LEARNED LISTENING

Spatial Hearing by Jens Blauert. MIT Press, $40.00.

There are two kinds of hearing. One kind is the hearing we are born with, a remarkable process extremely sensitive to small differences in the spectral content and timing of sounds. The other kind of hearing is learned, and develops as an elaborate form of pattern recognition in which the ear and brain cooperate. Among other functions, such as identifying a telephone caller from only the first few spoken syllables, this acquired hearing allow us to estimate the apparent direction, size and distance of sources of sound. The richness of this auditory space, much of it based on cues at or below the edge of conscious awareness, is explored in great detail by a distinguished expert in Spatial Hearing. It is a technically thorough, clearly presented review of almost everything known about that subject.

Spatial Hearing should be of particular interest to audio professionals because objective, standardized audio test procedures have largely ignored the spatial qualities of sound systems.

Perhaps this is because these qualities may have been thought to be too subjective to be measured rigorously. As the book shows, this is no more true of spatial than of spectral measurement; it is a matter of knowing what to measure. But it has been difficult, if not impossible, to review the large, rapidly growing and multi-lingual literature on the subject, distributed through many journals, conference proceedings and unpublished dissertations, in order to arrive at an understanding of how such standards might be created. Even salesmen at audio dealers' shops are often able to discuss the relative merits of measuring a loudspeaker's spectral output anechoically or in a reverberant room, because these subjects have been dealt with and debated in popular technical publications for decades.

On the other hand, authoritative information about the spatial aspects of hearing is only now becoming accessible to engineers and others who do not read the academic literature, but need and want such information. It is not too soon, as these effects have already been the basis of stereophony, binaural reproduction, so-called time-delay or ambience systems, and many other applications in electroacoustics.

Spatial Hearing should not be thought of as a popular book, but it contains an enormous amount of information essential to anyone working or writing professionally in audio. In fact, the book has been an important and useful reference to many such readers for as long as 10 years, but only for those able to read it in the German original or in its Russian translation.

Now, at last, this book has come to us in English, with a special bonus: A new 80-page section, added for the first time in the English edition, that is almost a small book in itself. In the new section, Blauert reviews the extensive research in the field carried out since the original manuscript was completed in 1972, including a number of experiments and theories with considerable significance to audio professionals.

Fortunately for such readers, the author has a more than passing personal interest in both sound recording and sound reproduction.

Although Spatial Hearing is a technical book, it is not difficult to read, partly because of the excellent translation by John Allen, and partly because of the large number of clear, well-produced diagrams, charts and tables that supplement the text. Nevertheless, most owners of the book will probably use it primarily as a reference or to learn about a specific aspect of the subject, as I have used my German edition, rather than as a text to sit down and read through. On the other hand, it is the kind of book that owners will want to pick up and read from time to time, because it is chock-full of interesting information.

Spatial Hearing is divided into three main parts, and each of these is organized quite systematically. There is also the newly added fourth and final section that brings the first three parts up to date.

The first part is a brief review, perhaps too brief and compact for readers who know little about the subject, covering the concepts and experimental methods used in psychophysical research. It is an adequate introduction to the material, however, and in it Blauert explains the terminology and methods used in psychoacoustics research, especially research dealing with spatial perception. There are references to a number of other books and papers dealing with psychological research methods, some useful information about the mathematical and perceptual properties of test signals, and a rather detailed description of the in-the-ear probe microphone apparatus used by Blauert and his students, complete with a schematic for an equalizer to be used with the B & K microphone capsule.

Apart from this latter topic, presented in unusual detail, the section is quite general and can, in fact, be profitably studied as an introduction to the science of psychoacoustics. (Those who want to pursue the subject should obtain Introduction to Hearing, by Prof. David Green of Harvard.) Among other subjects, there is a good, clear discussion of methods used to establish, with good statistical confidence, when a subject in an experiment has, in fact, heard something. My personal wish is that certain reviewers of audio equipment, specifically those who have been able to hear greatly clarified "depth of imaging" after setting their turntables on granite slabs, should be forced to memorize this section, preferably while suspended from their shock mounts.

It is in this first section that the reader is introduced to the firm scientific viewpoint that informs the entire book: It is certainly true that the position of the auditory event and the position of the vibrating body that radiates the sound waves (the sound source) frequently coincide. Nonetheless, the conclusion that the position of the sound source is also the intrinsically correct position of the auditory event is, at the very least, problematic. The sound source and the auditory event are both sensory objects, after all. If their positions differ, it is an idle question to ask which is false.

The point of this, of course, is that sounds do not automatically appear in the "right" place, and to make them do so in a recording or broadcast requires more than straightforward placement of a microphone. Blauert continues: The telecommunications engineer, of course, is especially interested in just those cases in which the positions of the sound source and the auditory event do not coincide. The telecommunications engineer seeks to reproduce the auditory events that occur at the point where a recording or transmission originates, using the smallest possible number of sound sources (e.g. loudspeakers). Sound events must be generated at the receiving end of the electroacoustic telecommunications chain in such a way that auditory events occur in the same directions and at the same distances as at the point of origin. Auditory events must, therefore, also occur at other positions besides those of the loudspeakers.

What Blauert does not say here is that recordings, including classical music recordings, now more and more frequently involve the producer and engineer in the creation of a space for the sake of the music, since such an auditory space does not arise automatically by following cookbook rules of microphone placement. The satisfying sound of a good recording is substantially dependent on an understanding of spatial cues that is not common.

The second part of the book deals with spatial hearing in the presence of a single sound source and covers what is basically known about localization, the process by which a listener assigns a position in space to the source of sound. Charts are used here, as everywhere in the book, to summarize research on specific topics, allowing the reader to see at a glance the different results obtained by different investigators of the same effect. On page 39, for example, one can quickly scan the findings of 12 major studies of localization blur-the change in direction of a sound source that listeners are just able to detect reliably. The range of lowest values goes from 0.75° (Klemm, 1920, using impulses) to 3.2° (Haustein and Schirmer, 1970, using broadband noise); values for speech, incidentally, appear to fall within this range as well. Turning over the page, we can see a diagram on page 41 that shows us exactly how the subject was positioned in relation to the sound sources for the 1970 study, with the notation that as many as 900 subjects were tested. This is typical of the book's clarity and thoroughness. Diagram after diagram amplifies the explanation in the text, showing experimental layout, equipment used and its method of interconnection, and graphs that compare one investigator's results to another's.

This section also reviews the physiology and acoustics of the external ear, and it contains a complete and illuminating review of work done on the measurement of the sound field at the ears of a listener for both a plane wave and a diffuse sound source. Beautifully drawn charts show how the spectrum of a sound at the entrance to the auditory canal varies with the direction of its incidence. The graphs are fascinating. How do the ear and brain of a listener retain the impression of an unchanging source of sound when so much of the sound changes with a simple turn or tilt of the head? Another extremely interesting section deals with the illusion that sounds of some frequencies, when emitted by a source centered on the midline of a listener's head, are coming from locations above or behind the subject.

Blauert reports his own surprising first encounter with this effect: Sitting before a loudspeaker radiating a sinusoidal signal whose frequency continuously rose from 200 Hz to 16 kHz, I noticed that the auditory event moved back and forth several times on a path from front to rear over the top of my head.

For many of us, the experience might have meant that the time had come to replace the tweeter and let it go at that.

This observation, however, fascinated Blauert and led him to undertake a series of experiments. The results provided useful clues about the way in which the shape of the outer ear allows a listener to establish the elevation of a sound source. As with other aspects of spatial hearing, each of us learns to associate subtle spectral and timing cues with the direction of a source, these cues being based on the shape, size and detailed form of our own ears.

The reflections from the folds and curls of skin, and the "personal resonances" these produce, allow us to find sound sources remarkably well, especially if we train ourselves to listen carefully.

When we listen with someone else's ears, on the other hand, or to a binaural recording made with sculpted artificial ears, the cues that we have learned by listening all our lives through our own ears are changed or are missing entirely. The most common result is the confusing illusion, when listening to binaural recordings, that sounds we know to have originated in front of the head seem to be coming from behind us.

Blauert describes a series of clever experiments in which subjects listened to sounds through modified outer ears, instead of their own. In the simplest of these experiments, short rubber hoses were inserted in the subjects' ears so as to eliminate the effects of acoustical cues provided by the pinna (the outer part of the ear that protrudes from the head). The result was that the subjects lost much of their ability to localize sounds. Other experimenters later used short brass tubes, then tried short brass funnels, then tubes with artificial pinnae facing backward. All of the changes, interestingly, usually produced rearward localization regardless of the angle of incidence of the sound. This front-back problem is familiar to anyone who has ever tried to make a binaural recording. I have only covered a small part of Blauert's review of work on this interesting problem, which is covered extensively and very clearly; an entire book could be written on this subject alone.

Other pages in this second section deal with ways in which differences in the signals at the two ears of a listener produce information about the location of the source. This is the heart of the subject of spatial hearing, of course, and it is treated very comprehensively, not only as regards coverage of experimental work but also in reviewing various theories of how auditory data is processed in the brain. Again, there are clear schematic diagrams to help the reader grasp these theories, and ample references for those who intend to look more closely into some of the work discussed.

The third part of the book concerns the perceived spatial qualities of complex sounds produced by multiple sources and in enclosed spaces.

These, of course, are the sound fields that normally form the setting for music recording and reproduction. A series of interesting diagrams show where listeners hear sound coming from, under many experimental conditions, in each of which some specific kind of cue has been studied. 'Studies of a listener's ability to center sounds produced by a pair of loudspeakers, for example, are followed by the effects on perception of many kinds of binaural and stereophonic presentations of echoes and delayed repetitions of signals from various directions.

One section describes an experiment in which Blauert tried an idea of his own, in which he attempted to make a sound appear to whirl around the listener's head. The method involved systematically modulating the delay and amplitude of the signals coming from six speakers. It worked, but with a strange and unexpected twist to the whirl; see page 273 of the book for the surprise. Another section deals with a half-dozen methods of producing pseudo-stereo from a monophonic signal. Looking a; the neatly laid out page of diagrams, clearly showing the operation of each of the systems under consideration, I could imagine many readers uttering an involuntary, "There goes my idea!" Many readers of Spatial Hearing will have this experience, I am sure. I would say there is a good chance that most of my best ideas that have not yet found in any of Olson's books are likely to turn up in Blauert's, one by one, especially those concerned with binaural/stereo developments. On the other hand, the book is bound to stimulate a torrent of new ideas, because it is so full of descriptions of experiments and effects that set one thinking.

The final section of the third part, originally the end of the book, covers multi-channel sound, including quadraphony. Among other material relevant to audio, there is a description of the TRADIS system, developed by Damaske and Mellert during the 1960s. In this system, left- and right-channel signals are phase-inverted, attenuated, delayed by about a half-millisecond, and then mixed with the signals of the opposite channels for loudspeaker presentation. When the values are just right, and especially if the room is acoustically "dead," the leakage around the head of the right channel to the left ear (and vice versa) are cancelled and the stereo loudspeaker sounds spring to life in a panorama across the room. For a while, in the early 1970s, Deutsche Grammophon apparently considered releasing stereo/binaural records with this type of processing already encoded in the two signals-at least I heard a test recording that had been circulated to evaluate listeners' reactions at that time.

Someone has said, but I am sure it's wrong, that since holography had not yet been invented at the time, there was nothing to name the process after, and that is why it was dropped by DGG. In any case, the history of the technique can be traced through Blauert's references, beginning with the announcement of the idea by the late Ben Bauer of CBS in 1961. It is interesting, to say the least, to see where these ideas have come from, and to discover how much intensive research work it has taken to gain an understanding of the perceptual processes involved. Too much of it has been hidden in places difficult to reach.

The last part of Spatial Hearing, "Progress and Trends Since 1972," covers an exceptionally productive and exciting period in the history of psychoacoustics, partially because of the enormous influence of computers and digital methods in generating test signals, especially low-noise delayed signals, and in analysis of data. New measurements of listeners' localization acuity are reviewed here, as are recent studies of the perceptibility of delayed echoes and comb-filter effects. There is a section on spaciousness and its importance to concert hall design, and the cues that contribute to it. Another section reviews the most recent theories of localization and of new, updated versions of classical theories.

Among these is a description and diagram of a modified version of the "coincidence model," based on an idea of the widely known teacher and theoretician Lloyd Jeffress. The theory, in the form shown, applies biological "and" gates and timing circuits in an elegant way to produce a mechanism that is simple but surprisingly effective in doing what ears and brains do when they locate a sound. In fact, the explanation is so clear that I was able to sit down at my microcomputer and program a simulation of the coincidence model just by looking at the diagram. What a fascinating subject! The section of the book on theories of binaural processing in the brain concludes with examples of "correlograms," plots that show the interaural cross-correlation of signals as a function of frequency, which are a kind of "snapshot" of the localization process; later work has turned these into "movies." The final section includes examples from architectural acoustics and closes with a good treatment of binaural recording. Blauert proposes a binaural mixing console that would allow the conversion and mixing of conventionally miked signals to binaural material, using signal processing methods.

He accepts and describes the conclusions of others who believe that the most successful dummy head recordings are obtained when no attempt is made to duplicate the ear canal or the eardrum impedance, and microphone capsules are simply coupled to the entrances of the auditory canals so as to achieve the best possible signal-to-noise ratio. But, as Blauert concludes, the definitive solution to binaural reproduction remains to be found, perhaps in the form of a personal "equalization box" that would allow each listener to add the time and frequency effects of his own outer ears to a standardized recording. In the era of the "walk-person" cassette player, this topic forms an appropriate last word on audio reproduction as well as spatial hearing.

An additional feature of the book that cannot be praised too highly is the immense bibliography that goes on for 45 pages. Scanning its pages, one sees familiar names among the unfamiliar: Edgar Villchur's work on earphone calibration, Floyd Toole's papers on localization, Ben Bauer's investigations of stereo. Here one can find the original source of every subject discussed in the book. Browsing through the bibliography is like walking along the shelves of one of the best technical libraries in the world, with the advantage that everything in it has been reviewed, even if only briefly in some cases, in this remarkable and valuable book.

Spatial Hearing is not light reading, as I have said. It is a carefully assembled technical text aimed at students and specialists in the field. However, it is understandable and is not too difficult to read, even if some paragraphs or the few mathematical expressions are challenging. It is likely to become a fixture on the bookshelf of everyone concerned with the design of loudspeakers or headphone listening systems, or involved in sound recording, who wants to know and use what science has learned about human spatial perception. Imaging, like everything else about human sensory development, is an endlessly interesting process, and that is what Spatial Hearing makes clear.

-Bob Berkovitz

(adapted from Audio magazine, Jan. 1985)

= = = =