Part 1 of this two-part series looked at the constant behind-the-scenes refinement of the technologies that enable the compression of audio and video files for online distribution. Part 2 delves into the specialized, somewhat arcane world of A/V compression and the role open source plays in standards development, and in the overall marketplace.
The choice of particular audio-video compression technology is of vital interest to commercial and not-for-profit organizations involved in producing and delivering digital multimedia products and services.
New York-based Paltalk’s choice to go with industry-standard H.264/Mpeg-4 AV for video and a combination of open source Speex and Siren as the audio codecs for its eponymous live Web video-teleconferencing service is a case in point.
The Right Choice
“Our codec choice proved to be wise indeed. We chose to go with H.264 several years ago when it was still fresh and new. Fortuitously, it is now being used almost everywhere, from 3G to HD. We have been able to integrate well with Flash now since they use the same compression we do. All around, it was a great choice, which paid back dividends for us later,” Paltalk founder and Chief Technical Officer Perry Scherer told LinuxInsider.
“Our choice of Speex for audio was a good one as well. Macromedia will be using Speex in their next release of client and server for audio conferencing,” Scherer added.
Open source development has provided the mechanism and means for individual and small groups of developers to rally round, affiliate and organize initiatives aimed at developing and promoting a wide range and variety of audio and video codecs and formats.
Bringing Media Access to the Masses
In so doing, open source’s collective potential affords individual software developers and smaller software development companies a means of competing against much larger, and more conventional, commercial brethren.
Free, broad public availability, quality equal to and in cases surpassing industry standards and often a broader range of features, open source audio and video formats have proven to be more than viable alternatives to proprietary industry standard formats, Mina Berman and Shy Keidar, project administrator and site/project maintainer, respectively, of Musepack, an open source audio codec and format, told LinuxInsider via e-mail.
Employing a radically different technology development and distribution model, free and open source (FOSS) audio-visual codecs and compression formats, notably FFMpeg, have been incorporated into a wide range of free open source multimedia players, as well as other products and services, both free and commercial.
“FFMpeg supports decoding of virtually every common, uncommon and totally obscure video and audio codec/format in existence,” explained project leader and site maintainer Michael Niedermayer.
“FFMpeg is probably the heart of all, or nearly all, free multimedia players today, and it is also used by many ‘non-free’ players. Some examples of free players would be Mplayer, Xine, Vlc. The latest Ffdshow build alone seems to have 11 million downloads, so FFMpeg is likely used by quite a lot of people,” he said.
Musepack’s audio codec and file formats are also proving popular. “Musepack is all about providing results its users require and expect,” Berman and Keidar maintain. “The one who benefits most of all is the user. People who use Musepack are people who are concerned about sound quality, their rights as users of the format, features and performance.
“Because it doesn’t require any licensing fees, is under an unrestrictive BSD-style (Berkeley Software Distribution) license, and offers outstanding sound quality at resource usage several times lower than that of industry standard formats, developers of several specialized professional audio applications have chosen Musepack for compression of their samples. Industry standard formats stand no chance when it comes to licensing and viability of the formats in specialized sectors. The more popularity the format gains due to what the industry standard formats fail to offer, the more users benefit,” they noted.
Cornering the Market
Development and widespread distribution of open source audio and video codecs, as in other sectors of computing and telecommunications software, is also proving more effective at guarding against market “lock-in,” the adoption of essentially proprietary compression technology developed and controlled by an individual company or group of companies as an industry standard.
The processes used by international industry associations to set technology standards haven’t changed much in recent years, according to Niedermayer, and they remain susceptible to undue influence from industry heavyweights. “Some, like Microsoft, seem to be trying to push standard committees to accept their own, often not-well-designed, proprietary technologies as international standards, the obvious example being SMPTE (Society of Motion Picture and Television Engineers) VC-1 [Microsoft’s Windows Media 9 high-definition video codec].”
Independent of such efforts, developing and promoting a proprietary, de facto standard is a well-worn and tried path to long-term commercial success, he noted, Adobe’s Flash technology being a case in point, Niedermayer continued.
Even more commonly, companies “have taken a standard or draft from one of the standard bodies, changed it slightly and ‘sold’ it to their customers. Examples of that are Real Video 1.0 and 2.0, which are practically [the same as] ITU H.263, Sorrenson’s SVQ3, which is based on a draft of ITU H.264 and Microsoft’s MSMpeg-4 and early Windows Media Video, which are somewhat modified versions of ISO Mpeg-4,” he added.
Guarding Against Lock-In
“All companies are aware of the public’s demand of support for open source audio formats, but only a few of them care,” Musepack’s Berman and Keidar contend. “Companies like Cowon, who care a great deal about their users, implement support for formats like Musepack, Ogg Vorbis and Flac in their players, even though they know that the majority of users will play MP3s on them.
“They know that there is a market for open source formats which the ‘big players’ are too old-fashioned and blind to see, and they know that supporting open source formats benefits everyone, including them, because if open source formats were more popular and industry standard formats less popular, they would have to pay less for licenses. The only way to get progress is to give the user a choice. If more companies supported open source formats in their players, more people would be able to escape the industry standards that are constantly shoved down their throats,” they added.
FFMpeg’s Niedermayer sees attempts at market and industry lock-in persisting despite greater transparency in industry standards setting organizations and processes. “In the past, companies often invented their own formats and codecs or modified some existing public standard and exercised control by being the only ones that knew the details of how to play/decode them.
“Today, the details about the formats and codecs are often publicly known while control is exercised by software patents. All these lock-ins, whether they are by non-publicized details or software patents, hinder healthy competition and lead to more expensive and limited products. The advantages are purely on the side of the ‘biggest fish’ controlling the ‘intellectual property.’ The end user always loses from lock-ins,” he added.
“I think that in terms of industry standards, the cream always rises,” Paltalk’s Scherer commented. “H.264 was simply so much better than anything before it that it rocketed to the top. In general, though, the fact that it is not royalty-free may hold it back a slight bit. The folks who build Lame — the ‘free’ MP3 encoder/decoder — are still in dispute with Fraunhofer/Thomson regarding intellectual property, so MP3 audio may not be ‘free’ to produce or convert in any sense. This does hamper the wider adoption of a codec, of course, but it did not stop us from licensing H.264. Pure quality and trendiness made the decision for us.”