Streaming media are media that are distributed directly through computer networks (such as the Internet). The process is called streaming. During streaming, a portion of the data is continuously placed in a buffer so that a program can play it. This allows the received media (video and audio) to be consumed directly without the entire broadcast being downloaded.
It may or may not be about livestreaming. If it is not live one speaks of a web video and video on demand.
Codecs ensure efficient distribution of image and sound. In general, codecs for streaming media are ‘lossy’ where sound and image information is limited for smooth distribution. The choice of a codec depends on the bandwidth of the connection and the computing power of the hardware for compression and playback. Streaming codecs scale from GSM (9.6 Kbps) to hd TV (4-10 Mbps). Other known codecs include MPEG-1 (including MP3), MPEG-2 and FLAC. Specific streaming media codes are RealVideo Windows Media, MPEG-4 (followed by H.264), FLV and AAC. On the Internet, mp3 and OGG exchange services are used for audio and Xvid for video.
The file format is the build-up and description of a file. In addition to (according to the codecs compressed according to the codecs mentioned) it also contains so-called metadata: technical descriptions of the audio and video files, substantive descriptions of the document, and finally, a file format offers possibilities for interactive functions. Examples are the . MOV QuickTime container layout, wherever the . MP4, MPEG-4, and .3GP file formats are based on it. 3GP is the standard for video on mobile phones. Other, less extensive file formats are . ASF (also known as . WMV) from Microsoft and .RM from RealNetworks. So-called DivX files use the simple . AVI file format containing a divx-coded video track and MP3 audio track. DivX 6 introduced its own file format (.divx) that enabled subtitles, markers and other interactive features, just like MPEG-4. The FLV format is a proprietary format of Macromedia and uses H.263 or ON2 VP6 as video codec and MP3 as audio codec.
Virtually all mobile phone manufacturers support the 3GPP format. This industry standard is a derivative of MPEG-4 with specific characteristics for handsets such as fixed screen size appointments and video and audio codecs. 3GPP is also the format used for DVB-H (Digital Video Broadcasting for Handhelds). The exception for now is the Microsoft Mobile-based smartphones, which represent a small market share and primarily support the Windows Media format: optionally, however, 3GPP playback software can be installed.
Recording sound and imagery requires a sound and/or video card with an input, such as composite video, S-Video, DV or SDI. The recorded audio or video is converted (or transcoded) to the desired file format.
To play, a media player that supports streaming media is used.
ITunes is especially popular by supporting podcasts and vodcasts.
Windows Media Player is often used in environments where foreclosure is required, due to DRM features.
QuickTime Player benefits from piggybacking distribution via digital cameras, the popular Apple iPod and the popular Apple iTunes. In addition to its own QuickTime format, more than 100 audio, video and image formats are supported. Apple encourages the use of the MPEG-4 standard over its own MOV format. The QuickTime Player has a Pro version, which offers a lot of additional features for a small fee, such as cutting and pasting at meta level and exporting audio and video to dozens of high-end and streaming formats. The QuickTime Player is often confused with QuickTime, which is an extensive multimedia architecture that runs as a framework on top of the OS, allowing third-party programs to use a rich library of multimedia features. Examples include iTunes, iMovie, Final Cut Pro, Word, Powerpoint, Premiere, Photoshop, Director, and hundreds of multimedia-oriented programs for both PC and Mac.
RealPlayer has built up a bad name among the public due to instability and alleged ‘spyware’. Due to the high cost of streaming distribution technology, there are still very few providers that use the Real format. As a result, the market share of this player is declining. RealPlayer is also building more and more on the underlying Windows Media and QuickTime architectures for being able to support various media formats.
There are software programs that use Windows Media Player, RealPlayer, and other players. These packages make it more user-friendly to zap between channels just like on normal television.
VideoLAN is an open source video player. It is currently under fire because, unlike other developers, the developers do not want to issue licenses for the use of MPEG-4. VideoLAN asks end users to take responsibility for paying a one-time fee for the use of technology to the MPEG-4 licensees, but does not facilitate this. In practice, end users overlook this provision in the VideoLAN EULA, and in addition, the licensees are not set on doing business with individuals but with technology developers.
In the living room, set-top boxes with hardware playback programs are more applicable.
Integrate into the browser
Embedding the media player plugin within an HTML environment is often done with ActiveX (Internet Explorer family) or EMBED (Mozilla family). The media player is not started as an external program, but is shown within the web page. This allows the provider to offer the video in its own designed html player.
Virtual cutting and pasting
It is possible to start randomly in VOD streams at any time. Instead of having to download the entire file first, the server can start playing from any time. The Virtual Slicer is a public web-based tool that makes it possible to virtually cut fragments from a VOD stream. No actual portion of the video is copied (it remains on the server in its entirety), but a metafile is generated that lists the VOD stream URL, start time, and lead time. This file automatically opens the correct media player and ensures that only the clip plays. Works with the most common media players.
Sometimes it is desired to shield audio or video content. The reasons for this may include privacy, secrecy, rights or commercial exploitation, such as renting an online feature film. There are multiple options to shield streams. Sometimes a combination is used:
- Obscurity: Not telling or hiding the URL (location) of the streaming file. A website tries to shield the source. This form of foreclosure is strongly discouraged because the URL can be quickly traced and can be spread quickly via Usenet, IRC and other social networks.
- Password protection: Files, or folders containing files are protected on the streaming server through a user rights management. Because distribution servers from different manufacturers maintain their own methodology for this, this is also not a recommended solution. Users must remember multiple passwords, or double enter, or administrators must build complex links and implement triple user management. Users can start sharing passwords. For some media servers, plug-ins are available that can be linked to user databases.
- Network shielding: A server is placed within a protected network (intranet, VPN), or access to the server is limited to a limited number of IP ranges. This allows for a relatively simple control for a particular group. However, foreclosure is not possible on a personal level (authentication).
- Barrier: A barrier is a server placed between the streaming server(s) and the network (often the Internet). The website generates a unique session code and pastes it behind the URL. The barrier only gives access to the URL if the session code is valid. After the course of the session, the user must request a new one via the website. A simple firewall can fill in this feature, but its performance is low. There is also a barrier solution at TCP/IP stack level, which offers higher performance. This barrier is a highly efficient, user-friendly and platform-independent shielding solution and is sufficient wherever encryption of the content is not required.
- DRM: The toughest foreclosure method is Digital Rights Management, which is covered in the following section.
Digital Rights Management
Digital Rights Management (DRM) makes data flows and files with a key unreadable. Only viewers with a suitable key are able to access the streams and files. However, it is not possible to copy, share or hand over the key to someone else. One can obtain one’s own key by, for example, first completing a transaction. This can be a financial transaction, but also regular authentication. The key contains so-called ‘business rules’. These rules determine how often and for how long the data stream or file can be viewed or listened to. The license server (which issued the key) can possibly check whether the key is still valid and whether the rules in the key may need to be changed. The PC can also temporarily take over the role of license server and sublicens the license to, for example, an mp3 player.
DRM allows the owner or publicist to provide access to this content only to those who have paid for content. In combination with streaming media servers, the viewing and listening behaviour can be precisely met. Microsoft and RealNetworks have an operational proprietary DRM solution that is intertwined with its own streaming media codes, coding programs, and playback software. The specifications for an industry standard MPEG-4-DRM solution are ready and applications are being developed at the time of writing. For mobile applications, there is the OMA DRM specification. Apple uses the proprietary FairPlay-DRM technique for the popular iTunes Music Store, but announced on January 5, 2009 that the entire Store would become DRM-free.
Hosting and distribution
Streaming requires a streaming server or an account on a streaming server from a carrier. The cost of streaming software varies widely. Some software is free and open source, but has a huge learning curve; other software (RealServer, Flash Media Server) can be stacked up to 400,000 euros. Streaming providers provide shared accounts (subscription), dedicated servers (rental), media server clusters, or distributed servers. Some providers seem cheap, but charge higher amounts for generated data traffic afterwards.
The bandwidth actually available must also be sufficient. Shared servers must have their own 100Mbps link to the Internet. Dedicated servers must have 100 Mbps or even one or more gigabit connections. The service level agreement (SLA) includes the terms of the service, including the guarantee of server availability,whether they are actively monitored and the speed at which a defect or crash is repaired.
The Flash Media Server II cost approximately $4,500 per 150 connections. If one wants to offer such a server unlimited to a gigabit connection (1000 simultaneous viewers at 1 Mbps), this costs the user a lot of money in software licenses. Therefore, some hosting providers offer FLV hosting as a progressive download. The disadvantage of this method is that no random access to fragments is possible, no proper measurements can be made. Meanwhile, innovative media hosting companies are launching streaming alternatives for FLV files that do offer random access and fast bursting. Adobe’s Flash Media Server III is split into an affordable ($4,500) unlimited media server and a more expensive (over $10,000) interactive server.
A disadvantage to the Flash Media Server is that it is only suitable for streaming to the Flash plug-in or players who can handle the Flash protocols. A new generation of multi-screen solutions is therefore on the rise. Multi-screen is the term used for being able to stream content to different types of devices, players, and plug-ins from the same server. Wowza Streaming Engine is an established name in the multi-screen world. The Open Source Server Mist server is an alternative. Unlike other open source server products (Red5, crtmpserver), this is the first product to offer a multi-screen solution in a Plug & Play format.
New is p2p distribution. Each viewer also becomes a transmitter. Companies like Octoshape have succeeded in making p2p suitable for live broadcasts as well.
In 2005, volume traffic on the AMS-IX doubled. The biggest growth was streaming traffic (by 80%). Of the 100Gbps data traffic, approximately 10 Gbps is already believed to be streams. Streaming traffic grew faster in 2005 than peer-to-peer, downloads, usenet, email and surfing combined. Growth is expected to accelerate in 2006. For this reason, streaming traffic is increasingly being processed via peering (via the AMS-IX node), or distributed decentralized. Traditionally, servers are placed centrally behind a backbone. However, by placing the servers decentralised (network-topologically) with Internet providers, the backbones are relieved and there are hardly any more costs for data traffic. In addition, it is possible to scale better and provide a quality guarantee, although some central clusters are very cheap and already offer high performance.
In order to manage the central servers properly, keep content management simple, to be able to distribute users among the many servers and to be able to process the log files centrally, a so-called ‘CDN’ management server is required. Most CDN operators develop and operate this technology exclusively for themselves, Akamai and Vitalstream are examples of this. Meanwhile, commercial CDN products have been launched, allowing any provider or broadcaster to roll out their own CDN at low cost and a short time to market.
History of streaming
The earliest known stream is cambridge’s webcast coffee pot camera, which was used to see if the coffee had gone through. It existed as early as 1991, long before the web gained worldwide fame.
Microsoft didn’t believe in the Internet. Manager Rob Glaser therefore took all the people with media knowledge and started RealNetworks. In 1996, the company launched RealAudio. For the first time, it was possible to distribute reasonable quality live audio via narrowband internet, through a proprietary codec, reflection server and media player. Other initiatives such as VDO were bought out by RealNetworks. In 1997, RealNetworks also came up with video support. In 1998, Microsoft licensed RealVideo technology to build it into its own Media Player. Shortly thereafter, RealNetworks launched RealVideo G2, leaving Microsoft’s license with little value. In 1999, Apple and Microsoft launched their streaming media products, based respectively on Apple’s QuickTime – which had been around since 1991 – and the Microsoft-developed Netshow later christened Windows Media.
Because none of the three product lines of RealNetworks, Microsoft and Apple were interchangeable, there were fears of a “media player war.” In order to promote consumer acceptance for streaming services, the Motion Pictures Expert Group (known from MPEG-1, MPEG-2 and therefore MP3) decided to invite the warring parties to participate in MPEG-4. RealNetworks proposed the RTSP/RTP protocol as the standard for distribution. That was accepted. Apple presented the QuickTime atom-based file format. That was accepted. Microsoft proposed its MPEG-4 video codec, but that was rejected. Partly because the codec code was stolen after publication and renamed DivX, Microsoft withdrew from the MPEG-4 standardization. Microsoft MPEG-4 is therefore not interchangeable with the ISO or ISMA MPEG-4 products on the market and has already been declared a maintenance version by Microsoft itself.
RealNetworks’ technology strategy is no longer based on its own RealVideo format, but on support for multiple formats, including those of Apple and Microsoft. The support of these formats is not official and is technologically one to two generations behind. Apple committed to the open MPEG-4 and 3GPP standards and Microsoft positioned the proprietary Windows Media format. The media player war has now been played out in areas outside the desktop PC with 3GPP dominant in the mobile market. In the HD DVD and set-top box market, the battle was undetermined between Windows Media and H.264 (MPEG-4). In 2008, Microsoft decided to add H.264 to Silverlight. In 2010, Google introduced its own codec WebM. But in 2012 Mozilla also prefers H.264 over WebM.