Development of the mp3 audio coding technique

Eduard Rhein, the former editor-in-chief of the ”Hörzu” magazine, was fascinated by technology throughout his life. Promoting ”audiovisual techniques for the benefit of viewers” was particularly close to his heart. Therefore it is probably appropriate to his line of thinking that the Eduard Rhein Foundation has frequently paid tribute to coding techniques in the past.Techniques for recording music and movies on CDs, DVDs and Blue Ray Discs were acknowledged with an award last year. The coding algorithm H264/VAC for animated pictures, used today for the transmission of videos on all smartphones, was honored four years earlier. Despite the fact that these techniques have revolutionized the whole media scene, only experts are familiar with the names of the underlying algorithms.

The winner of this year’s technology award is different: the coding technique mp3 is commonly known even to the public. This audio coding technique turned the music industry upside down in the 90’s and is today an integral part of our everyday life. The first ideas for mp3 emerged already in the 70’s. Professor Dieter Seitzer of Friedrich-Alexander-University in Erlangen, later also founding director of the Fraunhofer Institute for Integrated Circuits IIS in Erlangen, wanted to transmit music through telephone lines. The patent examiner declared this as impossible. In 1979 the young PhD student Karlheinz Brandenburg started in his group and got the assignment to investigate what was feasible. He got the hint to look deeper into the use of psychoacoustic principles for audio coding, which was a vital catalyst for the later mp3-standard. Towards the end of the 80’s the group experienced considerable growth: Karlheinz Brandenburg, Ernst Eberlein, Heinz Gerhäuser, Bernhard Grill, Jürgen Herre and Harald Popp, along with other employees of the Fraunhofer IIS and the University contributed significantly to the development of the mp3-standard.

While the development work in Erlangen was in full swing, ISO founded the working group ”Motion Picture Expert Group (MPEG)” in 1988. The experts were allotted with the task to develop a standard for digital coding of audio and video. As usual for standardization processes, several competing groups participated in the work – there were essentially four concepts. The consortium around Fraunhofer-Gesellschaft, AT&T and the German Thomson-Brandt succeeded in inserting the significant parts for one of the layers of the draft for the future standard: MPEG-Audio and thus also mp3 – MPEG-1 Layer III to be exact – was declared an ISO standard in 1992. Most of the research work was about the finishing touch, in order to make the last difficult pieces of music sound perfect. A tough nut to crack was the a-cappella recording of ”Tom’s Diner” by Suzanne Vega. This is why this piece of music is mentioned in connection with the development of mp3 even today.

The ultimate success of mp3 is linked closely to the introduction of the internet, rapidly gaining importance at the beginning of the 90’s. The limited transmission bandwidth, especially during the first few years, made efficient coding of audio data imperative. mp3 saves 90 percent of data, almost without audible loss. Today, this standard has become so important that practically no playback device can do without it. mp3-coded music titles can be found on PCs, laptops, tablets, smartphones, music players and, above all, in the internet.

Professor Karlheinz Brandenburg worked out the essential foundations for the audio coding while working as a PhD student at the University of Erlangen-Nürnberg. His ”OCF” (Optimum Coding in the Frequency Domain) can be regarded as the direct predecessor of mp3. From 1982 to 1989 he wrote his doctoral thesis at the University of Erlangen-Nürnberg. From the year 1987, while working at the university, he supported the team at Fraunhofer IIS, led by Professor Gerhäuser, in its work for further development of the audio coding techniques. As of 1993 he was responsible for continuation and marketing of the research activities in his role as department head at the Fraunhofer IIS. Since 2000 he has taught and conducts research in Thüringen as a professor at the Technical University of Ilmenau and director of the Fraunhofer Institute for Digital Media Technology IDMT.

The essential contributions of Dr. Bernhard Grill are related to the transformation of the basic algorithms into a practically applicable technique. These include for example the algorithms for handling difficult special challenges in the audio coding and multiple acceleration of the signal processing algorithms for real-time capability, as well as sonic adjustment of the complete system. He was also responsible for the integration of the first mp3-encoder in the world, for which he wrote most of the software. Today, Bernhard Grill heads the Audio & Multimedia department at the Fraunhofer IIS as vice director of the institute.

Professor Jürgen Herre contributed significantly to the development of the mp3 technique as a member of the core development team. Firstly, he was responsible for a part of the real time implementation of the mp3 codec for signal processors. Moreover, he developed the so-called ”common stereo coding” for the mp3-coder, enabling audio fidelity and efficient processing of stereo music – in addition to mono. This is achieved by simultaneous data processing of both stereo channels, taking into account the involved psychoacoustic phenomena. Today he is involved in teaching and research at the International Audio Laboratories in Erlangen, an institution commonly owned by Fraunhofer IIS and the University of Erlangen-Nürnberg.

mp3 is without doubt one of the most significant German innovations of the past few years. After the introduction of the standard, the Fraunhofer Gesellschaft very quickly started the distribution of the software through the internet. This worked well until an Australian student cracked the code using a stolen credit card and put it on the internet for free download. In retrospect, this was not even detrimental since it actually accelerated the adoption of mp3 for audio coding. In the following years Fraunhofer IIS concentrated its efforts on development of further audio coding techniques and related technologies, as well as exploitation of the standard-relevant patents. The generated income now flows back to research via a foundation. This completes the circle: approximately 200 scientists are working on novel audio technologies at Fraunhofer today.

Prof. Dr. Christoph Kutter
Fraunhofer EMFT and
Bundeswehr University, Munich