Physical Reverberation Research Articles

A dissociation between speech understanding and perceived reverberation

As direct-to-reverberant energy ratio (DRR) decreases or decay time increases, speech intelligibility tends to decrease for both normal-hearing and hearing-impaired listeners. Given this relationship, it is easy to assume that perceived reverberation (reverberance) would act as an intermediary—as physical reverberation increases, so does reverberance, and speech intelligibility decreases as a result. This assumption has not been tested explicitly. Two experiments were conducted to test this hypothesis. In Experiment 1, listeners performed a magnitude estimation task, reporting reverberance for speech stimuli that were convolved with impulse responses whose reverberant properties were manipulated. Listeners reported a decrease in reverberance when the DRR was increased at both ears (Natural Room condition), but not when it was increased at only the ear nearest the source (Hybrid condition). In Experiment 2, listeners performed a speech intelligibility task wherein noise-masked speech was convolved with a subset of the impulse responses from Experiment 1. As predicted by the speech transmission index (STI), speech intelligibility was good in cases where at least one ear received non-reverberant speech, including the Hybrid listening condition in Experiment 1. Thus, the Hybrid listening condition resulted simultaneously in high reverberance (Exp. 1) and high speech intelligibility (Exp. 2), demonstrating that reverberance and speech intelligibility can be dissociated.

Binaural perceptual weighting of reverberation level in normal hearing listeners

In general, perceived reverberation strength is related to physical reverberation level present at the two ears; however, this relation depends on listening condition. Pervious work using virtual auditory space techniques has demonstrated that when physical reverberation is reduced in one ear while leaving the other ear unchanged, listeners do not report a change in reverberation strength. Reducing physical reverberation equally in both ears under binaural listening or in the signal ear under monaural listening elicits a decrease in perceived reverberation strength. To better understand the relation between physical and perceived reverberation in different listening conditions, a perceptual weighting experiment was performed. Listeners reported perceived reverberation strength in a test stimulus using a magnitude estimate paradigm relative to a standard while listening over headphones. Test stimuli were generated by jittering reverberation level independently in the two channels of a binaural room impulse response (BRIR) before convolution with a speech token. BRIRs with reverberation levels of -12, -6, and 0 dB at each ear relative to the standard were tested. Results of a perceptual weighting analysis suggest that listeners use different weighting strategies in different listening conditions, and are generally consistent with a “better ear” hypothesis for reverberation listening.

Perceived amount of reverberation consistent with binaural summation model

Although perceived amount of reverberation (PAR) is known to be dependent on the physical amount of reverberation present at the two ears, the extent to which this information may be combined across the ears is not well understood. Previous work using virtual auditory space techniques has demonstrated that when physical reverberation is reduced in the ear nearest a sound source while the contralateral ear is left unchanged, listeners do not report a change in PAR. Reducing physical reverberation equally in both ears, however, elicits a decrease in PAR. To better understand this phenomenon, the present study examines how PAR is affected by three additional listening conditions: scaling the contralateral ear while leaving ipsilateral fixed, scaling the contralateral ear only with no ipsilateral signal (monaural), and scaling the ipsilateral ear only with no contralateral signal (monaural). This study also examines how PAR is affected by an increase in physical reverberation present at the ears in the two li...

Stages with high ceilings, pipe organs, and active acoustics

Two case studies are presented of rooms that have a high ceiling over the stage to accommodate a pipe organ. Svetlanov hall is the principal venue of Moscow's International Performing Arts Center. It has two to four rows of chorus seating on two levels at stage left and right. The upstage wall is occupied by the largest pipe organ in Russia. It has a physical reverberation time of 1.7 s, which is within the accepted range for symphonic music, but longer reverberation times would be preferred for some pipe organ repertoire. Christopher Cohan Performing Arts Center at California Polytechnic State University has a proscenium and thrust stage. The Forbes Pipe Organ is housed on the stage right wall of the thrust stage. The symphony usually performs on the thrust stage entirely in front of the proscenium with the solid decorative fire curtain down. Both rooms had poor communication on stage between performers, which led them to implement a solution using active acoustics. In both rooms all of the active acoustic elements overhead of the stage are on motors and can be retracted when not in use.

ARTIFICIAL AMBIANCE PROCESSING SYSTEM

An apparatus and method simulates more accurately the natural statistics of a physical reverberation process. A new filter design is provided having a comb shaped group delay. Gain minimums at a plurality of frequencies are combined with a delay line to create a constant reverberation time independent of frequency while allowing for temporal spreading. In addition, the connection topology between the plurality of energy transmission networks is temporally randomized to facilitate energy distribution within the reverberation apparatus. Both the temporal and spectral responses are actively changed on each iteration of the energy recirculation. By making the response have a high echo density and a lack of spectral coloration in the decay, the illusion of a natural process is enhanced.

Electro‐acoustic architecture: Is it Green?

Electro‐acoustic architecture systems offer a means of changing the acoustic properties of a room electronically. They are an alternative to physically variable acoustics, where acoustic properties are changed by retractable curtains and doors opening to reverberant chambers, etc. This paper will address the question of whether electro‐acoustic systems are a Green alternative to physically variable acoustics. In order to have a reverberation time appropriate for symphonic music, a room must have a large cubic volume and hard heavy surfaces. In contrast, in a room with an electro‐acoustic system, the physical reverberation time is desired to be low even at low frequencies. Therefore, the cubic volume can be low and surfaces can be made of a lightweight material over an airspace. This affects the amount of energy needed to deliver materials, construct the building, and operate HVAC and lighting systems. A lower volume building means less pollution and waste during construction, and allows for more open space around the building, and/or higher development density. Electro‐acoustic systems can easily be designed as a renovation, thus recycling existing buildings. Case studies will be reviewed demonstrating these issues.

Prediction and measurement of reverberation increase from electroacoustic architecture systems.

Electroacoustic architecture (EA) systems increase the reverberation time (RT) of rooms by regenerating the physical reverberation of the room, and/or adding electronic reverberation to the direct sound. In regenerative systems, the amount of regeneration is controlled by the loop gain of the system. Increasing loop gain is analogous to reducing the effective absorption in the room. The RT of the electronic reverberator is analogous to the decay time of a physically coupled room, and, therefore, analogous to adding cubic volume to the room. As the electronic RT is increased, there is a point above which the resulting reverberation will begin to be double sloped. This point depends on the gain. The higher the gain, the higher the electronic RT before the resulting reverberation double slopes. In previous work, the gain before feedback of EA systems has been predicted in a statistical approach based on the bulk properties of the room such as RT and cubic volume. This, in turn, can be used to predict the maximum possible power gain and RT increase before double sloping. In this paper, we measure systems in use in the field and compare the achieved performance with predictions.

Active field control (AFC) -electro-acoustic enhancement system using acoustical feedback control

AFC is an electro-acoustic enhancement system using FIR filters to optimize auditory impressions, such as liveness, loudness, and spaciousness. This system has been under development at Yamaha Corporation for more than 15 years and has been installed in approximately 50 venues in Japan to date. AFC utilizes feedback control techniques for recreation of reverberation from the physical reverberation of the room. In order to prevent coloration problems caused by a closed loop condition, two types of time-varying control techniques are implemented in the AFC system to ensure smooth loop gain and a sufficient margin in frequency characteristics to prevent instability. Those are: (a) EMR (electric microphone rotator) -smoothing frequency responses between microphones and speakers by changing the combinations of inputs and outputs periodically; (b) fluctuating-FIR -smoothing frequency responses of FIR filters and preventing coloration problems caused by fixed FIR filters, by moving each FIR tap periodically on time axis with a different phase and time period. In this paper, these techniques are summarized. A block diagram of AFC using new equipment named AFC1, which has been developed at Yamaha Corporation and released recently in the US, is also presented.