Scene Enhancement via Acoustical Environment Modification by Tom Guyette & Roy Rubinstein Scene Enhancement via Acoustical Environment Modification The theatrical experience is at its best when it envelops the audience in a completely fictional world: when lighting, sound effects, sets, and costumes are all combined seamlessly to create a completely artificial environment into which the audience allows itself to be placed. Oddly, one method of bringing the audience to another place has not been explored very thoroughly: acoustical environment modification. Human beings have a very keen sense of hearing. Everyone knows that we can tell whether a sound is in front of us, behind us, above or below, to the left or right. We are also very keyed into the general sound of rooms and locations. If a person were brought into a sewer pipe, even deprived of his vision (and his sense of smell), he would be able to identify the sewer simply with his sense of hearing. A sewer is a rather exaggerated example, but the underlying principle remains constant in all environments: people can tell where they are, simply by the sound of their voice in their environment. Or by the sound of someone else's voice. Or the sound of their own footsteps. It seems logical, then, that modifying the acoustical environment that is presented in a play's scene will enhance the illusion of being in another place. The problem is that there are many factors that contribute to a theater sounding the way it does. Most of these factors are impossible to change: wall material, room shape, room size, audience size, and the like, remain constant for practical reasons: although technical directors are wizards at moving scenery, rebuilding the hall several times during a play is most likely beyond even their sometimes-magical means. Thankfully, with recent advances in digital signal processing (DSP) technology, it is now possible to modify an acoustical environment without modifying the physical construction of the environment. Currently available DSP-based acoustical correction systems are relatively expensive and very complicated. The environment presented for a play need not sound completely natural; the audience is most likely not going to be listening as critically to the acoustics of the space as the audience at a concert. If the modifications are subtle, a relatively simple system can be very effective. This demonstration of our project makes use of a single microphone, which feeds a signal processor, which in turn feeds a pair of speakers. Early in the project, we made recordings of footsteps in various environments, and intended to analyze the frequency content of the signals over time. This would have allowed us to write algorithms for the signal processing device to process the sound of footsteps on a platform, and to make those footsteps sound exactly like footsteps on a number of surfaces in a number of environments. However, the software we intended to use for this analysis was very limited, and after about six hours trying to force it to do what we wanted, we aborted this attempt in favor of adjusting the processor by ear. The DSP box boasted a large number of algorithms, in a large number of possible configurations. After some experimentation, we decided on the configuration shown in Fig. 1: 旼컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴커 � � � 旼컴컴컴컴커 � � 旼컴컴컴컴커 旼큁late Rev.쳐� � � In캑 Comp. 쳐캑 읕컴컴컴컴켸 쳐컴횽ut � � 읕컴컴컴컴켸 � 旼컴컴컴컴커 � � � 읕퀿on-L Rev.쳐� � � 읕컴컴컴컴켸 � � � 쿑igure 1: The Configuration of our DSP system � 읕컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴켸 a compressor, which keeps loud signals at a reasonable level, and slightly raises the level of softer signals; a non-linear reverberation algorithm for simulating the first few reflections off of the walls of the simulated room; and a plate reverberation algorithm for simulating the 'reverberant mush' that exists approximately 80 milliseconds after a sound is made. We configured the two reverberation algorithms to simulate three distinct environments: the antechamber, which gives the impression of a small [15'x15'x10'] stone room, connected via a standard-sized doorway to another stone chamber; the Palace in London, which gives the impression of a large [75'x75'x35'] stone room; and the Palace in France, which gives the impression of a very long [20'x90'] stone hallway with a fairly low [15'] ceiling. 'Fuller' versions of this system could benefit from a larger number of speakers (4 to 8), and more precisely measured algorithms. However, for demonstration purposes, the effect produced by this simple system is dramatic. The effect produced by a more complex system would be nothing short of astonishing.