the transmission of music and video via streaming. The study compared the technical quality of offers from well-known streaming providers such as MTV , CNN and Tower Records with DVD, TV and CD quality. On the evaluation scale, the high-quality DVD achieved the maximum possible 10 points, while the streaming offers came at the bottom of the comparison with 1.19 to a maximum of 3.46 points. It was not the end customers' Internet connections but the networks of the Internet service providers (ISPs) that were identified as the bottleneck .

These qualitative restrictions apply above all to the transmission of music ; lower data rates are sufficient for human speech . On the other hand, the restrictions on Internet streaming are largely irrelevant for local streaming, for example from the home PC to the stereo system; Network bandwidths in the range of ten to 100 Mbit / s or higher are usually available here, which allow high-quality audio streams.

The Internet was further developed, the intercontinental connections expanded and the peering between the providers improved; Last but not least, broadband connections of the home PC established for large groups of users in Central Europe, as had been common in the USA for years with cable modems and various DSL technologies. The loss of quality in Internet streaming criticized by Keynote in 2000 has decreased and has now reached a sound quality that can definitely compete with a simple FM radio . Today, only a few Internet radios limit their streaming data rate to 64 kbit / s or less, but the basic problem remains with heavily frequented streaming audio and most streaming video offers.


The data to be streamed can either be delivered via a normal web server ( HTTP streaming ) or via a special streaming server with extended options. The listener needs a software ( streaming client ) or a streaming-capable device, which reassembles the data packets and makes the offers navigable (channel selection, starting and stopping the stream, etc.).

Since any streaming requires an available network, the streaming audio listener must either have a streaming provider in their local network or be connected to the Internet for the duration of the stream listening.

Differentiation areas

Every streaming method must be distinguished from downloading files, which results in a more or less clear time lag during downloading; this delay can be between a few seconds and several hours, depending on the size of the file. Such music downloads are possible, for example, via commercial payment services , but also via peer-to-peer networks.

Streaming audio is also to be distinguished from streaming video ; In this area, due to the higher data rate of digitized moving image sequences, there are even higher demands on the available bandwidth than for pure audio transmissions.

In practice, live streaming hardly differs from a conventional radio program for the end user. The user only needs other end devices - as a rule, this will be a personal computer and the broadest possible Internet connection - and in return receives a potentially considerably more diverse range of Internet radio "transmitters". This added value is bought primarily with a drastic loss of mobility: a PC with a mains connection is more sensitive, heavier and less portable than a portable radio .

On-demand streaming offers a greater variety of programs , in which the end user can access archived program contributions in a targeted and time-shifted manner. Theoretically, such a universal jukebox with millions of pieces of music and word contributions could be realized, but in practice this utopia is prevented by economic exploitation interests and legal barriers. The reality of on-demand streaming therefore limited to scattered program extracts that the streaming listeners actively using Multimedia - search engines gather together must. So-called " social software " offers alternative approaches . Streaming providers allow e.g. B. the creation of playlists that can be accessed at any time. A prime example is CBC Radio 3, the internet station of the Canadian CBC in cooperation with NMC, New Music Canada. Here you can put together your personal program from thousands of independent pieces and access it as a stream at any time.

Sophisticated combinations of streaming audio with other multimedia elements for web-based presentations or conferences usually go beyond the capabilities of simple streaming devices and require proprietary streaming clients and a computer. The possibilities of such combinations go far beyond simple radio or television offers, but are extremely costly to produce and are therefore only rarely implemented.

Areas of application and applications

There are many uses for streamed audio; Not all conceivable applications can already be refinanced; many applications are also on the verge of illegality. The fields of application can be differentiated according to the respective target groups.

B2B / B2E

In the area of B2B - i.e. commercial offers from commercial providers to the same customers - these are mainly financial communication, for example the transmission of general meetings, IPO-PKs as well as balance sheet press or analyst conferences, the transmission of congresses, content licensing ( syndication ), as well as training courses, web conferencing and web presentations. Corresponding applications also exist in the area of B2E , i.e. between the company and its employees.

In the business sector, pure audio offers have become rare after the collapse of the New Economy ; most frequently, combinations with presentations or synchronized texts are offered today . Interactive hybrid systems with the telephone network such as All-Streams Interactive could not establish themselves on the market and are rarely offered. In contrast, IP telephony is developing quite successfully, although it is not limited to commercial use.

B2C / C2B

In the area of B2C and C2B - i.e. commercial offers to consumers - a distinction can be made between the primary and secondary use of content, as well as various applications in the area of ​​e-commerce such as audio samples , audio-on-demand and pay-per-list (so-called paid content ), the interactive and / or personalized internet radio, the offer of additional and background information, online archives and various forms of online advertising such as brand renewal, cross-promotion and marketing, rich media ads (multimedia banner advertising) and preroll advertising (the import of advertising jingles before the actual stream).

In a kind of pilot experiment, the online music service Napster concluded between the end of 2003 and the beginning of 2004 with the US universities of Pennsylvania State University and the University of Rochester flat-rate contracts, which gave several thousand students access to the streaming offer from Napster; the offer was primarily intended to curb the spread of illegally reproduced music and was financed from the students' tuition fees . The streamed pieces of music were encoded in WMA format with a data rate of 32 to 96 kbit / s, protected by Digital Rights Management (DRM) and could be played for 30 days before the license expired.

Similar commercial music services are also offered by Apple ( iTunes Store ) and RealNetworks, and Microsoft ( MSN Music ) for US users. In October 2003 Apple had already sold 14 million pieces of music via iTunes from an initial range of 200,000 titles at a price of around 99 US cents each.

In Germany, a consortium of companies has been trying to establish the Phonoline sales platform since the end of 2003 , albeit with modest success; In the fall of 2004, plans became known to merge Phonoline with the Musicload download service from the Telekom subsidiary T-Online .

A business model for streaming digital music has been around for some time . With the revenue model, providers of this business model use the freemium principle. The customer can choose between a free advertising-financed and a paid version of the service. The paid premium offer, which is available for a monthly fixed amount of around five to ten euros, includes additional functions such as B. improved sound quality, access via an application for mobile devices including offline availability and advertising freedom. The largest providers in Germany include Spotify , Simfy and Deezer .

C2C / P2P

The areas of application in the C2C and P2P area - i.e. between consumers and other consumers - often go beyond what is legally permissible when streaming , since consumers usually only have extremely limited rights to use sound material; Streamed campus radios , for example, are virtually illegal in most countries around the world, and similar applications often operate in gray areas.

The open source programs Streamer by Ian McLeod (since 2001) and PeerCast by Gilles Goddard (since 2002) are an example of such a peer-to-peer application of Streaming Audio . Both can be used to operate Internet radios from the home computer; The file formats MP3 , Theora , Vorbis , WMA , WMV and NSV are supported , PeerCast uses the infrastructure of the Gnutella network to distribute the streams .

Commercial Internet radio operators such as Radio Free Virgin also rely on P2P technology; The program provider can thus save up to 75 percent of its bandwidth costs - most of the data is transported by the participants in the P2P networks.

A comparable commercial product is the P2P radio from Mercora ; In June 2004, the US provider presented software with which a global network of private MP3 and WMA webcasts can be set up; the internet radio client also offers an instant messenger , chat and blog functions as well as its own media player . The allegedly completely legal project is to be financed through an affiliate program.

Internet telephony

A special case of streamed audio data is IP telephony ; In principle, this is also about the continuous transmission of audio data, but the endpoints involved in the communication are end users, between whom a point-to-point connection must ultimately be established. Therefore, the requirements for the technical procedures (e.g. need for a signaling protocol) and the software clients with which the end user uses the service differ.

Functionality and procedure

Provider / sender side


In principle, streaming always comprises the following five steps in production, whereby the individual parts in the respective work processes can be differently pronounced:

  1. Pre-production : conception of the production, financing concept, rights clearing , possibly creation of a script and the distribution of tasks within the production company, the acquisition of advertising for preroll advertising, etc.
  2. Production : In streaming production , an output signal is converted into a streaming format with the encoder using special streaming codecs ;
  3. Distribution : The finished stream is marketed and, on the technical side, distributed either directly via a streaming server or via a more complex form of streaming distribution;
  4. Playback : The stream is received by a streaming client.
  5. Analysis : After the "broadcast", the server log files are evaluated in order to measure the success of the production and to settle accounts with advertising customers.

In the introductory pre-production phase and the final analysis phase , methods are used which are also used in a similar form in the production of sound carriers , radio or television broadcasts as well as web and multimedia productions ; therefore, only the specific characteristics of streaming audio are discussed below.


Technical sequence of a streaming production (schematic)

The technology behind streamed audio offers is basically quite simple: An analog audio signal is required, the audio quality of which does not have to be excessively high. This signal is digitized , cut and, if necessary, post-processed either in a commercially available computer or using special hardware .

The digital audio signal is then converted into a stream-capable format by an encoder during streaming production in the narrower sense ; Encoders are primarily used to compress the input signal as optimally as possible with the help of a special streaming codec and to convert it into a streaming format ; the strength of the compression and thus also the file size and the resulting bandwidth requirements to the end user can be controlled here. With the type of encoding, the output signal is tailor-made for the customer. Commercially available computers with the common Windows , GNU / Linux or Apple Mac OS operating systems are usually used as encoders . The encoder need software such as RealProducer of RealNetworks , which the work of the encode it done.

Passes the audio signal via a telephone line in the broadcast center , a special hardware, called a telephone hybrid like the Magic ISDN of AVT are used for processing, the already diverse spread coding has built. These devices automatically convert an incoming telephone call into an audio signal compatible with studio technology , which saves the streaming provider the need to configure and maintain dedicated encoder PCs.

The encoded file is stored on a file server in on-demand streaming or passed on directly to the streaming server in live streaming .

The basic technology behind the streaming is so simple up to this phase of the production that it can in principle nowadays be carried out by any experienced computer user and with any personal computer . However, the requirements increase disproportionately as soon as the complexity of the content to be streamed increases. For example, if you have to transmit several dozen live streams in parallel, you need a whole encoder park; If the streams are to be delivered not only to a handful of participants but to a mass audience at the same time, the capacities of the most powerful ISPs are no longer sufficient.


Streaming audio can - depending on the capabilities of the streaming server used and the associated client - in principle be transmitted via the popular HTTP and FTP protocols . Live streaming ( real time streaming ), on the other hand, basically requires real-time capable protocols such as RTP .

The most important requirement for special streaming protocols is fault tolerance ; In the case of a bad connection, at least five percent loss of data packets must be inaudible and around ten percent loss must be compensated for in an acceptable quality, i.e. interpolated.

A basic distinction must be made between unicast and multicast streaming during transmission. These basic functional modes of networks are differently well suited for the dissemination of data to several recipients: While with unicasting point-to-point connections always have to be established, with multicasting an arbitrarily large number of recipients can be supplied with one and the same data stream. Therefore, multicasting is theoretically far better suited for mass media streaming. However, it is not yet forwarded by all routers , which is why multicasting is currently mainly used in local networks (cable networks, VDSL from Deutsche Telekom), but rarely in the public Internet. In the future, IP multicast will increasingly replace the costly and resource- demanding unicast on the Internet .

Furthermore, a distinction must be made between wired and wireless streaming in terms of distribution . Wired streaming requires the same infrastructure as a physically wired local network , while wireless streaming largely decouples the playback device from the location of the PC. Starting in 2003 in particular, numerous solutions based on this technology came onto the market, for example the Wireless DJ from Logitech , the Devolo MicroLink product family or Apple's AirPort Express ; Wireless networks based on WLAN or Bluetooth technology are set up here, which can then be used for streaming.

Spreading Internet streaming to a larger audience either requires a very powerful ISP or the overlay network of a specialized streaming service provider such as Akamai. The load balancing specialist Akamai operates a global network with an “intelligent” operating software that analyzes requests from a streaming client for their geographical origin and forwards them to the nearest server. Content to be streamed is replicated with this server infrastructure and made available on a decentralized basis via a content delivery network (CDN). In addition to Akamai, there are regional and more specialized streaming service providers.

User / recipient side

Program selection

The streaming user can choose from a large number of offers, but there is currently no full electronic equivalent to a printed program guide. However, there are a number of relevant points of contact, especially the independent streaming portals , services such as Reciva and the self web deals of providers. Some streaming clients such as Microsoft's Media Player , the Real Player or Winamp have a comfortable search for Internet radio stations installed in For free and independent clients such as the VLC media player , the user has to find out the URL of the desired program himself and enter it into the client.


VLC media player on Linux when playing an HTTP stream

Streamed transmissions are played back using a streaming client; this can be software for the PC or hardware .

High- performance streaming boxes usually also take on numerous functions of a streaming server, while the simpler and cheaper models rely on a PC with often proprietary server software.

Advantages and disadvantages


User / recipient side:

  1. There is no time delay due to the complete download of a file.
  2. There is no need to download a complete file, it is possible to "listen in".

Provider / sender side:

  1. The initial investments are comparatively moderate in comparison to a classic broadcasting studio with an associated radio station.
  2. According to an analysis by Wolfgang Zieglmeier, a bandwidth of one ISDN B-channel (64 kbit / s) enables almost transparent mono audio transmission; One of the widespread broadband connections for end customers via DSL is now sufficient for high-quality audio streaming thanks to improved codecs .
  3. The flexibility in the design of the format increases compared to the download; both live broadcasts and, for example, preroll advertising are possible.
  4. A brand extension to new mobile devices opens up - at least in theory - new application possibilities compared to conventional radio.


User / recipient side:

  1. On the part of the listener, the playback quality is sometimes poor compared to FM radio; According to analyzes by the sound engineer Wolfgang Zieglmeier, the modem speed at very low bit rates (below 64 kbit / s) has a very large influence on the sound quality; at bit rates of up to 28.8 kbit / s, radio quality is not nearly achieved (also at 64 kbit / s, stereo, for example with ISDN ); Today, thanks to faster connections and modern broadcast and multicast processes, broadband audio streams (128 kbit / s and higher) are predominantly used.
  2. A fast and reliable internet connection is necessary, while wireless radio can be received anywhere with an inexpensive and handy radio. Ways out are WLAN or various radio networks and cellular networks such as UMTS , HSDPA , Bluetooth or WiMAX .
  3. The Internet connection must not be fully utilized and must offer bandwidth reserves, otherwise there will be a "network congestion " and the transmission will stall. This is especially true for narrow-band Internet access.
  4. The recipient creates data traces that allow attacks on his anonymity .

Provider / sender side:

  1. Especially when using the unicast protocol, a special streaming server is required on the part of the producer, which becomes more expensive due to license costs , the more simultaneous users are to be served; this problem is increasingly being minimized through flat-rate billing by collecting societies such as GEMA.
  2. There used to be numerous incompatible products available for streaming; sometimes there are significant qualitative differences between the individual systems with the same bit rate. Suppliers who want to reach as large a target group as possible therefore had to encode for several output formats, which multiplied the production effort; because open formats such as MP3 and MP4 / AAC are increasingly being used today, this problem has almost been overcome.
  3. The streamed signal requires large to enormous transmission bandwidths or an overlay network ; In any case, the more recipients are reached, the more expensive the transmission, while the broadcasting costs in conventional broadcasting are largely independent of the number of recipients. This is due to the fact that the provider has to send each client its own stream ( unicast ), while normal broadcasting only sends a signal once, which each client can record ( broadcast ). Broadcasting , peer-to-peer streaming and modern multicasting offer a way out of this dilemma .


Copyright / exploitation rights and rights management

Rights holders to content such as Sony BMG Music Entertainment in the music sector or DFL / DFB in the football Bundesliga are interested in protecting their economic interests. If content is licensed to commercial providers, third parties should not be able to participate free of charge. For this purpose, methods such as Digital Rights Management (DRM) have been developed to better control the use and distribution of content.

Streamed data is in principle better protected against illegal "secondary use" than downloads, but all copy protection methods of the popular streaming formats have so far been decrypted or circumvented. Basically, audio content can always be tapped in at least analog format as long as a transparent digital-analog conversion takes place somewhere , i.e. at the latest at the audio output of the sound card .

Free and commercial products are offered for recording ( ripping ) streamed content.

License fees

Webcasters have to pay license fees for their programs.

In Germany, the collecting society GEMA introduced flat monthly fees for web radios as early as June 2001.

The International Federation of the Phonographic Industry (IFPI) has been offering a worldwide license for web radios since the end of 2003.

A comparable international license has long existed for conventional radio stations that also offer their programs online: the so-called simulcast license.

In February 2004, various business interest groups reached an agreement with Internet radios on the amount of license fees to be paid: Web radios must pay around 0.07 US cents per music track and listener or, alternatively, 1.17 US cents per hour broadcast; As a third alternative, providers who market their programs as a subscription can transfer a flat rate of 10.9 percent of sales to the music industry.

According to the IFPI, at the end of 2003 around 1,250 Internet radio stations were broadcasting their programs with IFPI licenses in the USA alone.


In the software sector, too , more and more legal disputes arise due to trivial patents ; Acacia Research , for example, claims various patents under the name Digital Media Transmission (DMT) in the USA and several European countries ( EP patent 566662 , German patent 69230250 ) for the transmission of streaming offers via cable , satellite or LAN . The company has been trying to enforce these patent claims in court since July 2003. The entertainment and media group Disney with its subsidiaries ESPN and ABCNews is one of the most prominent licensees to date .

What Acacia registered as a patent back in 1991 is not a specific technology, but the idea of ​​decompressing audiovisual data in real time on the client side; Acacia claims intellectual property rights in streaming itself, not a particular implementation. In 2004 it was still being examined whether the patents applied for by Acacia would also become fully legally binding in Europe.

History and Development

The first attempts with streaming media in the narrower sense date back to 1994, when Progressive Networks , later RealNetworks , developed the first special streaming formats .

With the new economy, streaming also flourished ; numerous offers were developed and tried out; When the economic crisis of the Internet collapsed, however, disillusionment spread when it was discovered that most business models were not economically viable.

Today streaming has firmly established itself in some niche areas, for example in the form of Internet radio. Today, however, only strategically operating companies such as T-Online operate extensive streaming offers , which are trying to establish the more or less attractive broadband offers on the market in the medium term.

The intensive discussion about the type and extent of media convergence between traditional media and Internet-based offers is not yet over; Most of the indicators suggest, however, that in the near future there is less to be expected of the traditional mass media than of permanent coexistence .

A more detailed history and development of streaming media is elaborated elsewhere.

See also

  • SHOUTcast and Icecast , the most widely used open-system audio streaming platforms


  • Gerald Himmelein: Trouble-Free Streams. Enjoy audio streams without gaps despite the low bandwidth (practice). In: c't . Heise, Hanover 7.2002, p. 216
  • Tobias Künkel: Streaming Media in Practice. Technologies, standards, applications . Addison-Wesley, Munich 2001. ISBN 3827317983
  • Jürgen Mayer (Ed.): Streaming Media. Internet more moving, more colorful, louder (new technology). Markt + Technik, Munich 2001. ISBN 3827261430
  • Klaus J. Schäfer, Andreas Hensel, Franz Lehner: Video and audio streaming on the Internet . Universitäts-Verlag, Regensburg 2003. ISBN 3932345924
  • Axel Zerdick, Arnold Picot, Klaus Schrape and others: Die Internet-Ökonomie. Strategies for the digital economy (European Communication Council Report). Springer, Berlin - Heidelberg - New York a. a. 1999. ISBN 3-540-64915-8
  • Jonathan Dörr: Music as a Service - A new business model for digital music . epubli GmbH, Berlin 2012. ISBN 978-3844216714

English speaking:

  • José Alvear: Web Guide to Streaming Multimedia . Wiley, New York - Chichester - Weinheim et al. a. 1998. ISBN 0-471-24822-3
  • Eyal Menin: The Streaming Media Handbook . Prentice Hall, Upper Saddle River NJ 2002. ISBN 0130358134
  • Peggy Miles: Internet Guide to Webcasting . Wiley, New York - Chichester - Weinheim et al. a. 1998. ISBN 0-471-24217-9
  • Jeannie Novak, Pete Mankiewicz: Web Guide to Producing Live Webcasts . Wiley, New York - Chichester - Weinheim et al. a. 1998. ISBN 0-471-29409-8
  • Michael Topic: Streaming Media Demystified . McGraw-Hill, New York 2002. ISBN 007138877X

Web links

Wikibooks: Streaming Audio in the Home Area Network  - learning and teaching materials

Individual evidence

  1. Keynote Streaming 20 Streaming Media Performance Index ( Memento from December 5, 2000 in the Internet Archive )
  2. ^ Penn State's new online music service with Napster
  3. ^ University of Rochester, Napster to Provide Online Music to Students - First Private University in Nation to Sign Digital Music Agreement
  4. GEMA Info Brochure Webradio (PDF) ( Memento from January 26th, 2009 in the Internet Archive )
  5. IFPI: Recording industry announces new one-stop-shop for webcast licensing
  6. EP1335601 - Audio and video transmission and reception system - Acacia Media Technologies Corporation ( Memento of April 27, 2005 in the Internet Archive )
  7. Acacia Technologies Files Cable And Satellite TV Patent Infringement Lawsuit (PDF) ( Memento of March 26, 2005 in the Internet Archive )