<%BANNER%>

Effects of Temporal Distribution of Specular and Diffuse Reflections on Perceived Music Quality

HIDE
 Title Page
 Dedication
 Acknowledgement
 Table of Contents
 List of Tables
 List of Figures
 Abstract
 Introduction
 Literature review
 Materials and methods
 Results and discussion
 Conclusion and future work
 Appendices
 References
 Biographical sketch
 

PAGE 1

EFFECTS OF TEMPORAL DISTRIBUTION OF SPECULAR AND DIFFUSE REFLECTIONS ON PERCEIVED MUSIC QUALITY By PATTRA SMITTHAKORN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2006 1

PAGE 2

Copyright 2006 by Pattra Smitthakorn 2

PAGE 3

To my great mentors: Buddhadasa Bhikku, and Professor Li-Chien Shen. 3

PAGE 4

ACKNOWLEDGMENTS I would like to thank Professor James P. Sain, Professor Robert C. Stroh, Professor Bertram Y. Kinzey, Jr., and Professor Gary W. Siebein for their trust, support and constructive advice. I would like to thank Professor Martin A. Gold for his help on setting up the equipment for all experiments, providing opportunities for his students to participate in the listening tests, and his guidance on how to perform the listening tests. His open-mindedness to possibilities and challenges, his expertise, and his support have encouraged me to pursue such an impossible dream. There are no words that can describe my appreciation for his help and friendship. I am one of those few lucky students who have come to experience the true meaning of education. Professor Li-Chien Shen has taught me to humbly believe in the value of education that can transform ones life unimaginably without him saying anything. His wisdom has shone through and dispelled my years of darkness. The inspiration I have received from him and his colleagues, Professor Yuli Rudyak and Professor Alexander Berkovich, has changed my attitude toward life and my future career completely. I hope to be able to share this wonderful experience with my students in the future. I would like to thank my family and friends for all kinds of support and more importantly for believing in me. I am indebted to the Royal Thai Government for this life-transforming opportunity. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES...........................................................................................................................7 LIST OF FIGURES.........................................................................................................................8 ABSTRACT...................................................................................................................................11 CHAPTER 1 INTRODUCTION..................................................................................................................13 2 LITERATURE REVIEW.......................................................................................................15 Diffuse Reflections.................................................................................................................15 Temporal Diffusion.........................................................................................................18 Directional Diffusion.......................................................................................................21 Hedgehog Representation................................................................................................22 Sound Diffusivity Index..................................................................................................22 3 MATERIALS AND METHODS...........................................................................................25 Pilot Study I............................................................................................................................25 Pilot Study II...........................................................................................................................25 Final Experiment....................................................................................................................26 Objective..........................................................................................................................26 Method.............................................................................................................................27 4 RESULTS AND DISCUSSION.............................................................................................32 Source Characteristics............................................................................................................32 Orchestral Music.............................................................................................................32 Trumpet...........................................................................................................................33 Piano................................................................................................................................36 Initial Hypothesis: Effects from Path.....................................................................................39 Receiver..................................................................................................................................41 Results and Discussion...........................................................................................................41 Loudness..........................................................................................................................42 Clarity..............................................................................................................................43 Intimacy...........................................................................................................................45 Reverberance and Echoes................................................................................................48 Source Width...................................................................................................................48 Texture.............................................................................................................................51 Overall Impression..........................................................................................................51 5

PAGE 6

Conclusions.............................................................................................................................54 Specular Reflections........................................................................................................54 Diffuse Reflections..........................................................................................................54 Applications............................................................................................................................55 Comments...............................................................................................................................58 5 CONCLUSION AND FUTURE WORK...............................................................................59 Conclusion..............................................................................................................................59 Future Work............................................................................................................................59 APPENDIX A QUESTIONNAIRE................................................................................................................61 B RESULTS FROM LISTENING TEST..................................................................................63 C PILOT STUDY I....................................................................................................................71 D PILOT STUDY II...................................................................................................................78 Black-box Theater: Physical Model and Computer Modeling...............................................78 Impulse Response Analysis....................................................................................................83 LIST OF REFERENCES...............................................................................................................86 BIOGRAPHICAL SKETCH.........................................................................................................88 6

PAGE 7

LIST OF TABLES Table page 3-1 Absorption and scattering coefficients specified in CATT modeling...............................29 3-2 The 9 designed impulse responses showing various sets of arrival specular reflections generated by CATT-Acoustics.........................................................................30 3-3 The 9 designed impulse responses showing various sets of arrival diffuse reflections generated by CATT-Acoustics..........................................................................................31 C-1 Statistical results from the listening test, 4 represents clearly different, 0 represents that no perceived difference can be detected.....................................................................77 D-1 Impulse responses from painted model..............................................................................81 D-2 A comparison of the simplified impulse responses from painted model...........................82 7

PAGE 8

LIST OF FIGURES Figure page 2-1 Examples the impulse responses showing visual judgments of texture: (a) IRs from some of the best concert halls which have good texture, (b) IRs from hall with poorer texture.........................................................................................................19 2-2 Single-number criteria for the evaluation of echoes in an impulse response. After Niese..................................................................................................................................20 2-3 Single-number criteria for the evaluation of echoes in an impulse response. After Bolt and Doak....................................................................................................................21 2-4 Directional hedgehogs, for early and late-time arrival, at three locations in the Beethoven Hall of the Liederhalles, in Stuttgart................................................................23 2-5 The surface diffusivity index (SDI) of 31 concert halls. The highest rated halls have SDIs of 1.0, while those of the lowest rated halls fall in the range of 0.3 to 0.7..............24 3-1 The computer generated models created to produce five specular/diffuse reflections arrival to the receiver within 40 msec................................................................................29 4-1 Main directional radiation (0-3 dB) of the trumpet in the vertical plane.......................35 4-2 Sound spectra of a horn in different keys..........................................................................35 4-3 Reverberation time of a grand piano for some C-keys using the right pedal....................37 4-4 Patterns of directional radiation of a grand piano in a vertical plane................................38 4-5 Music preference of 47 subjects who participated in the listening test.............................41 4-6 Normalized perceived loudness of the three sources: Orchestral music, Trumpet, and Piano. The orange columns are anechoic signals of each group.......................................43 4-7 Clarity: normalized averaged ratings of perceived clarity of (a) Orchestral music, (b) Trumpet, and (c) Piano......................................................................................................44 4-8 Intimacy: normalized averaged ratings of perceived intimacy of the Piano source..........45 4-9 Reverberance: normalized averaged ratings of perceived reverberance of (a) Orchestral music, (b) Trumpet, and (c) Piano....................................................................46 4-10 Echoes: normalized averaged ratings of perceived echoes of (a) Orchestral music, (b) Trumpet, and (c) Piano......................................................................................................47 4-11 Source width: normalized averaged ratings of perceived source width of (a) Orchestral music, (b) Trumpet, and (c) Piano....................................................................49 8

PAGE 9

4-12 Texture: normalized averaged ratings of perceived texture of (a) Orchestral music, (b) Trumpet, and (c) Piano.................................................................................................50 4-13 A comparison on ratings of perceived acoustic qualities: (a) source width, (b) intimacy, (c) texture, and (d) overall impression of Orchestral music..............................52 4-14 A comparison on ratings of perceived acoustic qualities: (a) texture, and (b) overall impression of Trumpet.......................................................................................................53 4-15 A comparison on ratings of perceived acoustic qualities: (a) clarity, (b) intimacy, (c) texture, and (d) overall impression of Piano......................................................................53 4-16 Summary of the preferred arrival diffuse and specular reflections to provide improvement on perceived overall impression for 3 sourcesOrchestral music, Trumpet, and Piano............................................................................................................56 4-17 Three ellipsoids representing boundaries where reflecting panels can be located within to provide the 1st order reflections arriving to the front receiver within 3 time periods: 40, 80, and 160 msec after the direct sound.........................................................57 4-18 Three different boundaries for 3 different receiver locations where reflecting panels can be located to provide the 1st order reflections to arrive at the receivers with 40 msec after the direct sound................................................................................................58 B-1 Loudness: normalized averaged ratings of perceived loudness of (a) Orchestral music, (b) Trumpet, and (c) Piano.....................................................................................63 B-2 Clarity: normalized averaged ratings of perceived clarity of (a) Orchestral music, (b) Trumpet, and (c) Piano......................................................................................................64 B-3 Intimacy: normalized averaged ratings of perceived intimacy of (a) Orchestral music, (b) Trumpet, and (c) Piano.................................................................................................65 B-4 Reverberance: normalized averaged ratings of perceived reverberance of (a) Orchestral music, (b) Trumpet, and (c) Piano....................................................................66 B-5 Source width: normalized averaged ratings of perceived source width of (a) Orchestral music, (b) Trumpet, and (c) Piano....................................................................67 B-6 Texture: normalized averaged ratings of perceived texture of (a) Orchestral music, (b) Trumpet, and (c) Piano.................................................................................................68 B-7 Echoes: normalized averaged ratings of perceived echoes of (a) Orchestral music, (b) Trumpet, and (c) Piano......................................................................................................69 B-8 Overall impression: normalized averaged ratings of perceived overall impression of (a) Orchestral music, (b) Trumpet, and (c) Piano..............................................................70 9

PAGE 10

C-1 Computer simulation of the Charlotte auditorium with source (A0) and receiver locations specified: (a) without diffusers, (b) with diffusers on the walls and the ceiling.................................................................................................................................71 C-2 Image source method used to derive arrival time of specular and diffuse reflections in the impulse responses....................................................................................................72 C-3 Three dimensional image source method used to produce rooms IRs at selected receivers.............................................................................................................................73 C-4 Comparison of the rooms impulse reponses of the rooms 1st order diffusion at different receiver locations: (1) receiver 1, (3) receiver 3, (5) receiver 5..........................73 C-5 Hedgehog presentations at receiver 3: (a) without diffusers, (b) with diffusers................75 C-6 Average RT, IACC, LEF, and C80 of all receivers from 2 room conditions (with and without diffusers) across 6 frequency bands......................................................................76 C-7 The questionnaire used for the listening test developed from Torress.............................77 D-1 Two sets of 1:20 physical model tests were carried out with (a) painted, and (b) unpainted models. Three frequencies band-pass filtered IRs were measured.................78 D-2 Plan, section, and isometric of the base model with the source and the receivers locations.............................................................................................................................79 D-3 Six models with different sizes of diffusers for each set of test........................................80 D-4 The method of IR analysis: 1) find the boundary of energy fluctuation, 2) delete the data between the upper and the lower bounds, 3) Simplified IR is obtained.....................80 D-5 Comparisons of the energy bounds among 6 models IRs at 3 band-pass frequencies filtered of: (a) painted models, (b) unpainted models........................................................84 D-6 Comparisons of acoustical parameters of the 6 painted models simulated by CATT-Acoustic: (a) T30/RT, (b) IACC, (c) G10, (d) C80, (e) LFC /LF......................................85 10

PAGE 11

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EFFECTS OF TEMPORAL DISTRIBUTION OF SPECULAR AND DIFFUSE REFLECTIONS ON PERCEIVED MUSIC QUALITY By Pattra Smitthakorn December 2006 Chair: Gary W. Siebein Major Department: Design, Construction, and Planning The purpose of this study was to investigate the effects of the temporal distribution of diffuse and specular reflections on the perceived acoustic qualities of music performance. Sets of impulse responses were designed with different temporal distributions of early acoustic energy (specular and diffuse reflections). Then, three types of anechoic sound sourcesorchestral music, trumpet, and pianowere convolved with the designed impulse responses. The results from the listening tests revealed that different room environments were needed to acoustically support different source characteristics. The results show the following: 1) specular reflections arriving within 40 msec of the direct sound improved perceived clarity and intimacy; 2) specular reflections arriving between 40-80 msec after the direct sound improved perceived clarity for orchestral music; 3) specular reflections arriving later than 80 msec after the direct sound are not desirable; 4) large numbers of diffuse reflections arriving within 40 and 80 msec of the direct sound improved perceived intimacy, texture, and overall impression for all sound sources, heightened perceived clarity for trumpet and piano, and reduced perceived glare for trumpet; and 5) diffuse reflections arriving between 80-160 msec of the direct sound preserved perceived reverberance and reduced perceived echoes as opposed to specular 11

PAGE 12

reflections arriving in the same time period. The results of this study indicate that music performance halls should be designed to include diffuse reflections from surfaces within the 80 msec time period to achieve preferred texture, intimacy, clarity and overall impression and in the 160 msec time period to reduce echoes; specular reflections arriving within the 40 msec time period should be provided to enhance perceived clarity. 12

PAGE 13

CHAPTER 1 INTRODUCTION When architects design a concert hall, the chief goal that needs to be met, among many other functional requirements and design criteria, is to design a room that provides an excellent acoustic environment. Recommendations from acoustic consultants are given to architects to guide their designs to achieve this goal. The recommendations include the optimum range of acceptable acoustical parameters known to have major effects on the perceived acoustical qualities of the room. The parameters are listed below. Optimal Reverberation Time (RT) should be achieved; optimal RT varies depending on the purpose of the hall. This suggestion directly affects the volume and materials used inside the hall according to the RT formula. The Initial Time Delay Gap (ITDG) should be within 20 msec of the direct sound.1 The design of acoustic clouds and the first order reflectors are roughly guided by this requirement. Provide a number of early reflections to obtain clarity of sounds (high Clarity index-C80) perceived in the room. Limit the room width and provide a large number of early lateral reflections to obtain a high value of the Lateral Fraction (LF) which would enhance the sense of envelopment, spaciousness, and widen the source width.2 Avoid any acoustic defects such as echoes, sound focusing, and flutter echoes by locating absorbing materials or diffusers at the rear wall (rather than smooth reflective surfaces), avoid smooth large concave surfaces, and avoid smooth large parallel walls. Provide diffusing panels surrounding the orchestra (orchestra shell) for a sense of ensemble and blending of sounds from the sending end to the auditorium.3 Provide a large amount of various scales of surface irregularities to obtain a diffuse sound field.4 1 Beranek, L. (1962). Music, Acoustics and Architecture New York, Wiley. 2 Barron, M., & Marshall, A. H. (1981). "Spatial impression due to early lateral reflections in concert halls: The derivation of a physical measure." J. of Sound and Vibration 77: 211-232. 3 Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture NY, Springer. 4 Ibid. 13

PAGE 14

In the design of the interior of a concert hall with the knowledge that most of the materials used in the room should be reflective to preserve sound energy as much as possible, further questions are raisedwhat type of reflectors should be used: specular or diffuse reflectors? What type of diffuse reflectors (shape/form) should be selected? How many of the reflectors should be diffuse compared with specular? Where should these diffuse and specular reflectors be positioned in the hall? These questions thus far have not been answered with scientific support. There have been many experiments with vastly diverse approaches trying to find answers to questions concerning acoustical diffusion phenomena. This dissertation answers some of the questions raised by architects scientifically and provides a method that can be used to investigate these questions. The literature review contains discussion of several prior works which are thought to be related to and might have potentially influenced the experimental designs used in this study. A series of experiments with different approaches and settings have been designed and conducted to search for more evidence and to test new sets of hypotheses. The third experiment finally provides fruitful results, and hopefully this new approach can be further developed to study and search for knowledge in the field of room acoustics. 14

PAGE 15

CHAPTER 2 LITERATURE REVIEW Diffuse Reflections Scientific research of the role of diffuse reflections in concert halls was not begun until the 1980s, despite the fact that diffusing surfaces have always been considered (and included) as an important part of a concert hall design since their origination. Research and experiments in this area are all innovative and somewhat in their infancy. Diverse approaches of experiments are found in this area, which provides discontinuous pieces of knowledge. With their unsuccessful results, a pattern/trace to a future successful experimental design was not found. Therefore, past studies and investigations in this area seem to be scattered and inconclusive. Dalenback5, in his 1994 article on A macroscopic view of diffuse reflection, provides a widely accepted view and up-to-date information about diffusion. He defined the term diffuser as any surface that reflects sound in all or most directions, largely independent of the incidence angle. Such surfaces have a roughness (naturally or by design) with a size on the order of a wavelength in the frequency band where diffuse reflection is observed. Qualitative properties of diffusely reflecting surfaces have been described as follows:6 1) The diffusion capability of a diffuser is strongly frequency dependent. However, wide-band diffusion is possible depending on the magnitude of the surface depth. 2) The diffuser will give reflections outside the specular zone, thereby decreasing energy within the specular zone, which diminishes the comb-filter effects due to strong early specular reflections.7 With this characteristic, a diffuser can be used to enhance the sense of ensemble on stage. 5 Dalenback, B.-I., Kleiner, M., & Svensson, P. (1994). "A macroscopic view of diffuse reflection." J. Acoust. Soc. Am. 42(10): 793-806. 6 Ibid. 7 When there are equal intensities of primary sound and (too) early specular reflection interfere with one another the results are large linear distortions of the tone. Depending on the particular delay difference of the reflection, certain frequency ranges are very much reinforced whereas others are completely wiped out. (Haas,1972) 15

PAGE 16

3) Diffuse reflections can decrease the risk for dead spots and soften detrimental reflections since the reflected energy will be distributed throughout the whole space more quickly and more uniformly. This also diminishes sharp late echoes without the need for introducing absorbers. 4) A more uniform reverberant field will be created across the hall (by diffuse reflections) with an improved near exponential decay, giving a smooth impulse response. 5) Frequency content in reflections is affected by diffuse reflections, which results in sound coloration. 6) Diffuse reflections also create effects such as directivity smear, amplitude smoothing, and temporal smearing. Other research in the field of diffusion can be grouped into 2 categories: 1) analyses of diffusers (path), and 2) effects from diffuse reflections (receiver). First, a large number of research/studies are related to the search for a mathematical model to define and standardize diffusion coefficients8, methods of obtaining frequency dependent diffusion coefficients9, and optimizing diffuser design 10. D Antonio defined diffusion coefficient as a measure of the uniformity of the reflected sound which allows comparison of the performance of the surfaces, while scattering coefficient is a ratio of sound energy scattered in a non-specular manner to the total reflected sound energy. The latter measure corresponds to the theory used in geometrical room modeling software. In this case, misrepresentation of a triangular diffuser can be seen because it would scatter energy away from the specular zone and appear as a good diffuser, however this is a redirection of energy rather 8 Cox, T. J., & D' Antonio, P. (2004). Acoustic Absorbers and Diffusers: Theory, Design and Application NY, Spon Press. 9 Ibid. 10 Cox, T. J. (1996). "Designing curved diffusers for performance spaces." J. Audio Eng. Soc. 44(5): 354-364. 16

PAGE 17

than dispersion.11 Scattering coefficients and this procedure are what is used in CATT-Acoustic (Computer Aided Theater) software. Second, from the aspect of receiver, studies include: 1) the effects of diffusion on room acoustic qualities (auralization)12; 2) acoustical parameters that indicate the room diffusivity temporally (texture13, temporal diffusion14), and directionally (directional diffusion)15; and 3) a numerical rating of a space/hall using visual inspection called sound diffusivity index (SDI)16. This dissertation is another investigation in this category; therefore a closer review was made of the studies mentioned above. Torres and Kleiner17 (2000) conducted a listening test in search of the effects of surface diffusion on auralization at different frequency ranges. This was done by convolving 4 anechoic sources with selected room impulse responses obtained from assigning different (quasi-step function: high, low) diffusion coefficients to surfaces in the room (at 3 frequency ranges: low, mid, high). The 4 anechoic sources included two sustained (synthesized organ chord, pink noise) and two impulsive (string quartet with pizzicato, and the unconvolved binaural room impulse response (BRIR) alone) sources. Their results show the following: 11 Antonio, P. D., & Cox, T. (1998). "Two decades of sound diffusor design and development Part 2: prediction, measurement, and characterization." Ibid. 46(12): 1075-1091. 12 Torres, R. R., & Kleiner, M. (2000). "Audibility of diffusion in room acoustics auralization: an initial investigation." Acustica 86(6): 919-927. 13 Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture NY, Springer. 14 Bolt, R. H., & Doak, P. E. (1950). "A tentative criterion for the short-term transient response of auditoriums." J. Acoust. Soc. Am. 22: 507-509. 15 Meyer, E. (1954). "Definition and diffusion in rooms." Ibid. 26(5): 630-636. 16 Haan, C. H., & Fricke, F. R. (1993). Surface diffusivity as a measure of the acoustic quality of concert halls Proceedings of Conference of the Australia and New Zealand Architectural Science Association, Sydney. 17Torres, R. R. (2000). Studies of Edge Diffraction and Scattering: Applications to Room Acoustics and Auralization. Department of Applied Acoustics Gteborg, Chalmers University of Technology. 17

PAGE 18

1) For some signals, changes in the diffusion coefficient are clearly audible within a wide frequency region. Thus, diffuse reflections should be modeled in a frequency-dependent manner, although not all auralization programs currently do this. 2) The perception of these changes depends on the input signal. For sustained signals (e.g., an organ chord, pink noise), changes are strongly perceived as differences in coloration; for example, increasing low-frequency diffusion is perceived as decreasing the bass content or increasing the treble content of the signal. For impulsive signals (e.g., string pizzicato), coloration differences are less audible than for sustained signals, whereas spaciousness differences are relatively stronger. It is interesting that listeners, though uninformed of the differences between highor low-diffusion signals, give consistent answers regarding perceived changes in frequency coloration.18 Beranek19 defined texture as: Texture is the subjective impression that listeners derive from the patterns in which the sequences of early sound reflections arrive at their ears. In an excellent hall, those reflections that arrive soon after the direct sound follow in a more-or-less uniform sequence. In other halls there may be a considerable interval between the first and the following reflections or one reflection may overly dominate. Good texture requires a large number of early reflections, reasonably strong in amplitude, uniformly but not precisely spaced apart, and with no single reflection dominating the others. This quality can be determined visually from the impulse responses (IRs) as shown in Figure 2-1. Temporal Diffusion In the area of temporal diffusion, the patterns of impulse responses were observed in detail by Niese (1956), and Bolt & Doak (1950). Niese20 plotted the impulse response (|p(t)|) together with an exponential decay curve shown in Figure 2-2. He considered the areas of the impulses that, after the time limit t = 33 msec, lie above the decay curve as detrimental. He then proposed the criterion for the existence of echoes called the echo coefficient () by combining the detrimental areas S and useful areas N in the form shown in Equation 2-1. 18 Ibid. 19 Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture NY, Springer. 20 Niese, H. (1961). "Die Messung der Nutzschallund Echogradverteilung zur Beurteilung der Hrsamkeit in Rumen." Acustica 11: 202-213. 18

PAGE 19

(a) (b) Figure 2-1. Examples the impulse responses showing visual judgments of texture: (a) IRs from some of the best concert halls which have good texture, (b) IRs from hall with poorer texture.21 = NSS (Eq. 2-1) The drawback to this criterion is that the existence of a number of peaks that exceed the general decay only slightly would produce the same echo coefficient as a single strong reflection. The evaluation of flutter echoes proposed by Bolt and Doak (1950) utilizes the auto-correlation function (()) to analyze the randomness of the impulse energy (p(t)) above the touching decay curve, Figure 2-3, given as Equation 2-2. () = (Eq. 2-2) dtPPtt)(0)( 21 Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture NY, Springer. 19

PAGE 20

Where 1, 2, 3, are the time intervals of periodic sequences of impulses. Kuttruff22 proposed the criterion temporal diffusion, which is represented by shown in Equation 2-3. = )0(max)0( (Eq. 2-3) The higher the value of the more random the impulse response. This criterion is based on the nature of how flutter echoes are produced. Parallel walls reflected sound energy repeatedly back and forth is usually the cause of flutter echoes, not diffuse reflecting walls. This criterion gives a general measure of the randomness of a sequence of reflections assuming that any regular sequence of reflections is undesirable for hearing. However, describing the randomness quality of the reflections obtained from the impulse response is not enough, for it does not describe the smoothness of the IRs which is the room acoustical quality for texture. Figure 2-2. Single-number criteria for the evaluation of echoes in an impulse response. After Niese.23 22 Kuttruff, H. (1965/66). "ber Autokorrelationsmessungen in der Raumakustik." Acustica 16: 166-174. 23 Niese, H. (1961). "Die Messung der Nutzschallund Echogradverteilung zur Beurteilung der Hrsamkeit in Rumen." Acustica 11: 202-213. 20

PAGE 21

Figure 2-3. Single-number criteria for the evaluation of echoes in an impulse response. After Bolt and Doak. 24 Directional Diffusion Directional diffusion is a single number criterion that describes the directional distribution of the arriving sound at a particular location proposed by Thiele (1953)25. It is a measure of the deviation from average energy from all directions represented as follows. First the quantity is determined, the sum of the absolute differences between the incoming intensity J over all room directions within the solid angle of interest and the average incoming intensity J ; he normalized this sum to the average intensity J (Equation 2-4). = dJJJ1 (Eq. 2-4) This quantity is zero for a non-directional, isotropic sound field, and has its maximum value 0 in open air.26 The directional diffusion ( or DD) is given by Eq. 2-5. = 1 0 (Eq. 2-5) 24 Bolt, R. H., & Doak, P. E. (1950). "A tentative criterion for the short-term transient response of auditoriums." J. Acoust. Soc. Am. 22: 507-509. 25 Thiele, R. (1953). "Richtungsverteilung und Zeitfolge der Schallrckwrfe in Rumen." Ibid. 3: 291-302. 26 Cremer, L., & Muller, H. A. (1982). Principles and Applications of Room Acoustics NY, Applied Science Publishers Ltd. 21

PAGE 22

The directional diffusion is location dependent. The results from his study suggest that, for constant RT: 1) DD decreases with increasing volume, and 2) in a room with the same amount of absorbing, reflecting, and diffusing surfaces, DD varies depending on the location of those materials. The drawback to this criterion is that a single strong directional peak or several small peaks and valleys that are distributed equally over a solid angle might yield the same value for this measure. Hedgehog Representation The hedgehog representation studied by Junius (1959)27 shown in Figure 2-4 gives a clear picture of the spatial distribution of early (0-100 msec) and late energy (100msec). This representation shows that the assumptions on the diffuse sound field underlying statistical theory are poorly fulfilled in a large hall. Even in the later time period, the directional distribution is still far from uniform. However, the temporal distribution information gets lost in the hedgehog representation. Sound Diffusivity Index The criterion used by Beranek28 to evaluate the room diffusivity is the sound diffusivity index (SDI). SDI was proposed by Haan and Fricke (1993) using a visual inspection to classify the acoustical quality of the surface irregularities in a hall. The degree of diffusivity of surfaces can be categorized as high, medium and low diffusivity. The room with high diffusivity is the one that has a coffered ceiling with deep recesses or beams greater than 4 in. in depth, but not greater than 10 in., diffusing elements of sizable depth on the upper sidewalls, and fine scale irregularities of about 2 in. depth on the lower sidewalls 27 Junius, W. (1959). "Raumakustische Untersuchunger mit Neueren Messverfahren in der Liederhalle Stuttgart." Acustica 9: 289-303. 28 Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture NY, Springer. 22

PAGE 23

without absorbing surfaces. Medium diffusivity is assigned to broken surfaces of varying depth on the ceiling and sidewalls or shallow recesses (no greater than 2 in.) of ornamental decorative treatment, having no absorbing surfaces. A room with low diffusivity is one that contains large separate paneling, or smoothly curved surfaces, or large flat and smooth surfaces, or heavy absorptive treatment. Figure 2-4. Directional hedgehogs, for early and late-time arrival, at three locations in the Beethoven Hall of the Liederhalles, in Stuttgart.29 The numerical rating assigned to each degree of surface diffusivity is as follows: high equals 1.0, medium equals 0.5, and low equals 0. The SDI is obtained by dividing the summation of the weighted diffusing area by the actual total area for the ceiling and sidewalls. 29 Junius, W. (1959). "Raumakustische Untersuchunger mit Neueren Messverfahren in der Liederhalle Stuttgart." Acustica 9: 289-303. 23

PAGE 24

Beranek (2003) shows the plot of rank-ordering of 31 concert halls against SDI, Figure 2-5. The plot illustrates that the best halls have SDIs of 1.0; the medium quality halls, between 0.5 and 0.9; and the low quality halls between 0.3 and 0.7. He also mentioned that to obtain meaningful SDIs among those halls for rank-ordering, other acoustical parameters, for example, IACC, RT, ITDG, good stage design, and so on, must be taken into account simultaneously. The drawback for this method is that the judgment made by visual inspection is somewhat subjective. All proposed measures mentioned above have more or less contributed to the study of diffusion. However, due to their drawbacks, none of them have been further developed to become a standard measure for room diffusivity. The goal of this dissertation is to develop a series of studies to explore and understand the behavior of diffuse reflection and their effects on room acoustics. Figure 2-5. The surface diffusivity index (SDI) of 31 concert halls. The highest rated halls have SDIs of 1.0, while those of the lowest rated halls fall in the range of 0.3 to 0.7.30 30 Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture NY, Springer. 24

PAGE 25

CHAPTER 3 MATERIALS AND METHODS Pilot Study I Prior to the final experiment, two pilot studies were performed. The first pilot study was a comparison between two computer simulations of a chosen auditorium (Charlotte auditorium) with and without diffusers. The results obtained from this study show that: 1) the auralizations from the halls with and without diffusers are clearly discernible especially for speech; 2) no current quantitative measures can capture/explain this subtle perceived change; and 3) the fine structure of the room impulse responses (IRs) and the directional distribution of sound energy have been carefully observed to identify changes in arrival time of the 1st order specular and diffuse reflections. Further detail on this investigation can be found in Appendix C. Pilot Study II The second pilot study attempted to search for any (graphical) evidence of subtle changes in IRs provided by installing different types of diffusers in a room. A 1:20 physical model with six different types of diffusers installed on the rooms boundaries was built. Band-pass filtered impulse responses were recorded from six models. A careful graphical analysis had been conducted (see Appendix D). General results from the experiment are coherent with Sabines theories (RT) and initial expectations. However, concrete conclusions cannot be made due to the fact that there has not been any standard procedures developed to graphically evaluate the temporal distribution of strong reflections shown in the impulse responses. The lesson learned from the second pilot study strongly influenced the experimental design of the final experiment. On one hand, the temporal diffusivity of each diffuser has not yet been discovered. Only directional diffusivities of some diffusers are available. On the other hand, the room impulse responses do not take into account the room spatial information; they show only 25

PAGE 26

temporal energy distribution. Therefore, changing diffusers on the rooms surfaces provides no control over the temporal energy distribution. The final experiment then focused on taking control of the temporal distribution of the two types of energyspecular and diffuse reflections. Final Experiment Objective The objective of this experiment was to study the effects of the temporal distribution of early acoustic energy (specular and diffuse reflections). A set of variations of energy distribution was designed and a listening test was conducted to capture differences in perceived acoustical qualities. Three source types used in this study were orchestral music, trumpet, and piano. The focus on only the early part of the impulse responses was a result from Pilot Study I, which showed prominent variation in IRs characteristics affected by the first order diffuse reflections in the IRs not long before the energy reaches its uniformly diffuse sound field (when second and third order specular and diffuse reflections arrive). The three approximate time windows designed for this experiment are as follow: 1) between 20 and 40 msec of the direct sound, 2) between 40 and 80 msec of the direct sound, and 3) between 80 and 160 msec of the direct sound. The chosen time window of 40 msec was derived from the Haas Effect31, which suggests that specular reflections arriving within 40 msec of the direct sound are not perceived as echoes. The 80 msec limit came from the time limit for useful to detrimental sound energy of the room IRs for music proposed by Reichardt et al. in 1975. This time limit is now used in many acoustic parameters, for example, C80, IACC, and LF. 31 Haas, H. (1972). "The influence of a single echo on the audibility of speech." J. Audio Eng. Soc. 20(2): 146-159. 26

PAGE 27

The 160 msec limit approximates the arrival time of the first order specular and diffuse reflections of a medium to large concert hall, derived from Pilot Study I, and also completes a coherent logarithmic increment step for the whole experiment, 40, 80, and 160 msec. No acoustical parameters will be studied here because the previous experiments show no evidence of any existing acoustical parameters that can describe the detailed characteristic of energy distribution temporally and spatially. Instead, after the study of perceived acoustical quality was made, a careful look at how to capture and explain this quality quantitatively can be further investigated. Method Eighteen designed impulse responses were created using CATT-Acoustic V.8. CATT was used to model a highly absorptive room of the size 70m x 70m x 70m with the source and the receiver floating 14m apart in the middle of the room (Figure 3-1). All 6 surfaces of the room have absorption coefficients of 0.9 at 6 octave band frequencies (125Hz, 250, 500, 1 kHz, 2 kHz, 4 kHz) to simulate the anechoic environment. Five, 10, and 20 reflecting panels of the size 3m x 3m were located in the room to provide specular or diffuse reflections (from the upper-frontal hemisphere) arriving at the receiver within the three time-windows to get the designed room impulse responses as shown in Table 3-2, 3-3. The first reflection was set to arrive at 20 msec after the direct sound for all 18 impulse responses to control for the same ITDG. The absorption coefficients and scattering coefficients of the specular and diffuse reflectors are shown in Table 3-1. The scattering coefficients were chosen based on the recommendation from CATT-Acoustic manual.32 32 Dalenback, B.-I. (2002). CATT-Acoustic v8.0 User's Manual Gothenburg, Sweden. 27

PAGE 28

Then, these 18 impulse responses were convolved (excluding the late part impulse response) with three different sources of anechoic music (orchestral piece, trumpet, and piano) of approximately 20 seconds each. The (18 + 1 anechoic) x 3 convolved sound fields (SFs)/wave files were then played back to the subjects using supra-aural headphones SONY MDR-V600 through a 6 channel headphone bus. Averaged play back levels of 3 sources were set at (peak level) 87-88 dBA for orchestra (73 dBA-anechoic), 69-70 dBA for trumpet (66 dBA-anechoic), and 70-71 dBA for piano (68 dBA-anechoic). The playback levels were set according to an agreement from a few experienced listeners (acousticians); the peak level measurements of the sound fields were made to check if there could be any harm to the test subjects. Forty-five subjects had participated in this experiment; 15 subjects for each source type. Each subject evaluated 12 randomly picked sound fields + 1 anechoic + 3 learning samples (at the beginning). A total sample size of 10 for each sound field was used to obtain the final results. The questionnaire used was a seven-point bi-polar rating scale developed from Cervone33 and Gold34 shown in Appendix A. The additional perceived acoustic quality texture introduced in this study was defined according to Beraneks description35 mentioned in Chapter 2. 33 Cervone, R. P. (1990). Designing for Music: The Subjective and Objective Evaluation of Three Rooms for Music Listening. School of Architecture Gainesville, University of Florida. 34 Gold, M. A. (1994). An Evaluation of Perceived Acoustic Spatial ImpressionIbid. 35 Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture NY, Springer. 28

PAGE 29

Table 3-1. Absorption and scattering coefficients specified in CATT modeling. % Absorption Coefficients % Scattering Coefficients Surfaces 125 Hz 250 Hz 500 Hz 1 kHz 2 kHz 4 kHz 125 Hz 250 Hz 500 Hz 1 kHz 2 kHz 4 kHz Reflecting panels 1 1 1 1 1 1 10 10 10 10 10 10 Diffusing panels 1 1 1 1 1 1 30 40 50 90 90 90 Room envelopes 90 90 90 90 90 90 10 10 10 10 10 10 (a) (b) Figure 3-1. The computer generated models created to produce five specular/diffuse reflections arrival to the receiver within 40 msec. 29

PAGE 30

Table 3-2. The 9 designed impulse responses showing various sets of arrival specular reflections generated by CATT-Acoustics. Notes: S5R -stands forShort time delay with 5 specular Reflectors M10R -stands forMedium time delay with 10 specular Reflectors L20R -stands forLong time delay with 20 specular Reflectors 30

PAGE 31

Table 3-3. The 9 designed impulse responses showing various sets of arrival diffuse reflections generated by CATT-Acoustics. Notes: S5D -stands forShort time delay with 5 Diffuse reflectors M10D -stands forMedium time delay with 10 Diffuse reflectors L20D -stands forLong time delay with 20 Diffuse reflectors 31

PAGE 32

CHAPTER 4 RESULTS AND DISCUSSION The results obtained from the listening test can be evaluated according to five physical dimensions of room acoustic measurements: 1) intensity, 2) frequency, 3) time, 4) direction, and 5) distribution (temporally and directionally) of acoustical energy, which involves how the prior 4 acoustical characteristics interact. The latter is the additional dimension of room acoustic provided by the room environment. This experiment was set up to study the temporal distribution of acoustical energy and its effects to perceived acoustic qualities. The analysis of the experiment has been carried out through 3 aspects from acoustical point of viewsource, path, and receiver. Source Characteristics Meyer (1978) analyses the temporal structure of each note produced by different instruments into three parts: 1) The starting transient, this is the period during which the note develops out of complete silence until it reaches its stationary state. 2) The stationary state, this is the period during which the note is not subject to any change. 3) The decay, this is the period during which the note resounds between the end of the stimulation until complete silence. The duration of the starting transient can be measured from the very first beginning to the point when a level of 3 dB below the final level is reached. The decay time is measured over 60 dB, as is the reverberation time.36 The acoustical characteristics of the orchestral music, trumpet, and piano (sources used in this experiment) are briefly analyzed as follows. Orchestral Music An orchestra comprises a multitude of different instruments which interact partly in solo fashion and partly gathered in groups. A medium sized orchestra may generate a fortissimo sound power level of 118 dB On the other hand, the lower limit of the orchestral dynamic 36 Meyer, J. (1978). Acoustics and the Performance of Music Frankfurt/Main, Verlag Das Musikinstrument. 32

PAGE 33

range may be determined by only one instrument playing pianissimo, for example, a clarinet generating a sound power level of about 58 dB. This leads to an overall dynamic range of about 60 dB for a symphonic orchestra. 37 The orchestra occupies an extended stage area; a medium sized symphonic orchestra might take space on the order of 15 m wide and 12 m deep.38 The directivity of the orchestra depends on the instrument which is used and the arrangement of the multitude of instruments over the stage area. An orchestra shell plays an important role in blending the sounds from individual instruments providing a good sending end to the audience. At the same time it is crucial that the stage enclosure should provide a good distribution of sound throughout the stage for musicians to hear each other well for the accuracy of the ensemble. The temporal structure of the sound of the orchestra depends on the music piece and the instruments used to play/deliver the designed tempos. Similarly, the frequency spectrum of the sound of the orchestra also varies from simply a set of harmonics of notes played by only one instrument to very rich in tones when they are played in tutti of the entire orchestra. The stage enclosure and the room environment can enhance the perceived music experience in the hall when it is well designed to support attacks, and various tempos of the piece produced by different music instruments. Trumpet Brass instruments are the loudest instruments in the symphony orchestra. 39 The dynamic range of trumpet, measuring the sound pressure level when playing an extreme fortissimo 37 Meyer, J. (1993). "The sound of the orchestra." J. Audio Eng. Soc. 41(4): 203-212. 38 Ibid. 39 Meyer, J. (1978). Acoustics and the Performance of Music Frankfurt/Main, Verlag Das Musikinstrument. 33

PAGE 34

comparing with an extreme pianissimo, is about 30 phons for low notes (53 dBA-80 dBA) narrow to 13 phons for high notes (83 dBC-93 dBC), measured at a distance of 14 m.40 The shape and size of the bell and its immediate tubing determine the directional radiation of the trumpet, which is approximately rotationally symmetric around the bell axis. The relationship of the wavelength of the radiated sound and the bell dimension results in omnidirectional characteristics for frequencies below 500 Hz. At higher frequencies the higher intensity is radiated in the direction of the bell; at 2000 Hz and above, most of the sound energy is concentrated on the axis of the instrument41 as shown in Figure 4-1. The average starting transient of the high notes is about 40 msec; it gets longer at lower frequencies msec in the middle register and 180 msec in the low register. The decay time of the wind instruments is very short, between 70 and 150 msec, due to the fact that the energy stored in the vibrating air is very small; therefore in practice it is hard for one to hear any resonance decay.42 The frequency content of the notes produced by trumpet can be measured during its stationary state. Similar to other brass instruments, the trumpet produces rich overtones. The components of its harmonic sound reach up to very high frequencies; the upper limit of the spectrum extends up to 8000 Hz in the high register (Figure 4-2 shows the frequency spectrum of some notes produced by the horn which is very similar to the trumpets). Additionally the region where the strongest partial is found is relatively high in frequency. Therefore a bright, brilliant, and pure tone quality is produced. 40 Burghauser, J., und Spelda, A. (1971). Die akustischen Grundlagen der Instrumentation Rogensberg. 41 Meyer, J. (1978). Acoustics and the Performance of Music Frankfurt/Main, Verlag Das Musikinstrument. 42 Ibid. 34

PAGE 35

Note that the duration of a tone impulse affects the perceived loudness43; therefore the sound of the trumpet gets louder, the longer the duration of the tone is perceived. Figure 4-1. Main directional radiation (0-3 dB) of the trumpet in the vertical plane.44 Figure 4-2. Sound spectra of a horn in different keys.45 43 Zwicker, E., und Feldtkeller, R. (1967). Das Ohr als Nachrichtenempfnger Stuttgart. 44 Ibid. 35

PAGE 36

Piano The piano has a rather consistent dynamic range throughout the frequency range at about 35 phons (50 dBA-85 dBC) for bass and middle notes, and 33 phons (37 dBA-70 dBC) for higher notes, measured at 10 m distance.46 The vibrations of the sound board and the lid position are two major factors that determine the patterns of the directional radiation of a grand piano. Sound is radiated directly upward and is reflected off the lid and also downward and is reflected from the floor. Figure 4-4 shows directional characteristics of the grand piano for low, middle, and high registers as a polar diagram in a vertical plane. The sound radiation is relatively symmetrical in the low register; the directional characteristics affected by the lid begin to show in the middle register. Towards the high register, the influence of the lid on the directivity are pronounced, with a strong directionality between 15-35 shown in even the fundamentals of the notes (at 1000 Hz).47 The starting transient of the piano sound takes between 20 and 30 msec for low register and gets shorter towards the higher register about between 10 and 15 msec. No stationary sound situation is formed on the piano.48 A grand piano resonates for a very long time producing a decay time between 0.5 and 10 sec. Two decay slopes were found; first, the intensity rapidly decreases then remains decaying slowly for a long time. On average the decay time for the low register takes longer than that for the higher ones.49 Figure 4-3 shows the frequency dependent reverberation time of a grand piano for some C keys. 45 Ibid. 46 Burghauser, J., und Spelda, A. (1971). Die akustischen Grundlagen der Instrumentation Rogensberg. 47 Meyer, J. (1978). Acoustics and the Performance of Music Frankfurt/Main, Verlag Das Musikinstrument. 48 Ibid. 49 Fletcher, H., Blackham, E. D., & Stratton, R. (1962). "Quality of piano tones." J. Acoust. Soc. Am. 34: 749-761. 36

PAGE 37

The sound of a grand piano is mainly created by the resonant quality of the instrument. With a sudden change characteristic of the piano sound (without a stationary condition), an analysis of a sound into a set of harmonics is not possible.50 The frequency content is not formed of discrete harmonic single frequencies, but instead are continuous functions of frequency. The timbre of the piano sound therefore is characterized by the anharmonicity of the overtones and the noise component. The partials are underlaid by a continuum which has a percussive character and moreover, is colored by the instruments strongest resonances.51 Figure 4-3. Reverberation time of a grand piano for some C-keys using the right pedal.52 50 Meyer, J. (1978). Acoustics and the Performance of Music Frankfurt/Main, Verlag Das Musikinstrument. 51 Ibid. 52 Ibid. 37

PAGE 38

Figure 4-4. Patterns of directional radiation of a grand piano in a vertical plane.53 53 Ibid. 38

PAGE 39

Initial Hypothesis: Effects from Path Hypotheses were set up according to the characteristics of the temporal distribution of sound energy in the IRs regardless of the source type. These hypotheses are based on the definitions given for each acoustical quality presented in the questionnaire (see Appendix A) and many acoustical theories widely accepted as quantitative measures for those perceived acoustical qualities, for example, Haas Effect, Sabines theory (RT), G-strength (G10), clarity index (C80), initial time delay gap (ITDG), and lateral fraction (LF). Please refer to standard acoustic books for further explanation on these acoustic theories and parameters. 54,55,56,57,58 1) Loudness a. The larger the number of reflections the louder the signal (higher total energy). b. The sooner the arrival time of specular reflections the louder the signal (less energy reduction due to distance). c. Diffuse reflections might soften the attacks and therefore reduce the perceived loudness. 2) Clarity a. The sooner the arrival of specular reflections the clearer the signal (high C80 energy within 80 msec). 54 Mehta, M., Johnson J., & Rocafort, J. (1999). Architectural Acoustics: Principles and Design NJ, Prentice-Hall, Inc. 55 Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture NY, Springer. 56 Cremer, L., & Muller, H. A. (1982). Principles and Applications of Room Acoustics NY, Applied Science Publishers Ltd. 57 Barron, M. (1993). Audirorium Acoustics and Architectural Design NY, Routledge. 58 Egan, M. D. (1988). Architectural Acoustics NY, McGraw-Hill, Inc. 39

PAGE 40

b. The larger the number of specular reflections the clearer the signal (higher total energy, however these reflections must arrive within the useful time period). 3) Intimacy a. The sooner and denser the arrival of specular and diffuse reflections the more intimate the perceived signal. 4) Reverberance a. The later the arrival time of specular reflections the more reverberance (RT gets longer). b. The larger the number of specular reflections the clearer the perceived reverberance (more energy). c. Early diffuse reflections may not play any role on the perceived reverberance. 5) Echoes a. The later the arrival time of specular reflections the clearer the perceived echoes. b. The larger the number of late arrival specular reflections the clearer the perceived echoes. c. Diffuse reflections should soften the effect of echoes (reduce the intensity of the long delay reflections). 6) Source width a. Diffuse reflections should widen the source width. 7) Texture a. The diffuse reflections should smooth the perceived texture. 40

PAGE 41

Receiver The fifty subjects who participated in this experiment are architectural students whose ages ranged from 18-30 years old. Fifteen of the 50 students play some instruments or are musicians. Figure 4-5 shows that a large number of the subjects prefer Rock music; Classical, Jazz, and Pop music are second in popularity among them. These subjects represent a young generation of general audiences. Data from 47 out of 50 were used for the results analyses, due to the fact that there were a few students with mild to moderate hearing loss. Music Preference0510152025303540ClassicalJazzBluesRockPopRapCountryReaggaeNo. of Subjects (47 total) Figure 4-5. Music preference of 47 subjects who participated in the listening test. Results and Discussion The test results can be divided into 3 source groupsorchestral music, trumpet, and piano; 19 sound fields (SFs) in each group have been evaluated separately. Normalized data shown in this section were produced by first, for each rated perceived quality, the average of all 19 SFs was called total average, then the averaged data of each SF was subtracted by the total average. The graphs show the total average on the vertical axis set as . In Figure 4-6, the ratings of perceived loudness of the three source types were compared together in one graph. 41

PAGE 42

The process is similar to what was mentioned above only the total average in this case was the averaged data across all three sources. Each graph shows 19 sound fields arranged from left to rightanechoic sound field, S5R, S10R, S20R, M5R, M10R, M20R, L5R, L10R, L20R, S5D, S10D, S20D, M5D, M10D, M20D, L5D, L10D, L20D. S, M, or L stand for Short, Medium, or Long time delay. Numbers 5R, 10R, 20R, 5D, 10D, or 20D stand for the number of reflecting or diffusing panels provided to obtain each simulated sound field. Please refer to Chapter 3 for more information on the method. Due to the fact that the perceived differences among sound fields caused by specular and diffuse reflections are very subtle, the standard statistical significant tests cannot be applied in this context. Instead, the trends of the resulting data were observed to understand and explain perceived acoustical qualities due to specular and diffuse reflections. Loudness The results from Figure 4-6 show that, for orchestral music and piano, additional reflections (regardless of being either specular or diffuse reflections) provide an increase in perceived loudness as compared to the anechoic signal, which was expected. However, in the case of the trumpet, the anechoic sound field was perceived as louder then when diffuse or specular reflections were added. This unexpected result does not agree with the law of the addition of energies.59 The explanation to this phenomenon is not known. Finally, there is no trend on the difference in perceived loudness among the designed impulse responses as initially hypothesized (according to the temporal energy distribution characteristics). The assertions can be made, first, the receiver in this simulation might be 59 Haas, H. (1972). "The influence of a single echo on the audibility of speech." J. Audio Eng. Soc. 20(2): 146-159. 42

PAGE 43

located too close to the source (14 m), therefore, the direct sound might dominate the effect on perceived loudness. If the receiver were to be located farther away, the effects on the perceived loudness due to the additional specular reflections might emerge. Second, the number of 5 large reflecting panels (3x3m) might produce minimally sufficient energy comparable to that of 10 and 20 reflecting panels. The complete results are available in Appendix B. Figure 4-6. Normalized perceived loudness of the three sources: Orchestral music, Trumpet, and Piano. The orange columns are anechoic signals of each group. Clarity The results show a subtle increase in perceived clarity for orchestra and piano when early specular reflections were added (within 40 msec of the direct sound) (Figure 4-7a, b). For orchestral music, specular reflections within 160 msec of the direct sound are beneficial for perceived clarity. What is interesting is the trend showing the improvement of perceived clarity for piano and trumpet when a large number of early diffuse reflections arrived sooner (within 40 msec) as shown in Figure 4-7b,c. This finding is important because it shows that for the same amount of early sound energy (specular or diffuse reflections) present within the same time windows, the 43

PAGE 44

improvement in perceived clarity provided by diffuse reflections rather than specular reflections can not be captured by clarity index (C80). Clarity (O)-1.5-1.0-0.50.00.51.01.52.0Normalized Preference (a) A nechoic S 05 R S 20 D M 05 D M 10 D M 20 D M 05 R M 10 R M 20 R L 05 R L 20 D L 05 D L 10 D S 05 D S 10 D L 10 R L 20 R S 10 R S 20 R Clarity (T)-1.5-1.0-0.50.00.51.01.52.0Normalized Preference (b) Clarity (P)-1.5-1.0-0.50.00.51.01.52.0Normalized Preference (c) A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D Figure 4-7. Clarity: normalized averaged ratings of perceived clarity of (a) Orchestral music, (b) Trumpet, and (c) Piano. 44

PAGE 45

Intimacy The initially set up hypothesis fails for orchestra and trumpet. Short time arrival of a large number of specular reflections does not help improve perceived intimacy of the orchestra and trumpet. This might be because of the control setting for the same initial time delay gap (ITDG) at 20 msec for all sound fields except the anechoic one. However, the results from piano show strong improvement (Figure 4-8) on perceived intimacy when the specular or diffuse reflections arrive sooner. Moreover, early diffuse reflections (within 40 msec) improve perceived intimacy very significantly while later arrival of specular reflections reduces the perceived intimacy of the piano sound. Intimacy (P)-1.5-1.0-0.50.00.51.01.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D Figure 4-8. Intimacy: normalized averaged ratings of perceived intimacy of the Piano source. 45

PAGE 46

Reverberance (O)-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference (a) Reverberance (T)-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference (b) Reverberance (P)-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D A nechoic S 05 R (c) Figure 4-9. Reverberance: normalized averaged ratings of perceived reverberance of (a) Orchestral music, (b) Trumpet, and (c) Piano. 46

PAGE 47

Echoes (O)-2.0-1.5-1.0-0.50.00.51.01.52.02.5Normalized Preference (a) Echoes (T)-2.0-1.5-1.0-0.50.00.51.01.52.02.5Normalized Preference (b) Echoes (P)-2.0-1.5-1.0-0.50.00.51.01.52.02.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D A nechoic S 05 R A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (c) Figure 4-10. Echoes: normalized averaged ratings of perceived echoes of (a) Orchestral music, (b) Trumpet, and (c) Piano. 47

PAGE 48

Reverberance and Echoes The results shown in Figure 4-9 seem to support all the hypotheses set up earlier that the perceived reverberance increases as the arrival time of specular reflections increases. The perceived reverberance also gets clearer when the number of specular reflections increase regardless of the source type. In addition, the time delay and the amount of diffuse reflections do not affect perceived reverberance. Perceived echoes were evaluated similarly to perceived reverberance as shown in Figure 4-10. A long time delay and a large number of specular reflections increase the perceived echoes while diffuse reflections do not produce echoes in any time period except for the trumpet sound source. The trumpet was particularly sensitive to both specular and diffuse reflections arriving between 80 and 160 msec after the direct sound. Source Width Additional reflections, either specular or diffuse, arriving at any time period help widen the perceived source width of the sound fields as compared to the anechoic sound field (Figure 4-11). Only the trumpet source produces results with a clear trend. It shows that the longer the delay of arrival of specular reflections the wider the perceived source width. An improvement in perceived source width also appears when diffuse reflections arriving between 80-160 msec after the direct sound were added to the sound field. The perceived source width for trumpet is strongly affected according to the change in energy distribution patterns; this might be due to the high directionality, and the very short decay time characteristics of the trumpet. 48

PAGE 49

Source Width (O)-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference (a) Source Width (T)-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference (b) Source Width (P)-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D A nechoic S 05 R (c) Figure 4-11. Source width: normalized averaged ratings of perceived source width of (a) Orchestral music, (b) Trumpet, and (c) Piano. 49

PAGE 50

Texture (O)-2.0-1.5-1.0-0.50.00.51.01.52.0Normalized Preference (a) Texture (T)-2.0-1.5-1.0-0.50.00.51.01.52.0Normalized Preference (b) Texture (P)-2.0-1.5-1.0-0.50.00.51.01.52.0Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D A nechoic S 05 R (c) Figure 4-12. Texture: normalized averaged ratings of perceived texture of (a) Orchestral music, (b) Trumpet, and (c) Piano. 50

PAGE 51

For orchestral music and piano, the results are not in accord with the initial expectation. One would expect that the behavior of diffuse reflections might help improve the perceived source width to some extent due to the temporal and directional smearing effects produced by diffuse reflections. This might be the case of the wrong choice of words. The same study should be repeated with the rating of perceived spaciousness rather than source width to see if diffuse reflections help improve perceived spaciousness of the sound field. Texture The results from Figure 4-12 show that overall diffuse reflections help improve perceived texture of all source typesorchestra, trumpet and piano, as expected. For orchestral music and trumpet, perceived texture was rated highest when diffuse reflections arrive within 40-80 msec of the direct sound. Diffuse reflections arriving prior to or later than the 40-80 msec period, are not rated as high. A clear trend is shown in the case of piano music, late arrival of specular reflections reduced rated perceived texture. Diffuse reflections in general are more preferred than specular reflections; the sooner arrival of diffuse reflections the better-perceived texture. The large amount of diffuse reflections arriving within 40 msec of the direct sound was rated highest for perceived texture (Figure 4-12c). Overall Impression When the ratings of perceived overall impression were compared with other perceived music qualities, some similar patterns were found. The following comparisons reveal evidence that special attention on particular perceived music quality should be addressed differently for different types of music playing in a room due to the dissimilar temporal structures and acoustical characteristics of the sources. The trends for orchestra and trumpet are very weak; only piano sound shows strong trend. 51

PAGE 52

Source Width (O)-2.5-2.0-1.5-1.0-0.50.00.51.01.5_anechoi1S 05 R1S 10 R1S 20 R2M 05 R2M 10 R2M 20 R3L 05 R3L 10 R3L 20 R4S 05 D4S 10 D4S 20 D5M 05 D5M 10 D5M 20 D6L 05 D6L 10 D6L 20 DNormalized Preference Intimacy (O)-1.5-1.0-0.50.00.51.01.5_anechoi1S 05 R1S 10 R1S 20 R2M 05 R2M 10 R2M 20 R3L 05 R3L 10 R3L 20 R4S 05 D4S 10 D4S 20 D5M 05 D5M 10 D5M 20 D6L 05 D6L 10 D6L 20 DNormalized Preference (a) (b) Texture (O)-2.0-1.5-1.0-0.50.00.51.01.52.0_anechoi1S 05 R1S 10 R1S 20 R2M 05 R2M 10 R2M 20 R3L 05 R3L 10 R3L 20 R4S 05 D4S 10 D4S 20 D5M 05 D5M 10 D5M 20 D6L 05 D6L 10 D6L 20 DNormalized Preference Overall Impression (O)-2.0-1.5-1.0-0.50.00.51.01.52.0_anechoi1S 05 R1S 10 R1S 20 R2M 05 R2M 10 R2M 20 R3L 05 R3L 10 R3L 20 R4S 05 D4S 10 D4S 20 D5M 05 D5M 10 D5M 20 D6L 05 D6L 10 D6L 20 DNormalized Preference (c) (d) Figure 4-13. A comparison on ratings of perceived acoustic qualities: (a) source width, (b) intimacy, (c) texture, and (d) overall impression of Orchestral music. Piano The results shown in Figure 4-15 express a strong similar trend among the ratings of perceived clarity, intimacy, texture, and overall impression of the piano sound. All of these 4 perceived acoustic qualities were rated in favor of early arrival of diffuse reflections (within 40 msec of the direct sound). Due to the impulsive characteristic and the long resonant decay of the piano sound, this trend shows that there is a need for early diffuse reflections from room acoustic to support the transient notes, however there is no need for later arrival specular reflections to interfere with the instruments own reverberation (pianos resonant decay). In addition, the ratings of perceived texture and overall impression correspond well to one another for all 3 sources which means that the perceived texture of the sound field has a direct impact on the perceived overall impression (Figure 4-13, 4-14, 4-15). 52

PAGE 53

Texture (T)-2.0-1.5-1.0-0.50.00.51.01.52.0_anechoi1S 05 R1S 10 R1S 20 R2M 05 R2M 10 R2M 20 R3L 05 R3L 10 R3L 20 R4S 05 D4S 10 D4S 20 D5M 05 D5M 10 D5M 20 D6L 05 D6L 10 D6L 20 DNormalized Preference Overall Impression (T)-2.0-1.5-1.0-0.50.00.51.01.52.0_anechoi1S 05 R1S 10 R1S 20 R2M 05 R2M 10 R2M 20 R3L 05 R3L 10 R3L 20 R4S 05 D4S 10 D4S 20 D5M 05 D5M 10 D5M 20 D6L 05 D6L 10 D6L 20 DNormalized Preference (a) (b) Figure 4-14. A comparison on ratings of perceived acoustic qualities: (a) texture, and (b) overall impression of Trumpet. Clarity (P)-1.5-1.0-0.50.00.51.01.52.0_anechoi1S 05 R1S 10 R1S 20 R2M 05 R2M 10 R2M 20 R3L 05 R3L 10 R3L 20 R4S 05 D4S 10 D4S 20 D5M 05 D5M 10 D5M 20 D6L 05 D6L 10 D6L 20 DNormalized Preference Intimacy (P)-1.5-1.0-0.50.00.51.01.5_anechoi1S 05 R1S 10 R1S 20 R2M 05 R2M 10 R2M 20 R3L 05 R3L 10 R3L 20 R4S 05 D4S 10 D4S 20 D5M 05 D5M 10 D5M 20 D6L 05 D6L 10 D6L 20 DNormalized Preference (a) (b) Texture (P)-2.0-1.5-1.0-0.50.00.51.01.52.0_anechoi1S 05 R1S 10 R1S 20 R2M 05 R2M 10 R2M 20 R3L 05 R3L 10 R3L 20 R4S 05 D4S 10 D4S 20 D5M 05 D5M 10 D5M 20 D6L 05 D6L 10 D6L 20 DNormalized Preference Overall Impression (P)-2.0-1.5-1.0-0.50.00.51.01.52.0_anechoi1S 05 R1S 10 R1S 20 R2M 05 R2M 10 R2M 20 R3L 05 R3L 10 R3L 20 R4S 05 D4S 10 D4S 20 D5M 05 D5M 10 D5M 20 D6L 05 D6L 10 D6L 20 DNormalized Preference (c) (d) Figure 4-15. A comparison on ratings of perceived acoustic qualities: (a) clarity, (b) intimacy, (c) texture, and (d) overall impression of Piano. 53

PAGE 54

Conclusions The effects of temporal distribution of specular and diffuse reflections can be concluded as follows. Specular reflections Specular reflections arriving within 40 msec of the direct sound help improve perceived clarity for all 3 source types and improve perceived intimacy for orchestral and piano music. Specular reflections arriving between 40-80 msec of the direct sound help improve perceived clarity for orchestral music. Specular reflections arriving between 80-160 msec of the direct sound are not desirable. Strong long delayed reflections provide harsh texture for all source types, decrease perceived clarity of trumpet and piano sounds and decrease perceived intimacy for piano. These strong reflections also increase perceived echoes and reverberance. Diffuse reflections Diffuse reflections arriving within 80 msec of the direct sound improve perceived texture and intimacy for all 3 sources, improve perceived clarity for trumpet and piano, and reduce perceived glare (loudness) for trumpet. Diffuse reflections arriving between 80-160 msec of the direct sound help widen the perceived source width for orchestra and trumpet, increase perceived intimacy for orchestra, and reduce perceived echoes for all 3 sources. Overall, all diffuse sound fields improved perceived texture and preserved perceived reverberance while not providing echoes for all 3 sources. Also the study shows a strong relationship between the improvement of perceived texture and overall impression. Figure 4-16 shows the types of reflections and time delays that would improve perceived overall impression for 3 sourcesorchestral music, trumpet, and piano. 54

PAGE 55

Orchestral music needs a lot of support from both specular and diffuse reflections arriving within 80 msec of the direct sound to provide strength for attacks and fullness of tone. A large number of diffuse reflections arriving after 80 msec are desirable. Trumpet music needs a large number of diffuse reflections to arrive a little later (between 40-80 msec after the direct sound) to prolong its decay time without increasing its loudness and widen its source width. Piano music prefers a large number of diffuse reflections to arrive early (within 40 msec) to support its impulsive sound providing support for running notes without adding excessive reverberation to the instruments own long resonant decay. Applications In addition to the required acoustic criteria, the results from this study can be used partly to suggest designers how to locate reflectors or diffusers in their amphitheaters or concert halls as shown in Figures 4-17 and 4-18. Figure 4-17 shows 3 ellipsoids representing boundaries where reflecting panels can be located to provide the 1st order reflections arriving to the front receiver within 3 time periods40, 80, and 160 msec after the direct sound. Figure 4-18 shows 3 different boundaries for 3 different receiver locations where reflecting panels can be located to provide the 1st order reflections to arrive at the receivers within 40 msec after the direct sound. Another application is for recording purpose; electronic sound effects should be added to the anechoic sound differently for different sound sources due to their dissimilar temporal structures to provide proper perceived music preference. 55

PAGE 56

Specular reflections Diffuse reflections within 40 80 160 msec Best Best Specular reflections Diffuse reflections 40 80 160 msec Best Specular reflections Diffuse reflections 40 80 160 msec Best --Provide support to improve perceived overall impression 5 reflections 10 reflections 20 reflections Orchestra 5 reflections 10 reflections 20 reflections Trumpet Piano 5 reflections 10 reflections 20 reflections Figure 4-16. Summary of the preferred arrival diffuse and specular reflections to provide improvement on perceived overall impression for 3 sourcesOrchestral music, Trumpet, and Piano. 56

PAGE 57

Finally, knowledge gained from this study provides rough guidance for a search for a quantitative measure that should relate well to the perceived texture. The new measure should be: 1) frequency dependent at mid and high frequenciesdue to the fact that human ears are more spatially sensitive to high frequency sounds60; 2) time delay dependentthe perceived texture of sounds in rooms improves as time goes by however the time windows which need to be carefully treated for each instrument are different; 3) providing acceptable range of perceived texture (degree of smoothness) for each time windowtoo much diffuse reflections are sometimes not desirable, the spatial and frequency variation in time which supports perceived spaciousness can get lost if perfectly diffuse sound field occurs too soon. Figure 4-17. Three ellipsoids representing boundaries where reflecting panels can be located within to provide the 1st order reflections arriving to the front receiver within 3 time periods: 40, 80, and 160 msec after the direct sound. 60 Due to the fact that 1) human ear is more sensitive to high frequencies, 2) head shadow effect becomes more prominent at higher frequencies, and 3) higher tones (frequencies) are highly directional due to their productiontheir wavelengths are smaller than the instrumentss dimension in which they were produced. 57

PAGE 58

Figure 4-18. Three different boundaries for 3 different receiver locations where reflecting panels can be located to provide the 1st order reflections to arrive at the receivers with 40 msec after the direct sound. Comments Some comments provided by the subjects who had listened to 16 sound fields (SFs) have been collected. They had listened to each SF twice and evaluated each SF until they finished evaluating 16 of them. The comments are as follows: 1) The SFs are too short (20 sec. long) especially for orchestral music. 2) Too many acoustical qualities to evaluate for each SF. 3) Confusing range bi-polar rating scale, some are ranged between bad-good, some are ranged between two extremes (bad-good-bad). 4) The sound level is too soft for orchestral music. 5) Listening to the same short piece twice for 16 times exhausts the ears resulting in meaningless evaluations in some SFs. 6) Listeners dont know how to evaluate the first few SFs, they dont know what to compare it to (mostly compare to the previous one). There are some suggestions on choosing a longer piece of music and play each once. 58

PAGE 59

CHAPTER 5 CONCLUSION AND FUTURE WORK Conclusion The differences in perceived acoustical effects produced by specular reflections and diffuse reflections are very subtle. No significant tests according to statistical norms can be used in this context. For those who are looking for numerical description and discussion should be disappointed to read my work and see only, some said, casual explanations. I think that making something simple look difficult, complicated, and hard to comprehend by showing parades of numbers to convince others does not take much effort. On the contrary, being able to present the most complicated and mysterious issue in the simplest form that everyone can understand requires a higher level of understanding, originality, and creativity. There are many ways to look at a problem, many approaches and methods that can be used to search for knowledge. The works of many researchers (including the authors) in the field of diffuse reflections presented here are just some of the various approaches used to address the same problem. Going through a series of failed experiments can be truly painful and heartbroken; however, to be able to contemplate and appreciate the facts, logic, and reasons why the experiments failed is something that cannot be shared or talked about. A series of experiments presented in Appendix C and D are the authors prior works which more or less have provided some clues to finally shape the well-designed experiment for this dissertation. Readers are welcome to walk through my anguished search for knowledge in those Appendices for their own appreciation or choose to skip them. Future Work This dissertation provides some logical evidence to the understanding of diffuse reflections phenomena. Many assertions and explanations made here still call for more concrete scientific 59

PAGE 60

clarification. The final investigation was set to simplify the room acoustic situations to accentuate the effects of the temporal distribution on perceived music quality. Many results shown here might disappear when the impulse responses get too crowded by rooms reflections as happens in real room situations. The author hopes that this initial study will open the floor for many questions to rise and challenge for many more creative investigations to emerge to prove our curiosity and imaginations. Some of the future investigations that should be carried out include: 1) A revision of the final experiment presented here according to the listeners comments and with better equipment (better headphones); a revision of the questionnairereplace perceived source width with spaciousness, a rearrangement of the mixture of arrival specular and diffuse reflections at each time window, and an addition of the rooms reverberant tail to the SFs. 2) Study the impulse responses of different room shapes to search for the rooms signature which should provide answers to why the traditional shoe-box shape provides good room acoustic. Upon knowing the room shapes signatures, designers can then be able to alter their room signatures if needed. 3) A search for quantitative measure/parameter that can describe the characteristics of the distribution of energy of the IRs that can relate well with the perceived acoustical quality of texture. 60

PAGE 61

APPENDIX A QUESTIONNAIRE ACOUSTICS EVALUATION SHEET DEFINITIONS LOUDNESS: The overall loudness or strength of the sound that you are hearing CLARITY: The degree to which notes or words are distinctly separated in time and clearly heard. INTIMACY: The auditory impression of the apparent closeness of the orchestra. REVERBERANCE: The blending of sounds into subsequent following sounds. The persistence of sound in the space. SOURCE WIDTH: The apparent widening of the sound source. The source occupying an area as opposed to a single point in space. TEXTURE: The smoothness or harshness of sound persisting in a space ECHOES: Long delayed reflections that are clearly audible. OVERALL IMPRESSION: Overall impression of the acoustical quality that you are hearing. PERSONAL DATA (optional) Music preference (circle one or more): Classical Jazz Blues Rock Pop Rap Other. Number of orchestral concerts attended (circle one). None 1-10 11-20 21-30 30 or more Are you a musician? Yes No Do you have normal hearing? Yes No 61

PAGE 62

62 VIRTUAL SOUND FIELD EVALUATION SHEET LOUDNESS 1 2 3 4 5 6 7 0 Not loud too loud cannot enough tell CLARITY 1 2 3 4 5 6 7 0 Not clear extremely cannot enough clear tell INTIMACY 1 2 3 4 5 6 7 0 Not intimate extremely cannot enough intimate tell REVERBERANCE 1 2 3 4 5 6 7 0 Not reverberant too cannot enough reverberant tell SOURCE WIDTH 1 2 3 4 5 6 7 0 Not wide extremely cannot wide tell TEXTURE 1 2 3 4 5 6 7 0 Not smooth extremely cannot (harsh) smooth tell ECHOES 1 2 3 4 5 6 7 0 None detected clearly heard cannot tell OVERALL 1 2 3 4 5 6 7 0 IMPRESSION very bad very good cannot tell COMMENTS:

PAGE 63

63 APPENDIX B RESULTS FROM LISTENING TEST Loudness (O)-3.0-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference (a) A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D Loudness (T)-3.0-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference (b) A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D Loudness (P)-3.0-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference (c) A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D Figure B-1. Loudness: normalized averaged ratings of perceived loudness of (a) Orchestral music, (b) Trumpet, and (c) Piano.

PAGE 64

64 Clarity (O)-1.5-1.0-0.50.00.51.01.52.0Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (a) Clarity (T)-1.5-1.0-0.50.00.51.01.52.0Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (b) Clarity (P)-1.5-1.0-0.50.00.51.01.52.0Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (c) Figure B-2. Clarity: normalized averaged ratings of perceived clarity of (a) Orchestral music, (b) Trumpet, and (c) Piano.

PAGE 65

65 Intimacy (O)-1.5-1.0-0.50.00.51.01.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (a) Intimacy (T)-1.5-1.0-0.50.00.51.01.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (b) Intimacy (P)-1.5-1.0-0.50.00.51.01.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (c) Figure B-3. Intimacy: normalized averaged ratings of perceived intimacy of (a) Orchestral music, (b) Trumpet, and (c) Piano.

PAGE 66

66 Reverberance (O)-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (a) Reverberance (T)-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (b) Reverberance (P)-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (c) Figure B-4. Reverberance: normalized averaged ratings of perceived reverberance of (a) Orchestral music, (b) Trumpet, and (c) Piano.

PAGE 67

67 Source Width (O)-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (a) Source Width (T)-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (b) Source Width (P)-2.5-2.0-1.5-1.0-0.50.00.51.01.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (c) Figure B-5. Source width: normalized averaged ratings of perceived source width of (a) Orchestral music, (b) Trumpet, and (c) Piano.

PAGE 68

68 Texture (O)-2.0-1.5-1.0-0.50.00.51.01.52.0Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (a) Texture (T)-2.0-1.5-1.0-0.50.00.51.01.52.0Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (b) Texture (P)-2.0-1.5-1.0-0.50.00.51.01.52.0Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (c) Figure B-6. Texture: normalized averaged ratings of perceived texture of (a) Orchestral music, (b) Trumpet, and (c) Piano.

PAGE 69

69 Echoes (O)-2.0-1.5-1.0-0.50.00.51.01.52.02.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (a) Echoes (T)-2.0-1.5-1.0-0.50.00.51.01.52.02.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (b) Echoes (P)-2.0-1.5-1.0-0.50.00.51.01.52.02.5Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (c) Figure B-7. Echoes: normalized averaged ratings of perceived echoes of (a) Orchestral music, (b) Trumpet, and (c) Piano.

PAGE 70

70 Overall Impression (O)-2.0-1.5-1.0-0.50.00.51.01.52.0Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (a) Overall Impression (T)-2.0-1.5-1.0-0.50.00.51.01.52.0Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (b) Overall Impression (P)-2.0-1.5-1.0-0.50.00.51.01.52.0Normalized Preference A nechoic S 05 R S 10 R S 20 R M 05 R M 10 R M 20 R L 05 R L 10 R L 20 R S 05 D S 10 D S 20 D M 05 D M 10 D M 20 D L 05 D L 10 D L 20 D (c) Figure B-8. Overall impression: normalized averaged ratings of perceived overall impression of (a) Orchestral music, (b) Trumpet, and (c) Piano.

PAGE 71

APPENDIX C PILOT STUDY I Computer Modeling of Charlotte Auditorium Charlotte auditorium was chosen to study the behavior of diffuse reflections at 12 receivers. Two computer models were built using CATT-Acoustic v.8: one having pyramid diffusing panels on the side and rear walls, and curved diffusers as acoustic clouds; another one without any diffusing surfaces as shown in Figure C-1. (a) (b) Figure C-1. Computer simulation of the Charlotte auditorium with source (A0) and receiver locations specified: (a) without diffusers, (b) with diffusers on the walls and the ceiling. 71

PAGE 72

An omnidirectional sound source (A0) is located 4 feet high from the stage level; twelve receivers were selected for this study as shown in Figure B-1a. This study includes 3 major parts: 1) The study of effective time windows of the first order specular and diffuse reflections by looking at the impulse responses and hedgehog representations at 12 receiver locations. 2) A comparison of room acoustic parameters of the two room conditions (with and without diffusers) to search for evidence of the acoustical effects from diffusers. 3) Listening test: four anechoic sounds including speech were convolved with the impulse responses (including both early part and reverberant tail) from both halls (with and without diffusers) to search for any perceived differences between these two room conditions. First, impulse responses of the first order specular and diffuse reflections (obtained by the image source method) at 6 receiver locations were plotted by hand to compare with the impulse responses of the first order diffuse reflections and up-to third order specular reflections obtained from CATT-Acoustic as shown in Figure C-4. Figure C-2. Image source method used to derive arrival time of specular and diffuse reflections in the impulse responses. 72

PAGE 73

Figure C-3. Three dimensional image source method used to produce rooms IRs at selected receivers. Figure C-4. Comparison of the rooms impulse reponses of the rooms 1st order diffusion at different receiver locations: (1) receiver 1, (3) receiver 3, (5) receiver 5. (b) 3rd order reflections & 1st order diffusion of the room w/o diffusers produced by CATT-Acoustic (a) 1st order reflections & diffusion (c) 3rd order reflections & 1st order diffusion of the room with diffusers produced by CATT-Acoustic (1) (3) (5) 73

PAGE 74

The results show: 1) the time window where the first order specular and diffuse reflections appear in the IRs get narrower to within 70 msec of the direct sound as receivers move toward the rear of the hall while the effective time windows of the receivers in the front area are wider, as long as within 125 msec of the direct sound. 2) Early diffuse reflections provide continuous energy distribution and bridge the gaps among the early specular reflections and late reverberant tail (temporal smearing). The hedgehog representations show the effect on directional smearing. The plots show that uniform directional energy distribution appears sooner in the room with diffusers than in the room without diffusers. The time window of the directional smearing corresponds well to the first order specular and diffuse reflections shown in the IRs. In another words, not long after the effective time window of the 1st order diffuse reflections, the sound field reaches its (roughly) spatially uniform distribution. Figure C-5. 74

PAGE 75

(a) (b) Figure C-5. Hedgehog presentations at receiver 3: (a) without diffusers, (b) with diffusers. 75

PAGE 76

Second, no significant differences are shown through the acoustical parameters generated by CATT-Acoustic. Only IACC improves at high frequency for the hall with diffusers. No_diff C-80 Diff C-80 Average IACC00.20.40.60.811.21252505001k2k4k8k16kOctave Band Center Frequency (Hz)% Correlation No_diff IACC Diff IACC Average RT (T-30)0123451252505001k2k4kOctave Band Center Frequency (Hz)Time (sec) No_diff T-30 Diff T-30 Average LEF051015202530351252505001k2k4kOctave Band Center Frequency (Hz) No_diff LEF Diff LEF Average C-80-2-1012341252505001k2k4kOctave Band Center Frequency (Hz)dB Figure C-6. Average RT, IACC, LEF, and C80 of all receivers from 2 room conditions (with and without diffusers) across 6 frequency bands. Finally, four anechoic pieces: speech, guitar, cello, and trumpet were convolved with the impulse responses obtained from the two halls (with and without diffusers). The listening test was conducted using the questionnaire developed by Torres (2000)61 (Figure C-7). The results show that subtle differences can be detected especially for speech due to its impulsive high frequency. The difference in the case of cello is harder to detect. 61 Torres, R. R., & Kleiner, M. (2000). "Audibility of diffusion in room acoustics auralization: an initial investigation." Acustica 86(6): 919-927. 76

PAGE 77

Figure C-7. The questionnaire used for the listening test developed from Torress.62 Table C-1. Statistical results from the listening test, 4 represents clearly different, 0 represents that no perceived difference can be detected. Variable N Mean StDev SE Mean 95% CI Coefficient of Variation Speech 9 2.00 1.12 0.37 (1.14, 2.86) 56% Guitar 9 1.33 1.32 0.44 (0.32, 2.35) 99% Cello 9 0.89 1.05 0.35 (0.08, 1.70) 118% Trumpet 9 1.33 0.50 0.17 (0.95, 1.72) 38% 62 Torres, R. R. (2000). Studies of Edge Diffraction and Scattering: Applications to Room Acoustics and Auralization. Department of Applied Acoustics Gteborg, Chalmers University of Technology. 77

PAGE 78

APPENDIX D PILOT STUDY II Black-box Theater: Physical Model and Computer Modeling The second approach to study the effects of diffusion was done by altering the path, i.e., different sizes of diffusers were installed in a model room, and observations of the change in the rooms impulse responses and acoustical parameters were made. The Black Box Theater at the Center of Performing Arts of the size 11.8 x 18 x 9.3 m (or 38.7 x 59 x 30.5 ft) has been simplified and built as a base model. Four sizes of (square-rod) diffusers were designed with their dimensions corresponding to the wavelength of the sound to scatter the energy at four frequency bands: 500 Hz, 1 kHz, 2 kHz, 4 kHz. Six models with diffusers installed on 3 walls and a ceiling were constructed by: 1) 1:20 physical model to observe the impulse responses, and 2) computer modeling to observe the change in acoustical parameters. First, 1:20 physical models were built, all diffusers were made of 3-ply chipboard (painted, and unpainted). Band-pass filtered impulse responses were obtained from each model in 3 frequency bands. Impulse Response 250 Hz 1 kHz Impulse Response 500 Hz 2 kHz Impulse Response 1 4 kHz (a) (b) Figure D-1. Two sets of 1:20 physical model tests were carried out with (a) painted, and (b) unpainted models. Three frequencies band-pass filtered IRs were measured. 78

PAGE 79

Section Isometric Plan Figure D-2. Plan, section, and isometric of the base model with the source and the receivers locations. 79

PAGE 80

Figure D-3. Six models with different sizes of diffusers for each set of test. Model 1 Bare room Model 2 3 walls and the ceilin g covered with diffusers, dimension ~ the wavelength of 500 Hz Model 4 diffusers dimension ~ the wavelen g th of 2 kHz Model 5 diffusers dimension ~ the wavelength of 4 kHz Mixture of diffusers dimension ~ 1 & 2 kHz Original Impulse response Energy bound Simplified Impulse response Model 3 diffusers dimension ~ the wavelength of 1 kHz Model 6 Figure D-4. The method of IR analysis: 1) find the boundary of energy fluctuation, 2) delete the data between the upper and the lower bounds, 3) Simplified IR is obtained. 80

PAGE 81

Table D-1. Impulse responses from painted model. IR (250 Hz-1 kHz) IR (500 Hz-2 kHz) IR (1-4 kHz) Model 1 (Bare room) Model 2 (500 Hz diffusers ) Model 3 (1 kHz diffusers ) Model 4 (2 kHz diffusers ) Model 5 (4 kHz diffusers ) Model 6 (1+2 kHz diffusers ) 81

PAGE 82

Table D-2. A comparison of the simplified impulse responses from painted model. IR (250 Hz-1 kHz) IR (500 Hz-2 kHz) IR (1-4 kHz) Model 1 (Bare room ) Model 2 (500 Hz diffusers ) Model 3 (1 kHz diffusers ) Model 4 (2 kHz diffusers ) Model 5 (4 kHz diffusers ) Model 6 (1+2 kHz diffusers ) 82

PAGE 83

Impulse Response Analysis Analyses of impulse responses characteristics were made in each test by graphically approximating the range of energy fluctuation and average energy decay. Then the simplified impulse responses were created by using the average upper and lower bounds of the energy fluctuation among 6 models as a guided energy band, any strong reflections outside this band were left unerased. Figure D-4 shows the method of analysis of the impulse response. Results from physical model testing show that: 1) The sound energy decays faster in the unpainted models which corresponds well with the differences of materials absorption coefficients, which follows Sabines RT theory. 2) The average energy decay curves among the six painted models are quite similar which means that there are approximately the same RTs among six painted models. 3) Wider energy fluctuations are found at low frequency band-pass IRs. 4) The only conclusion that can be drawn from the simplified impulse responses is that Models 3-6 show improvements from Model 1 for the number of strong repetitive spikes in Model 3-6 are much less than in Model 1s. The lesson learned from this experiment is that changing the types of diffusers at the room boundaries cannot control the characteristics of the impulse responses temporally. Due to the fact that there is no standard procedure to graphically evaluate the temporal distribution of those spikes (strong reflections) shown in the impulse responses, further conclusions from these data cannot be drawn. 83

PAGE 84

Energy bounds 250 Hz 1 kHz Energy bounds 500 Hz 2 kHz Energy bounds 1 4 kHz (a) (b) Figure D-5. Comparisons of the energy bounds among 6 models IRs at 3 band-pass frequencies filtered of: (a) painted models, (b) unpainted models. The same six models were created in CATT-Acoustic. The results from the computer model show: 1) RTs of 6 models correspond well with the hand-calculations and with 1:20 physical model tests. 2) Strong G10 is found in the bare room (strong reflections). 3) C80s vary according to the size of diffusers (large diffusers, less ITDG, higher C80). 4) No conclusion can be drawn from the IACCs and LEF/LFs data. 84

PAGE 85

No strong relationship can be found between diffusion and current room acoustic parameters (generated by CATT-Acoustic). Test A Comparison of IACC0.000.200.400.600.801.001252505001 k2 k4 kOctave Band Center Frequency (Hz) Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Test A Comparison of T-30 0.000.501.001.502.002.501252505001 k2 k4 kOctave Band Center Frequency (Hz)Time (sec.) Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 (a) (b) Test A Comparison of G-100.05.010.015.020.01252505001 k2 k4 kOctave Band Center Frequency (Hz)dB Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Test A Comparison of C-80-4.0-3.0-2.0-1.00.01.02.03.0Octave Band Center Frequency (Hz)dB Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 (c) (d) Test A Comparison of LFC25.030.035.040.045.050.01252505001 k2 k4 kOctave Band Center Frequency (Hz) Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 (e) Figure D-6. Comparisons of acoustical parameters of the 6 painted models simulated by CATT-Acoustic: (a) T30/RT, (b) IACC, (c) G10, (d) C80, (e) LFC /LF. 85

PAGE 86

LIST OF REFERENCES Antonio, P. D., & Cox, T. (1998). "Two decades of sound diffusor design and development Part 2: prediction, measurement, and characterization." J. Audio Eng. Soc. 46(12): 1075-1091. Barron, M. (1993). Audirorium Acoustics and Architectural Design NY, Routledge. Barron, M., & Marshall, A. H. (1981). "Spatial impression due to early lateral reflections in concert halls: The derivation of a physical measure." J. Sound Vibration. 77: 211-232. Beranek, L. (1962). Music, Acoustics and Architecture New York, Wiley. Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture NY, Springer. Bolt, R. H., & Doak, P. E. (1950). "A tentative criterion for the short-term transient response of auditoriums." J. Acoust. Soc. Am. 22: 507-509. Burghauser, J., und Spelda, A. (1971). Die akustischen Grundlagen der Instrumentation Rogensberg. Cervone, R. P. (1990). Designing for Music: The Subjective and Objective Evaluation of Three Rooms for Music Listening. Masters Thesis, School of Architecture Gainesville, University of Florida Cox, T. J. (1996). "Designing curved diffusers for performance spaces." J. Audio Eng. Soc. 44(5): 354-364. Cox, T. J., & D' Antonio, P. (2004). Acoustic Absorbers and Diffusers: Theory, Design and Application NY, Spon Press. Cremer, L., & Muller, H. A. (1982). Principles and Applications of Room Acoustics NY, Applied Science Publishers Ltd. Dalenback, B.-I. (2002). CATT-Acoustic v8.0 User's Manual Gothenburg, Sweden. Dalenback, B.-I., Kleiner, M., & Svensson, P. (1994). "A macroscopic view of diffuse reflection." J. Acoust. Soc. Am. 42(10): 793-806. Egan, M. D. (1988). Architectural Acoustics NY, McGraw-Hill, Inc. Fletcher, H., Blackham, E. D., & Stratton, R. (1962). "Quality of piano tones." J. Acoust. Soc. Am. 34: 749-761. Gold, M. A. (1994). An Evaluation of Perceived Acoustic Spatial Impression. Masters Thesis, School of Architecture Gainesville, University of Florida. 86

PAGE 87

Haan, C. H., & Fricke, F. R. (1993). Surface diffusivity as a measure of the acoustic quality of concert halls Proceedings of Conference of the Australia and New Zealand Architectural Science Association, Sydney. Haas, H. (1972). "The influence of a single echo on the audibility of speech." J. Audio Eng. Soc. 20(2): 146-159. Junius, W. (1959). "Raumakustische Untersuchunger mit Neueren Messverfahren in der Liederhalle Stuttgart." Acustica 9: 289-303. Kuttruff, H. (1965/66). "ber Autokorrelationsmessungen in der Raumakustik." Acustica 16: 166-174. Mehta, M., Johnson J., & Rocafort, J. (1999). Architectural Acoustics: Principles and Design NJ, Prentice-Hall, Inc. Meyer, J. (1978). Acoustics and the Performance of Music Frankfurt/Main, Verlag Das Musikinstrument. Meyer, J. (1993). "The sound of the orchestra." J. Audio Eng. Soc. 41(4): 203-212. Niese, H. (1961). "Die Messung der Nutzschallund Echogradverteilung zur Beurteilung der Hrsamkeit in Rumen." Acustica 11: 202-213. Reichardt, W., Abdel Alim, O., und Schmidt, W. (1975). "Definition und Megrundlage eines objektiven Maes zur Ermittlung der Grenze zwischen brauchbarer und unbrauchbarer Durchsichtigkeit bei Musikdarbietung." Acustica 32: 126-139. Thiele, R. (1953). "Richtungsverteilung und Zeitfolge der Schallrckwrfe in Rumen." Acustica 3: 291-302. Torres, R. R. (2000). Studies of Edge Diffraction and Scattering: Applications to Room Acoustics and Auralization. Department of Applied Acoustics Gteborg, Chalmers University of Technology. Torres, R. R., & Kleiner, M. (2000). "Audibility of diffusion in room acoustics auralization: an initial investigation." Acustica 86(6): 919-927. Zwicker, E., und Feldtkeller, R. (1967). Das Ohr als Nachrichtenempfnger Stuttgart. 87

PAGE 88

BIOGRAPHICAL SKETCH Pattra Smitthakorn received her bachelors degree in architecture from Chulalongkorn University, Bangkok, Thailand, in 1994. In 1997, she was awarded a scholarship from the Royal Thai Government to study abroad; in 2000, she received her Master of Science in Architectural Studies majoring in architectural acoustics at the University of Florida. She enrolled in a PhD program in Building Performance Diagnostics at Carnegie Mellon University in 1999. Two years of education at CMU incubated her critical thinking in mathematics with the help of a tutor, Aram Tangboonduangjit. In 2001, she transferred back to the University of Florida to continue her education in architectural acoustics. She found her inspiration in 2004 when studying mathematics with Professor Li-Chien Shen; together with Professor Alexander Berkovich and Professor Yuli Rudyak, they rigorously trained her to logically prove assertions, thoughts, and problems of various kinds. She then proceeded to work on the most difficult issue in the field of architectural acoustics and received her PhD degree in 2006. She will return to her home country after her graduation and become a professor at King Mongkuts University of Technology Thonburi, Bangkok, Thailand. 88


Permanent Link: http://ufdc.ufl.edu/UFE0016560/00001

Material Information

Title: Effects of Temporal Distribution of Specular and Diffuse Reflections on Perceived Music Quality
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0016560:00001

Permanent Link: http://ufdc.ufl.edu/UFE0016560/00001

Material Information

Title: Effects of Temporal Distribution of Specular and Diffuse Reflections on Perceived Music Quality
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0016560:00001


This item has the following downloads:


Table of Contents
    Title Page
        Page 1
        Page 2
    Dedication
        Page 3
    Acknowledgement
        Page 4
    Table of Contents
        Page 5
        Page 6
    List of Tables
        Page 7
    List of Figures
        Page 8
        Page 9
        Page 10
    Abstract
        Page 11
        Page 12
    Introduction
        Page 13
        Page 14
    Literature review
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
    Materials and methods
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
    Results and discussion
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
    Conclusion and future work
        Page 59
        Page 60
    Appendices
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
    References
        Page 86
        Page 87
    Biographical sketch
        Page 88
Full Text





EFFECTS OF TEMPORAL DISTRIBUTION OF SPECULAR AND DIFFUSE
REFLECTIONS ON PERCEIVED MUSIC QUALITY




















By

PATTRA SMITTHAKORN


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2006
































Copyright 2006

by

Pattra Smitthakorn
































To my great mentors: Buddhadasa Bhikku, and Professor Li-Chien Shen.









ACKNOWLEDGMENTS

I would like to thank Professor James P. Sain, Professor Robert C. Stroh, Professor

Bertram Y. Kinzey, Jr., and Professor Gary W. Siebein for their trust, support and constructive

advice. I would like to thank Professor Martin A. Gold for his help on setting up the equipment

for all experiments, providing opportunities for his students to participate in the listening tests,

and his guidance on how to perform the listening tests. His open-mindedness to possibilities and

challenges, his expertise, and his support have encouraged me to pursue such an impossible

dream. There are no words that can describe my appreciation for his help and friendship.

I am one of those few lucky students who have come to experience the true meaning of

education. Professor Li-Chien Shen has taught me to humbly believe in the value of education

that can transform one's life unimaginably without him saying anything. His wisdom has shone

through and dispelled my years of darkness. The inspiration I have received from him and his

colleagues, Professor Yuli Rudyak and Professor Alexander Berkovich, has changed my attitude

toward life and my future career completely. I hope to be able to share this wonderful

experience with my students in the future.

I would like to thank my family and friends for all kinds of support and more importantly

for believing in me. I am indebted to the Royal Thai Government for this life-transforming

opportunity.









TABLE OF CONTENTS



A C K N O W L E D G M E N T S ..............................................................................................................4

LIST OF TABLES ................. .. ....................... .. ........... ...................................... .. 7

L IST O F F IG U R E S ......................................................................... ................................... . 8

A B S T R A C T ............... ......................................................... ................................................. 1 1

CHAPTER

1 INTRODUCTION .................................. .. ........... ..................................... 13

2 L ITE R A TU R E R E V IE W .............. ..................................................................... 15

Diffuse Reflections ..................................... .. .......... ............................. 15
T em p oral D iffu sion ......................................................................................................... 18
Directional Diffusion ....... ............... .. ........... .....................................21
H edgehog R presentation .. ...................................................................... ................ 22
Sound D iffusivity Index .................................................. .............. ................ 22

3 M A TERIAL S AN D M ETH OD S .......................................... ......................... ................ 25

P ilo t S tu d y I ............................................................................................................................2 5
Pilot Study II ................................................ .............................. 25
F in a l E x p e rim e n t ....................................................................................................................2 6
Objective ............................................... ............................... 26
M ethod ................................................. ............................. 27

4 R E SU L T S A N D D ISC U SSIO N ............................................................................................. 32

S ou rce C h aracteristics .......... ......................................... ....................................... .......... ....... 32
O orchestral M music .............................................................................................................32
T ru m p e t ...........................................................................................................................3 3
Piano..................................................................... ...... ..................... 36
Initial H hypothesis: E effects from Path ....................... ......................................... ...............39
R e c e iv e r ............................................................................................................... .. ............4 1
R results and D discussion .................................................................................................... 41
Loudness .................................................... .................. 42
C la rity ................................................................................................... ........ . ....... 4 3
In tim a c y ........................................................................................................................... 4 5
R everberance and E choes .............................................................................................48
Source W idth ......................................... ............. .................. 48
Texture ................................................. ............................. 51
O v erall Im p re ssio n .......................................................................................................... 5 1









Conclusions ................................................. ............................. 54
Specular R efl sections ............... .. ................... ................ .......................... ............... 54
D iffuse R efl sections .............. ................................. ...... ............ ........ .... ............... 54
Applications ..................................................... .................. 55
C om m ents .................................................................................. ...... ..... ..................... 58

5 CONCLUSION AND FUTURE W ORK ................. ..................................................... 59

C o n c lu sio n .............................................................................................................................. 5 9
Future W ork ................................................ ............................... 59

APPENDIX

A Q U E ST IO N N A IR E .............. ........................................................................ 6 1

B RESULTS FROM LISTEN IN G TEST ........................................ ..................... ............... 63

C PILOT STUD Y I ..... ....... .. .. .. ................................. 71

D P IL O T ST U D Y II ........................................................ ................................................... 78

Black-box Theater: Physical Model and Computer Modeling.........................................78
Im pulse R response A naly sis .................................................. ............................................ 83

L IST O F R EFE R EN C E S ............................................................................................. 86

B IO G R A PH IC A L SK E T C H .............. ...................................................................... 88


























6









LIST OF TABLES


Table page

3-1 Absorption and scattering coefficients specified in CATT modeling. .............................29

3-2 The 9 designed impulse responses showing various sets of arrival specular
reflections generated by CA TT-A coustics.................................................... ................ 30

3-3 The 9 designed impulse responses showing various sets of arrival diffuse reflections
generated by CA TT-A coustics. ................. ............................................................. 31

C-i Statistical results from the listening test, 4 represents clearly different, 0 represents
that no perceived difference can be detected ................................................ ................ 77

D -1 Im pulse responses from painted m odel......................................................... ................ 81

D-2 A comparison of the simplified impulse responses from painted model........................82









LIST OF FIGURES


Figure page

2-1 Examples the impulse responses showing visual judgments of "texture": (a) IR's
from some of the best concert halls which have good "texture", (b) IR's from hall
w ith poorer "texture". .. ............................................... ......... ........ ................ 19

2-2 Single-number criteria for the evaluation of echoes in an impulse response. After
N ie se .............................................................................................................. ....... .. 2 0

2-3 Single-number criteria for the evaluation of echoes in an impulse response. After
B o lt a n d D o a k ................................................................................................................. ... 2 1

2-4 Directional "hedgehogs", for early and late-time arrival, at three locations in the
Beethoven H all of the Liederhalles, in Stuttgart........................................... ................ 23

2-5 The surface diffusivity index (SDI) of 31 concert halls. The highest rated halls have
SDI's of 1.0, while those of the lowest rated halls fall in the range of 0.3 to 0.7..............24

3-1 The computer generated models created to produce five specular/diffuse reflections
arrival to the receiver w within 40 m sec ........................................................... ................ 29

4-1 Main directional radiation (0... -3 dB) of the trumpet in the vertical plane....................35

4-2 Sound spectra of a horn in different keys. ....................... ......................................... 35

4-3 Reverberation time of a grand piano for some C-keys using the right pedal. .................37

4-4 Patterns of directional radiation of a grand piano in a vertical plane ..............................38

4-5 Music preference of 47 subjects who participated in the listening test. ..........................41

4-6 Normalized perceived loudness of the three sources: Orchestral music, Trumpet, and
Piano. The orange columns are anechoic signals of each group..................................43

4-7 Clarity: normalized averaged ratings of perceived clarity of (a) Orchestral music, (b)
T rum pet, and (c) P iano ..................................................................................................... 44

4-8 Intimacy: normalized averaged ratings of perceived intimacy of the Piano source ..........45

4-9 Reverberance: normalized averaged ratings of perceived reverberance of (a)
Orchestral m usic, (b) Trum pet, and (c) Piano............................................... ................ 46

4-10 Echoes: normalized averaged ratings of perceived echoes of (a) Orchestral music, (b)
T rum pet, and (c) P iano ..................................................................................................... 47

4-11 Source width: normalized averaged ratings of perceived source width of (a)
Orchestral m usic, (b) Trum pet, and (c) Piano............................................... ................ 49









4-12 Texture: normalized averaged ratings of perceived texture of (a) Orchestral music,
(b) T rum pet, and (c) P iano ............................................................................... ................ 50

4-13 A comparison on ratings of perceived acoustic qualities: (a) source width, (b)
intimacy, (c) texture, and (d) overall impression of Orchestral music. ............................ 52

4-14 A comparison on ratings of perceived acoustic qualities: (a) texture, and (b) overall
impression of Trumpet ... ............... .. ........... ........................................53

4-15 A comparison on ratings of perceived acoustic qualities: (a) clarity, (b) intimacy, (c)
texture, and (d) overall im pression of Piano ................................................ ................ 53

4-16 Summary of the preferred arrival diffuse and specular reflections to provide
improvement on perceived "overall impression" for 3 sources-Orchestral music,
T ru m p et, an d P ian o ................ ................................ ...............................................5 6

4-17 Three ellipsoids representing boundaries where reflecting panels can be located
within to provide the 1st order reflections arriving to the front receiver within 3 time
periods: 40, 80, and 160 m sec after the direct sound.................................... ................ 57

4-18 Three different boundaries for 3 different receiver locations where reflecting panels
can be located to provide the 1st order reflections to arrive at the receivers with 40
m sec after the direct sound. ............................................. ............. ................ 58

B-I Loudness: normalized averaged ratings of perceived loudness of (a) Orchestral
m usic, (b) T rum pet, and (c) Piano ...................................... ...................... ................ 63

B-2 Clarity: normalized averaged ratings of perceived clarity of (a) Orchestral music, (b)
T rum pet, and (c) P iano ..................................................................................................... 64

B-3 Intimacy: normalized averaged ratings of perceived intimacy of (a) Orchestral music,
(b) T rum pet, and (c) P iano ............................................................................... ................ 65

B-4 Reverberance: normalized averaged ratings of perceived reverberance of (a)
Orchestral m usic, (b) Trum pet, and (c) Piano............................................... ................ 66

B-5 Source width: normalized averaged ratings of perceived source width of (a)
Orchestral m usic, (b) Trum pet, and (c) Piano............................................... ................ 67

B-6 Texture: normalized averaged ratings of perceived texture of (a) Orchestral music,
(b) T rum pet, and (c) P iano ............................................................................... ................ 68

B-7 Echoes: normalized averaged ratings of perceived echoes of (a) Orchestral music, (b)
T rum pet, and (c) P iano ..................................................................................................... 69

B-8 Overall impression: normalized averaged ratings of perceived overall impression of
(a) Orchestral music, (b) Trum pet, and (c) Piano. ........................................ ................ 70









C-i Computer simulation of the Charlotte auditorium with source (AO) and receiver
locations specified: (a) without diffusers, (b) with diffusers on the walls and the
ceiling ............................................................................................ ........ .. 7 1

C-2 Image source method used to derive arrival time of specular and diffuse reflections
in the impulse responses. .................................. ........ .... ............... 72

C-3 Three dimensional image source method used to produce room's IR's at selected
receive ers ......................................................................................................... . ....... .. 7 3

C-4 Comparison of the room's impulse reponses of the room's 1st order diffusion at
different receiver locations: (1) receiver 1, (3) receiver 3, (5) receiver 5.......................73

C-5 Hedgehog presentations at receiver 3: (a) without diffusers, (b) with diffusers .............75

C-6 Average RT, IACC, LEF, and C80 of all receivers from 2 room conditions (with and
w without diffusers) across 6 frequency bands................................................. ................ 76

C-7 The questionnaire used for the listening test developed from Torres's .......................... 77

D-1 Two sets of 1:20 physical model tests were carried out with (a) painted, and (b)
unpainted models. Three frequencies band-pass filtered IR's were measured..............78

D-2 Plan, section, and isometric of the base model with the source and the receiver's
lo catio n s ......................................................................................................... . ....... .. 7 9

D-3 Six models with different sizes of diffusers for each set of test. ..................................80

D-4 The method of IR analysis: 1) find the boundary of energy fluctuation, 2) delete the
data between the upper and the lower bounds, 3) Simplified IR is obtained.................. 80

D-5 Comparisons of the energy bounds among 6 models' IR's at 3 band-pass frequencies
filtered of: (a) painted models, (b) unpainted models................................... ................ 84

D-6 Comparisons of acoustical parameters of the 6 painted models simulated by CATT-
Acoustic: (a) T30/RT, (b) IACC, (c) G10, (d) C80, (e) LFC /LF.................................85










Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

EFFECTS OF TEMPORAL DISTRIBUTION OF SPECULAR AND DIFFUSE
REFLECTIONS ON PERCEIVED MUSIC QUALITY

By

Pattra Smitthakorn

December 2006

Chair: Gary W. Siebein
Major Department: Design, Construction, and Planning

The purpose of this study was to investigate the effects of the temporal distribution of

diffuse and specular reflections on the perceived acoustic qualities of music performance. Sets

of impulse responses were designed with different temporal distributions of early acoustic energy

(specular and diffuse reflections). Then, three types of anechoic sound sources-orchestral

music, trumpet, and piano-were convolved with the designed impulse responses. The results

from the listening tests revealed that different room environments were needed to acoustically

support different source characteristics. The results show the following: 1) specular reflections

arriving within 40 msec of the direct sound improved perceived "clarity" and "intimacy"; 2)

specular reflections arriving between 40-80 msec after the direct sound improved perceived

"clarity" for orchestral music; 3) specular reflections arriving later than 80 msec after the direct

sound are not desirable; 4) large numbers of diffuse reflections arriving within 40 and 80 msec of

the direct sound improved perceived "intimacy", "texture", and "overall impression" for all

sound sources, heightened perceived "clarity" for trumpet and piano, and reduced perceived

"glare" for trumpet; and 5) diffuse reflections arriving between 80-160 msec of the direct sound

preserved perceived "reverberance" and reduced perceived "echoes" as opposed to specular









reflections arriving in the same time period. The results of this study indicate that music

performance halls should be designed to include diffuse reflections from surfaces within the 80

msec time period to achieve preferred texture, intimacy, clarity and overall impression and in the

160 msec time period to reduce echoes; specular reflections arriving within the 40 msec time

period should be provided to enhance perceived clarity.









CHAPTER 1
INTRODUCTION

When architects design a concert hall, the chief goal that needs to be met, among many

other functional requirements and design criteria, is to design a room that provides an excellent

acoustic environment. Recommendations from acoustic consultants are given to architects to

guide their designs to achieve this goal. The recommendations include the optimum range of

acceptable acoustical parameters known to have major effects on the perceived acoustical

qualities of the room. The parameters are listed below.

* Optimal Reverberation Time (RT) should be achieved; optimal RT varies depending on the
purpose of the hall. This suggestion directly affects the volume and materials used inside
the hall according to the RT formula.

* The Initial Time Delay Gap (ITDG) should be within 20 msec of the direct sound.1 The
design of acoustic clouds and the first order reflectors are roughly guided by this
requirement.

* Provide a number of early reflections to obtain clarity of sounds (high Clarity index-C80)
perceived in the room.

* Limit the room width and provide a large number of early lateral reflections to obtain a
high value of the Lateral Fraction (LF) which would enhance the sense of envelopment,
spaciousness, and widen the source width.2

* Avoid any acoustic defects such as echoes, sound focusing, and flutter echoes by locating
absorbing materials or diffusers at the rear wall (rather than smooth reflective surfaces),
avoid smooth large concave surfaces, and avoid smooth large parallel walls.

* Provide diffusing panels surrounding the orchestra (orchestra shell) for a sense of
ensemble and blending of sounds from the sending end to the auditorium.3

* Provide a large amount of various scales of surface irregularities to obtain a diffuse sound
field.4


1 Beranek, L. (1962). Music, Acoustics and Architecture. New York, Wiley.
2 Barron, M., & Marshall, A. H. (1981). "Spatial impression due to early lateral reflections in concert halls: The
derivation of a physical measure." J. of Sound and Vibration 77: 211-232.

3 Beranek, L. (2003). Concert Halls and Opera Houses: Music. Acoustics and Architecture. NY, Springer.
4 Ibid.









In the design of the interior of a concert hall with the knowledge that most of the materials

used in the room should be reflective to preserve sound energy as much as possible, further

questions are raised-what type of reflectors should be used: specular or diffuse reflectors?

What type of diffuse reflectors (shape/form) should be selected? How many of the reflectors

should be diffuse compared with specular? Where should these diffuse and specular reflectors

be positioned in the hall? These questions thus far have not been answered with scientific

support. There have been many experiments with vastly diverse approaches trying to find

answers to questions concerning acoustical diffusion phenomena. This dissertation answers

some of the questions raised by architects scientifically and provides a method that can be used

to investigate these questions.

The literature review contains discussion of several prior works which are thought to be

related to and might have potentially influenced the experimental designs used in this study. A

series of experiments with different approaches and settings have been designed and conducted

to search for more evidence and to test new sets of hypotheses. The third experiment finally

provides fruitful results, and hopefully this new approach can be further developed to study and

search for knowledge in the field of room acoustics.









CHAPTER 2
LITERATURE REVIEW

Diffuse Reflections

Scientific research of the role of diffuse reflections in concert halls was not begun until the

1980s, despite the fact that diffusing surfaces have always been considered (and included) as an

important part of a concert hall design since their origination. Research and experiments in this

area are all innovative and somewhat in their infancy. Diverse approaches of experiments are

found in this area, which provides discontinuous pieces of knowledge. With their unsuccessful

results, a pattern/trace to a future successful experimental design was not found. Therefore, past

studies and investigations in this area seem to be scattered and inconclusive.

Dalenback5, in his 1994 article on "A macroscopic view of diffuse reflection", provides a

widely accepted view and up-to-date information about diffusion. He defined the term diffuser

as "any surface that reflects sound in all or most directions, largely independent of the incidence

angle. Such surfaces have a roughness (naturally or by design) with a size on the order of a

wavelength in the frequency band where diffuse reflection is observed." Qualitative properties

of diffusely reflecting surfaces have been described as follows:6

1) The diffusion capability of a diffuser is strongly frequency dependent. However, wide-band
diffusion is possible depending on the magnitude of the surface depth.

2) The diffuser will give reflections outside the specular zone, thereby decreasing energy within
the specular zone, which diminishes the comb-filter effects due to strong early specular
reflections.7 With this characteristic, a diffuser can be used to enhance the sense of
ensemble on stage.


5 Dalenback, B.-I., Kleiner, M., & Svensson, P. (1994). "A macroscopic view of diffuse reflection." J. Acoust. Soc.
Am. 42(10): 793-806.
6 Ibid.

7 When there are equal intensities of primary sound and (too) early specular reflection interfere with one another the
results are large linear distortions of the tone. "Depending on the particular delay difference of the reflection,
certain frequency ranges are very much reinforced whereas others are completely wiped out." (Haas, 1972)









3) Diffuse reflections can decrease the risk for "dead spots" and "soften" detrimental reflections
since the reflected energy will be distributed throughout the whole space more quickly and
more uniformly. This also diminishes sharp late echoes without the need for introducing
absorbers.

4) A more uniform reverberant field will be created across the hall (by diffuse reflections) with
an improved near exponential decay, giving a smooth impulse response.

5) Frequency content in reflections is affected by diffuse reflections, which results in sound
coloration.

6) Diffuse reflections also create effects such as directivity smear, amplitude smoothing, and
temporal smearing.

Other research in the field of diffusion can be grouped into 2 categories: 1) analyses of

diffusers (path), and 2) effects from diffuse reflections (receiver).

First, a large number of research/studies are related to the search for a mathematical model

to define and standardize diffusion coefficients8, methods of obtaining frequency dependent

diffusion coefficients9, and optimizing diffuser design 10. D' Antonio defined "diffusion

coefficient" as a measure of the uniformity of the reflected sound which allows comparison of

the performance of the surfaces, while "scattering coefficient" is a ratio of sound energy

scattered in a non-specular manner to the total reflected sound energy. The latter measure

corresponds to the theory used in geometrical room modeling software. In this case,

misrepresentation of a triangular diffuser can be seen because it would scatter energy away from

the specular zone and appear as a good diffuser, however this is a redirection of energy rather







8 Cox, T. J., & D' Antonio, P. (2 14). Acoustic Absorbers and Diffusers: Theory, Design and Application. NY, Spon
Press.
9 Ibid.
10 Cox, T. J. (1996). "Designing curved diffusers for performance spaces." J. Audio Eng. Soc. 44(5): 354-364.









than dispersion."1 Scattering coefficients and this procedure are what is used in CATT-Acoustic

(Computer Aided Theater) software.

Second, from the aspect of receiver, studies include: 1) the effects of diffusion on room

acoustic qualities (auralization)12; 2) acoustical parameters that indicate the room diffusivity

temporally texturee1, temporal diffusion14), and directionally (directional diffusion)15; and 3) a

numerical rating of a space/hall using visual inspection called sound diffusivity index (SDI)16.

This dissertation is another investigation in this category; therefore a closer review was made of

the studies mentioned above.

Torres and Kleiner17 (2000) conducted a listening test in search of the effects of surface

diffusion on auralization at different frequency ranges. This was done by convolving 4 anechoic

sources with selected room impulse responses obtained from assigning different (quasi-step

function: high, low) diffusion coefficients to surfaces in the room (at 3 frequency ranges: low,

mid, high). The 4 anechoic sources included two sustained (synthesized organ chord, pink noise)

and two impulsive (string quartet with pizzicato, and the unconvolved binaural room impulse

response (BRIR) alone) sources. Their results show the following:



11 Antonio, P. D., & Cox, T. (1998). "Two decades of sound diffusor design and development Part 2: prediction,
measurement, and characterization." Ibid. 46(12): 1075-1091.
12 Torres, R. R., & Kleiner, M. (2000). "Audibility of diffusion in room acoustics auralization: an initial
investigation." Acustica 86(6): 919-927.

13 Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture. NY, Springer.
14 Bolt, R. H., & Doak, P. E. (1950). "A tentative criterion for the short-term transient response of auditoriums." J.
Acoust. Soc. Am. 22: 507-509.
15 Meyer, E. (1954). "Definition and diffusion in rooms." Ibid. 26(5): 630-636.

16 Haan, C. H., & Fricke, F. R. (1993). Surface diffusivity as a measure of the acoustic quality of concert halls.
Proceedings of Conference of the Australia and New Zealand Architectural Science Association, Sydney.
17Torres, R. R. (2000). Studies of Edge Diffraction and Scattering: Applications to Room Acoustics and
Auralization. Department of Applied Acoustics. G6teborg, Chalmers University of Technology.









1) For some signals, changes in the diffusion coefficient are clearly audible within a wide
frequency region. Thus, diffuse reflections should be modeled in a frequency-dependent
manner, although not all auralization programs currently do this. 2) The perception of
these changes depends on the input signal. For sustained signals (e.g., an organ chord,
pink noise), changes are strongly perceived as differences in coloration; for example,
increasing low-frequency diffusion is perceived as "decreasing the bass" content or
"increasing the treble" content of the signal. For impulsive signals (e.g., string pizzicato),
coloration differences are less audible than for sustained signals, whereas spaciousness
differences are relatively stronger. It is interesting that listeners, though uninformed of the
differences between high- or low-diffusion signals, give consistent answers regarding
perceived changes in frequency coloration.18

Beranek19 defined "texture" as:

Texture is the subjective impression that listeners derive from the patterns in which the
sequences of early sound reflections arrive at their ears. In an excellent hall, those
reflections that arrive soon after the direct sound follow in a more-or-less uniform
sequence. In other halls there may be a considerable interval between the first and the
following reflections or one reflection may overly dominate. Good texture requires a large
number of early reflections, reasonably strong in amplitude, uniformly but not precisely
spaced apart, and with no single reflection dominating the others.

This quality can be determined visually from the impulse responses (IRs) as shown in

Figure 2-1.

Temporal Diffusion

In the area of temporal diffusion, the patterns of impulse responses were observed in detail

by Niese (1956), and Bolt & Doak (1950). Niese20 plotted the impulse response (|p(t)|) together

with an exponential decay curve shown in Figure 2-2. He considered the areas of the impulses

that, after the time limit t = 33 msec, lie above the decay curve as "detrimental". He then

proposed the criterion for the existence of echoes called the "echo coefficient" (s) by combining

the "detrimental" areas S and "useful areas" N in the form shown in Equation 2-1.



18 Ibid.
19 Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture. NY, Springer.
20 Niese, H. (1961). "Die Messung der Nutzschall- und Echogradverteilung zur Beurteilung der Horsamkeit in
Raiumen." Acustica 11: 202-213.
















TT Tokyo, New Nat'L TIter







BA Buemm AuirTentr Col6a







VS Vienne, Stasaoper


CC Chicago, Civic Opera


............. Berlin, De ..... r .. .
BD Berlin, Deutscheoper
P:?'fflIN1Uf1Sm I N. L


0 TIMtse) oo a T n am,) ngo
(a) (b)
Figure 2-1. Examples the impulse responses showing visual judgments of "texture": (a) IR's
from some of the best concert halls which have good "texture", (b) IR's from hall
with poorer "texture".21

S
S S- (Eq. 2-1)
S+N

The drawback to this criterion is that the existence of a number of peaks that exceed the

general decay only slightly would produce the same echo coefficient as a single strong reflection.

The evaluation of flutter echoes proposed by Bolt and Doak (1950) utilizes the auto-

correlation function (4('c)) to analyze the randomness of the impulse energy (p(t)) above the

touching decay curve, Figure 2-3, given as Equation 2-2.


f(T) = P(P(t+)dt (Eq. 2-2)
0



21 Beranek, L. (2003). Concert Halls and Opera Houses: Music. Acoustics and Architecture. NY, Springer.









Where Ti, T2, C3, ... are the time intervals of periodic sequences of impulses.

Kuttruff22 proposed the criterion "temporal diffusion", which is represented by A shown in

Equation 2-3.


A = 0() (Eq. 2-3)
Omax (Z 0)

The higher the value of A the more random the impulse response. This criterion is based on the

nature of how flutter echoes are produced. Parallel walls reflected sound energy repeatedly back

and forth is usually the cause of flutter echoes, not diffuse reflecting walls. This criterion gives a

general measure of the randomness of a sequence of reflections assuming that any regular

sequence of reflections is undesirable for hearing.

However, describing the randomness quality of the reflections obtained from the impulse

response is not enough, for it does not describe the smoothness of the IRs which is the room

acoustical quality for "texture".


a-


0 100 200 ms 300

Figure 2-2. Single-number criteria for the evaluation of echoes in an impulse response. After
Niese.23




22 Kuttruff, H. (1965/66). "Uber Autokorrelationsmessungen in der Raumakustik." Acustica 16: 166-174.

23 Niese, H. (1961). "Die Messung der Nutzschall- und Echogradverteilung zur Beurteilung der Horsamkeit in
Raiumen." Acustica 11: 202-213.














0 100 200 ms 3001
10



-20
-30
-40




Figure 2-3. Single-number criteria for the evaluation of echoes in an impulse response. After









1-4
Bolt and Doak. (Eq.

Directional Diffusion

Directional diffusion is a single number criterion that describes the directional distribution

of the arriving sound at a particular location proposed by Thiele (1953)25. It is a measure of the

deviation from average energy from all directions represented as follows. First the quantity [t is

determined, "the sum of the absolute differences between the incoming intensity J over all room

directions within the solid angle of interest Q and the average incoming intensity J7; he

normalized this sum to the average intensity J, (Equation 2-4).

[t JJ Ifi- 7/ (Eq. 2-4)



This quantity is zero for a non-directional, isotropic sound field, and has its maximum

value uo in open air.",26 The "directional diffusion" (0 or DD) is given by Eq. 2-5.


0 = 1- (Eq. 2-5)


24 Bolt, R. H., & Doak, P. E. (1950). "A tentative criterion for the short-term transient response of auditoriums." J_.
Acoust. Soc. Am. 22: 507-509.
25 Thiele, R. (1953). "Richtungsverteilung und Zeitfolge der Schallrtickwuirfe in Rdiumen." Ibid. 3: 291-302.

26 Cremer, L., & Muller, H. A. (1982). Principles and Applications of Room Acoustics. NY, Applied Science
Publishers Ltd.









The "directional diffusion" is location dependent. The results from his study suggest that,

for constant RT: 1) DD decreases with increasing volume, and 2) in a room with the same

amount of absorbing, reflecting, and diffusing surfaces, DD varies depending on the location of

those materials. The drawback to this criterion is that a single strong directional peak or several

small peaks and valleys that are distributed equally over a solid angle might yield the same value

for this measure.

Hedgehog Representation

The hedgehog representation studied by Junius (1959)27 shown in Figure 2-4 gives a clear

picture of the spatial distribution of early (0-100 msec) and late energy (100-3c msec). This

representation shows that the assumptions on the diffuse sound field underlying statistical theory

are poorly fulfilled in a large hall. Even in the later time period, the directional distribution is

still far from uniform. However, the temporal distribution information gets lost in the hedgehog

representation.

Sound Diffusivity Index

The criterion used by Beranek28 to evaluate the room diffusivity is the "sound diffusivity

index" (SDI). SDI was proposed by Haan and Fricke (1993) using a visual inspection to classify

the "acoustical quality" of the surface irregularities in a hall. The "degree of diffusivity" of

surfaces can be categorized as high, medium and low diffusivity.

The room with high diffusivity is the one that has a coffered ceiling with deep recesses or

beams greater than 4 in. in depth, but not greater than 10 in., diffusing elements of sizable depth

on the upper sidewalls, and fine scale irregularities of about 2 in. depth on the lower sidewalls


27 Junius, W. (1959). "Raumakustische Untersuchunger mit Neueren Messverfahren in der Liederhalle Stuttgart."
Acustica 9: 289-303.
28 Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture. NY, Springer.











without absorbing surfaces. Medium diffusivity is assigned to broken surfaces of varying depth

on the ceiling and sidewalls or shallow recesses (no greater than 2 in.) of ornamental decorative

treatment, having no absorbing surfaces. A room with low diffusivity is one that contains large

separate paneling, or smoothly curved surfaces, or large flat and smooth surfaces, or heavy

absorptive treatment.


0 100 msec

100- oo msec

LONGITUDINAL
SECTION


00 m~ec
X~ T'~C


'or





L
oL


is
(- *SOURCE

0- -

is(


N


PLAN


L
r"N


100 msec

00 mnsec


6:;


L C






L C (S SOURCE




J1
I Q

x -


Figure 2-4. Directional "hedgehogs", for early and late-time arrival, at three locations in the
Beethoven Hall of the Liederhalles, in Stuttgart.29

The numerical rating assigned to each degree of surface diffusivity is as follows: "high"

equals 1.0, "medium" equals 0.5, and "low" equals 0. The SDI is obtained by dividing the

summation of the weighted diffusing area by the actual total area for the ceiling and sidewalls.



29 Junius, W. (1959). "Raumakustische Untersuchunger mit Neueren Messverfahren in der Liederhalle Stuttgart."
Acustica 9: 289-303.










Beranek (2003) shows the plot of rank-ordering of 31 concert halls against SDI, Figure 2-

5. The plot illustrates that the best halls have SDIs of 1.0; the medium quality halls, between 0.5

and 0.9; and the low quality halls between 0.3 and 0.7. He also mentioned that to obtain

meaningful SDIs among those halls for rank-ordering, other acoustical parameters, for example,

IACC, RT, ITDG, good stage design, and so on, must be taken into account simultaneously. The

drawback for this method is that the judgment made by visual inspection is somewhat subjective.

All proposed measures mentioned above have more or less contributed to the study of

diffusion. However, due to their drawbacks, none of them have been further developed to

become a standard measure for room diffusivity. The goal of this dissertation is to develop a

series of studies to explore and understand the behavior of diffuse reflection and their effects on

room acoustics.


SURFACE DIFFUSIVITY INDEX, SDI

10


08





0.4


0.2 CM I ii N 1 W R M P
Thirty-One Concert Halls


Figure 2-5. The surface diffusivity index (SDI) of 31 concert halls. The highest rated halls have
SDI's of 1.0, while those of the lowest rated halls fall in the range of 0.3 to 0.7.30







30 Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture. NY, Springer.









CHAPTER 3
MATERIALS AND METHODS

Pilot Study I

Prior to the final experiment, two pilot studies were performed. The first pilot study was a

comparison between two computer simulations of a chosen auditorium (Charlotte auditorium)

with and without diffusers. The results obtained from this study show that: 1) the auralizations

from the halls with and without diffusers are clearly discernible especially for speech; 2) no

current quantitative measures can capture/explain this subtle perceived change; and 3) the fine

structure of the room impulse responses (IRs) and the directional distribution of sound energy

have been carefully observed to identify changes in arrival time of the 1st order specular and

diffuse reflections. Further detail on this investigation can be found in Appendix C.

Pilot Study II

The second pilot study attempted to search for any (graphical) evidence of subtle changes

in IRs provided by installing different types of diffusers in a room. A 1:20 physical model with

six different types of diffusers installed on the room's boundaries was built. Band-pass filtered

impulse responses were recorded from six models. A careful graphical analysis had been

conducted (see Appendix D). General results from the experiment are coherent with Sabine's

theories (RT) and initial expectations. However, concrete conclusions cannot be made due to the

fact that there has not been any standard procedures developed to graphically evaluate the

temporal distribution of strong reflections shown in the impulse responses.

The lesson learned from the second pilot study strongly influenced the experimental design

of the final experiment. On one hand, the temporal diffusivity of each diffuser has not yet been

discovered. Only directional diffusivities of some diffusers are available. On the other hand, the

room impulse responses do not take into account the room spatial information; they show only









temporal energy distribution. Therefore, changing diffusers on the room's surfaces provides no

control over the temporal energy distribution. The final experiment then focused on taking

control of the temporal distribution of the two types of energy-specular and diffuse reflections.

Final Experiment

Objective

The objective of this experiment was to study the effects of the temporal distribution of

early acoustic energy (specular and diffuse reflections). A set of variations of energy distribution

was designed and a listening test was conducted to capture differences in perceived acoustical

qualities. Three source types used in this study were orchestral music, trumpet, and piano.

The focus on only the early part of the impulse responses was a result from Pilot Study I,

which showed prominent variation in IRs characteristics affected by the first order diffuse

reflections in the IRs not long before the energy reaches its uniformly diffuse sound field (when

second and third order specular and diffuse reflections arrive).

The three approximate time windows designed for this experiment are as follow:

1) between 20 and 40 msec of the direct sound,

2) between 40 and 80 msec of the direct sound, and

3) between 80 and 160 msec of the direct sound.

The chosen time window of 40 msec was derived from the Haas Effect31, which suggests

that specular reflections arriving within 40 msec of the direct sound are not perceived as echoes.

The 80 msec limit came from the time limit for useful to detrimental sound energy of the

room IRs for music proposed by Reichardt et al. in 1975. This time limit is now used in many

acoustic parameters, for example, C80, IACC, and LF.



31 Haas, H. (1972). "The influence of a single echo on the audibility of speech." J. Audio Eng. Soc. 20(2): 146-159.









The 160 msec limit approximates the arrival time of the first order specular and diffuse

reflections of a medium to large concert hall, derived from Pilot Study I, and also completes a

coherent logarithmic increment step for the whole experiment-20, 40, 80, and 160 msec.

No acoustical parameters will be studied here because the previous experiments show no

evidence of any existing acoustical parameters that can describe the detailed characteristic of

energy distribution temporally and spatially. Instead, after the study of perceived acoustical

quality was made, a careful look at how to capture and explain this quality quantitatively can be

further investigated.

Method

Eighteen designed impulse responses were created using CATT-Acoustic V.8. CATT was

used to model a highly absorptive room of the size 70m x 70m x 70m with the source and the

receiver floating 14m apart in the middle of the room (Figure 3-1). All 6 surfaces of the room

have absorption coefficients of 0.9 at 6 octave band frequencies (125Hz, 250, 500, 1 kHz, 2 kHz,

4 kHz) to simulate the anechoic environment. Five, 10, and 20 reflecting panels of the size 3m x

3m were located in the room to provide specular or diffuse reflections (from the upper-frontal

hemisphere) arriving at the receiver within the three time-windows to get the designed room

impulse responses as shown in Table 3-2, 3-3. The first reflection was set to arrive at 20 msec

after the direct sound for all 18 impulse responses to control for the same ITDG. The absorption

coefficients and scattering coefficients of the specular and diffuse reflectors are shown in

Table 3-1. The scattering coefficients were chosen based on the recommendation from CATT-

Acoustic manual.32




32 Dalenback, B.-I. (2002). CATT-Acoustic v8.0 User's Manual. Gothenburg, Sweden.









Then, these 18 impulse responses were convolved (excluding the late part impulse

response) with three different sources of anechoic music (orchestral piece, trumpet, and piano) of

approximately 20 seconds each.

The (18 + 1 anechoic) x 3 convolved sound fields (SFs)/wave files were then played back

to the subjects using supra-aural headphones SONY MDR-V600 through a 6 channel headphone

bus. Averaged play back levels of 3 sources were set at (peak level) 87-88 dBA for orchestra

(73 dBA-anechoic), 69-70 dBA for trumpet (66 dBA-anechoic), and 70-71 dBA for piano (68

dBA-anechoic). The playback levels were set according to an agreement from a few experienced

listeners acousticianss); the peak level measurements of the sound fields were made to check if

there could be any harm to the test subjects.

Forty-five subjects had participated in this experiment; 15 subjects for each source type.

Each subject evaluated 12 randomly picked sound fields + 1 anechoic + 3 learning samples (at

the beginning). A total sample size of 10 for each sound field was used to obtain the final

results. The questionnaire used was a seven-point bi-polar rating scale developed from

Cervone33 and Gold34 shown in Appendix A. The additional perceived acoustic quality "texture"

introduced in this study was defined according to Beranek's description35 mentioned in

Chapter 2.









33 Cervone, R. P. (1990). Designing for Music: The Subjective and Objective Evaluation of Three Rooms for Music
Listening. School of Architecture. Gainesville, University of Florida.
34 Gold, M. A. (1994). An Evaluation of Perceived Acoustic Spatial ImpressionIbid.
35 Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture. NY, Springer.










Table 3-1. Absorption and scattering coefficients specified in CATT modeling.

% Absorption Coefficients % Scattering Coefficients
Surfaces
125 250 500 1 2 4 125 250 500 1 2 4
Hz Hz Hz kHz kHz kHz Hz Hz Hz kHz kHz kHz

Reflecting panels 1 1 1 1 1 1 10 10 10 10 10 10

Diffusing panels 1 1 1 1 1 1 30 40 50 90 90 90

Room envelopes 90 90 90 90 90 90 10 10 10 10 10 10


rn


(a) (b)
Figure 3-1. The computer generated models created to produce five specular/diffuse reflections
arrival to the receiver within 40 msec.









Table 3-2. The 9 designed impulse responses showing various sets of arrival specular reflections
generated by CATT-Acoustics.
Time arrival of V'ihin 40 msec after Between 40-80 msec Between 80-160 msec
specular reflections the direct sound after the direct sound after the direct sound



5 reflecting
panels


0 40 O 160 0 4 80 160 0 40 80 160
SSR M5R L5R



10 reflecting
panels


0 40 D 160 0 40 80 160 0 40 80 160
SIOR M10R L10R



20 reflecting
panels 11L


0 40 80 160 0 4 80 160 0 40 80 160
S20R M20R L20R


S5R -stands for-
M10R -stands for-
L20R -stands for-


Short time delay with 5 specular Reflectors
Medium time delay with 10 specular Reflectors
Long time delay with 20 specular Reflectors


Notes:










Table 3-3. The 9 designed impulse responses showing various sets of arrival diffuse reflections
generated by CATT-Acoustics.
Time arrival of Vuthin 40 msec after Between 40-80 msec Between 80-160 msec
diffuse reflections the direct sound after the direct sound after the direct sound




5 diffusing
panels



0 40 80 160 0 40 80 160 0 40 80 160
SSD ______________S M5D L5D



10 diffusing
panels



0 40 80 160 40 80 160 0 40 80 160
S10D MIOD L1iD



20 diffusing
panels 1 Il



0 40 80 160 0 40 80 160 0 40 80 160
S20D M20D L20D


-stands for-
-stands for-
-stands for-


Short time delay with 5 Diffuse reflectors
Medium time delay with 10 Diffuse reflectors
Long time delay with 20 Diffuse reflectors


Notes:


S5D
M10D
L20D









CHAPTER 4
RESULTS AND DISCUSSION

The results obtained from the listening test can be evaluated according to five physical

dimensions of room acoustic measurements: 1) intensity, 2) frequency, 3) time, 4) direction, and

5) distribution temporallyy and directionally) of acoustical energy, which involves how the prior

4 acoustical characteristics interact. The latter is the additional dimension of room acoustic

provided by the room environment. This experiment was set up to study the temporal

distribution of acoustical energy and its effects to perceived acoustic qualities. The analysis of

the experiment has been carried out through 3 aspects from acoustical point of view-source,

path, and receiver.

Source Characteristics

Meyer (1978) analyses the temporal structure of each note produced by different

instruments into three parts:

1) The starting transient, this is the period during which the note develops out of complete
silence until it reaches its stationary state. 2) The stationary state, this is the period during
which the note is not subject to any change. 3) The decay, this is the period during which
the note resounds between the end of the stimulation until complete silence.... The
duration of the starting transient can be measured from the very first beginning to the point
when a level of 3 dB below the final level is reached. The decay time is measured over 60
dB, as is the reverberation time.36

The acoustical characteristics of the orchestral music, trumpet, and piano (sources used in

this experiment) are briefly analyzed as follows.

Orchestral Music

An orchestra comprises a multitude of different instruments which interact partly in solo

fashion and partly gathered in groups. "A medium sized orchestra may generate a fortissimo

sound power level of 118 dB... On the other hand, the lower limit of the orchestral dynamic

36 Meyer, J. (1978). Acoustics and the Performance of Music. Frankfurt/Main, Verlag Das Musikinstrument.









range may be determined by only one instrument playing pianissimo, for example, a clarinet

generating a sound power level of about 58 dB. This leads to an overall dynamic range of about

60 dB for a symphonic orchestra. ,37

The orchestra occupies an extended stage area; a medium sized symphonic orchestra might

take space on the order of 15 m wide and 12 m deep.38 The directivity of the orchestra depends

on the instrument which is used and the arrangement of the multitude of instruments over the

stage area.

An orchestra shell plays an important role in blending the sounds from individual

instruments providing a good sending end to the audience. At the same time it is crucial that the

stage enclosure should provide a good distribution of sound throughout the stage for musicians to

hear each other well for the accuracy of the ensemble.

The temporal structure of the sound of the orchestra depends on the music piece and the

instruments used to play/deliver the designed tempos. Similarly, the frequency spectrum of the

sound of the orchestra also varies from simply a set of harmonics of notes played by only one

instrument to very rich in tones when they are played in tutti of the entire orchestra.

The stage enclosure and the room environment can enhance the perceived music

experience in the hall when it is well designed to support attacks, and various tempos of the piece

produced by different music instruments.

Trumpet

Brass instruments are the loudest instruments in the symphony orchestra. 39 The dynamic

range of trumpet, measuring the sound pressure level when playing an extreme fortissimo


37 Meyer, J. (1993). "The sound of the orchestra." J. Audio Eng. Soc. 41(4): 203-212.
38 Ibid.

39 Meyer, J. (1978). Acoustics and the Performance of Music. Frankfurt/Main, Verlag Das Musikinstrument.









comparing with an extreme pianissimo, is about 30 phons for low notes (53 dBA-80 dBA)

narrow to 13 phons for high notes (83 dBC-93 dBC), measured at a distance of 14 m.40

The shape and size of the bell and its immediate tubing determine the directional radiation

of the trumpet, which is approximately rotationally symmetric around the bell axis. The

relationship of the wavelength of the radiated sound and the bell dimension results in

omnidirectional characteristics for frequencies below 500 Hz. At higher frequencies the higher

intensity is radiated in the direction of the bell; at 2000 Hz and above, most of the sound energy

is concentrated on the axis of the instrument41 as shown in Figure 4-1.

The average starting transient of the high notes is about 40 msec; it gets longer at lower

frequencies-150 msec in the middle register and 180 msec in the low register. The decay time

of the wind instruments is very short, between 70 and 150 msec, due to the fact that the energy

stored in the vibrating air is very small; therefore in practice it is hard for one to hear any

42
resonance decay.4

The frequency content of the notes produced by trumpet can be measured during its

stationary state. Similar to other brass instruments, the trumpet produces rich overtones. The

components of its harmonic sound reach up to very high frequencies; the upper limit of the

spectrum extends up to 8000 Hz in the high register (Figure 4-2 shows the frequency spectrum of

some notes produced by the horn which is very similar to the trumpet's). Additionally the region

where the strongest partial is found is relatively high in frequency. Therefore a bright, brilliant,

and pure tone quality is produced.



40 Burghauser, J., und Spelda, A. (1971). Die akustischen Grundlagen der Instrumentation. Rogensberg.

41 Meyer, J. (1978). Acoustics and the Performance of Music. Frankfurt/Main, Verlag Das Musikinstrument.
42 Ibid.











Note that the duration of a tone impulse affects the perceived loudness43; therefore the


sound of the trumpet gets louder, the longer the duration of the tone is perceived.



550Hz

000Hz










oo1000 H 1250 Hz











1500 2500 Hz 4000 15000 Hz









Figure 4-1. Main directional radiation (0... -3 dB) of the trumpet in the vertical plane.44

dB '- -,- --







o Wf 20 30 m *O 00 40 20 30 40 5O so 20 00 40 o
frequency
Figure 4-2. Sound spectra of a horn in different keys.45


43 Zwicker, E., und Feldtkeller, R. (1967). Das Ohr als Nachrichtenempffnger. Stuttgart.

44 Ibid.









Piano

The piano has a rather consistent dynamic range throughout the frequency range at about

35 phons (50 dBA-85 dBC) for bass and middle notes, and 33 phons (37 dBA-70 dBC) for

higher notes, measured at 10 m distance.46 The vibrations of the sound board and the lid position

are two major factors that determine the patterns of the directional radiation of a grand piano.

Sound is radiated directly upward and is reflected off the lid and also downward and is reflected

from the floor. Figure 4-4 shows directional characteristics of the grand piano for low, middle,

and high registers as a polar diagram in a vertical plane. The sound radiation is relatively

symmetrical in the low register; the directional characteristics affected by the lid begin to show

in the middle register. Towards the high register, the influence of the lid on the directivity are

pronounced, with a strong directionality between 15-35 shown in even the fundamentals of the

notes (at 1000 Hz).47

The starting transient of the piano sound takes between 20 and 30 msec for low register

and gets shorter towards the higher register about between 10 and 15 msec. No stationary sound

situation is formed on the piano.48 A grand piano resonates for a very long time producing a

decay time between 0.5 and 10 sec. Two decay slopes were found; first, the intensity rapidly

decreases then remains decaying slowly for a long time. On average the decay time for the low

register takes longer than that for the higher ones.49 Figure 4-3 shows the frequency dependent

reverberation time of a grand piano for some C keys.


45 Ibid.

46 Burghauser, J., und Spelda, A. (1971). Die akustischen Grundlagen der Instrumentation. Rogensberg.

47 Meyer, J. (1978). Acoustics and the Performance of Music. Frankfurt/Main, Verlag Das Musikinstrument.
48 Ibid.

49 Fletcher, H., Blackham, E. D., & Stratton, R. (1962). "Quality of piano tones." J. Acoust. Soc. Am. 34: 749-761.










The sound of a grand piano is mainly created by the resonant quality of the instrument.

With a sudden change characteristic of the piano sound (without a stationary condition), an

analysis of a sound into a set of harmonics is not possible.50 The frequency content is not formed

of discrete harmonic single frequencies, but instead are continuous functions of frequency. The

timbre of the piano sound therefore is characterized by the anharmonicity of the overtones and

the noise component. "The partial are underlaid by a continuum which has a percussive

character and moreover, is colored by the instrument's strongest resonances."51






10 -- -- -- --
N-I
20- -------






0 nw 200 )00 oo 00 cow

Figure 4-3. Reverberation time of a grand piano for some C-keys using the right pedal.52


















50 Meyer, J. (1978). Acoustics and the Performance of Music. Frankfurt/Main, Verlag Das Musikinstrument.

51 Ibid.

52 Ibid.

























I I I


I I i I I I I I


4 30 20 k0 0 V 20 30 dB 40
lid of grand piano opened
............ lid of grand piano closed


1r .. rr ro" o-






Figure 4-4. Patterns of directional radiation of a grand piano in a vertical plane.53








53 Ibid.









Initial Hypothesis: Effects from Path

Hypotheses were set up according to the characteristics of the temporal distribution of

sound energy in the IRs regardless of the source type. These hypotheses are based on the

definitions given for each acoustical quality presented in the questionnaire (see Appendix A) and

many acoustical theories widely accepted as quantitative measures for those perceived acoustical

qualities, for example, Haas Effect, Sabine's theory (RT), G-strength (G10), clarity index (C80),

initial time delay gap (ITDG), and lateral fraction (LF). Please refer to standard acoustic books

for further explanation on these acoustic theories and parameters. 54,55,56,57,58

1) Loudness

a. The larger the number of reflections the louder the signal (higher total

energy).

b. The sooner the arrival time of specular reflections the louder the signal (less

energy reduction due to distance).

c. Diffuse reflections might soften the attacks and therefore reduce the perceived

loudness.

2) Clarity

a. The sooner the arrival of specular reflections the clearer the signal (high C80

energy within 80 msec).



54 Mehta, M., Johnson J., & Rocafort, J. (1999). Architectural Acoustics: Principles and Design. NJ, Prentice-Hall,
Inc.
55 Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture. NY, Springer.
56 Cremer, L., & Muller, H. A. (1982). Principles and Applications of Room Acoustics. NY, Applied Science
Publishers Ltd.
57 Barron, M. (1993). Audirorium Acoustics and Architectural Design. NY, Routledge.
58 Egan, M. D. (1988). Architectural Acoustics. NY, McGraw-Hill, Inc.









b. The larger the number of specular reflections the clearer the signal (higher

total energy, however these reflections must arrive within the useful time

period).

3) Intimacy

a. The sooner and denser the arrival of specular and diffuse reflections the more

intimate the perceived signal.

4) Reverberance

a. The later the arrival time of specular reflections the more reverberance (RT

gets longer).

b. The larger the number of specular reflections the clearer the perceived

reverberance (more energy).

c. Early diffuse reflections may not play any role on the perceived reverberance.

5) Echoes

a. The later the arrival time of specular reflections the clearer the perceived

echoes.

b. The larger the number of late arrival specular reflections the clearer the

perceived echoes.

c. Diffuse reflections should soften the effect of echoes (reduce the intensity of

the long delay reflections).

6) Source width

a. Diffuse reflections should widen the source width.

7) Texture

a. The diffuse reflections should smooth the perceived texture.









Receiver

The fifty subjects who participated in this experiment are architectural students whose ages

ranged from 18-30 years old. Fifteen of the 50 students play some instruments or are musicians.

Figure 4-5 shows that a large number of the subjects prefer Rock music; Classical, Jazz, and Pop

music are second in popularity among them. These subjects represent a young generation of

general audiences. Data from 47 out of 50 were used for the results analyses, due to the fact that

there were a few students with mild to moderate hearing loss.


Music Preference


40
35
30
25
S20
5 15
0

5
0
Classical Jazz Blues Rock Pop Rap Country Reaggae

Figure 4-5. Music preference of 47 subjects who participated in the listening test.

Results and Discussion

The test results can be divided into 3 source groups-orchestral music, trumpet, and piano;

19 sound fields (SFs) in each group have been evaluated separately. Normalized data shown in

this section were produced by first, for each rated perceived quality, the average of all 19 SFs

was called "total average", then the averaged data of each SF was subtracted by the "total

average". The graphs show the "total average" on the vertical axis set as "0". In Figure 4-6, the

ratings of perceived loudness of the three source types were compared together in one graph.









The process is similar to what was mentioned above only the "total average" in this case was the

averaged data across all three sources.

Each graph shows 19 sound fields arranged from left to right-anechoic sound field, S5R,

S10R, S20R, M5R, M10R, M20R, L5R, L10R, L20R, S5D, S10D, S20D, M5D, M10D, M20D,

L5D, L10D, L20D.

S, M, or L stand for Short, Medium, or Long time delay. Numbers 5R, 10R, 20R, 5D,

10D, or 20D stand for the number of reflecting or diffusing panels provided to obtain each

simulated sound field. Please refer to Chapter 3 for more information on the method.

Due to the fact that the perceived differences among sound fields caused by specular and

diffuse reflections are very subtle, the standard statistical significant tests cannot be applied in

this context. Instead, the trends of the resulting data were observed to understand and explain

perceived acoustical qualities due to specular and diffuse reflections.

Loudness

The results from Figure 4-6 show that, for orchestral music and piano, additional

reflections (regardless of being either specular or diffuse reflections) provide an increase in

perceived loudness as compared to the anechoic signal, which was expected.

However, in the case of the trumpet, the anechoic sound field was perceived as louder then

when diffuse or specular reflections were added. This unexpected result does not agree with the

law of the addition of energies.59 The explanation to this phenomenon is not known.

Finally, there is no trend on the difference in perceived loudness among the designed

impulse responses as initially hypothesized (according to the temporal energy distribution

characteristics). The assertions can be made, first, the receiver in this simulation might be


59 Haas, H. (1972). "The influence of a single echo on the audibility of speech." J. Audio Eng. Soc. 20(2): 146-159.









located too close to the source (14 m), therefore, the direct sound might dominate the effect on

perceived loudness. If the receiver were to be located farther away, the effects on the perceived

loudness due to the additional specular reflections might emerge. Second, the number of 5 large

reflecting panels (3x3m) might produce minimally sufficient energy comparable to that of 10 and

20 reflecting panels. The complete results are available in Appendix B.


Loudness

Orchestra Trumpet Piano

S1.5 -,
'* 1

0.5


0-


S Piano. The orange columns are anechoic signals of each groupI
The results show a subtle.5 increase in perceived clarity for orchestra and piano when early
C 1


-2.
S -/ -.u------------------------
-3

Figure 4-6. Normalized perceived loudness of the three sources: Orchestral music, Trumpet, and
Piano. The orange columns are anechoic signals of each group.

Clarity

The results show a subtle increase in perceived clarity for orchestra and piano when early

specular reflections were added (within 40 msec of the direct sound) (Figure 4-7a, b). For

orchestral music, specular reflections within 160 msec of the direct sound are beneficial for

perceived clarity.

What is interesting is the trend showing the improvement of perceived clarity for piano and

trumpet when a large number of early diffuse reflections arrived sooner (within 40 msec) as

shown in Figure 4-7b,c. This finding is important because it shows that for the same amount of

early sound energy (specular or diffuse reflections) present within the same time windows, the










improvement in perceived clarity provided by diffuse reflections rather than specular reflections


can not be captured by "clarity index" (C80).


Clarity (0)


2.0
1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5




2.0
1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5




2.0
1.5
1.0
0.5
0.0
-0.5
-1.0
-1 5


Clarity (T)


Clarity (P)


Figure 4-7. Clarity: normalized averaged ratings of perceived clarity
Trumpet, and (c) Piano.


(c)
of (a) Orchestral music, (b)


I .i I I*
-jun1 ***~ III


liii


m*I El


I


-
-
-
-
-
-
-








Intimacy

The initially set up hypothesis fails for orchestra and trumpet. Short time arrival of a large

number of specular reflections does not help improve perceived intimacy of the orchestra and

trumpet. This might be because of the control setting for the same initial time delay gap (ITDG)

at 20 msec for all sound fields except the anechoic one.

However, the results from piano show strong improvement (Figure 4-8) on perceived

intimacy when the specular or diffuse reflections arrive sooner. Moreover, early diffuse

reflections (within 40 msec) improve perceived intimacy very significantly while later arrival of

specular reflections reduces the perceived intimacy of the piano sound.

Intimacy (P)

1.5 -.
i 1.0
S0.5

"I 0.0 I i
-0.5


-1.5 a

Figure 4-8. Intimacy: normalized averaged ratings of perceived intimacy of the Piano source.











Reverberance (0)


1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0
-2.5






1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0
-2.5






1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0
-2.5


(b)

Reverberance (P)


(c)
Figure 4-9. Reverberance: normalized averaged ratings of perceived reverberance of (a)
Orchestral music, (b) Trumpet, and (c) Piano.


i Hi II~ EmEEMhE


lull
.I-I I~


.-~- I.
lEEhhIll


-
-
-
-
-
-
-
-


-
-
-
-
-
-
-


-
-
-
-
-
-
-
-


(a)

Reverberance (T)








Echoes (0)


(a)
Echoes (T)






- 1-.11 1 n i


(b)
Echoes (P)


1.5
1.0
0.5
0o.o I l l

2.0


(c)
Figure 4-10. Echoes: normalized averaged ratings of perceived echoes of (a) Orchestral music,
(b) Trumpet, and (c) Piano.


,i,, -I II


-Mm 11Mmm









Reverberance and Echoes

The results shown in Figure 4-9 seem to support all the hypotheses set up earlier that the

perceived reverberance increases as the arrival time of specular reflections increases. The

perceived reverberance also gets clearer when the number of specular reflections increase

regardless of the source type. In addition, the time delay and the amount of diffuse reflections do

not affect perceived reverberance.

Perceived echoes were evaluated similarly to perceived reverberance as shown in Figure 4-

10. A long time delay and a large number of specular reflections increase the perceived echoes

while diffuse reflections do not produce echoes in any time period except for the trumpet sound

source. The trumpet was particularly sensitive to both specular and diffuse reflections arriving

between 80 and 160 msec after the direct sound.

Source Width

Additional reflections, either specular or diffuse, arriving at any time period help widen the

perceived source width of the sound fields as compared to the anechoic sound field

(Figure 4-11).

Only the trumpet source produces results with a clear trend. It shows that the longer the

delay of arrival of specular reflections the wider the perceived source width. An improvement in

perceived source width also appears when diffuse reflections arriving between 80-160 msec after

the direct sound were added to the sound field. The perceived source width for trumpet is

strongly affected according to the change in energy distribution patterns; this might be due to the

high directionality, and the very short decay time characteristics of the trumpet.









Source Width (0)


(a)
Source Width (T)


'~I 2 2 2 ~< ~< ~< -4 -4 -4 2


S-t -t -t 1--4 1-- 1--


(b)
Source Width (P)


(c)
Figure 4-11. Source width: normalized averaged ratings of perceived source width of (a)
Orchestral music, (b) Trumpet, and (c) Piano.


SIII 1 11 11.











Texture (0)

2.0
1.5
1.
2 1.0
.$ 0.5
0.0 -ME M I
& -0.5 *
E -1.0
o
2 -1.5
-2.0 -


(a)
Texture (T)

2.0
1.5
01




-1.0




S0.5
E -1.

-2.0 -. -, -









-2.0









(c)
Figure 4-12. Texture: normalized averaged ratings of perceived texture of (a) Orchestral music,

(b) Trumpet, and (-1.0c) Piano.
o
z -1.5
-2.0


(c)
Figure 4-12. Texture: normalized averaged ratings of perceived texture of (a) Orchestral music,
(b) Trumpet, and (c) Piano.









For orchestral music and piano, the results are not in accord with the initial expectation.

One would expect that the behavior of diffuse reflections might help improve the perceived

source width to some extent due to the temporal and directional smearing effects produced by

diffuse reflections. This might be the case of the wrong choice of words. The same study should

be repeated with the rating of perceived "spaciousness" rather than "source width" to see if

diffuse reflections help improve perceived "spaciousness" of the sound field.

Texture

The results from Figure 4-12 show that overall diffuse reflections help improve perceived

texture of all source types-orchestra, trumpet and piano, as expected.

For orchestral music and trumpet, perceived texture was rated highest when diffuse

reflections arrive within 40-80 msec of the direct sound. Diffuse reflections arriving prior to or

later than the 40-80 msec period, are not rated as high.

A clear trend is shown in the case of piano music, late arrival of specular reflections

reduced rated perceived texture. Diffuse reflections in general are more preferred than specular

reflections; the sooner arrival of diffuse reflections the better-perceived texture. The large

amount of diffuse reflections arriving within 40 msec of the direct sound was rated highest for

perceived texture (Figure 4-12c).

Overall Impression

When the ratings of perceived overall impression were compared with other perceived

music qualities, some similar patterns were found. The following comparisons reveal evidence

that special attention on particular perceived music quality should be addressed differently for

different types of music playing in a room due to the dissimilar temporal structures and

acoustical characteristics of the sources. The trends for orchestra and trumpet are very weak;

only piano sound shows strong trend.











Source Width (0)


W 15 W 15
U0
1 0 0


!| I .m .. i..III os .
050 5
V -1 5
-20 E -1 0
0 0
Z -2 5 -1 5

(a) (b)

Texture (0) Overall Impression (0)

) 20 ) 20
U U
15 15
1 0 1 0
05 I05
10 10

N -0 51N -0 5

S-15 -1 5
0 0
Z -2 0 Z -2 0


(c) (d)
Figure 4-13. A comparison on ratings of perceived acoustic qualities: (a) source width, (b)
intimacy, (c) texture, and (d) overall impression of Orchestral music.

Piano

The results shown in Figure 4-15 express a strong similar trend among the ratings of

perceived "clarity", "intimacy", "texture", and "overall impression" of the piano sound. All of

these 4 perceived acoustic qualities were rated in favor of early arrival of diffuse reflections

(within 40 msec of the direct sound). Due to the impulsive characteristic and the long resonant

decay of the piano sound, this trend shows that there is a need for early diffuse reflections from

room acoustic to support the transient notes, however there is no need for later arrival specular

reflections to interfere with the instrument's own reverberation (piano's resonant decay).

In addition, the ratings of perceived "texture" and "overall impression" correspond well to

one another for all 3 sources which means that the perceived "texture" of the sound field has a

direct impact on the perceived "overall impression" (Figure 4-13, 4-14, 4-15).


Intimacy (0)











Overall Impression (T)


20
15
10
o* == I II I
05 -

-10
-15


2U
15
10 U
. I I ii IIIi.
05 lI" II "

1 0
1 5


Figure 4-14. A comparison on ratings of perceived acoustic qualities: (a) texture, and (b) overall
impression of Trumpet.


Clarity (P)


Intimacy (P)


(a)
Texture (P)






*I JII 1| 11-


(b)
Overall Impression (P)


2 20
S15
S10
S05
(-
O 00
- 05
M -10
-15
0
Z -20


Figure 4-15. A comparison on ratings of perceived acoustic qualities: (a) clarity, (b) intimacy,
(c) texture, and (d) overall impression of Piano.


l .I'- 1111 i.. I.
L.I-ll


Texture (T)









Conclusions

The effects of temporal distribution of specular and diffuse reflections can be concluded as

follows.

Specular reflections

Specular reflections arriving within 40 msec of the direct sound help improve perceived

"clarity" for all 3 source types and improve perceived "intimacy" for orchestral and piano music.

Specular reflections arriving between 40-80 msec of the direct sound help improve perceived

"clarity" for orchestral music. Specular reflections arriving between 80-160 msec of the direct

sound are not desirable.

Strong long delayed reflections provide harsh "texture" for all source types, decrease

perceived "clarity" of trumpet and piano sounds and decrease perceived "intimacy" for piano.

These strong reflections also increase perceived "echoes" and "reverberance".

Diffuse reflections

Diffuse reflections arriving within 80 msec of the direct sound improve perceived "texture"

and "intimacy" for all 3 sources, improve perceived "clarity" for trumpet and piano, and reduce

perceived glare (loudness) for trumpet.

Diffuse reflections arriving between 80-160 msec of the direct sound help widen the

perceived "source width" for orchestra and trumpet, increase perceived "intimacy" for orchestra,

and reduce perceived "echoes" for all 3 sources.

Overall, all diffuse sound fields improved perceived "texture" and preserved perceived

"reverberance" while not providing echoes for all 3 sources. Also the study shows a strong

relationship between the improvement of perceived "texture" and "overall impression".

Figure 4-16 shows the types of reflections and time delays that would improve perceived

"overall impression" for 3 sources-orchestral music, trumpet, and piano.









Orchestral music needs a lot of support from both specular and diffuse reflections

arriving within 80 msec of the direct sound to provide strength for attacks and fullness of tone.

A large number of diffuse reflections arriving after 80 msec are desirable.

Trumpet music needs a large number of diffuse reflections to arrive a little later (between

40-80 msec after the direct sound) to prolong its decay time without increasing its loudness and

widen its source width.

Piano music prefers a large number of diffuse reflections to arrive early (within 40 msec)

to support its impulsive sound providing support for running notes without adding excessive

reverberation to the instrument's own long resonant decay.

Applications

In addition to the required acoustic criteria, the results from this study can be used partly to

suggest designers how to locate reflectors or diffusers in their amphitheaters or concert halls as

shown in Figures 4-17 and 4-18.

Figure 4-17 shows 3 ellipsoids representing boundaries where reflecting panels can be

located to provide the Ist order reflections arriving to the front receiver within 3 time periods-

40, 80, and 160 msec after the direct sound.

Figure 4-18 shows 3 different boundaries for 3 different receiver locations where reflecting

panels can be located to provide the 1st order reflections to arrive at the receivers within 40 msec

after the direct sound.

Another application is for recording purpose; electronic sound effects should be added to

the anechoic sound differently for different sound sources due to their dissimilar temporal

structures to provide proper perceived music preference.










Specular reflections

within 40 80 160 msec
5 reflections / / V

10 reflections / / /

20 reflections / Best

Specular reflections

40 80 160 msec
5 reflections

10 reflections /

20 reflections


Specular reflections

40 80 160 msec
5 reflections V

10 reflections V

20 reflections /


Diffuse reflections





/ / Best




/ /





V Best

Diffuse reflections


/



Best


/
/

/


/ --Provide support to improve perceived "overall impression"



Figure 4-16. Summary of the preferred arrival diffuse and specular reflections to provide
improvement on perceived "overall impression" for 3 sources-Orchestral
music, Trumpet, and Piano.


Orchestra









Trumpet








Piano









Finally, knowledge gained from this study provides rough guidance for a search for a

quantitative measure that should relate well to the perceived "texture". The new measure should

be: 1) frequency dependent at mid and high frequencies-due to the fact that human ears are

more spatially sensitive to high frequency sounds60; 2) time delay dependent-the perceived

texture of sounds in rooms improves as time goes by however the time windows which need to

be carefully treated for each instrument are different; 3) providing acceptable range of perceived

"texture" (degree of smoothness) for each time window-too much diffuse reflections are

sometimes not desirable, the spatial and frequency variation in time which supports perceived

"spaciousness" can get lost if perfectly diffuse sound field occurs too soon.


Figure 4-17. Three ellipsoids representing boundaries where reflecting panels can be located
within to provide the 1st order reflections arriving to the front receiver within 3 time
periods: 40, 80, and 160 msec after the direct sound.




60 Due to the fact that 1) human ear is more sensitive to high frequencies, 2) head shadow effect becomes more
prominent at higher frequencies, and 3) higher tones (frequencies) are highly directional due to their production-
their wavelengths are smaller than the instruments's dimension in which they were produced.































Figure 4-18. Three different boundaries for 3 different receiver locations where reflecting panels
can be located to provide the 1st order reflections to arrive at the receivers with 40
msec after the direct sound.

Comments

Some comments provided by the subjects who had listened to 16 sound fields (SFs) have

been collected. They had listened to each SF twice and evaluated each SF until they finished

evaluating 16 of them. The comments are as follows:

1) The SFs are too short (20 sec. long) especially for orchestral music.

2) Too many acoustical qualities to evaluate for each SF.

3) Confusing range bi-polar rating scale, some are ranged between bad-good, some are ranged
between two extremes (bad-good-bad).

4) The sound level is too soft for orchestral music.

5) Listening to the same short piece twice for 16 times exhausts the ears resulting in
meaningless evaluations in some SFs.

6) Listeners don't know how to evaluate the first few SFs, they don't know what to compare it
to (mostly compare to the previous one).

There are some suggestions on choosing a longer piece of music and play each once.


NV

t i








CHAPTER 5
CONCLUSION AND FUTURE WORK

Conclusion

The differences in perceived acoustical effects produced by specular reflections and diffuse

reflections are very subtle. No significant tests according to statistical norms can be used in this

context. For those who are looking for numerical description and discussion should be

disappointed to read my work and see only, some said, "casual explanations". I think that

making something simple look difficult, complicated, and hard to comprehend by showing

parades of numbers to convince others does not take much effort. On the contrary, being able to

present the most complicated and mysterious issue in the simplest form that everyone can

understand requires a higher level of understanding, originality, and creativity.

There are many ways to look at a problem, many approaches and methods that can be used

to search for knowledge. The works of many researchers (including the author's) in the field of

diffuse reflections presented here are just some of the various approaches used to address the

same problem. Going through a series of failed experiments can be truly painful and

heartbroken; however, to be able to contemplate and appreciate the facts, logic, and reasons why

the experiments failed is something that cannot be shared or talked about. A series of

experiments presented in Appendix C and D are the author's prior works which more or less

have provided some clues to finally shape the well-designed experiment for this dissertation.

Readers are welcome to walk through my anguished search for knowledge in those Appendices

for their own appreciation or choose to skip them.

Future Work

This dissertation provides some logical evidence to the understanding of diffuse reflections

phenomena. Many assertions and explanations made here still call for more concrete scientific









clarification. The final investigation was set to simplify the room acoustic situations to

accentuate the effects of the temporal distribution on perceived music quality. Many results

shown here might disappear when the impulse responses get too crowded by room's reflections

as happens in real room situations. The author hopes that this initial study will open the floor for

many questions to rise and challenge for many more creative investigations to emerge to prove

our curiosity and imaginations. Some of the future investigations that should be carried out

include:

1) A revision of the final experiment presented here according to the listeners' comments and
with better equipment (better headphones); a revision of the questionnaire-replace
perceived "source width" with "spaciousness", a rearrangement of the mixture of arrival
specular and diffuse reflections at each time window, and an addition of the room's
reverberant tail to the SFs.

2) Study the impulse responses of different room shapes to search for the room's signature
which should provide answers to why the traditional shoe-box shape provides good room
acoustic. Upon knowing the room shapes' signatures, designers can then be able to alter
their room signatures if needed.

3) A search for quantitative measure/parameter that can describe the characteristics of the
distribution of energy of the IRs that can relate well with the perceived acoustical quality of
"texture".









APPENDIX A
QUESTIONNAIRE

ACOUSTICS EVALUATION SHEET DEFINITIONS


LOUDNESS:


CLARITY:


INTIMACY:


REVERBERANCE:


SOURCE WIDTH:


TEXTURE:


ECHOES:


The overall loudness or strength of the sound that you are hearing


The degree to which notes or words are distinctly separated in time and
clearly heard.

The auditory impression of the apparent closeness of the orchestra.


The blending of sounds into subsequent following sounds. The
persistence of sound in the space.

The apparent widening of the sound source. The source occupying an area
as opposed to a single point in space.

The smoothness or harshness of sound persisting in a space


Long delayed reflections that are clearly audible.


OVERALL IMPRESSION: Overall impression of the acoustical quality that you are hearing.


PERSONAL DATA (optional)

Music preference (circle one or more):

Classical Jazz Blues

O th er .........................

Number of orchestral concerts attended (circle one).

None 1-10 11-20

Are you a musician? Yes

Do you have normal hearing? Yes


Rock


Pop


Rap


30 or more


21-30

No

No










VIRTUAL SOUND FIELD EVALUATION SHEET


1
Not loud
enough

1
Not clear
enough


1 2
Not intimate
enough


REVERBERANCE 1 2
Not reverberant
enough

SOURCE WIDTH 1 2
Not wide


TEXTURE



ECHOES


OVERALL
IMPRESSION


1
Not smooth
(harsh)

1
None detected


1
very bad


2 3 4 5 6 7
too loud


2 3 4 5 6 7
extremely
clear


3 4 5 6 7
extremely
intimate

3 4 5 6 7
too
reverberant

3 4 5 6 7
extremely
wide


2 3 4 5 6 7
extremely
smooth

2 3 4 5 6 7
clearly heard


2 3 4 5 6 7
very good


COMMENTS:


LOUDNESS



CLARITY


INTIMACY


0
cannot
tell

0
cannot
tell

0
cannot
tell

0
cannot
tell

0
cannot
tell

0
cannot
tell

0
cannot
tell

0
cannot
tell










APPENDIX B
RESULTS FROM LISTENING TEST


Loudness (0)


o 0.5
:-

2 -0.5
0.
* -1.0
' -1.5
E -2.0
o
z -2.5


0.5
S0.0
2 -0.5
0.
* -1.0-
= -1.5
E -2.0
0
25


Loudness (T)


- '- '


Loudness (P)


o 0.5
I I m i1 I I
2 -0.5
S-1.0
5 -1.5
E -2.0
o
z -2.5 -
-3.0
(c)
Figure B-1. Loudness: normalized averaged ratings of perceived loudness of (a) Orchestral
music, (b) Trumpet, and (c) Piano.


**~**.* I~. is.....


-- m I- h-I-,-, '


-0.U


-,3.U










Clarity (0)


2.0
1.5
S 1.0
U
0.5
Eoo U. l I III
-0.5
z -1.0 -
-15

(a)
Clarity (T)

2.0
1.5
5 1.0

Ie I i ="
-0.5



0 -.0
-1.5 E



(b)
Clarity (P)

2.0
8 1.5
1.0

0.5
I I I





-1.5

(c)
Figure B-2. Clarity: normalized averaged ratings of perceived clarity of (a) Orchestral music, (b)
Trumpet, and (c) Piano.








Intimacy (0)


(a)
Intimacy (T)


I I I I.1


mu Illilli.


- I .- lQ
(b)
Intimacy (P)
1.5
1.0
0.5
I0.0!I I
-0.5 -
-1.05


Figure B-3. Intimacy: normalized averaged ratings of perceived intimacy of (a) Orchestral
music, (b) Trumpet, and (c) Piano.


I 'I'*I ** miElimi II











Reverberance (0)


1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0
-2.5






1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0
-2.5






1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0
-2.5


(b)

Reverberance (P)


(c)
Figure B-4. Reverberance: normalized averaged ratings of perceived reverberance of (a)
Orchestral music, (b) Trumpet, and (c) Piano.


i Hi II~ EmEEMhE


lull
.I-I I~


.-~- I.
lEEhhIll


-
-
-
-
-
-
-
-


-
-
-
-
-
-
-


-
-
-
-
-
-
-
-


(a)

Reverberance (T)









Source Width (0)


(a)
Source Width (T)


'~I 2 2 2 ~< ~< ~< -4 -4 -4 2


S-t -t -t 1--4 1-- 1--


(b)
Source Width (P)


(c)
Figure B-5. Source width: normalized averaged ratings of perceived source width of (a)
Orchestral music, (b) Trumpet, and (c) Piano.


SIII 1 11 11.











Texture (0)

2.0
1.5
1.
2 1.0
0.5
0.0 M -E l .I E ..
N -0.5
E -1.0
o
2 -1.5
-2.0 -


(a)
Texture (T)

2.0
1.5
01




-1.0




S0.5
E -1.

-2.0 -. -, -









-2.0









(c)
Figure B-6. Texture: normalized averaged ratings of perceived texture of (a) Orchestral music,

(b) Trumpet, and (-1.0c) Piano.
o
z -1.5
-2.0


(c)
Figure B-6. Texture: normalized averaged ratings of perceived texture of (a) Orchestral music,
(b) Trumpet, and (c) Piano.








Echoes (0)


(a)
Echoes (T)






1-.11 1 n i


(b)
Echoes (P)




,I U


-- -* -^ ,
'Q &Q &Q -, ]^ ]& Q Q ^ ^ ^ ^ ]


(c)
Figure B-7. Echoes: normalized averaged ratings of perceived
(b) Trumpet, and (c) Piano.


echoes of (a) Orchestral music,


,il,_ .1 1


-. M 1 ,











Overall Impression (0)


2.0
1.5
2 1.0
02 0.5
0.0 mmmm
N -0.5
E -1.0
o0
z -1.5
-2.0

(a)

Overall Impression (T)

2.0
1.5
S1.0
0.5
oo I II In .
I II
o0o 0-
N. -0.5
E -1.0
0
z -1.5 .
-2.0 -_-

(b)

Overall Impression (P)

2.0
1.5

S 0.5








(c)
Figure B-8. Overall impression: normalized averaged ratings of perceived overall impression of
(a) Orchestral music, (b) Trumpet, and (c) Piano.









APPENDIX C
PILOT STUDY I

Computer Modeling of Charlotte Auditorium

Charlotte auditorium was chosen to study the behavior of diffuse reflections at 12

receivers. Two computer models were built using CATT-Acoustic v.8: one having pyramid

diffusing panels on the side and rear walls, and curved diffusers as acoustic clouds; another one

without any diffusing surfaces as shown in Figure C-1.

7 9 11








(a)












0Y (b)








Figure C-1. Computer simulation of the Charlotte auditorium with source (AO) and receiver
locations specified: (a) without diffusers, (b) with diffusers on the walls and the
ceiling.









An omnidirectional sound source (AO) is located 4 feet high from the stage level; twelve

receivers were selected for this study as shown in Figure B-la. This study includes 3 major

parts:

1) The study of effective time windows of the first order specular and diffuse reflections by
looking at the impulse responses and hedgehog representations at 12 receiver locations.

2) A comparison of room acoustic parameters of the two room conditions (with and without
diffusers) to search for evidence of the acoustical effects from diffusers.

3) Listening test: four anechoic sounds including speech were convolved with the impulse
responses (including both early part and reverberant tail) from both halls (with and without
diffusers) to search for any perceived differences between these two room conditions.

First, impulse responses of the first order specular and diffuse reflections (obtained by the

image source method) at 6 receiver locations were plotted by hand to compare with the impulse

responses of the first order diffuse reflections and up-to third order specular reflections obtained

from CATT-Acoustic as shown in Figure C-4.

Image source Source








SReceiver 1
......... Specular sector limit

Figure C-2. Image source method used to derive arrival time of specular and diffuse reflections
in the impulse responses.


















W


Figure C-3. Three dimensional image source method used to produce room's IR's at selected
receivers.


~ 'VP


tr-7


- -- ihilIhw i -





hj l iU _, ..:


.Ini


i^1 1I


(a) 1st order reflections & (b) 3rd order reflections & (c) 3rd order reflections &
diffusion 1st order diffusion of the room w/o 1st order diffusion of the room
diffusers produced by CATT- with diffusers produced by
Acoustic CATT-Acoustic
Figure C-4. Comparison of the room's impulse reponses of the room's 1st order diffusion at
different receiver locations: (1) receiver 1, (3) receiver 3, (5) receiver 5.


^^7









The results show:

1) the time window where the first order specular and diffuse reflections appear in the IR's get
narrower to within 70 msec of the direct sound as receivers move toward the rear of the hall
while the effective time windows of the receivers in the front area are wider, as long as
within 125 msec of the direct sound.

2) Early diffuse reflections provide continuous energy distribution and bridge the gaps among
the early specular reflections and late reverberant tail (temporal smearing).

The hedgehog representations show the effect on directional smearing. The plots show

that uniform directional energy distribution appears sooner in the room with diffusers than in the

room without diffusers. The time window of the directional smearing corresponds well to the

first order specular and diffuse reflections shown in the IR's. In another words, not long after the

effective time window of the 1st order diffuse reflections, the sound field reaches its (roughly)

spatially uniform distribution. Figure C-5.

















r 100% 100% r 74% ~, 7'7% r 82% 'p 80%


00 < t < 20.0 ms
'p 53%















1.0 < t < 110.0 ms


0.00 < t < 20.0 ms
S80%















80.0 < t < 110.0 ms


.0 < t < 50.0 ms


.0 < t < 150.0 ms


r. 78%


tf-4rI\


.0 < t < 50.0 ms


.p 78%







110.0 < t < 150.0 ms


vp 75%






50.0 < t < 80.0 me
rF 80% up 81%


75%






150.0 < t < 200.0 me


r 90% 'p 80%








a 66%






a 50.0 < t < 80.0 ME

P= 77% up 79%

//^ \ 's


150.0 < t < 200.0 me


Figure C-5. Hedgehog presentations at receiver 3: (a) without diffusers, (b) with diffusers.


r- 100%


u- 100%


r- 74%


- 77%


r= 82%


u, 80%












Second, no significant differences are shown through the acoustical parameters generated


by CATT-Acoustic. Only IACC improves at high frequency for the hall with diffusers.


Average IACC


12
1
'-- 08
-- No _diff T-30 08
S06
-- DlffT-30 04

S02


--No diff IACC
-Diff ACCO


125 250 500 1k 2k 4k 8k 16k
Octave Band Center Frequency (Hz)


Average C-80


125 250 500 1k 2k
Octave Band Center Frequency (Hz)


4
3
- No diff LEF 2
- Diff LEF V 1
0
-1
-2


Octave Band Center Frequency (Hz)


Figure C-6. Average RT, IACC, LEF, and C80 of all receivers from 2 room conditions (with
and without diffusers) across 6 frequency bands.





Finally, four anechoic pieces: speech, guitar, cello, and trumpet were convolved with the


impulse responses obtained from the two halls (with and without diffusers). The listening test


was conducted using the questionnaire developed by Torres (2000)61 (Figure C-7). The results


show that subtle differences can be detected especially for speech due to its impulsive high


frequency. The difference in the case of cello is harder to detect.










61 Torres, R. R., & Kleiner, M. (2000). "Audibility of diffusion in room acoustics auralization: an initial

investigation." Acustica 86(6): 919-927.


Average RT (T-30)


125 250 500 1k 2k 4k
Octave Band Center Frequency (Hz)


Average LEF












Name:
Do you have a normal hearing'
Are you an experience listener? Musician?

Compare the difference between the pairs of sounds.

Speech S _A: S I_B

Not different 2--- --- 4 Clearly different

Character of difference





Figure C-7. The questionnaire used for the listening test developed from Torres' s.62



Table C-1. Statistical results from the listening test, 4 represents clearly different, 0 represents
that no perceived difference can be detected.


Variable N Mean StDev SE Mean 95% CI Coefficient of
Variation

Speech 9 2.00 1.12 0.37 (1.14, 2.86) 56%

Guitar 9 1.33 1.32 0.44 (0.32, 2.35) 99%

Cello 9 0.89 1.05 0.35 (0.08, 1.70) 118%

Trumpet 9 1.33 0.50 0.17 (0.95, 1.72) 38%


62 Torres, R. R. (2000). Studies of Edge Diffraction and Scattering: Applications to Room Acoustics and
Auralization. Department of Applied Acoustics. Goteborg, Chalmers University of Technology.









APPENDIX D
PILOT STUDY II

Black-box Theater: Physical Model and Computer Modeling

The second approach to study the effects of diffusion was done by altering the path, i.e.,

different sizes of diffusers were installed in a model room, and observations of the change in the

room's impulse responses and acoustical parameters were made. The Black Box Theater at the

Center of Performing Arts of the size 11.8 x 18 x 9.3 m (or 38.7 x 59 x 30.5 ft) has been

simplified and built as a base model. Four sizes of (square-rod) diffusers were designed with

their dimensions corresponding to the wavelength of the sound to scatter the energy at four

frequency bands: 500 Hz, 1 kHz, 2 kHz, 4 kHz. Six models with diffusers installed on 3 walls

and a ceiling were constructed by: 1) 1:20 physical model to observe the impulse responses, and

2) computer modeling to observe the change in acoustical parameters.

First, 1:20 physical models were built, all diffusers were made of 3-ply chipboard

(painted, and unpainted). Band-pass filtered impulse responses were obtained from each model

in 3 frequency bands.
Impulse Response Impulse Response Impulse Response
250 Hz- 1 kHz 500 Hz- 2 kHz 1 4 kHz







(a)






(b)

Figure D-1. Two sets of 1:20 physical model tests were carried out with (a) painted, and (b)
unpainted models. Three frequencies band-pass filtered IR's were measured.





























90 crn


Spark source


U
Ln

Microphone 1
II- I


Plan


Section


^^^^d^U^^*1_-ff^^^^^^^^^^_


Isometric


Figure D-2. Plan, section, and isometric of the base model with the source and the receiver's
locations.



















Model 2
3 walls and the ceiling covered with
diffusers, dimension ~ the
wavelength of 500 Hz


Model 3
diffusers dimension ~ the
wavelength of 1 kHz


Model 4
diffusers dimension ~ the wavelength
of 2 kHz


Model 5
diffusers dimension ~ the
wavelength of 4 kHz


Model 6
Mixture of diffusers dimension ~ 1
& 2 kHz


Figure D-3. Six models with different sizes of diffusers for each set of test.


Energy
bound


Simplified
Impulse response


Figure D-4. The method of IR analysis: 1) find the boundary of energy fluctuation, 2) delete the
data between the upper and the lower bounds, 3) Simplified IR is obtained.


Model 1
Bare room










Table D-1. Impulse responses from painted model.

IR IR
(250 Hz-1 kHz) (500 Hz-2 kHz)


Model 1
(Bare room)


Model 2
(500 Hz
diffusers)


Model 3
(1 kHz
diffusers)






Model 4
(2 kHz
diffusers)


Model 5
(4 kHz
diffusers)






Model 6
(1+2 kHz
diffusers)


,i ~I


"' '~ i~


IR
(1-4 kHz)









Table D-2. A comparison of the simplified impulse responses from painted model.


IR
(250 Hz-1 kHz)


IR
(500 Hz-2 kHz)


IR
(1-4 kHz)


Model 1
(Bare
room)





Model 2
(500 Hz
diffusers)





Model 3
(1 kHz
diffusers)


Model 4
(2 kHz
diffusers)





Model 5
(4 kHz
diffusers)





Model 6
(1+2 kHz
diffusers)


K...





~1









Impulse Response Analysis

Analyses of impulse responses' characteristics were made in each test by graphically

approximating the range of energy fluctuation and average energy decay. Then the simplified

impulse responses were created by using the average upper and lower bounds of the energy

fluctuation among 6 models as a guided energy band, any strong reflections outside this band

were left unerased. Figure D-4 shows the method of analysis of the impulse response.

Results from physical model testing show that:

1) The sound energy decays faster in the unpainted models which corresponds well with the
differences of materials' absorption coefficients, which follows Sabine's RT theory.

2) The average energy decay curves among the six painted models are quite similar which
means that there are approximately the same RTs among six painted models.

3) Wider energy fluctuations are found at low frequency band-pass IR's.

4) The only conclusion that can be drawn from the simplified impulse responses is that Models
3-6 show improvements from Model 1 for the number of strong repetitive spikes in Model
3-6 are much less than in Model l's.

The lesson learned from this experiment is that changing the types of diffusers at the room

boundaries cannot control the characteristics of the impulse responses temporally. Due to the

fact that there is no standard procedure to graphically evaluate the temporal distribution of those

spikes (strong reflections) shown in the impulse responses, further conclusions from these data

cannot be drawn.























200 400 600 msec


200 400 600 msec


1B




00






260 400 600 msec


NM 4
odelIF













S6








Moel 6





Made 6


(a) (b)
Figure D-5. Comparisons of the energy bounds among 6 models' IR's at 3 band-pass
frequencies filtered of: (a) painted models, (b) unpainted models.


The same six models were created in CATT-Acoustic. The results from the computer


model show:


1) RTs of 6 models correspond well with the hand-calculation's and with 1:20 physical model
test's.


2) Strong G10 is found in the bare room (strong reflections).


3) C80's vary according to the size of diffusers (large diffusers, less ITDG, higher C80).


4) No conclusion can be drawn from the IACC's and LEF/LF's data.


Energy bounds
250 Hz 1 kHz


Energy bounds
500 Hz 2 kHz


Energy bounds
1- 4 kHz


30


20










30


20





200 400 600 msec


dB

30


20 00 400 600 e






200 400 600 Msec


&" 2

















Modd 4
Hodfl
MoeI














No strong relationship can be found between diffusion and current room acoustic



parameters (generated by CATT-Acoustic).


Test A Comparison of T-30


Test A Comparison of IACC


125 250 500 1 k 2 k 4 k
Octave Band Center Frequency (Hz)


(a)

Test A Comparison of G-10


-s- Model 1
Model 2
Model 3
-X- Model 4
-i- Model 5
-*- Model 6


125 250 500 1 k 2k
Octave Band Center Frequency (Hz)


(b)

Test A Comparison of C-80


250

200 -

g 150

g 100-
i-
0 50

0 00









200

150

100

50

00


125 250 500 1 k 2 k 4 k
Octave Band Center Frequency (Hz)


SU
20
10
00
1 0

-20
-30


Octave Band Center Frequency (Hz)


Test A Comparison of LFC


125 250 500 1 k 2k 4k
Octave Band Center Frequency (Hz)


-*--Model 1
Model 2
Model 3
---Model 4
--Model 5
---Model 6


(e)


Figure D-6. Comparisons of acoustical parameters of the 6 painted models simulated by CATT-

Acoustic: (a) T30/RT, (b) IACC, (c) G10, (d) C80, (e) LFC /LF.


-- Model 1
Model 2
Model 3
---Model 4
-- Model 5
-*-o-Model 6


-*- Model 1
Model 2
Model 3
-x- Model 4
- Model 5
-*-Model 6


-*--Model 1
Model 2
Model 3
--Model 4
--Model 5
-a--Model 6









LIST OF REFERENCES


Antonio, P. D., & Cox, T. (1998). "Two decades of sound diffusor design and development Part
2: prediction, measurement, and characterization." J. Audio Eng. Soc. 46(12): 1075-1091.

Barron, M. (1993). Audirorium Acoustics and Architectural Design. NY, Routledge.

Barron, M., & Marshall, A. H. (1981). "Spatial impression due to early lateral reflections in
concert halls: The derivation of a physical measure." J. Sound Vibration. 77: 211-232.

Beranek, L. (1962). Music, Acoustics and Architecture. New York, Wiley.

Beranek, L. (2003). Concert Halls and Opera Houses: Music, Acoustics and Architecture. NY,
Springer.

Bolt, R. H., & Doak, P. E. (1950). "A tentative criterion for the short-term transient response of
auditoriums." J. Acoust. Soc. Am. 22: 507-509.

Burghauser, J., und Spelda, A. (1971). Die akustischen Grundlagen der Instrumentation.
Rogensberg.

Cervone, R. P. (1990). Designing for Music: The Subjective and Objective Evaluation of Three
Rooms for Music Listening. Masters Thesis, School of Architecture. Gainesville, University of
Florida

Cox, T. J. (1996). "Designing curved diffusers for performance spaces." J. Audio Eng. Soc.
44(5): 354-364.

Cox, T. J., & D' Antonio, P. (2004). Acoustic Absorbers and Diffusers: Theory, Design and
Application. NY, Spon Press.

Cremer, L., & Muller, H. A. (1982). Principles and Applications of Room Acoustics. NY,
Applied Science Publishers Ltd.

Dalenback, B.-I. (2002). CATT-Acoustic v8.0 User's Manual. Gothenburg, Sweden.

Dalenback, B.-I., Kleiner, M., & Svensson, P. (1994). "A macroscopic view of diffuse
reflection." J. Acoust. Soc. Am. 42(10): 793-806.

Egan, M. D. (1988). Architectural Acoustics. NY, McGraw-Hill, Inc.

Fletcher, H., Blackham, E. D., & Stratton, R. (1962). "Quality of piano tones." J. Acoust. Soc.
Am. 34: 749-761.

Gold, M. A. (1994). An Evaluation of Perceived Acoustic Spatial Impression. Masters Thesis,
School of Architecture. Gainesville, University of Florida.









Haan, C. H., & Fricke, F. R. (1993). Surface diffusivity as a measure of the acoustic quality of
concert halls. Proceedings of Conference of the Australia and New Zealand Architectural
Science Association, Sydney.

Haas, H. (1972). "The influence of a single echo on the audibility of speech." J. Audio Eng. Soc.
20(2): 146-159.

Junius, W. (1959). "Raumakustische Untersuchunger mit Neueren Messverfahren in der
Liederhalle Stuttgart." Acustica 9: 289-303.

Kuttruff, H. (1965/66). "Uber Autokorrelationsmessungen in der Raumakustik." Acustica 16:
166-174.

Mehta, M., Johnson J., & Rocafort, J. (1999). Architectural Acoustics: Principles and Design.
NJ, Prentice-Hall, Inc.

Meyer, J. (1978). Acoustics and the Performance of Music. Frankfurt/Main, Verlag Das
Musikinstrument.

Meyer, J. (1993). "The sound of the orchestra." J. Audio Eng. Soc. 41(4): 203-212.

Niese, H. (1961). "Die Messung der Nutzschall- und Echogradverteilung zur Beurteilung der
Horsamkeit in Raumen." Acustica 11: 202-213.

Reichardt, W., Abdel Alim, 0., und Schmidt, W. (1975). "Definition und Mef3grundlage eines
objektiven MaBes zur Ermittlung der Grenze zwischen brauchbarer und unbrauchbarer
Durchsichtigkeit bei Musikdarbietung." Acustica 32: 126-139.

Thiele, R. (1953). "Richtungsverteilung und Zeitfolge der Schallruckwurfe in Raumen." Acustica
3: 291-302.

Torres, R. R. (2000). Studies of Edge Diffraction and Scattering: Applications to Room
Acoustics and Auralization. Department of Applied Acoustics. Goteborg, Chalmers University
of Technology.

Torres, R. R., & Kleiner, M. (2000). "Audibility of diffusion in room acoustics auralization: an
initial investigation." Acustica 86(6): 919-927.

Zwicker, E., und Feldtkeller, R. (1967). Das Ohr als Nachrichtenempfanger. Stuttgart.









BIOGRAPHICAL SKETCH

Pattra Smitthakorn received her bachelor's degree in architecture from Chulalongkorn

University, Bangkok, Thailand, in 1994. In 1997, she was awarded a scholarship from the Royal

Thai Government to study abroad; in 2000, she received her Master of Science in Architectural

Studies majoring in architectural acoustics at the University of Florida. She enrolled in a PhD

program in Building Performance Diagnostics at Carnegie Mellon University in 1999. Two

years of education at CMU incubated her critical thinking in mathematics with the help of a

tutor, Aram Tangboonduangjit. In 2001, she transferred back to the University of Florida to

continue her education in architectural acoustics. She found her inspiration in 2004 when

studying mathematics with Professor Li-Chien Shen; together with Professor Alexander

Berkovich and Professor Yuli Rudyak, they rigorously trained her to logically prove assertions,

thoughts, and problems of various kinds. She then proceeded to work on the most difficult issue

in the field of architectural acoustics and received her PhD degree in 2006. She will return to her

home country after her graduation and become a professor at King Mongkut's University of

Technology Thonburi, Bangkok, Thailand.