Anthony Parks – parksa@rpi.edu
Jonas Braasch – braasj@rpi.edu
Program in Architectural Acoustics
School of Architecture
Rensselaer Polytechnic Institute
110 8th Street
Troy, NY 12180 USA
Popular version of paper 1aAA12
Presented Monday morning, October 31, 2011
162nd ASA Meeting, San Diego, Calif.
Harry Haller, the protagonist of Herman Hesse's Steppenwolf, encounters a rather curious situation. His fictional lover, Hermine, introduces Harry to Pablo, a musician, who gives Harry a unique opportunity: to visit The Magic Theatre, where Harry's inner fantasies spring to life.
But what if Herman Hesse had been an acoustician? What form might The Magic Theatre have taken? What if I proposed to you that the modern concert hall is nothing more than a special kind of magic theater, one in which listeners can step inside an auditory fantasy?
As outlandish as it sounds, this might not be too far from the truth. The concert hall provides an environment that gives rise to a number of perceived auditory phenomena that are engendered by the room itself. For one, the architecture of the room, including, but not limited to, the spacing of the walls, the materials used, even the miniature details of the balconies, all provide surfaces where incident sound waves can reflect and scatter across the whole audience area.
One such perceptual phenomenon that is most readily apparent to even the most casual concertgoer is that of spaciousness. Acousticians generally define spaciousness as that sense of spatial impression that a hall provides to both performers and listeners that consists of two principle components: apparent broadness of the sound (known as apparent source width) and the degree to which one perceives themselves as being surrounded by the sound (known as listener envelopment).
In the context of classical music performed in a hall like Avery Fischer Hall at Lincoln Center in New York or Boston Symphony Hall, these two aspects, source width and envelopment, give the appearance that the sound, which generally comes from one or several performers on stage, comes from a source much larger than it appears visually and from a multiplicity of directions which the directly incident sound does not appear to come from.
One thing is certain; musicians and listeners alike prefer spaciousness in classical music. However, what it takes from an architectural standpoint to foster a sense of spaciousness in a hall
is a matter of hot scholarly debate among acousticians. Several competing theories exist, corroborated by carefully controlled listening tests. In one theory, Bradley and Souloudre (1995) believe that the proportion of energy arriving at both sides of the listener's head (that is to say, laterally) has a great influence on the apparent width of the sound source on stage, while Morimoto and Okubo's experiments have shown that the proportion of sound energy arriving from behind the listener may foster a sense of envelopment.
But, until very recently, almost all of the listening tests that have been conducted in an effort to correlate concert hall design parameters with perceived spaciousness relied on one totally unrealistic condition: that the listener participant in the test keep her head absolutely still.
From an experimental standpoint, this makes a certain amount of sense. Why? Because if listeners keep their heads fixed, the inter-aural time and level differences (ITDs and ILDs) are kept constant, and this creates a controlled variable for the experiment. But what happens if we let listeners move their heads freely? Does this in any way alter our perception of the spaciousness of a concert hall?
To test this hypothesis, myself and Dr. Jonas Braasch designed a listening test where we could compare differences in perceived spaciousness while listeners moved their heads, and while they kept
their heads fixed. To do this, we created fifteen different virtual concert halls using Virtual Microphone Control (realtime spatialization software) that each contained varying amounts of coherence and front-to-back energy, parameters which we knew from several other researchers' previous work which had shown these parameters to correlate strongly with subjective spaciousness. The virtual concert halls were played back over an eight-channel surround loudspeaker setup in our Communication Acoustics and Aural Architecture Research Lab at Rensselaer Polytechnic Institute.
We asked fourteen listeners to assign ratings between 1 (least enveloping, narrow) and 7 (most enveloping, very broad) to each of the concert halls under two conditions. In the first condition, the listeners rated each virtual concert hall while they were asked to keep their heads were fixed, and in the second, listeners rated the halls while they were instructed to move their heads freely. Listeners, throughout the duration of every test, wore a special headband that tracked the positions and angles of rotation of their heads.
The results of the experiment were, on first glance, unsurprising. For one, listeners do indeed move their heads (when they're told they can). Furthermore, when listeners are asked to keep their heads fixed, they still move their heads (a little bit)! This if anything perhaps (partially) questions the results of previous listening tests where listeners were instructed to keep their heads fixed, because the assumption that ITDs and ILDs are kept constant simply cannot be true.
More interestingly, however, it was found that listeners, when allowed to move their heads, assigned higher ratings to the apparent source width of the virtual halls (but not envelopment) with statistical significance as compared to when they were asked to keep their heads still. This possibly implies that, since listeners tend to perform better on sound source localization tasks when they move their heads (see Perrett and Noble 1997), that the phenomenon of apparent source width may be a matter of more precise localization, while listener envelopment, on the other hand, may be a phenomenon of de-localization (or 'becoming lost' in the sound field).
This research attempts, in part, to foster within the scientific community a more dynamic view of how we evaluate our sonic environment. By examining the multiplicity of cues that we use in human audition, we can create more perceptually accurate models of listening to better understand the mysteries that lay between the sound around us and the receiver within.
Bibliography