Keywords:Future tele-collaboration systems need to support seamless spatial sound reproduction to achieve realistic immersion. In this chapter, we present techniques based on binaural signal processing, capable of encoding and rendering sound sources accurately in three-dimensional space using only two recording/playback channels, thus creating the illusion of a virtual source in space. The technique exploits the mechanisms that help to decode spatial information in the auditory system. An overview of the encoding and decoding mechanisms is given and their practical application to the design of virtual spatial sound (VSS) systems is discussed.Binaural Hearing, Binaural Technology, Virtual Sound, Spatial Sound, HRTF, HRIR, Crosstalk Cancellation
INTRODUCTIONIn this chapter, we are concerned with the rendering of sound sources in threedimensional space using only two playback channels. We focus our study on a family of techniques based on the synthesis of binaural signals. Given that the human hearing mechanism is binaural, all spatial information available to the listener is encoded within two acoustic signals (i.e. one to each ear). Thus in principle, only two playback channels are necessary and sufficient to realistically render sound sources localized at any point in space. Since the sound is not perceived as coming from any given loudspeaker the technique creates the illusion of a virtual sound source in free space.There are many ways to create the impression that a sound source is located at a given point in space away from the loudspeakers. One familiar example is the playback over two speakers commonly known as stereo reproduction, where a sound source can be positioned within the angle subtended by the loudspeakers by simply manipulating the gains assigned to the source in each channel (i.e.
346Audio Signal Processing amplitude panning). If the desired rendering position of a source is outside of this region, then the angle between speakers needs to be increased or more rendering channels (and corresponding gain reassignments) have to be added. The success of amplitude panning depends highly on the location of the listener with respect to the loudspeakers. The location where amplitude panning (or any other playback technique) is most effective is commonly known as the sweet spot. Another technique, which renders virtual spatial sound accurately and has a large sweet spot, but which requires multiple playback channels is wavefield synthesis. This technique is described in Chapter 12.The advantage of binaural systems over other techniques is that only two playback channels are required, and a virtual sound source can be positioned at any desired point in space with great accuracy. This advantage does not come without difficulties. The binaural signal has to be delivered to the listener's ears with great fidelity to be realistic. Generally headphones are used to render binaural signals, but in some applications this is not practical. As discussed later, techniques using standard stereo loudspeaker se...