Preface

The physical and practical limitations that have to be considered in the field of audio engineering include limited directionality of microphones, limited number of audio channels, the need for backwards compatibility, storage and transmission channel constraints, loudspeaker positioning, and cost, which may dictate other restrictions.

There is a long history of interaction between the fields of audio engineering and psychoacoustics. In many cases, the reason for this interaction is to achieve a certain goal given the imposed limitations. For example, in the 1970s the first widely used cinema surround systems applied psychoacoustic knowledge to improve their perceptual performance. Specifically, due to technical limitations the method of representing multi-channel audio often caused a lot of crosstalk between various loudspeaker signals, resulting in a risk that the front dialogue was occasionally perceived from behind. To prevent this, delays were applied motivated by the psychoacoustic ‘law of the first wavefront’ (sound is often perceived from the direction from which the ‘first wavefront’ arrives and delayed reflections arriving from different directions are not perceived explicitly). Another example is that the spectral content of the center dialogue loudspeaker in cinemas is modified such that the dialogue is perceived from above the center loudspeaker, i.e. from the center of the screen as is desired.

A more recent example of how psychoacoustic knowledge benefits audio engineering is perceptual audio coding. Invented in the late 1980s at Bell Laboratories, perceptual audio coders reduce the precision of audio waveforms to the minimum such that the error is just not perceived. Due to the psychoacoustic phenomenon of (monaural) masking, i.e. that one sound can render other sounds inaudible, the precision of an audio signal can be reduced in a signal-adaptive manner with hardly any audible impairment. The most prominent example of such a perceptual audio coder is MP3.

Spatial audio coding and processing, the focus of this book, comprises processing of audio signals considering spatial aspects of sound in relation with the abilities and limitations of the human hearing system. As explained in this book, the amount of information required to represent stereo or multi-channel audio can be significantly reduced by considering how humans perceive the spatial aspect of sound. More generally, this book also gives many examples of audio signal processing, considering spatial hearing, for achieving desired results, such as binaural audio processing and two to N-channel audio upmix.

The authors would gratefully like to acknowledge the help, support, valuable insights, comments, suggestions and observations by the following people (in alphabetical order): Frank Baumgarte, Frans de Bont, Bert den Brinker, Thomas Eisele, Jüurgen Herre, Gerard Hotho, Armin Kohlrausch, Jeroen Koppens, Peter Kroon, Juha Merimaa, Francois Myburg, Fabian Nater, Werner Oomen, Mykola Ostrovskyy, Erik Schuijers, Michel van Loon, Leon van de Kerkhof, Steven van de Par, and Martin Vetterli. Furthermore, the authors would like to thank their colleagues from Agere Systems, Coding Technologies, Fraunhofer IIS and Philips for their support in developing and exploiting various audio coders based on the technology explained in this book.

Jeroen Breebaart
Christof Faller

May 2007