The MP3, its origins, and its impact

Johnny Ryan

I've been thinking on the theme of disruption to the music industry as an example of the Internet's impact on media in my forthcoming book.

First, lets get a sense of the scale of the disruption. In 1999, the top 10 albums in the United States generated sales worth over 54.6 $M. By the end of 2008, that figure had fallen to just under $18.7 $M. Yet, 2008 marked a peak of four consecutive years of growth in overall sales of music. What is happening? In 2008, music purchases including both individual tracks and albums of popular and niche musicians totaled 1.5 $ Billion, up from 1.4 $B in 2007, 1.2 $B in 2006, and 1$ in 2005. While sales of the top 10 albums, which of necessity cater to the most widely applicable common audience, have plummeted, the new distribution environment has benefitted consumers who prefer to select specific tracks, rather than entire pre packaged albums. Chris Andersen realized in 2004 that unlimited choice will expand demand rather than contract it. The Internet seems to benefit musicians who do not naturally reside in the top 10 albums category.

The record industry, which relies on the medium of recording as much as it does on musicians, has undergone successive upsets since the nineteenth century phonographs. Edison’s first recording was ‘Mary had a little lamb’ in 1877 on a tinfoil cylinder phonograph. Thereafter, the recording industry has undergone successive periods of turmoil arising from technical transition, first from cylinder to disk, and then from acoustic to electric recording, then to magnetic tape, to the cassette, the 5” compact disk in 1982, the standardization of MP3 (MPEG layer 3) in 1992, and the release of the iPod in 2001. In 2008, sales of individual music tracks sold online surpassed 1 $Billion for the first time, having grown 27% over 2007 sales.

In 1996, the song ‘Until it sleeps’ by Metallica became the first track to be illegally copied from CD, encoded as an MP3, and made available on the Internet by a user operating under the nickname ‘NetFrack’. NetFrack announced the revolution in music piracy to Affinity disk magazine, a diskette based news sheet circulated in the computer underground:

I’ve thought of the idea of somehow pirating, music. … The problem in the past … was HD [hard disk] space. … We eliminated the size constraints. We use a new format to compress our music. The MP3 format.

The development of the MP3 marked a pivotal moment in the history of the recording industry, ruining distribution and business models, and challenging the future of the major labels. This revolution, as with much else in the digital upheaval, started in the early 1970s.

A researcher in Germany named Dieter Seitzer began researching the problem of transmitting high quality speech in digital form using standard phone lines. Sound waves generated by the human voice were generally relayed along telephone cable in anolog form. One could digitize them using pulse-code modulation (PCM), a system that recorded a digital copy of an analog signal by taking samples of it at even distances. This results in a digital version that sounds almost exactly the same as the analog version, and which can be reproduced indefinitely and perfectly. Yet, digitization required the transfer of far more data – so much, in fact, that standard telephone lines could not support it, and Seitzer’s initial patent application to do so was apparently rejected as infeasible. As Seitzer and his students at Erlangen-Nuremberg University worked on the problem a new generation of telecommunications infrastructure solved the problem for him. At the end of the 1970s, ISDN lines and fiber optic cabling began to provide far more bandwidth, and meant that digitized speech could be relayed. Seitzer turned his attentions to the transmission of high quality music, which, despite the higher capacity of the infrastructure, remained beyond its capacity to transmit.

One of Seitzer’s students, Karlheinz Brandenburg, approached the problem using psychoacoustics. Brandenburg realized that the parts of sound imperceptible by the human ear could usefully be excluded from recorded music, thereby reducing the amount of data required to reproduce a record. In 1987, Eureka, an intergovernmental European research fund, supported further development of digital sound encoding at Erlangen-Nuremberg University, now partnered and lead by the Fraunhofer Institute for Integrated Circuits. In 1989 Brandenburg submitted his PhD thesis on ‘optimum coding in the frequency (OCF) domain’, which contributed to an encoding system that could allow the transmission of music in real-time at a reasonable quality over telephone lines.

The ear generally picks up sounds between 20 Hz and 20,000 Hz, though young adults may have an upper limit of 17,000 Hz and one’s useful lower limit might actually be around 60Hz. In addition, the human ear is incapable of deciphering sounds when masked by another sound occurring at the same time, or, depending on the sound, by a sound that makes a sound that follows it inaudible. Sounds beyond the threshold of audibility were excluded. ‘Simultaneous’ and ‘temporal masking’ can also be excluded from the digital recording. Since modern microphones are capable of picking up a range of frequencies far beyond the human ear’s threshold of audibility, the result when the surplus sound is removed is a far smaller set of digital data that represents a very high quality record of the original music. As NetFrack put it in describing the virtues of MP3 to his peers in the computer underground in 1996, ‘if you rip a song from CD … then you would end up with a 50 MB WAV’ file. To put this in context, the average modem on sale in 1996 was capable of making a connection that could transfer 33.6 kilobytes of data per second (Kbps). Downloading a single track on even a perfect connection could take between three and four hours. Yet from 1996 on, ‘that same 50 MB WAV would be 3 to 4MB in MP3’.

However, working on the initial concept in 1987, the researchers were using computers so slow in relative terms to those that would appear a decade later that sound compression took many hours on high performance machines. The team could initially only test the effect of this compression method on short sound samples, until May 1987, when they developed a system that could handle real-time implementation of an experimental music compressor and decompress or (“codec”). In 1991, the Moving Picture Experts Group (MPEG) of the ISO considered a number of proposals for standards to adopt for the compressed encoding of digital audio. AT&T, Thomson and Hannover University refined the OCF system developed by the Fraunhofer Institute, which the MPEG adopted within a family of three codecs. It was adopted as ISO MPEG Layer 3 in December 1991. The codec made its debut as a high-end format for TV and radio stations. Fraunhofer built and supplied powerful studio equipment using the codec to allow the stations to transfer audio material by ISDN for use in broadcasts. By the mid nineties personal computers had become sufficiently powerful to perform audio decoding and allow their users to listen to music from the computer’s speakers. In July 1995, Fraunhofer decided on a file extension for its MPEG Layer 3 codec. Following an internal poll Fraunhofer settled upon “.MP3”.

To demonstrate MP3 to personal computer users who now had machines adequately powerful to listen to them, Fraunhofer released WinPlay3, a demonstration piece of software, which was promptly cracked and illegally released for free by a student in Australia. Within two years, the combination of efficient computers, free MP3 codec software, and increasingly faster Internet speeds brought the MP3 format into the public consciousness. In May 1997, USA Today announced ‘Sound advances open doors to bootleggers’. In August the previous year, NetFrack had founded the first crew of MP3 pirates. In March 1998, Saehan Information Systems of Korea released the MPMan F10, the world’s first portable MP3 player. It had a capacity of 32MB, sufficient for eight or nine tracks. On 23 October 2001, Apple Computer released the first iPod, an MP3 player that ‘1,000 songs in your pocket’. In September 1998, Diamond Multimedia released a 32MB MP3 device called The Rio, and sweetened the deal by teaming up with new website MP3.com to offer music. However, the Recording Industry Association of America (RIAA) immediately sued Diamond Multimedia on the grounds that the device violated the US Home Recordings Act, about which, more anon…

As an Independent forum, the Institute does not express any opinions of its own. The views expressed in the article are the sole responsibility of the author.