A COMPARISION OF ACTIVE ACOUSTIC SYSTEMS FOR ARCHITECTURE

Transcription

1 A COMPARISION OF ACTIVE ACOUSTIC SYSTEMS FOR ARCHITECTURE A BRIEF OVERVIEW OF THE MOST WIDELY USED SYSTEMS Ron Freiheit 3 July 2001

2 A Comparison of Active Acoustic System for Architecture A BRIEF OVERVIEW OF THE MOST WIDELY USED SYSTEMS INTRODUCTION With the advent of electronics in the twenty century and its use related to acoustics (microphones, amplifiers, mixers, speakers, etc.) there has been a dream that these elements could work together to improve room acoustics related to music performances of orchestral concerts, choral concerts, opera and many others when these performances are limited to facilities that are less than ideal acoustically. Many times these facilities must be shared with a variety of events that would preclude the benefit of a purpose-designed venue providing optimum acoustics. This task of creating or altering acoustics thought the use of electronics has been generally referred to as active acoustics. Since development of active acoustics has been based largely on progress made by successive systems, it s of some benefit to briefly mention some of most notable systems in leading to the present day offerings. The pursuit of this dream has made steady progress during the last half of the twentieth century as will be seen from the review. Following the historical perspective a brief overview of the current commercially available active acoustics systems will be presented. The selection criteria use for choosing the most widely used commercial systems today was based on the number of installations that have been completed since This would exclude systems that were installed purely for research or that are not offered commercially. Also not included are new systems that have been introduced, but are not widely installed as of this date. In doing the research there has emerged some common themes related to active acoustics for architecture. One of the strongest themes is that the system and its components must be hidden from the audience in terms of visual impact. Another common theme is the need for the system to 2

3 be transparent acoustically and indistinguishable from the perceived natural acoustics of the space. This places a high emphasis on not being able to localize the speaker locations of a particular installation, no matter where the seat location. Also of great importance is the low noise criterion for the system since natural reverberation does not generate noise. The goal of a neutral enhanced sound field is equally important. The systems should not impart an undesirable timbre of their own (i.e. metallic sounding), but faithfully represent the acoustics of the performance in the enhanced space. Many of these basic principles have been know for many years 1 BREIF HISTORY OF ACTIVE ACOUSTIC SYSTEMS FOR ARCHITECTURE One of the earliest experiments done by Harold Burris-Meyer 2 in 1940 related to improving the acoustics on stage for the performer by using an amplified speaker at a fixed distance from the performer (creating a delay) while hiding the system from the audience both visual and aurally. The theory being that if the performer heard themselves better, they would in turn provide a better performance. Harry F. Olson 3 conducted an experiment at RCA labs creating an Acoustoelectric Auditorium in An auditorium was designed with and array of directional microphones hidden in the ceiling above the stage. Ceiling speakers and a signal delay system provided timing of the sound to the appropriate speakers based on their distance from the stage (source). The system provided natural sounding reinforcement of the acoustics in the room without the use of a traditional dedicated public address system. The principles of directional microphones to avoid pickup of the enhanced sound field, hidden speakers and signal processing (a magnetic tape recorder with spaced tape playback heads) were utilized to enhance the natural acoustics of the room. Multi-channel systems proposed as a way to improve the problem with feedback by increasing the number of independent channels. Franssen 4 proposed a system with a large number of 3

4 independent speakers and microphones. The location for each microphone was outside the reverberation distance. With a large number of independent channels the open loop gain can be held low enough to reduce the risk of feedback or coloration of the sound. The drawback to such a system was the large number of channels needed. Typically 50 channels must be used to double the reverberation time. There is also the problem of sensitivity of the open loop gain in each channel as the reverberation increases due to the fluctuations between the microphone, room and loudspeakers. This necessitates the use of some signal filtering for each channel. Eventually this became the Multi- Channel Reverberation (MCR) that was commercially available from Phillips. One of the most famous installations was the Assisted Resonance (AR) System at Royal Festival Hall in London. The hall suffered from inadequate reverberation time for musical performances. The solution suggested by Parkin and Morgan 5 was similar to the multi-channel system but restricted the frequency range of each channel for improved stability. The microphones and speakers were specifically mounted in the antinodes of the room at a particular resonance. Each channel covered a very narrow frequency band (1/3 octave) so a large number of channels were necessary to affect the desired reverberation bandwidth. Many years were spent using and learning how the acoustics could be changed through this technique. A further development by Chris Jaffe was the Early Reflected Energy System 6 (ERES), which sought to provide higher levels of the early reflections from the front of the hall with the use of microphones, digital delay lines and speakers. The ERES system was taken a step further with the inclusion of a Reverberation on Demand System (RODS) by Peter Barnett. Sound from microphones in the direct field are continually fed to digital reverberators, however there is no output from the system to the room as long as the input level is steady or rising. Once the level drops, the incoming signal is terminated and the output from the reverberator is coupled to the room through loudspeakers. Reverberant energy is added only when there is a drop in level such as the 4

5 end of a note, thus reducing the problem with recirulation of the sound through the system, which decreases the risk of feedback. The system relies on the masking of the running level of the music to hide the lack of reverberance in the hall only turning on when the reverberation would obviously be heard. One of the most recent systems developed has been the Variable Room Acoustics Systems 7 (VRAS), which simulates a coupled room utilizing a multi-channel reverberation matrix to create a secondary room. The result of the matrix is feedback to the to the main room via a loop of loudspeakers. The goal of the system is to provide high reverberation gain with low loop gain avoiding coloration. This system is very new and no large numbers of installations are known to date. COMMON PROBLEMS WITH EARLY SYSTEMS During the research there emerged common problems for all the early systems, which were very difficult to overcome. In most of these systems the microphones where located in field of the reverberant energy. This usually led to systems that would provide very low actual gain in reverberant level which is arguably as important as the desired increase in reverberant time (what s the use of increasing the reverberation time if you can not hear it?). Lack of actual gain in the reverberant level was usually due to the instability of the systems as the gain was increased due to the microphones operating within the enhanced reverberant field. M. Schroeder 8 had found that time variance could be unutilized improve the gain before feedback of a system. Another reoccurring problem is that of the quality of the sound, which is related to the issue of system gain. If the systems were operated too close to their gain limit, coloration of the reverberant field would be heard. It was not until the advent of high-speed digital audio processing that these issues could be dealt with on an effective basis. 5

6 OVERVIEW OF SELECTED SYSTEMS ACOUSTIC CONTROL SYSTEM (ACS), SYSTEM FOR IMPROVED ACOUSTIC PERFORMANCE (SIAP) LEXICON ACOUSTIC REVERBERATION ENHANCEMENT SYSTEM (LARES) OVERVIEW OF ACOUSTIC CONTROL SYSTEM (ACS) ACS is based on theory of wave field synthesis as described by Berkhout, de Vries and Vogel. In their paper on Acoustic control by wave field synthesis 9 the theory as applied to room acoustics is to create an electroacoustics wave front by picking up the source by a microphone array and feeding the signal to a system capable of calculating a wave field extrapolation based on the microphone locations and a planar loudspeaker array and the desired enhanced parameters for the room. For each microphone there exists one subsystem for one wave front. Benefits of this system are the ability to preserve the spatial information of the source to the audience area thus creating a more natural enhanced sound field. The practical implementation of this theory is by the use of a spaced array of hypercardiod microphones located 5 to 8 meters in above the stage area for sound pick up. The signals from the microphone array are and feed to a processing unit for calculation of the wave front extrapolation feeding the speaker array. Since a planar speaker array is not practical for most applications, a linear array on the front wall above the stage is utilized preserving the wave front in the lateral dimension. These speakers along with additional speakers located along sidewalls provide a means of creating image sources for early reflections. A separate processing unit and set of speakers are used to create the reverberant field for the room. The processing units contain the parameters of the desired space that will be replicated by the wave front creating early reflections and reverberation. Initially the reverberation part was created by allowing a controlled amount of acoustic coupling between the reverberant field and some of the microphone inputs. In recent 6

7 history this concept has been updated utilizing a digital reverberation system that does not rely on coupling from the reverberant field of the room. One of the drawbacks of the system is the large number of microphones typically used. Many installations contain 12 to 24 microphones that may decrease the stability level by 10 * log (number) db compared to a single microphone system. A reduction in level of 10dB to 14 db would have to be accommodated or the system would experience feedback or severe coloration. To this end time variance is included to control coloration due to recirulation in the system. Hypercardiod microphones are used minimize the coupling of the reverberant energy of the room to the system. Many installed systems contain 18 to 48 microphones, which is also the number of channels that will be needed, and 36 to 84 loudspeakers. The number of channels dictates the number of separate amplification channels needed. Speakers near the front of the auditorium are utilized for creating the early reflections from system (depending on the settings) and speakers located near the rear of the auditorium (walls and ceiling) are used to create the reverberant energy. It s important that the speakers be located at least 2.5 meters away from the listeners ears to minimize localization. The placement of microphones is sometimes limited due to stability issues related to proximity to loudspeakers. OVERVIEW OF SYSTEM FOR IMPROVED ACOUSTIC PERFORMANCE (SIAP) The basic concept of the SIAP system is that each performance space has its own unique acoustic character and the system will assist with correcting the deficiencies in the performance space, but does not synthesize a space that is fundamentally different (i.e. it does not create a cathedral sound for a space the size of a small auditorium). The designers of the system feel that if a simulated space is aurally excessively different from the visual space, there will be a disassociation between the two for the listener. The space is analyzed for acoustic deficiencies such as missing 7

8 reflections, low reverberation times, low levels of reverberation, etc. Once the analysis has been completed, microphones, speakers and processors are added to correct the deficiencies of the performance space. Speakers are located in the areas deemed to have acoustic deficiencies such as missing reflections. Larger numbers of speakers are added for increased in reverberation time or level. From this concept it is readily seen that each design is unique to each application. Typical systems utilize 4 to 8 super cardiod microphones 10 to 12 meters above the front of the stage. The microphone signals are fed to an input module for A/D conversions, 10-band parametric equalization, high and low pass shelving and optionally decorrelation and delay processing. The signal is then feed an acoustic processor (DSP), which controls early reflection and reverberation parameters. The signal is then fed to an output matrix router for appropriate distribution to speakers and amplifiers. Many systems utilizes between 50 and 150 speakers depending on application. A typical system contains 2 dual input modules (four input channels) 4 acoustic processor modules and 14 dual output modules. Each output channel requires 120W to 150W RMS to drive the speaker system. This allows for 448 (4 4 28) independent, decorrelated reflection patterns, providing 26.5 db (10 log # of independent channels) gain before feedback compared to a single channel system with an omni microphone. Time variance appears to be used, but not in connection with the reverberation section of the system. OVERVIEW OF LEXICON ACOUSTIC REVERBERATION ENHANCEMENT SYSTEM (LARES) The LARES concept provides an enhancement system to create an acoustic response for the room. This is done through the addition of the enhancement system components and if necessary absorption to minimize room artifacts that are not desirable (reduces coupling of the artifacts to the enhancement system). The LARES system is based on the use of completely time variant reverberators providing high reverberation levels with microphone locations 15 meters from the stage without coloration from acoustic feedback 10. The success of LARES systems has its 8

9 foundation in the ability to provide significant levels of early reflections and reverberation, which has been difficult to achieve with earlier systems without the resulting feedback or coloration. Much of the LARES development focused on the improvement in of the loop gain of the system before coloration of feedback. A typically system utilizing 16 reverberators and two cardiod microphones provides a total of 22.8 db of gain. The system is reduced by 12 db to provide stability and then a further 6 db below the onset of ringing to prevent obvious coloration. A net result of 4.8 db gain is available for the enhancement system without coloration or instability. The increase in gain allows for significant changes in the reflected and reverberant energy in the room which allows for synthesis of rooms that are not possible within the physical structure of the existing room. A basic system requires at least two cardiod microphones, 8 reverberators, and 4 output channels. More reverberators may be added to the system to provide addition gain to the system (for every doubling of the number of reverberators there is a 3 db gain in the system). Additionally there are amplifiers to drive each of the output channels. The design of the speaker coverage for the audience area has a criterion of ± 1.5 db averaged over the various locations. It s important that each speaker location not be adjacent to the speaker of the same channel (this maximizes the decorrelation of the system). Microphones are typically located 7 to 15 meters of the stage. One of the greatest challenges with this type of system is the successful implementation of the time-variant strategy without noticeably affecting pitch. The time delay in a digital system is quantized in units that are inversely related to the sample rate, so interpolation must be performed on each moving delay otherwise the output would contain an unacceptable amount of noise. Given that the delays are randomly changing there is a probability that at some point in time the delays will all shift in the same direction at some point in time causing the pitch to go sharp or flat. A special algorithm is implemented to minimize this pitch shift, but it is dependent on having multiple 9

10 channels in the system. The software is also optimized to improve coloration when the system is subjected to higher levels of acoustics feedback. COMPARISONS SYSTEM CONCEPT COMPLEXITY FLEXILBITY COSTS ACS SIMULATION VIA WAVEFRONT HIGH LARGE NUMBERS OF MICROPHONE AND OUTPUT CHANNELS FIXED WIRING PLAN HIGH COST DUE TO TOTAL NUMBER OF INPUT AND OUTPUT CHANNELS SIAP REPLACEMENT OF MISSING ACOUSTIC ELEMENTS HIGH LARGE NUMBERS OF OUTPUT CHANNELS FLEXIBLE OUTPUT MATRIX HIGH COST DUE TO NUMBER OF INPUT AND OUTPUT CHANNELS LARES SIMULATION VIA TIME VARIANCE LOW TYPICALLY TWO INPUTS AND 16 OUTPUTS FIXED WIRING PLAN LOWER COST LEAST NUMBER OF INPUT AND OUTPUT CHANNELS CONCLUSIONS The selection of the best system will be directly related to the desired result. If correction only is desired for a room, then the SIAP system may be preferred over ACS or LARES. If simulations of other spaces are desired then the selection may be between LARES and ACS. As is the case for almost all decisions, cost will play a significant factor in system selection. 10

11 During the last decade there have been significant advances made in the performance of active acoustic systems for architecture primarily driven by the improvements in high speed digital signal processing. Improvements in gain and stability of systems through new and innovative methods have helped increase the acceptance of systems for use in music performance. Sound quality has also improved due to the ability to provide higher levels without approaching the coloration limits. Improvements in other components such as lower noise microphones, digital filters, improved amplifiers have all contributed to better systems. Speaker development has also improved greatly in the same time period offering speakers with better control patterns allowing designers to provide consistent audience coverage. As digital signal processing continues to evolve and improve, active acoustic systems will continue to provide higher levels of performance creating an ever-widening sphere of acceptance. Active acoustics will become more commonplace as systems are included in performance venue designs that require increasing acoustic flexibility. REFERENCES 1 Heinrich Kuttruff, Room Acoustics, second edition, Applied Science Publishers, Chapter X.5, pages , Harold Burris-Meyer, The Control of Acoustic Conditions on the Concert Stage, J. Acoust. Soc. Am., volume 12, Harry F. Olson, Acoustoelectronic Auditorium, J. Acoust. Soc. Am., volume 31, number 7, Heinrich Kuttruff, Room Acoustics, second edition, Applied Science Publishers, Chapter X.5, pages , P. H. Parkin and K. Morgan, Assisted Resonance in the Royal Festival Hall, London: , J. Acoust. Soc. Am., volume 48, number 5, (part 1), Wim Prinssen and Dr. Peter D Antonio, The History of Electronic Architecture and Variable Acoustics, SIAP, pages Mark Poletti, An Assisted Reverberation System for Controlling Apparent Room Absorption and Volume, Audio Engineering Society, Presented at the 101 st Convention, November 8-11, M. R. Schroeder, Improvement of Acoustic-Feedback Stability by Frequency Shifting, J. Acoust. Soc. Am., volume 36, number 9,

12 9 A.J. Berkhout, D. de Vries, and P. Vogel, Acoustic control by wave field synthesis, J. Acoust. Soc. Am., volume 93, number 5, David Griesinger, Improving room acoustics through time-variant synthetic reverberation, Audio Engineering Society, Presented at the 90 th Convention February 19-22,