Foreword

It is a pleasure and great honor for me to contribute a foreword to this fascinating book on the state-of-the-art in stereo and multi-channel audio processing. Given my own research interest in binaural hearing, it is exciting to follow the detailed description of how scientific insights into human spatial perception in combination with digital signal processing techniques have enabled a major step forward in audio coding. I also feel honored to have a close relationship with the two authors. Both are young scientists with already an impressive output of scientific publications and also patents, and they both have made significant contributions to international standards. The book that they present here documents their deep insights in auditory perception and their ability to translate these into real-time algorithms for the digital representation of multi-channel sounds. I was lucky to follow many of the described developments from close by.

A remarkable aspect of the authors' careers is that both are or have been related to research environments with a long history in using perception insights to steer technological developments. Christof Faller was for many years affiliated with the Bell Laboratories of Lucent Technologies and later with Agere, a spin-off of Bell Laboratories and Lucent. The history of psychoacoustics at Bell Labs started around 1914, when Harvey Fletcher initiated a research program on speech and hearing with the clear goal to improve the design of telephone systems. Since then, many important contributions to audio and speech signal processing applications have come out of this laboratory. Jeroen Breebaart got his academic training at the former Institute for Perception Research (IPO), for many years a joint research institution of Philips Research and the Technische Universiteit Eindhoven. The transfer of his psychoacoustic modeling knowledge into algorithmic applications started after he had joined the Philips Research Laboratories in Eindhoven. Hearing research has been a topic at Philips since the 1930s when the technical possibility of stereophonic sound reproduction required a deeper understanding of spatial hearing. Among the many studies done in that period by, among others, K. de Boer were early applications of a dummy head to study and improve interpersonal communication and support hard-of-hearing subjects.

The book nicely demonstrates how close coding applications have come to present-day psychoacoustic research. When perceptual audio coders were first realized in the 1980s, the bit-rate reduction mainly exploited the concept of spectral masking. This concept was included in the encoder by so-called spectral spreading functions, an approximation which had been known in basic research for at least 30 years. Although using only a very crude description of perceptual processes, these early coders became essential in enabling the internet to be used for music distribution. Up to very recently, audio coding did not take much advantage of the redundancy between channels. The problem here is that, despite the perceptual similarity between different channels in a recording, the similarity in terms of interchannel correlation is often very low. Using a signal-based analysis thus does not give much room for redundancy removal. In order to capture and remove this redundancy, a time–frequency analysis and parametrization of the perceptually relevant spatial parameters is needed. The insights into these perceptual relations is of very recent origin, and the authors were able to apply them so quickly in audio processing because they had, in part, been involved in generating this perceptual knowledge.

Thus, the book is a remarkable document of what can be achieved by combining such complementary knowledge areas as psychoacoustics and digital signal processing. Given the economic and societal impact of audio compression, I hope that this example will help to attract future students to delve into this, certainly not simple, but highly rewarding research domain. Reading this book will certainly help readers to come to such a career choice.

Armin Kohlrausch

Research Fellow,
Philips Research, Eindhoven
Professor for Auditory and Multisensory Perception,
TU Eindhoven, the Netherlands
Eindhoven, May 2007