AAEESS112266th CCoonnvveenn ttiioonn PPrroo ggrraamm May 7 – 10, 2009 M,O,C, Center, Munich

The AES has launched a new opportunity to recognize Chair: Damian Murphy , University of York, York, UK student members who author technical papers. The Stu - dent Paper Award Competition is based on the preprint 09:00 manuscripts accepted for the AES convention. Forty student-authored papers were nominated. The P1-1 20 Things You Should Know Before excellent quality of the submissions has made the selec - Migrating Your Audio Network to IP— Simon tion process both challenging and exhilarating. Daniels , APT, Belfast, Northern Ireland, UK The award-winning student paper will be honored dur - For many years, synchronous networks have ing the convention, and the student-authored manuscript been considered the industry standard for audio will be published in a timely manner in the Journal of the transport worldwide. Balanced analog copper Audio Engineering Society . circuits, microwave, and synchronous based Nominees for the Student Paper Award were required systems such as V.35/X.21 or T1/E1 have been to meet the following qualifications: the traditional choice for studio transmitter and (a) The paper was accepted for presentation at the inter-studio links in professional audio broadcast AES 126th Convention. networks. Readily available from all major ser - (b) The first author was a student when the work was vice providers, the popularity of synchronous conducted and the manuscript prepared. links has been largely due to the fact that they (c) The student author’s affiliation listed in the manu - offer dedicated, reliable, point-to-point and bi- script is an accredited educational institution. directional communication at guaranteed data and (d) The student will deliver the lecture or poster pre - error rates. However, the reign of synchronous sentation at the Convention. links as the preferred choice for STLs is currently coming under threat from a new challenger, in the * * * * * form of IP-based network technology. The winner of the 126th AES Convention Convention Paper 7651 Student Paper Award is: Modeling of External Ear Acoustics for Insert 09:30 Headphone Usage— Marko Hiipakka (Presenting Author), Miikka Tikander, P1-2 Deploying Large Scale Audio IP Networks— Matti Karjalainen Kevin Campbell , APT, Belfast, Northern Ireland, Convention Paper 7739 UK This paper will examine the key considerations for To be presented on Friday, May 8 in Session P15 those interested in deploying large-scale IP audio —Posters: Hearing networks. It will include an overview of the main * * * * * challenges and draw on the experience of national public broadcasters who have already migrated to Student Event/Career Development IP. We will provide an overview of the key con - STUDENT SCIENCE SPOT cerns such as jitter, delay, and link reliability that Thursday, May 7 through Sunday, May 10 are valid for an IP network of any size. However, Hall 2 Foyer this paper will focus mainly on the issues arising Ongoing presentations of student's non-commercial sci - from the greater complexity and scale of large na - entifically challenging works will be given throughout the tional and country-wide deployments. The paper Convention in an accessible way. They will offer proto - will use illustrations and network applications from types or demonstrations, including a student poster ses - real-world deployments to illustrate the points. sion. Within the scope of their presentation, the students Convention Paper 7652 will have the opportunity to present themselves and their Paper presented by Hartmut Foerster institutes in an appropriate way. The SSS will be hosted by the AES Student Section Graz (AES SSG), supported 10:00 by the SDA and managed in cooperation with the stu - dents who share their projects at the booth. P1-3 A Spatial Filtering Approach for Directional Audio Coding— Markus Kallinger, Henning Ochsenfeld, Giovanni Del Galdo, Fabian Kuech, Session P1 Thursday, May 7 Dirk Mahne, Richard Schultz-Amling, Oliver 09:00 – 11:30 Room K3 Thiergart , Fraunhofer Institute for Integrated AUDIO FOR TELECOMMUNICATIONS Circuits IIS, Erlangen, Germany ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 1 Technical Progra m In hands-free telephony, spatial filtering tech - Session P2 Thursday, May 7 niques are employed to enhance intelligibility of 09:00 – 11:30 Room K4 speech. More precisely, these techniques aim at reducing the reverberation of the desired speech AUDIO FOR GAMES AND INTERACTIVE MEDIA signal and attenuating interferences. Additional - ly, it is well-known that the spatially separate reproduction of desired and interfering sources Chair: Michael Kelly , Sony Computer Entertainment, enhance intelligibility of speech. For the latter UK task, Directional Audio Coding (DirAC) has proven to be an efficient method to capture and 09:00 reproduce spatial sound. In this paper we pro - pose a spatial filtering processing block, which P2-1 Viable Distribution of Multichannel Audio- works in the parameter domain of DirAC. Simu - over-IP for Live and Interactive “Voice lation results show that compared to a standard Talent”-Based Gaming Using High-Quality, beamformer the novel technique offers signifi - Low-Latency Audio Codec Technology— cantly higher interference attenuation, while Gregory Massey , APT, Belfast, Northern Ireland, introducing comparably low distortion of the UK desired signal. Additional subjective tests of The delivery of multichannel audio—from mono speech intelligibility confirm the instrumentally to surround sound—in real-time over public IP obtained results. networks for the purpose of interactive crowd- Convention Paper 7653 participant gaming presents a significant design engineering challenge to games developers, 10:30 console manufacturers, ISPs, and CDNs. Lever - aging expertise gained in professional broad - P1-4 A New Bandwidth Extension for Audio casting and recording studio postproduction, Signals without Using Side-Information— Kha APT has developed a robust and scalable audio Le Dinh, Chon Tam Le Dinh, Roch Lefebvre , codec technology that meshes with popular Université de Sherbrooke, Sherbrooke, Quebec, gaming systems to realize low-latency distribu - Canada tion of high-quality audio for immersive, instanta - The use of narrow bandwidth (300 – 3400 Hz) in neous audio experiences in massively multi- the current telephone network limits the percep - player online games involving interactive tual quality of telephone conversations. Chang - audience responses to vocal/singing talent. ing to wideband network is a solution that can Convention Paper 7656 help to improve quality, but it will need a long Paper presented by David Trainor time to upgrade. Thus, bandwidth extension can be seen as an alternative solution during the 09:30 transition time. A new bandwidth extension P2-2 Elevator: Emotional Tracking Using Audio/Vi - method is presented in this paper. Without using sual Interaction— Basileios Psarras, Andreas any side-information, the proposed method can Floros, Marianna Strapatsakis , Ionian University, be applied as a post-processing step at the ter - Corfu, Greece minal devices, maintaining the compatibility to the current telephone network, and thus, no The research interest on modeling everyday hu - modification is needed in the network nodes. man emotions and controlling them through typi - Experimental results show that the proposed cal multimedia content (i.e., audio and video solution can help to improve significantly the per - data) has recently increased. In this paper an in - ceptual quality of narrowband telephone signal. teractive methodology is introduced for detect - Convention Paper 7654 ing, controlling, and tracking emotions. Based on the above methodology, an interactive audiovi - 11:00 sual installation termed “Elevator” was realized, aiming to analyze and manipulate simple emo - P1-5 Feature Selection vs. Feature Space tions of the participants (such as anger) using Transformation in Music Genre Classification simplified emotion detection audio signal pro - Framework— Hanna Lukashevich , Fraunhofer cessing techniques and specifically selected Institute for Digital Media Technology IDMT, combined audio/visual content. As a result, the Ilmenau, Germany human emotions are “elevated” to pre-defined Automatic classification of music genres is an levels and appropriately mapped to visual con - important task in music information retrieval tent, which corresponds to the emotional research. Nearly all state-of-the-art music genre “thumbnail” of the participants. recognition systems start from the feature ex - Convention Paper 7657 traction block. The extracted acoustical features often could tend to be correlated or/and redun - 10:00 dant, which can cause various difficulties in the P2-3 Applications of Bending Wave Technology classification stage. In this paper we present a in Human Interface Devices— Neil Harris , comparative analysis on applying supervised New Transducers Ltd. (NXT), Cambridge, UK Feature Selection (FS) and Feature Space Transformation (FST) algorithms to reduce the The application of bending-waves to so-called feature dimensionality. We discuss pros and “flat panel loudspeakers” has often been the cons of the methods and weigh the benefits of topic of papers at AES Conventions. This each one against the others. paper looks at other interesting applications of Convention Paper 7655 the technology that have, or are beginning to

2 Audio Engineering Society 126th Convention Program, 2009 Spring have commercial pull. These applications are Panelists: Thomas Sporer, Fraunhofer Institute for also part of the interface between human and Digital Media Technology IDMT, Ilmenau, machine, but focus on the sense of touch Germany rather than of hearing. The idea of a touch Florian Wickelmaier, University of Tübingen, screen is not new but is only now becoming Tübingen, Germany ubiquitous with a new generation of devices, typified by the i-Phone. If touch sensors are Listening tests have become an important part of the the analog of the microphone, then haptic development of audio systems (CODECS, loudspeak- feedback generators are the analog of the ers, etc.). Unfortunately, even the simplest statistics loudspeaker. Bending waves are beginning to (mean and standard deviation) are often misused. find application here too. This workshop will start with a basic introduction to statis- Convention Paper 7658 tics, but room will be given to discuss the pertinence of some commonly-used tests, and alternative methods will be proposed, thereby making it interesting for more 10:30 experienced statisticians as well. The following topics will be P2-4 Designing Auditory Display Menu Interfaces covered (among others): experimental design, distributions, —Cues for Users' Current Location in hypothesis testing, confidence intervals, analysis of paired Extensive Menus— Erik Sikström, Jan Berg , comparisons and ranking data, and common pitfalls Luleå University of Technology, Luleå, Sweden potentially leading to wrong conclusions.

This paper reviews the current research in audi - Tutorial 1 Thursday, May 7 tory display in search for design guidelines for 09:00 – 11:00 Room K1 presenting the contents in audio-only menu interfaces. The aim of the review is to find new A NEW DOWNMIX ALGORITHM OPTIMIZING directions for auditory display menu interface COMB FILTER PERFORMANCE design. Among several techniques for represent - ing individual menu items the preliminary results Presenter: Jörg Deigmöller, IRT Munich show that the spearcon seems to be the most suitable method. For the layout of menu items, Downmixing 5.1 surround sound to 2.0 stereo is a neces- studies have shown that spatial separation, dif - sity in current challenging production environments. ferent timbres, and staggering onset between Straightforward downmix coefficients, as simple as they the items improves recognition rates, particularly are to execute, result in comb filtering effects dependent for concurrently presented items. A remaining on the correlation of the signals that are downmixed. A issue to be investigated is how to remind the new algorithm is presented that tackles this issue with users of their current location in the menus of optimized decorrelation pre-processing, leading to much extensive menu interfaces. improved timbre of the downmixed signal. Convention Paper 7659 Tutorial 2 Thursday, May 7 11:00 09:00 – 11:00 Room K2

P2-5 Symmetry Model-Based Key Finding— Markus AN INTRODUCTION TO DIGITAL AUDIO EFFECTS Mehnert ,1 Gabriel Gatzsche ,2 Daniel Arndt 3 1Technische Universität Ilmenau, Ilmenau, Chair: Christoph M. Musialik, Algorithmix GmbH Germany 2Fraunhofer Institute for Digital Media Panelists: Joshua D. Reiss, Queen Mary, University Technology IDMT, Ilmenau, Germany of London, London, UK 3Fraunhofer IIS, Ilmenau, Germany Udo Zölzer, Helmut-Schmidt-Universität, Hamburg, Germany In this paper we introduce a new key finding algorithm that is based on the symmetry model In this tutorial we discuss the ways by which signal process- introduced by Gatzsche et al. The algorithm con - ing techniques are used to produce effects acting on digital sists of two parts. First, the most probable dia - audio signals. The audio effects are systematically classified tonic pitch class set of the musical piece is rec - and discussed, with emphasis on how and why they are ognized. Second, using one of the subspaces of used. Practical examples of common effects are provided, the symmetry model the mode of the piece is along with block diagrams, pseudo-code, and sound exam- estimated. The algorithm is evaluated with 100 ples. During the tutorial, a few effects will be created from Beatles songs, 90 newer “Pop and Rock” songs, scratch and the audience will be provided with the basic and 252 classical pieces from the Naxos data - background knowledge to design their own effects. base. The results will be compared to the algo - rithms of Lerch, Zhu et al., and an algorithm Thursday, May 7 09:00 Room D120 based on binary major and minor chord profiles. Technical Committee Meeting on Acoustics The new algorithm has the highest overall key and Sound Reinforcement finding MIREX’05 score of 82.9 percent. Convention Paper 7660 Session P3 Thursday, May 7 10:00 – 11:30 K4 Foyer

Workshop 1 Thursday, May 7 POSTERS: RECORDING, REPRODUCTION, 09:00 – 11:00 Room D123 AND DELIVERY

LIES, DAMN LIES, AND STATISTICS 10:00 Chair: Sylvain Choisel , Philips Consumer Lifestyle, P3-1 Audio Content Annotation, Description, and Leuven, Belgium Management Using Joint Audio Detection, ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 3 Technical Progra m Segmentation, and Classification Techniques sound field. The first objective is to determine —Christos Vegiris, Charalambos Dimoulas, which form of panning is consistently preferred George Papanikolaou , Aristotle University of for panning sources around the loudspeaker Thessaloniki, Thessaloniki, Greece array. The second objective and main focus of the paper is localizing sources within the loud - The current paper focuses on audio content man - speaker array. We seek to determine if the agement by means of joint audio segmentation and sound sources can be located without move - classification. We concentrate on the separation of ment or a secondary reference source. The typical audio classes, such as silence/background authors compare various techniques based on noise, speech, crowded speech, music, and their ambisonics, vector base amplitude panning and combinations. A compact feature-vector subset is time delay based panning. We report on sub - selected by a Correlation feature selection subset jective listening tests that show which method evaluation algorithm after the use of EM clustering of panning is preferred by listeners and rate the algorithm on an initial audio data set. Time and success of panning within a 3-D loudspeaker spectral parameters are extracted using filter-banks array. and wavelets in combination with sliding windows Convention Paper 7663 and exponential moving averaging techniques. Features are extracted on a point-to-point basis, using the finest possible time resolution, so that 10:00 each sample can be individually classified to one of P3-4 The Effect of Listening Room on Audio the available groups. Clustering algorithms like EM Quality in Ambisonics Reproduction— Olli or Simple K-means are tested to evaluate the final Santala ,1 Heikki Vertanen ,1,2 Jussi Pekonen ,1 point-to-point classification result, therefore the joint Jan Oksanen ,1 Ville Pulkki 1 audio detection-classification indexes. The extract - 1Helsinki University of Technology, Espoo, ed audio detection, segmentation, and classifica - Finland tion results can be incorporated into appropriate 2University of Helsinki, Helsinki, Finland description schemes that would annotate audio events/segments for content description and man - In multichannel reproduction of spatial audio with agement purposes. first-order Ambisonics the loudspeaker signals Convention Paper 7661 are relatively coherent, which produces promi - nent coloration. The coloration artifacts have 10:00 been suggested to depend on the acoustics of the listening room. This dependency was P3-2 Ambience Sound Recording Utilizing Dual researched with subjective listening tests in an MS (Mid-Side) Microphone Systems Based anechoic chamber with an octagonal loudspeak - upon Frequency Dependent Spatial Cross er setup. Different virtual listening rooms were Correlation (FSCC) [Part 3: Consideration of created by adding diffuse reverberation with 0.25 Microphones’ Locations]— Teruo Muraoka, seconds RT60 using a 3-D 16-channel loud - Takahiro Miura, Tohru Ifukube , University of speaker setup. In the test, the subjects com - Tokyo, Tokyo, Japan pared the audio quality in the virtual rooms. The In order to achieve ambient and exactly sound- results suggest that optimal audio quality was localized musical recording with fewer number of obtained when the virtual room effect and the microphones, we studied sound acquisition per - direct sound were on equal level at the listening formances of microphone arrangements utilizing position. their Frequency Dependent Spatial Cross Corre - Convention Paper 7664 lation (FSCC). The result is that an MS micro - phone is best for this purpose. The setting of the 10:00 microphone's directional azimuth at 132 degrees P3-5 Ontology-Based Information Management in is the best for ambient sound acquisition and Music Production— Gyorgy Fazekas, Mark setting of that at 120 degrees is best for Sandler , Queen Mary, University of London, on-stage sound acquisition. We conducted actu - London, UK al concert recordings with a combination of those MS microphones (Dual MS microphone In information management, ontologies are used systems) and obtained satisfactory results. Suc - for defining concepts and relationships of a cessively, we studied the proper setting posi - domain in question. The use of a schema per - tions of those microphones. For ambient sound mits structuring, interoperability, and automatic acquisition, suspending the microphone at the interpretation of data, thus allows accessing center of a concert hall is favorable, and for on- information by means of complex queries. In this stage sound acquisition, locating it at almost paper we use ontologies to associate metadata, above the conductor’s position will also be satis - captured during music production, with explicit factory. Process of the studies will be reported. semantics. The collected data is used for finding Convention Paper 7662 audio clips processed in a particular way, for instance, using engineering procedures or acoustic 10:00 signal features. As opposed to existing metadata P3-3 A Comparative Approach to Sound standards, our system builds on the Resource De - Localization within a 3-D Sound Field— scription Framework, the data model of the Seman - Martin J. Morrell, Joshua D. Reiss , Queen Mary, tic Web, which provides flexible and open-ended University of London, London, UK knowledge representation. Using this model, we demonstrate a framework for managing informa - In this paper we compare different methods tion, relevant in music production. for sound localization around and within a 3-D Convention Paper 7665

4 Audio Engineering Society 126th Convention Program, 2009 Spring Thursday, May 7 10:00 Room D120 at BMW for more than 20 years. His speech will highlight Technical Committee Meeting on Archiving, many aspects of perception and acoustics from an un - Restoration, and Digital Libraries usual point of view: What does a driver in a car need to hear, what should not hear, and how can the acoustics Thursday, May 7 11:00 Room D120 and sounds of a car help to significantly enhance driving Technical Committee Meeting on Coding of Audio pleasure and safety? The title of his address is “Human- Signals Oriented Acoustics—Targets for Future Developments in Vehicle Acoustics.” Student Event/Career Development OPENING AND STUDENT DELEGATE ASSEMBLY Thursday, May 7 13:30 Room D104 MEETING – PART 1 Standards Committee Meeting on SC-02-02 Digital Thursday, May 7, 11:15 – 12:00 Room K2 Input/Output Interfacing Chair: Misato Yamada Workshop 2 Thursday, May, 7 Vice Chair: Miroslav Jakovljevic 13:30 – 15:30 Room K2

The first Student Delegate Assembly (SDA) meeting is New Technologies for Audio Over IP the official opening of the Convention’s student program and a great opportunity to meet with fellow students from all corners of the world. This opening meeting of the Stu - Chair: Jeremy Cooperstock , McGill University, dent Delegate Assembly will introduce new events and Montreal, Quebec, Canada election proceedings, announce candidates for the com - Panelists: Steve Church , Telos ing year’s election for the Europe/International Regions, Christian Diehl , Mayah announce the finalists in the recording competition cate - Manfred Lutzky , Fraunhofer IIS gories, hand out the judges’ sheets to the nonfinalists, Greg Massey , APT and announce any upcoming events of the Convention. Students and student sections will be given the opportu - This workshop is intended to provide an “under the hood” nity to introduce themselves and their activities, in order discussion of various low-latency codecs as well as a to stimulate international contacts. comparison of their pros and cons for different applica - All students and educators are invited to participate in tions. Codecs including AAC-ELD and ULD will be dis - this meeting. Election results and Recording Competition cussed, along with techniques such as adaptive jitter and Poster Awards will be given at the Student Delegate buffer management. Assembly Meeting–Part 2 on Saturday, May 9, at 17:30. Session P4 Thursday, May 7 Special Event 14:00 – 18:30 Room K3 AWARDS PRESENTATION AND KEYNOTE ADDRESS Thursday, May 7, 12:00 – 13:3 RECORDING, REPRODUCTION, AND DELIVERY Room K3 Opening Remarks :• Executive Director Roger Furness Chairs: Jörg Wuttke , Jörg Wuttke Consultancy, Pfinztal, • President Jim Anderson Germany, Schoeps GmbH, Karlsruhe, Germany • Convention Chair Martin Wöhr Siegfried Linkwitz , Linkwitz Lab, Corte Madera, CA, USA Program: • AES Awards Presentation • Introduction of Keynote Speaker 14:00 by Convention Chair Martin Wöhr • Keynote Address by Gerhard Thoma P4-1 An Expert in Absentia: A Case-Study for Awards Presentation Using Technology to Support Recording Pl ease join us as the AES presents special awards to those Studio Practice — Andrew King , University who have made outstanding contributions to the Society in of Hull, Scarborough, North Yorkshire, UK such areas of research, scholarship, and publications, as This paper examines the use of a Learning well as other accomplishments that have contributed to the Technology Interface (LTI) to support the enhancement of our industry. The awardees are: completion of a recording workbook with audio Bronze Medal Award: examples over a ten-week period. The VLE provid - • Ivan Stamac ed contingent support to studio users for tech nical problems encountered in the completion of four Fellowship Award: recording tasks. Previous research has • Martin Wöhr investigated how students collaborate and prob - Board of Governors Award lem-solve during a short session in the recording • Jan Berg studio using technology as a contingent support • Kimio Hamasaki tool. In addition, online message boards have • Shinji Koyano been used to record problems encountered when • Tapio Lokki completing a prescribed task (critical-incident • Jiri Ocenasek recording). A mixed-methods case study approach • John Oh was used in this study. The students interactions • Jan Abildgaard Pedersen within the LTI were logged (i.e., frequency, time, • Joshua Reiss duration, type of support) and their feedback was elicited via a user questionnaire at the end of the Keynote Speaker project. Data for this study demonstrates that This year’s Keynote Speaker is Gerhard Thoma . Thoma learning technology can be a successful support has been leading the Department of Acoustics Projects tool and also highlights the frequency and themes ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 5 Technical Progra m concerning the types of recording practice infor - 3Fraunhofer Institute for Building Physics IBP, mation accessed by the learners. Stuttgart, Germany Convention Paper 7669 The new audio laboratory rooms of the Fraunhofer IIS and their technical design are presented here. 14:30 The vision behind them is driven by the very high P4-2 Recording and Reproduction over Two demands of a leading edge audio research organi - Loudspeakers as Heard Live—Part 1: zation with more than 100 scientists and engineers. Hearing, Loudspeakers, and Rooms— The 300 m 2 sound studio complex was designed Siegfried Linkwitz ,1 Don Barringer 2 with the intention of providing capabilities that are 1Linkwitz Lab, Corte Madera, CA, USA in combination far more extensive than those 2Linkwitz Lab, Arlington, CA, USA available in common audio research or production facilities. The reproduction room for listening tests Innate hearing processes define the realism that follows the strict recommendations of ITU-R BS can be obtained from reproduced sound. An 1116. The results of the qualification measure - unspecified system with two loudspeakers in a ments regarding direct sound, reflected sound, and room places considerable limitations upon the steady state sound field will be shown and the con - degree of auditory realism that can be obtained. struction efforts needed to achieve these values It has been observed that loudspeakers and are explained. The connection from all the room must be hidden from the auditory scene computers in the server room to more than 70 that is evoked in the listener’s brain. Require - loudspeakers in the reproduction rooms, other au - ments upon the polar response and the output dio interfaces, and the projection screens is done volume capability of the loudspeaker will be dis - by an audio and video routing system. The archi - cussed. Problems and solutions in designing a tecture of the advanced control software of this three-way, open baffle loudspeaker with piston routing system is presented. It allows easy and drivers will be presented. Loudspeakers and lis - flexible access for each class of user to all the tener must be symmetrically placed in the room possibilities made available by this completely new to minimize the effects of reflections upon the system. auditory illusion. Convention Paper 7672 Convention Paper 7670 16:00 15:00 P4-5 Advances in National Broadcaster Networks: P4-3 Recording and Reproduction over Two Exploring Transparent High Definition IPTV— Loudspeakers as Heard Live—Part 2: Matthew O’Donnell , British Sky Broadcasting, Recording Concepts and Practices— Upminster, UK Don Barringer ,1 Siegfried Linkwitz 2 1Linkwitz Lab, Arlington, CA, USA British commercial broadcasters are increasing 2Linkwitz Lab, Corte Madera, CA, USA their ability to determine the quality of distribu - tion of audio-over-IP by acquiring and installing For a half century, the crucial interaction next generation national Gigabit networks. This between recording engineer and monitor loud - paper explores how broadcasters can use the speakers during two-channel stereophonic advances in broadband technology to transpar - recording has not been resolved, leaving the ently integrate supplemental on-demand IPTV engineer to cope with uncertainties. However, re - services with traditional broadcasting transport, cent advances in defining and improving this loud - which has led to broadcasters being confident in speaker-room-listener interface have finally al - achieving scalable carrier-class quality of service lowed objectivity to inform and shape the for delivery of high definition media direct to the engineer’s choices. The full potential of the customer’s set top box. two-channel format is now accessible to the Convention Paper 7673 recording engineer, and in a room that is just as normal as most consumers’ rooms. The 16:30 improved reproduction has also allowed a deeper understanding of the merits and limits of spaced P4-6 Multi-Perspective Surround Sound Audio and coincident/near-coincident microphone arrays. Recording— Mark J. Sarisky , The University As a result of these and earlier observations, a of Texas at Austin, Austin, TX, USA four-microphone array was conceived that exploits With the advent of Blu-Ray Disc Audio (BD- natural hearing processes to achieve greater audi - Audio), high resolution uncompressed audio tory realism from two loudspeakers. A number of recordings can be presented as a consumer insights have emerged from the experiments. product in a variety of surround sound formats. Convention Paper 7671 This paper proposes a new take on the recording of live and studio music in surround sound that 15:30 allows the consumer to benefit from the large capacity of the BD-Audio disc and enjoy the P4-4 Vision and Technique behind the New recording from multiple listening perspectives. Studios and Listening Rooms of the Convention Paper 7674 Fraunhofer IIS Audio Laboratory— Andreas 1 1 1 Silzle, Stefan Geyersberger, Gerd Brohasga , 17:00 Dieter Weninger ,1,2 Michael Leistner 3 1Fraunhofer Institute for Integrated Circuits IIS, P4-7 Sound Intensity-Based Three-Dimensional Erlangen, Germany Panning— Akio Ando, Kimio Hamasaki , NHK 2Innovationszentrum für Telekommunikation Science and Technical Research Laboratories, stechnik GmbH IZT, Erlangen, Germany Setagaya, Tokyo, Japan

6 Audio Engineering Society 126th Convention Program, 2009 Spring Three-dimensional (3-D) panning equipment is es - 1Philips Research Europe, Eindhoven, sential for the production of 3-D audio content. We The Netherlands have already proposed an algorithm to enable 2Technical University of Eindhoven, Eindhoven, such panning. It generates the input signal to be The Netherlands; fed into multichannel loudspeakers so as to realize the same physical properties of sound at the Formulas are presented for acoustical quantities receiving point as those created by a single loud - of a harmonically excited resilient, flat, circular speaker model of the virtual source. A sound pres - loudspeaker in an infinite baffle. These quanti - sure vector is used as the physical property. This ties are the sound pressure on-axis, far-field, paper proposes a new method that uses sound directivity and the total radiated power. These intensity instead of the sound pressure vector and quantities are obtained by expanding the velocity shows that both conventional “vector base ampli - distribution in terms of orthogonal polynomials. tude panning” and our previous method come very For rigid and non-rigid radiators, this yields close to achieving coincidence of sound intensity. explicit, series expressions for both the on-axis A new panning method using four loudspeakers is and far-field pressure. In the reverse direction, a also proposed. method of estimating velocity distributions from Convention Paper 7675 (measured) on-axis pressures by matching in terms of expansion coefficients is described. Together with the forward far-field computation 17:30 scheme, this yields a method for assessment of P4-8 A Practical Comparison of Three Tetrahedral loudspeakers in the far-field and of the total radi - Ambisonic Microphones— Dan Hemingson, ated power from (relatively near-field) on-axis Mark Sarisky , The University of Texas at Austin, data (generalized Keele scheme). Austin, TX, USA Convention Paper 7678 This paper compares two low-cost tetrahedral 14:30 ambisonic microphones, an experimental micro - phone, and a Core Sound TetraMic with a P5-2 Testing and Simulation of a Thermoacoustic Soundfield MKV or SPS422B serving as a stan - Transducer Prototype— Fotios Kontomichos, dard for comparison. Recordings were made in Alexandros Koutsioubas, John Mourjopoulos, natural environments of live performances, in a Nikolaos Spiliopoulos, Alexandros Vradis, recording studio, and in an anechoic chamber. Stamatis Vassilantonopoulos , University of The results of analytical and direct listening tests Patras, Patras, Greece of these recordings are discussed in this paper. A description of the experimental microphone Thermoacoustic transduction is the transforma - and the recording setup is included. tion of thermal energy fluctuations into sound. Convention Paper 7676 Devices fabricated by appropriate materials uti - lize such a mechanism in order to achieve acoustic wave generation by direct application of 18:00 an electrical audio signal and without the use of P4-9 A New Reference Listening Room for any moving components. A thermoacoustic Consumer, Professional, and Automotive transducer causes local vibration of air mole - Audio Research— Sean Olive , Harman cules resulting in a proportional pressure International, Northridge, CA, USA change. The present paper studies an imple - mentation of this alternative audio transduction This paper describes the features, scientific technique for a prototype developed on silicon rationale, and acoustical performance of a new wafer. Measurements of the performance of this reference listening room designed for the pur - hybrid solid state device are presented and com - pose of conducting controlled listening tests and pared to the theoretical principles of its opera - psychoacoustic research for consumer, profes - tion, which are evaluated via simulations. sional, and automotive audio products. The main Convention Paper 7679 features of the room include quiet and adjustable room acoustics, a high-quality calibrated play - 15:00 back system, an in-wall loudspeaker mover, and complete automated control of listening tests P5-3 Analysis of Viscoelasticity and Residual performed in the room. Strains in an Electrodynamic Loudspeaker— Convention Paper 7677 Ivan Djurek, 1 Antonio Petosic ,1 Danijel Djurek 2 1University of Zagreb, Zagreb, Croatia 2 Session P5 Thursday, May 7 Alessandro Volta Applied Ceramics (AVAC), 14:00 – 18:30 Room K4 Zagreb, Croatia An electrodynamic loudspeaker was analyzed in LOUDSPEAKERS three steps: (a) as a device supplied by the mar - ket, (b) removed upper suspension, and (c) dis - Chair: John Vanderkooy , University of Waterloo, mantled assembly consisting only of vibrating Waterloo, Ontario, Canada spider and voice coil. In three steps, resonant fre - 14:00 quency and stiffness were measured dynamically for driving currents up to 100 mA, whereas stiff - P5-1 Estimating the Velocity Profile and ness was also measured quasi-statically by the Acoustical Quantities of a Harmonically use of calibrated masses. It was found that wide - Vibrating Loudspeaker Membrane from ly quoted effect of decreasing resonant frequen - On-Axis Pressure Data— Ronald M. Aarts ,1,2 cy, as plotted against driving current, comes from Augustus J. Janssen 1 the residual strain in the vibrating material, and ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 7 Technical Progra m significant contribution is associated with the spi - outcome of this is that the electrostatic forces act der. When driving current increases residual predominantly within the inner and outer turn of strain is gradually compensated, giving rise to the the capacitor body, while all of the forces acting minimum of stiffness, and further increase of res - within the body of the capacitor are balanced onant frequency is attributed to a common non - almost to zero. Experimental results where reso - linearity in the forced vibrating system. nant acoustic emissions have been measured and Convention Paper 7680 analyzed are presented and discussed in the con - text of the model proposed. 15:30 Convention Paper 7682

P5-4 Integrated Conventional and Directional 16:30 Loudspeaker— Changuang Cai ,1 Koray Ozcan ,2 Jani Nurminen ,3 Xuesheng Li, 4 Limei Xu ,4 P5-6 On the Use of Motion Feedback as Used in Salo Antti ,5 Xia Wang ,1 Leon Xu 1 4th Order Systems— Stefan Willems ,1 Guido 1Nokia Research Center (Beijing), Beijing, China D’Hoogh 2 2Nokia (UK) Ltd., Hampshire, UK 1Denon & Marantz Holding, Premium Sound 3Nokia Devices R&D, Tampere, Finland Solutions, Leuven Belgium 4University of Electronic Science and 2Retired Technology of China, Chengdu Sichuan, China 5Nokia Research Center, Helsinki, Finland Class D amplification allows the design of com - pact very high power amplifiers with a high This paper presents a novel loudspeaker efficiency. Those amplifiers are an excellent can - concept that realizes both conventional and didate for being used in compact high-powered directional loudspeaker functions using a single subwoofers. The drawback of compact sub - transducer component. The transducer compo - woofers is the nonlinear compression of the air nent works in such a way that its low frequency inside the (acoustically) small box. Fourth order characteristics satisfy the requirements of the systems are beneficial over 2nd order systems conventional loudspeaker, while one of its natural due to their increased efficiency. To combine the frequencies is located in the ultrasonic range. best of both worlds, 4th order design and When the input exciting signal is a conventional acoustically small enclosures, a feedback mech - audio signal, the transducer can produce conven - anism has been developed to reduce the nonlin - tional audible waves. When the input signal is an ear distortion found in compact high-powered ultrasonic signal modulated by an audible signal, subwoofers. Acceleration feedback on woofer it can produce the ultrasound by working near the systems is traditionally used in 2nd order sys - specific natural frequency of the transducer com - tems. This paper discusses the use of an accel - ponent. Then, the nonlinear interaction of ultra - eration and velocity feedback system applied to sound in air can produce directional audible a 4th order system. sound. In this study a demonstrator is built for Convention Paper 7683 testing the feasibility of the concept. The experi - ment results indicate that this concept is feasible 17:00 even though the performance of this demonstra - tor should be further improved and optimized. P5-7 Mapping of the Loudspeaker Emission by the Convention Paper 7681 Use of Anemometric Method— Danijel Djurek ,1 Paper not presented but available for purchase. Ivan Djurek, 2 Antonio Petosic 2 1Alessandro Volta Applied Ceramics (AVAC), 16:00 Zagreb, Croatia 2University of Zagreb, Zagreb, Croatia P5-5 Forces in Cylindrical Metalized Film Audio Capacitors— Philip J. Duncan ,1 Nigel Williams, 2 Lateral wire anemometry (LWA) has been devel - Paul S. Dodds 2 oped for recording of air vibration. Standard 1University of Salford, Greater Manchester, UK anemometry is founded upon the hot wire 2ICW Ltd., Wrexham, Wales, UK method, and wire temperature changes in the oscillating air velocity in the range 800 –1000 °C, This paper is concerned with the analysis of forces which is less suitable because of the proper heat acting in metalized polypropylene film capacitors in emission from the wire. LWA deals only with the use in loudspeaker crossover circuits. Capacitors initial slope of the changing wire resistance, and have been subjected to rapid discharge measure - subsequent Fourier analysis enables measure - ments to investigate mechanical resonance of the ments of periodic air velocity. The probe has capacitor body and the electrical forces that drive been developed for precise mapping of the air the resonance. The force due to adjacent flat cur - velocity field in the front of the membrane, and rent sheets has been calculated in order that the local power emission of the membrane may be magnitude of the electro-dynamic force due to the evaluated in the region fitted to 0.15 cm 2. discharge current can be calculated and compared Convention Paper 7684 with the electrostatic force due to the potential dif - ference between the capacitor plates. The electro - 17:30 static force is found to be dominant by several orders of magnitude, contrary to assumptions in P5-8 Flat Panel Loudspeaker Consisting of an previous work where the electro-dynamic force is Array of Miniature Transducers— Daniel assumed to be dominant. The capacitor is then Beer ,1 Stephan Mauer ,1 Sandra Brix ,1 modeled as a series of concentric cylindrical con - Jürgen Peissig 2 ductors and the distribution of forces within the 1Fraunhofer Institute for Digital Media body of the capacitor is considered. The primary Technology IDMT, Ilmenau, Germany

8 Audio Engineering Society 126th Convention Program, 2009 Spring 2Sennheiser Electronic GmbH & Co. KG, 14:00 Wedemark, Germany P6-2 A Study of MPEG Surround Configurations Multichannel audio reproduction systems like the and Its Performance Evaluation— Evelyn Wave Field Synthesis (WFS) use a large number Kurniawati, Samsudin Ng, Sapna George , of small and closely spaced loudspeakers. The ST Microelectronics Asia Pacific Pte. Ltd., successful use of WFS requires, among other Singapore things, the ability of an "invisible” integration of The standardization of MPEG Surround in 2007 loudspeakers in a room. Flat panel loudspeakers opens a new range of possibility for low bit rate compared with conventional cone loudspeakers multichannel audio encoding. While ensuring provide advantages in the space saved room inte - backward compatibility with legacy decoder, gration because of their low manufactured depth. MPEG Surround offers various configurations to In this way flat panel loudspeakers can be found in upmix to the desired number of channels. The furniture, media devices, or like pictures hung on downmix stream, which can be in mono or the wall. Besides the integration, flat loudspeakers stereo format, can be passed to transform, should provide at least the same good acoustical hybrid, or any other types of encoder. These performance as conventional loudspeakers. This options give us more than one possible combi - is indeed a problem, because the low depth nega - nation to encode a multichannel stream at a spe - tively influences the acoustical quality of reproduc - cific bit rate. This paper presents a comparative tion in the lower and middle frequency range. This study between those options in terms of their paper demonstrates a new flat panel loudspeaker quality performance that will help us choose the consisting of an array of miniature transducers. most suitable configuration of MPEG Surround Convention Paper 7685 in a range of operating bit rate. Convention Paper 7667 18:00 14:00 P5-9 Subwoofer Loudspeaker System with Dynamic Push-Pull Drive— Drazenko Sukalo , P6-3 Lossless Compression of Spherical DSLab–Device Solution Laboratory, Munich, Microphone Array Recordings— Erik Hellerud, Germany U. Peter Svensson , Norwegian University of Sci - ence and Technology, Trondheim, Norway This paper examines the influence of mutual coupling between two driver-diaphragms driven The amount of spatial redundancy for recordings by two electrical signals, each with a 90° phase from a spherical microphone array is evaluated shift on the voice-coil impedance curve. A new using a low delay lossless compression scheme. model of the system is described, and the The original microphone signals, as well as sig - effects are observed using the electrical circuit nals transformed to the spherical harmonics do - simulator PSpice. Finally, predicted and mea - main, are investigated. It is found that the corre - sured values are presented. lation between channels is, as expected, very Convention Paper 7686 high for the microphone signals, in several differ - ent acoustical environments. For the signals in the spherical harmonics domain, the compres - Session P6 Thursday, May 7 sion gain from using inter-channel prediction is 14:00 – 15:30 K4 Foyer reduced, since this conversion results in many channels with low energy. Several alternatives POSTERS: MULTICHANNEL CODING for reducing the coding complexity are also investigated. 14:00 Convention Paper 7668

P6-1 Adaptive Predictive Modeling of Stereo LPC Workshop 3 Thursday, May, 7 with Application to Lossless Audio 14:00 – 16:00 Room D123 Compression— Florin Ghido, Ioan Tabus , Tampere University of Technology, Tampere, INTELLIGENT DIGITAL AUDIO EFFECTS Finland Chair: Christian Uhle , Fraunhofer Institute for We propose a novel method for exploiting the Integrated Circuits IIS, Erlangen, Germany redundancy of stereo linear prediction coeffi - cients by using adaptive linear prediction for the Panelists: Alexander Lerch , Z-Plane, Berlin, Germany coefficients themselves. We show that an impor - Josh Reiss , Queen Mary, University of tant proportion of the stereo linear prediction London, London, UK coefficients, on both the intrachannel and the Udo Zölzer , Helmut-Schmidt-Universität, interchannel parts, still contains important redun - Hamburg, Germany dancy inherited from the signal. We can there - fore significantly reduce the amplitude range of Intelligent Digital Audio Effects (I-DAFx) process audio those LP coefficients by using adaptive linear signals in a signal-adaptive way by using some kind of prediction with orders up to 4, separately on the high-level analysis of the input. Beat tracking, for exam - intrachannel and intrachannel parts. When inte - ple, enables the automated adaption of time delays or of grated into asymmetrical OptimFROG, the new the LFO rate in tremolos, auto-wahs, and vibrato effects. method obtains on average 0.29 percent A harmonizer can adapt the additional intervals to the improvement in compression with negligible melody that is played. Automatic mixing is approached increase in decoder complexity. by analyzing the signal content in all channels to control Convention Paper 7666 the panning. These are examples of techniques that are ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 9 Technical Progra m in the scope of this workshop. It presents an overview of DSP processor. The complexity and quality are I-DAFx and of methods of semantic audio analysis used compared with a conventional head-related in these devices. Practical examples are described and transfer function (HRTF) filtering method and sound examples are demonstrated. Dolby headphone that are the most current in commercial binaural rending technology, demon - Tutorial 3 Thursday, May 7 strating significant complexity reduction and com - 14:00 – 16:30 Room K1 parable sound quality to the Dolby headphone. Convention Paper 7687 DESIGN OF HIGH-PERFORMANCE BALANCED AUDIO INTERFACES 16:30 P7-2 Optimal Filtering for Focused Sound Field Presenter: Bill Whitlock , Jensen Transformers, Inc., Reproductions Using a Loudspeaker Array— Chatsworth, CA, USA Youngtae Kim, Sangchul Ko, Jung-Woo Choi, High signal-to-noise ratio is an important goal for most Jungho Kim , SAIT, Samsung Electronics Co., audio systems. However, AC power connections Ltd., Gyeonggi-do, Korea unavoidably create ground voltage differences, magnetic This paper describes audio signal processing tech - fields, and electric fields. Balanced interfaces, in theory, niques in designing multichannel filters for repro - are totally immune to such interference. For 50 years, ducing an arbitrary spatial directivity pattern with a virtually all audio equipment used transformers at its bal - typical loudspeaker array. In designing the multi - anced inputs and outputs. Their high noise rejection was channel filters, some design criteria based on, for taken for granted and the reason for it all but forgotten. example, least-squares methods and the maxi - The transformer's extremely high common-mode imped - mum energy array are introduced as non- ance—about a thousand times that of its solid-state iterative optimization techniques with a lower com - equivalents —is the reason. Traditional input stages will putational complexity. The abilities of the criteria be discussed and compared. A novel IC that compares are first evaluated with a given loudspeaker config - favorably to the best transformers will be described. uration for reproducing a desired acoustic property Widespread misunderstanding of the meaning of balance in a spatial area of interest. Also, additional con - as well as the underlying theory has resulted in all-too- straints are considered to impose for minimizing common design mistakes in modern equipment and seri - the error between the amplitudes of actual and the ously flawed testing methods. Therefore, noise rejection desired spatial directivity pattern. Their limitations in in today's real-world systems is often inadequate or mar - practical applications are revealed by experimental ginal. Other topics will include tradeoffs in output stage demonstrations, and finally some guidelines are design, effects of non-ideal cables, and the pin 1 problem. proposed in designing optimal filters. Convention Paper 7688 Thursday, May 7 14:00 Room D120 Technical Committee Meeting on Fiber Optics 16:30 (Formative Meeting) P7-3 Single-Channel Sound Source Distance Estimation Based on Statistical and Source- Thursday, May 7 15:00 Room D120 Specific Features — Eleftheria Georganti ,1,2 Technical Committee Meeting on Audio for Games Tobias May ,3 Steven van de Par, 1 Aki Härmä ,1 John Mourjopoulos 2 Thursday, May 7 16:00 Room D120 1Philips Research Europe, Eindhoven, The Technical Committee Meeting on Hearing Netherlands and Hearing Loss Prevention 2University of Patras, Patras, Greece 3Technische Universiteit Eindhoven, Eindhoven, Session P7 Thursday, May 7 The Netherlands; 16:30 – 18:00 K4 Foyer In this paper we study the problem of estimating the distance of a sound source from a single POSTERS: SPATIAL AUDIO PROCESSING microphone recording in a room environment. The room effect cannot be separated from the 16:30 problem without making assumptions about the properties of the source signal. Therefore, it is P7-1 Low Complexity Binaural Rendering for necessary to develop methods of distance esti - Multichannel Sound— Kangeun Lee, mation separately for different types of source Changyong Son, Dohyung Kim , Samsung signals. In this paper we focus on speech sig - Advanced Institute of Technology, Suwon, Korea nals. The proposed solution is to compute a num - The current paper is concerned with an effective ber of statistical and source-specific features from method to emulate the multichannel sound in a the speech signal and to use pattern recognition portable environment where low power is techniques to develop a robust distance estimator required. The goal of this paper is to show the for speech signals. Experiments with a database complexity of binaural rendering of the multichan - of real speech recordings showed that the nel to stereo sound systems in cases of portable proposed model is capable of estimating source devices. To achieve this, we proposed the modi - distance with acceptable performance for applica - fied discrete cosine transform (MDCT) based bin - tions such as ambient telephony. aural rendering, combined with the Dolby Digital Convention Paper 7689 decoder (AC-3) that is a multichannel audio 16:30 decoder. A reverberation algorithm is added to the proposed algorithm for closing to real sound. P7-4 Implementation of DSP-Based Adaptive This combined structure is implemented on a Inverse Filtering System for ECTF 10 Audio Engineering Society 126th Convention Program, 2009 Spring Equalization— Masataka Yoshida, Haruhide source excitation conditions, offering an estimate Hokari, Shoji Shimada , Nagaoka University of on the achieved accuracy. Sound source forward Technology, Nagaoka, Niigata, Japan propagation models can be applied in case that confident localization accuracy has achieved, to The Head Related Transfer Function (HRTF) visualize the corresponding sound field. Other - and the inverse Ear Canal Transfer Function wise, 3-D audio/surround sound reproduction (ECTF) must be accurately determined if stereo simulation can be used instead. In any case, earphones are realized out-of-head sound local - sound level distribution colormap-videos and ization (OHL) with high presence. However, the highlighting images can be extracted. MPEG-7 characteristics of ECTF depend on the type of adapted description schemes are proposed for earphone used and the number of earphone spatial-audio audiovisual content description and mounting and demounting operations. There - management, facilitating a variety of user-interac - fore, we present a DSP-based adaptive inverse tive postprocessing applications. filtering system for ECTF equalization in this Convention Paper 7692 paper. The buffer composition and size of DSP were studied so as to implement operation processing. As a result, we succeeded in con - Workshop 4 Thursday, May 7 structing a system that was able to work in the 16:00 – 18:30 Room K2 audio-band of 15 kHz with the sampling frequen - cy of 44.1 kHz. Listening tests clarified that the MICROPHONES—WHAT TO LISTEN FOR—WHAT effective estimation error of the adaptive inverse- SPECS TO LOOK FOR ECTF for OHL was less than –11 dB with con - vergence time of about 0.3 seconds. Chair: Eddy B. Brixen , EBB-consult, Smørum, Convention Paper 7690 Denmark Panelists: Jean-Marie Greijsen , Polyhymnia International 16:30 David Josephson , Josephson Engineering P7-5 Improved Localization of Sound Sources Douglas McKinnie , Middle Tennessee State Using Multi-Band Processing of Ambisonic University Components — Charalampos Dimoulas, Mikkel Nymand , DPA Microphones George Kalliris, Konstantinos Avdelidis, Ossian Ryner , DR, Danish Broadcasting George Papanikolaou , Aristotle University of Thessaloniki, Thessaloniki, Greece When selecting microphones for a specific music recording, it is worth knowing what to expect and what The current paper focuses on the use of multi- to listen for. Accordingly it is good to know what specifi - band ambisonic-processing for improved sound cations that would be optimum for that microphone. source localization. Energy-based localization This workshop explains the process of selecting a mi - can be easily delivered using soundfield micro - crophone both from the aesthetical as well as the tech - phone pairs, as long as free field conditions and nical point of view. Also explained and demonstrated: the single omni-directional-point-source model what to expect when placing the microphone. This is apply. Multi-band SNR-based selective process - not a “I feel like . . .” presentation. All presenters on the ing improves the noise tolerance and the local - panel are serious and experienced engineers and ton - ization accuracy, eliminating the influence of meisters. The purpose of this workshop is to encourage reverberation and background noise. Band-relat - and teach young engineers and students to take advan - ed sound-localization statistics are further tage by taking a closer look at the specifications the exploited to verify the single or multiple sound- next time they are going to pick a microphone for a job. sources scenario, while continuous spectral fingerprinting indicates the potential arrival of a Workshop 5 Thursday, May 7 new source. Different sound-excitation scenarios 16:30 – 18:30 Room D123 are examined (single /multiple sources, narrow - band / wideband signals, time-overlapping, noise, PROFESSIONAL AUDIO NETWORKING IN SOUND reverberation). Various time-frequency analysis REINFORCEMENT AND BROADCAST APPLICATIONS schemes are considered, including filter-banks, [CANCELLED] windowed-FFT and wavelets with different time resolutions. Evaluation results are presented. Convention Paper 7691 Tutorial 4 Thursday, May 7 17:00 – 18:30 Room K1 16:30 THE GROWING IMPORTANCE OF MASTERING P7-6 Spatial Audio Content Management within IN THE HOME STUDIO ERA the MPEG-7 Standard of Ambisonic Localization and Visualization Descriptions— Presenter: Andrés Mayo , Andrés Mayo Mastering Charalampos Dimoulas, George Kalliris, & DVD, Argentina Kostantinos Avdelidis, George Papanikolaou , Aristotle University of Thessaloniki, Thessaloniki, Artists and producers are widely using their home studios Greece for music production, with a better cost/benefit ratio. But they usually lack technical resources and the acoustic The current paper focuses on spatial audio response of their rooms is unknown. Therefore, there is video/imaging and sound field visualization using greater need for a professional mastering service in order ambisonic-processing, combined with MPEG-7 to achieve the so-called "standard commercial quality." description schemes for multi-modal content This tutorial presents a list of common mistakes that can description and management. Sound localization be found in homemade mixes with real-life audio can be easily delivered using multi-band ambison - examples taken directly from recent mastering sessions. ic processing under free-field and single point- What can and what can’t be fixed at the mastering stage. ¯ Audio Engineering Society 126th Convention Program, 2009 Spring 11 Technical Progra m Thursday, May 7 16:00 Room D120 toring around airports, it is now one of the world’s leading Technical Committee Meeting on Semantic Audio companies in acoustic front-ends and transducers forming a Analysis wide range of general purpose and specialized micro - phones, electro-acoustic measurement devices such as ear Special Event couplers, precision calibration tools and multi-dimensional OPEN HOUSE OF THE TECHNICAL COUNCIL sound intensity probes. The title of his lecture is, “The AND THE RICHARD C. HEYSER MEMORIAL LECTURE Reproduction of Sound Starts at the Microphone.” Thursday, May 7, 18:30 – 19:30 The microphones may be developed for many specific Room K3 purposes: for communication, recording or precision mea - surements. Quality may have different meaning for different Lecturer: Gunnar Rasmussen applications. Price may be a dominating factor. Carbon mi - The Heyser Series is an endowment for lectures by emi - crophones were dominating up to the 1950s. Electret micro - nent individuals with outstanding reputations in audio engi - phones have taken the place of carbon microphones with neering and its related fields. The series is featured twice great improvement in quality and performance at low prices. annually at both the United States and European AES The MEMS microphones are on the way. Conventions. Established in May 1999, The Richard C. The challenge in the high quality microphone development Heyser Memorial Lecture honors the memory of Richard is to match or exceed the human ear in perception of sound Heyser, a scientist at the Jet Propulsion Laboratory, who for measurement purposes. Without measurements we can - was awarded nine patents in audio and communication not qualify our progress. We are still trying to match the fre - techniques and was widely known for his ability to clearly quency band, the dynamic range, the phase linearity of the present new and complex technical ideas. Heyser was also human ear and to obtain very good reproducibility in all situa - an AES governor and AES Silver Medal recipient. tions where humans are involved. We need microphones for The Richard C. Heyser distinguished lecturer for the development, for standardized measurements and for legal 126th AES Convention is Gunnar Rasmussen, a pioneer in related measurements. Where are we today? the construction of acoustic instrumentation, particularly of Rasmussen ’s presentation will be followed by a recep - microphones, transducers, vibration and related devices. He tion hosted by the AES Technical Council. was employed at Brüel & Kjær Denmark as an electronics engineer immediately after his graduation in 1950. After Session P8 Friday, May 8 holding various positions in development, testing, and quali - 09:00 – 11:30 Room K3 ty control, he spent one year in the United States working for Brüel & Kjær in sales and service. ROOM ACOUSTICS After his return to Denmark in the mid-1950s he began Chair: Ronald Aarts , Philips Research Europe, the development of a new measurement microphone. This Eindhoven, The Netherlands resulted in a superior mechanical stability, increased tem - perature, and long term stability. The resulting one-inch 09:00 pressure microphone soon became the de facto standard P8-1 Phase Velocity and Group Velocity in microphone for acoustical measurements to replace the Cylindrical and Spherical Waves— Ian M. famous W.E. 640AA standardized microphone. Dash ,1 Fergus R. Fricke 2 The optimized mechanical design of the new generation 1Australian Broadcasting Corporation, Sydney, of measurement microphones opened up the possibility for NSW, Australia reducing the size of the microphones, first to a ½” micro - 2University of Sydney, Sydney, NSW, Australia phone and then to ¼” and 1/8” microphones with essentially the same superior mechanical, temperature and long term Closed-form expressions are derived for phase stability. Notably the ½” microphone is still the most widely velocity and group velocity in cylindrical and spheri - used measurement tool today. Since the beginning of the cal sound waves. These are plotted and compared 1960s, this microphone design has been preferred for all for orders 0, 1, and 2, but the expressions are gen - types of acoustic measurements and has formed the basis eral and may be applied to waves of any order. for the IEC 1094 series of international standards for mea - Dispersion characteristics of these waves are surement microphones. examined and discussed. The implications for ther - Gunnar Rasmussen received the Danish Design Award modynamic applicability of the wave equation and in 1969 for his novel design of the microphones that were for application of Huygens’ principle are discussed. exhibited at the New York Museum of Modern Art. He also Convention Paper 7693 developed the first acoustically optimized sound level meter, where the shape of the body was designed to minimize the 9:30 effect of reflections from the casing to the microphone. This P8-2 Selection of Loudspeaker Positions for type 2203 Sound Level meter was for many years seen as Reverberation Time and Sound Field the archetype of sound level meters and its characteristic Measurements— Elena Prokofieva , Napier shape became the symbol of a sound level meter. University, Edinburgh, UK Other major inventions and designs include the Delta Shear accelerometer, the dual piston pistonphone calibrator According to the various building standards, the for precision calibration, the face-to-face sound intensity source loudspeakers and receiving microphones probe and hydrophones, occluded ears, artificial mouth, etc. during the internal noise level measurements Rasmussen is also the author of numerous papers on can be placed “in any convenient position,” with acoustics and vibration and has served as chairman and just some distance restrictions from the nearest vice-chairman of various international organizations and reflecting surfaces. In rooms of different shapes standard committees. In 1990 he received the CETIM and volumes the location of the source and medal for his contribution to the field of intensity techniques. receiver microphone may significantly affect the He is also a Fellow of the Acoustical Society of America. measured results. If the difference between the In 1994 Rasmussen started his own company, G.R.A.S. reverberation times or noise levels measured for Sound and Vibration. Originally a company specializing in two positions in the same room exceeds 10 per - precision Outdoor Microphones for permanent noise moni - cent, they cannot be averaged. The simulation

12 Audio Engineering Society 126th Convention Program, 2009 Spring program is created to recommend the most suit - quality of output and computational demands, able locations for microphone and loudspeakers are given along with the details of how they are in tested room for reverberation time measure - incorporated into the hybrid system within which ments. The results of series of tests are analyzed they are employed. to confirm the results of the simulation. Convention Paper 7697 Convention Paper 7694 Session P9 Friday, May 8 10:00 09:00 – 12:30 Room K4 P8-3 A Rehearsal Hall with Virtual Acoustics for Symphony Orchestras— Tapio Lokki ,1 Jukka SIGNAL ANALYSIS, MEASUREMENTS, Pätynen ,1 Timo Peltonen ,2 Olli Salmensaari 2 RESTORATION 1Helsinki University of Technology, Espoo, Chair: Jan Abildgaard Pedersen , Lyngdorf Audio, Finland Skive, Denmark 2Akukon Consulting Engineers Ltd., Helsinki, Finland 09:00 A solution for constructing a small rehearsal hall, P9-1 Some Improvements of the Playback Path the acoustics of which resembles the stage of a of Wire Recorders— Nadja Wallaszkovits ,1 large concert hall is presented. The implemented Heinrich Pichler 2 system was evaluated both objectively with mea - 1Phonogrammarchiv Austrian Academy of surements and subjectively by collecting feedback Sciences, Vienna, Austria from musicians. The subjective opinions were very 2Audio Consultant, Vienna, Austria positive and encouraging and the main objective was achieved. The electroacoustically enhanced The archival transfer of wire recordings to the digi - rehearsal space sounded like a much bigger hall, tal domain is a highly specialized process that although the sound pressure level increased less incorporates a wide range of specific challenges. than one decibel. The presented solution is applic - One of the basic problems is the format incompati - able in all spaces, which are not very reverberant bility between different manufacturers and models. by nature and where the height of the room is at The paper discusses the special design philoso - least twice the standard room height. phy, using the tone control network in the record Convention Paper 7695 path as well as in the playback path. This tone control circuit causes additional phase and group 10:30 delay distortions. The influence and characteristics of the tone control (which was not a priori present P8-4 Sound Field Characterization and Absorption with every model) is discussed and analog phase Measurement of Wideband Absorbers— correction networks are described. The correction Soledad Torres-Guijarro ,1 Antonio Pena, 2 2 of phase errors is outlined. As this format has Alfonso Rodríguez-Molares, Norberto Degara- been obsolete for many decades, a high quality Quintela 2 1 archival transfer can only be reached by modifying Laboratorio Oficial de Metroloxía de Galicia dedicated equipment. The authors propose some (LOMG), possible main modifications and improvements of Ourense, Spain 2 the playback path of wire recorders, such as signal Universidad de Vigo, Vigo, Spain pickup directly after the playback head, introducing Wideband absorbers are a fundamental part of a high quality preamplifier, followed by analog non-environment control rooms. They consist of phase correction and correction of the amplitude huge angled hanging panels in conjunction with characteristics. Alternatively signal pickup directly a multilayer wall or ceiling. Their absorption after the playback head, introducing a high quality capacity is very noticeable, mostly in the low fre - preamplifier, followed by digital signal processing quency range. In this paper the mechanisms of to optimize the output signal is discussed. absorption of the wideband absorbers of the rear Convention Paper 7698 wall of the control room at the Universidad de Vigo will be studied. Conclusions will be drawn 09:30 from the analysis of pressure, velocity volume, P9-2 Acoustics of the Crime Scene as Transmitted and intensity measurements performed in the by Mobile Phones— Eddy B. Brixen , EBB- vicinity of the panels, and from the computation consult, Smorum, Denmark of the normal specific acoustic impedance and the normal absorption coefficient. One task for the audio forensics engineer is to Convention Paper 7696 extract background information from audio recordings. A major problem is the assessment 11:00 of analyzed telephone calls in general and P8-5 Temporal Matching of 2-D and 3-D Wave- mobile phones (LPC-algorithms) in particular. In Based Acoustic Modeling for Efficient and this paper the kind of acoustic information to be Realistic Simulation of Rooms— Jeremy J. extracted from a recorded phone call is initially Wells, Damian T. Murphy, Mark Beeson , explained. The parameters used for the charac - University of York, York, UK terization of the various acoustic spaces and events in question are described. It is discussed Methods for adapting the output of a two-dimen - how the acoustical cues should be assessed. sional Kirchoff-variable digital waveguide mesh The validity of acoustic analyses carried out in to better match that of a 3-D mesh, both of which the attempt to provide crime scene information are intended to model the same acoustic space, like reverberation time is presented. are presented. Details of the methods, including Convention Paper 7699 ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 13 Technical Progra m 10:00 cal analysis of signal level has been performed by its statistical distribution, cumulative distribu - P9-3 Silence Sweep: A Novel Method for Measur - tion, effective value within complete duration of ing Electroacoustical Devices— Angelo Farina , piece, mean level value, and level value corre - University of Parma, Parma, Italy sponding to maximum of the statistical distribu - This paper presents a new method for measur - tion. The parameter L1, L10, L50, and L99 were ing some properties of an electroacoustical sys - extracted from cumulative distributions as tem, for example a loudspeaker or a complete numerical indicators of dynamic properties. The sound system. Coupled with the already estab - paper contains detailed statistical data and aver - lished method based on Exponential Sine aged data for all observed genres, as well as Sweep, this new Silence Sweep method pro - quantitative data about dynamic range and crest vides a quick and complete characterization of factor of various music signals. not linear distortions and noise of the device Convention Paper 7702 under test. The method is based on the analysis of the distortion products, such as harmonic dis - 11:30 tortion products or intermodulation effects, occurring when the system is fed with a wide- P9-6 Multi-Band Generalized Harmonic Analysis band signal. Removing from the test signal a (MGHA) and its Fundamental Characteristics small portion of the whole spectrum, it becomes in Audio Signal Processing— Takahiro Miura, possible to collect and analyze the not-linear Teruo Muraoka, Tohru Ifukube , University of response and the noise of the system in that Tokyo, Tokyo, Japan “suppressed” band. Changing continuously the suppressed band over time, we get the Silence One of the main problems in sound restoration Sweep test signal, which allows for quick mea - of valuable historical recordings includes the surement of noise and distortion over the whole noise reduction. We have been proposing and spectrum. The paper explains the method with a continuing to improve the noise reduction number of examples. The results obtained for method utilized by inharmonic analysis such as some typical devices are presented, compared GHA (Generalized Harmonic Analysis). Algo - with those obtained with a standard, state-of-the- rithm of GHA frequency extraction enables us to art measurement system. extract arbitrary frequency components. In this Convention Paper 7700 paper we aimed at more accurate frequency identification from noisy signals to divide ana - lyzed frequency section into multi-bands before 10:30 analysis: this algorithm is named as Multi-Band P9-4 Pitch and Played String Estimation in Classic GHA (MGHA). The simulation of frequency and Acoustic Guitars— Isabel Barbancho, analysis in a noise-free condition indicated that Lorenzo Tardón, Ana M. Barbancho, Simone MGHA is more effective than GHA for the Sammartino , Universidad de Málaga, Málaga, extraction of low frequency components in the Spain condition of both lower window length and amount of frequency components. However, In classic and acoustic guitars that use standard excluding the case of both lower window length tuning, the same pitch can be produced at differ - and amount of frequency components, GHA ent strings. The aim of this paper is to present a identifies frequency components more precisely. method based on the time and frequency- Furthermore the result of frequency analysis in domain characteristics of the recorded sound to condition with steady noise shows that MGHA determine, not only the pitch but also the string can be more effectively applied to the case of of the guitar that has been played to produce short window length, many frequency compo - that pitch. This system will provide information nents, and low S/N. not only of the pitch of the notes played, but also Convention Paper 7703 about how those notes were played. This specif - ic information can be valuable to identify the style of the player and can be used in teaching 12:00 to play the guitar. Convention Paper 7701 P9-7 Automatic Detection of Salient Frequencies— Joerg Bitzer ,1 Jay LeBoeuf 2 1 11:00 University of Applied Science Oldenburg, Oldenburg, Germany P9-5 Statistical Properties of Music Signals— 2Imagine Research, Inc., San Francisco, CA, USA Miomir Mijic, Drasko Masovic, Dragana Sumarac-Pavlovic , Faculty of Electrical In this paper we present several techniques to Engineering, Belgrade, Serbia find the most significant frequencies in recorded audio tracks. These estimated frequencies could This paper is concerned with the results of a be used as a starting point for mixing engineers complex approach to statistical properties of var - in the EQing process. In order to evaluate the ious music signals based on 412 musical pieces results, we compare the detected frequencies classified in 12 different genres. Analyzed sig - with a list of reported salient frequencies from nals contain more than 24 hours of music. For audio engineers. The results show that automat - each piece time variation of the signal level was ic detection is possible. Thus, one of the more found, performed with a 10 ms period of integra - boring tasks of a mixing engineer can be auto - tion in rms calculation and with 90 percent over - mated, which gives the mixing engineer more lap, making a new signal representing the level time to do the artistic part of the mixing process. as a function of time. For each piece the statisti - Convention Paper 7704

14 Audio Engineering Society 126th Convention Program, 2009 Spring Workshop 6 Friday, May 8 Presenter: Friedrich Gierlinger , IRT Munich 09:00 – 11:00 Room K1 Synchronization of audio and video through the whole THE OPERA BEHIND THE OPERA—CHALLENGES broadcast chain became ever more challenging with the AND REAL WORLD SOLUTIONS IN TRANSMISSIONS switch to fully digital production methods, wireless OF BIG OPERA PRODUCTIONS picture acquisition practices, omnipresent different pro - cessing latency, and an unavoidable delay in consumer Presenters: Billy Henningsen , Senior Sound displays that may or may not be able to be compensat - Engineer, Norwegian TV, Oslo, Norway ed. This tutorial systematically examines the various rea - Joseph Schütz , Senior Engineer, Austrian sons why lipsync problems occur, quantifies the possible Radio, Vienna, Austria error at any given stage, and recommends best practices Peter Urban , Senior Engineer, Bavarian to improve the situation. Radio, Munich, Germany Friday, May 8 09:00 Room D120 In this workshop the aesthetics and concepts of an ideal Technical Committee Meeting on Spatial Audio opera sound will be introduced, theoretically. Be swept away by the realities of awkward acoustics, directors and Friday, May 8 09:00 Room D104 producers with challenging stage directions and last Standards Committee Meeting on SC-02-08 Audio- minute ideas. Both Bayreuth and Vienna, the new opera File Transfer and Exchange of Oslo, and all the other big stages have had their trans - missions and you will hear how the challenges have been met. In stereo and surround. Student Event/Career Development MENTORING SESSIONS Friday, May 8, 10:00 – 12:00 Workshop 7 Friday, May 8 K4 Foyer 09:00 – 11:00 Room K2 Students are invited to sign-up for an individual meeting AUDIO CODING FOR BROADCASTING—STATUS with a distinguished mentor from the audio industry. The REPORT opportunity to sign up will be given at the end of the Chair: Gerhard Stoll , IRT Munich opening SDA meeting. Any remaining open spots will be posted in the student area. All students are encouraged Panelists: Bjørn Aarseth , NRK, Norway to participate in this exciting and rewarding opportunity Udo Appel , Bayerischer Rundfunk, Germany for individual discussion with industry mentors . Dave Marston , BBC, UK Roger Miles , EBU Switzerland Friday, May 8 10:00 Room D120 Thomas Saner , SRG SSR idée suisse, Technical Committee Meeting on Signal Processing Switzerland Karl M. Slavik , Broadcast Consultant Dolby Session P10 Friday, May 8 Laboratories, Inc./Artecast, Austria 10:30 – 12:00 K4 Foyer Both Digital Radio and Digital TV are relying on new, effi - cient audio codecs in order to supply the listener with lots POSTERS: AUDIO FOR TELECOMMUNICATIONS of audio channels. Lots of audio channels may have a different meaning for TV and radio. Whereas TV, in par - 10:30 ticular HDTV, is going toward mainly multichannel audio, P10-1 Harmonic Representation and Auditory digital radio, here in particular terrestrial radio, is aiming Model-Based Parametric Matching and its for more and more radio programs with stereo as the Application in Speech/Audio Analysis— main audio format. In this context several questions need Alexey Petrovsky, 1 Elias Azarov ,1 Alexander to be answered, e.g.: Petrovsky 2 • Do we have the appropriate audio coding systems 1Belarusian State University of Informatics and for stereo and multichannel audio? Radioelectronics, Minsk, Belarus • What kind of audio quality do we need together with 2Bialystok Technical University, Bialystok, Poland HDTV? The paper presents new methods for the selection • Is 5.1 multichannel audio still the suitable format for of sinusoids and transients components in hybrid HDTV? sinusoidal modeling of speech/audio. The instan - • What happens with cascading multichannel audio taneous harmonic parameters (magnitude, codecs, which occur in real-life situations? frequency, and phase) are calculated as the result • Is broadcast quality, as defined by EBU in 1990, of the narrow band filtering of speech/audio. The when DAB had been developed, still an issue for radio frequency-modulated filters synthesis with the broadcasting? closed form impulse response has been proposed. • How many radio programs can be provided by new The filter frequency bounds can be determined radio broadcast systems, e.g., DAB+? during the components frequency tracking and • What is the role of 5.1 multichannel audio in radio can be adjusted according to the fundamental fre - broadcasting? quency modulations. It can be implemented Several broadcast experts will talk about their experience speech/audio harmonic/noise decomposition. The with low bit-rate audio in a modern broadcast infrastructure transient components modeling are presented by and will try to provide answers to the questions above. matching pursuit with frame-based psychoacoustic optimized wavelet packet dictionary. The choice of Tutorial 5 Friday, May 8 most relevant coefficients is based on maximizing 09:00 – 10:30 Room D123 the matching between the auditory excitation scalograms of original and modeled signals. LIPSYNC IN THE LATENCY AGE Convention Paper 7705 ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 15 Technical Progra m 10:30 derive omnidirectional and dipole signals from stereo microphones for directional analysis are P10-2 Perceptual Compression Methods for presented and their applicability is discussed. Metadata in Directional Audio Coding Convention Paper 7708 Applied to Audiovisual Teleconference— 1 2 2 Toni Hirvonen , Jukka Ahonen, Ville Pulkki 10:30 1Institute of Computer Science (ICS) of the Foundation for Research and Technology, P10-5 Robust Noise Reduction Based on Hellas, Greece Stochastic Spatial Features— Mitsunori 2TKK, Espoo, Finland Mizumachi , Kyushu Institute of Technology, Fukuoka, Japan In teleconferencing application of Directional Audio Coding, the transmitted data consists of This paper proposes a robust noise reduction monophonic audio signal and directional meta - method relying on stochastic spatial features. data measured in frequency bands depending Almost all of noise reduction methods have both on time. In reproduction, each frequency chan - strong and weak sides in the real world. In this nel of the signal is reproduced to corresponding paper time evolution of direction of arrival (DOA) direction with corresponding diffuseness. This and its stochastic reliability are the clues for paper examines methods for reducing the data selecting a suitable approach of noise reduction rate of the metadata. The compression methods under time-variant noisy environments, where a are based on psychoacoustic studies about the DOA is an important spatial feature in beam - accuracy of directional hearing, and further forming for noise reduction. On the other hand, developed and validated. Informal tests with single channel approaches for noise reduction one-way reproduction, as well as usability test - may be reasonable when DOA estimates are not ing where an actual teleconference was reliable. Then, either spectral subtraction or arranged, were utilized for this purpose. The beamforming is selected out for achieving robust results indicate that the data rate can be as low noise reduction depending on a DOA estimate as approximately 3 kbit/s without a significant and its reliability. The proposed method had an loss in the reproduced spatial quality. advantage in noise reduction compared with a Convention Paper 7706 conventional approach. Convention Paper 7709 10:30 P10-3 Speaker Detection and Separation with Small Tutorial 6 Friday, May 8 Microphone Arrays— Maximo Cobos, Jose J. 11:00 – 13:30 Room D123 Lopez, David Martinez , Universidad Politécnica de Valencia, Valencia, Spain LOUDNESS—LIGHT AT THE END OF THE TUNNEL

Small microphone arrays are desirable for many Chair: Florian Camerer , ORF, EBU Group P/LOUD practical speech processing applications. In this paper we describe a system for detecting sever - Presenters: Eelco Grimm , Dutch Loudness al sound sources in a room and enhancing a Committee predominant target source using a pair of close Mike Kahsnitz , RTW microphones. The system consists of three main Ralph Kessler , Pinguin Engineering steps: time-frequency processing of the input Thomas Lund , tc electronic signals, source localization via model fitting, and Andrew Mason, BBC R&D time-frequency masking for interference reduc - Audio levels in broadcasting have become increasingly tion. Experiments and results using recorded diverse and different over the last decades. Despite clear signals in real scenarios are discussed. guidelines and recommended practices the general use Convention Paper 7707 of peak measurement in audio metering and the devel - opment of more and more sophisticated level processors 10:30 have led to over-compression of audio signals with the P10-4 Directional Audio Coding with Stereo questionable aim of being louder than the competition. Microphone Input— Jukka Ahonen, 1 Ville This attitude has especially impacted the audio quality of Pulkki ,1 Fabian Kuech, 2 Giovanni Del Galdo, 2 advertisements and promos with very little dynamic Markus Kallinger, 2 Richard Schultz-Amling 2 range. Already considered a hopeless situation, the intro - 1TKK,Espoo, Finland duction of loudness level metadata and especially the 2Fraunhofer Institute for Integrated Circuits IIS, introduction of an international standard of loudness Erlangen, Germany measurement (ITU-R BS.1770) is a light at the end of the tunnel. A few broadcasters and even whole countries The use of stereo microphone configuration as have addressed the loudness issue thoroughly, and their input to teleconference application of Directional experience shows that it is possible to solve that problem Audio Coding (DirAC) is presented. DirAC is a to the advantage of the consumer. It is long overdue to method for spatial sound processing, in which the establish a new paradigm in audio leveling: the switch direction of the arrival of sound and diffuseness are from peak normalization to loudness normalization. With analyzed and used for different purposes in repro - widespread adoption of this approach consistent loud - duction. So far, omnidirectional microphones ness not only within a channel, but also between differ - arranged in an array have been used to generate ent channels will be within reach—thus finding the “Holy input signals for one- and two-dimensional sound Grail” of audio broadcasting. field analysis in DirAC processing. In this study the In this session the current situation from the perspec - possibility to use domestic stereo microphones with tive of the EBU Group “P/LOUD” will be examined. Ven - DirAC analysis is investigated. Different methods to dors will present their approaches to loudness metering.

16 Audio Engineering Society 126th Convention Program, 2009 Spring Friday, May 8 11:00 Room D120 and reproduction has grown in areas such as sound field Technical Committee Meeting on Microphones recording, acoustic research, sound field simulation, and Applications audio for electronic games, music listening, and artificial reality. Each of these technologies has its own technical Friday, May 8 11:00 Room D105 concerns involving transducers, environmental simula - Standards Committee Meeting on SC-03-02 Transfer tion, human perception, position sensing, and signal pro - Technologies cessing. This tutorial will cover the underlying principles germane to binaural perception, simulation, recording, Exhibitor Seminar Friday, May 8 and reproduction. It will include live demonstrations as 11:00 – 12:00 Exhibit Floor Booth 2303 well as recorded audio/visual examples. ES1 OPTOCORE –A Friday, May 8 11:30 Room D104 TECHNICAL INTRODUCTION TO OPTOCORE Standards Committee Meeting on SC-02-12 Audio Presenters: Martin Barbour , OPTOCORE Support Applications of IEEE 1394 Engineer Andreas Kaspar , OPTOCORE Support Exhibitor Seminar Friday, May 8 Engineer 11:30 – 12:30 Room D124 Thorsten Schulze , OPTOCORE Support ES2 KRK (USA) and Product Manager The OPTOCORE Synchronous Fibre Network is intro - ROOM OPTIMIZATION WITH ERGO duced. Content will include an Introduction/Overview, AND STUDIO MONITOR DESIGN design of the network, synchronicity, latency, topology, Presenter: Nils Karsten redundancy, protocol, and devices. Every room has his own characteristics. Especially these Workshop 8 Friday, May 8 days it is important, to optimize the acoustics of the lis - 11:30 – 13:30 Room K1 tening environment. An acoustically well treated environ - ment is needed to ensure a well translating mix. ERGO WE HAVE YOU SURROUNDED—MASTERING makes any monitor system more accurate by reducing FOR MULTICHANNEL “bad” room sound. Besides the room acoustic also the speaker design has a fundamental impact on the radia - Chair: Darcy Proper , Senior Mastering Engineer, tion of the sound. The KRK design is known to built bullet Galaxy Studios, Mol, Belgium proof tools for the recording and mastering engineer to ensure to receive a well translating mix. Panelists: Simon Heyworth , Owner/Chief Engineer, Super Audio Mastering, Devon, UK Friday, May 8 12:00 Room D120 Thor Levgold , Owner/Chief Engineer, Sonovo Technical Committee Meeting on Loudspeakers Mastering, Stavanger, Norway and Headphones With the increasing adoption of home cinema systems, DVD releases of performing artists as well as the EBU specifying 5.1 surround audio as standard for HDTV, the Exhibitor Seminar Friday, May 8 need for professional mastering in multichannel formats 12:00 – 13:00 Exhibit Floor Booth 2303 is growing. For many, working in surround represents a ES3 OPTOCORE –B change of paradigm and brings with it unique challenges BASICS OF FIBRE OPTICS and requirements. The panelists in this workshop will share some of the types of challenges they face in the Presenters: Martin Barbour , OPTOCORE Support course of their work and some practical tips for handling Engineer them when they arise in stereo, surround, and multime - Andreas Kaspar , OPTOCORE Support dia productions. Engineer Thorsten Schulze , OPTOCORE Support Tutorial 7 Friday, May 8 and Product Manager 11:30 – 13:30 Room K2 Technical, economical, safety aspects, and advantages of fibre optics is the topic of this seminar. Content will BINAURAL AUDIO TECHNOLOGY—HISTORY, include fibre advantages, reliability, types of fibre, para - CURRENT PRACTICE, AND EMERGING TRENDS meters of fibres (attenuation, bandwidth length product, latency), connector types, cleaning, and testing. Presenter: Robert Schulein , RBS Consultants During the winter and spring of 1931-32, Bell Telephone Laboratories, in cooperation with Leopold Stokowski and Friday, May 8 12:30 Room D105 the Philadelphia Symphony Orchestra, undertook a Standards Committee Meeting on SC-03-04 Storage series of tests of musical reproduction using the most and Handling of Media advanced apparatus obtainable at that time. The objec - tives were to determine how closely an acoustic facsimile of an orchestra could be approached using both stereo Session P11 Friday, May 8 loudspeakers and binaural reproduction. Detailed docu - 13:00 – 16:30 Room K4 ments discovered within the Bell Telephone archives will serve as a basis for describing the results and problems AUDIO CODING revealed while creating the binaural demonstrations. Since these historic events, interest in binaural recording Chair: Nick Zacharov , DELTA, Hørsholm, Denmark ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 17 Technical Progra m 13:00 14:00 P11-1 A Novel Scheme for Low Bit Rate Unified P11-3 A Phase Vocoder Driven Bandwidth Speech and Audio Coding—MPEG RM0— Extension Method with Novel Transient Max Neuendorf ,1 Philippe Gournay ,2 Markus Handling for Audio Codecs— Frederik Nagel ,1 Multrus, 1 Jérémie Lecomte ,1 Bruno Bessette ,2 Sascha Disch ,2 Nikolaus Rettelbach 1 Ralf Geiger, 1 Stefan Bayer, 1 Guillaume Fuchs, 1 1Fraunhofer Institute for Integrated Circuits IIS, Johannes Hilpert, 1 Nikolaus Rettelbach, 1 Erlangen, Germany Frederik Nagel, 1 Julien Robilliard ,1 Redwan 2Leibniz Universitaet Hanover, Hanover, Salami ,3 Gerald Schuller ,4 Roch Lefebvre ,2 Germany Bernhard Grill 1 Storage or transmission of audio signals is of - 1Fraunhofer Institute for Integrated Circuits IIS, ten subject to strict bit-rate constraints. This is Erlangen, Germany; accommodated by audio encoders that encode 2Université de Sherbrooke, Sherbrooke, the lower frequency part in a waveform pre - Quebec, Canada serving way and approximate the high frequen - 3VoiceAge Corporation, Montreal, Quebec, cy signal from the lower frequency data by Canada using a set of reconstruction parameters. This 4Fraunhofer IDMT, Ilmenau, Germany so called bandwidth extension can lead to Coding of speech signals at low bit rates, such roughness and other unpleasant auditory sen - as 16 kbps, has to rely on an efficient speech sations. In this paper the origin of these arti - reproduction model to achieve reasonable facts is identified, and an improved bandwidth speech quality. However, for audio signals not extension method called Harmonic Bandwidth fitting to the model this approach generally fails. Extension (HBE) is outlined avoiding auditory On the other hand, generic audio codecs, roughness in the reconstructed audio signal. designed to handle any kind of audio signal, Since HBE is based on phase vocoders, and tend to show unsatisfactory results for speech thus intrinsically not well suited for transient signals, especially at low bit rates. To overcome signals, an enhancement of the method by a this, a process was initiated by ISO/MPEG, aim - novel transient handling approach is presented. ing to standardize a new codec with consistent A listening test demonstrates the advantage of high quality for speech, music, and mixed con - the proposed method over a simple phase tent over a broad range of bit rates. After a for - vocoder approach. mal listening test evaluating several proposals Convention Paper 7711 MPEG has selected the best performing codec as the reference model for the standardization 14:30 process. This paper describes this codec in detail and shows that the new reference model P11-4 Efficient Cross-Fade Windows for Transitions reaches the goal of consistent high quality for all between LPC-Based and Non-LPC-Based signal types. Audio Coding— Jérémie Lecomte ,1 Philippe Convention Paper 7713 Gournay ,2 Ralf Geiger ,1 Bruno Bessette ,2 Max Neuendorf 1 1Fraunhofer Institute for Integrated Circuits IIS, 13:30 Erlangen, Germany P11-2 A Time-Warped MDCT Approach to Speech 2Université de Sherbrooke, Sherbrooke, Transform Coding— Bernd Edler, 1 Sascha Quebec, Canada; 1 2 2 Disch , Stefan Bayer, Guillaume Fuchs, Ralf The reference model selected by MPEG for the Geiger 2 1 forthcoming unified speech and audio codec Leibniz Universität Hannover, Hannover, (USAC) switches between a non-LPC-based Germany 2 coding mode (based on AAC) operating in the Fraunhofer Institute for Integrated Circuits IIS, transform-domain and an LPC-based coding Erlangen, Germany (derived from AMR-WB+) operating either in the The modified discrete cosine transform (MDCT) time domain (ACELP) or in the frequency is often used for audio coding due to its domain (wLPT). Seamlessly switching between critical sampling property and good energy these different coding modes required the compaction, especially for harmonic tones with design of a new set of cross-fade windows constant fundamental frequencies (pitch). How - optimized to minimize the amount of overhead ever, in voiced human speech the pitch is information sent during transitions between LPC- time-varying and thus the energy is spread over based and non-LPC-based coding. This paper several transform coefficients, leading to a presents the new set of windows that was reduction of coding efficiency. The approach designed in order to provide an adequate trade- presented herein compensates for pitch off between overlap duration and time/frequency variation in each MDCT block by application of resolution, and to maintain the benefits of critical time-variant re-sampling. A dedicated signal sampling through all coding modes. adaptive transform window computation Convention Paper 7712 ensures the preservation of the time domain aliasing cancellation (TDAC) property. Re-sampling 15:00 can be designed such that the duration of the processed blocks is not altered, facilitating the P11-5 Low Bit-Rate Audio Coding in Multichannel replacement of the conventional MDCT in exist - Digital Wireless Microphone Systems— ing audio coders. Stephen Wray , APT Licensing Ltd., Belfast, Convention Paper 7710 Northern Ireland, UK

18 Audio Engineering Society 122nd Convention Program, 2007 Spring Despite advances in voice and data communica - diction often gives more coding gain than mere tions in other domains, sound production for live M/S stereo coding, both in terms of signal-to- events (concerts, theater, conferences, sports, noise ratio and perceptual entropy. worship, etc.) still largely depends on spectrum- Convention Paper 7716 inefficient forms of analog wireless microphone technology. In these live scenarios, low-latency transmission of high-quality audio is mission criti - Friday, May 8 13:00 Room D120 cal. However, while demand increases for wire - Technical Committee Meeting on Audio for less audio channels (for microphones, in-ear Telecommunications monitoring, and talkback systems), some of the radio bands available for “Program Making and Special Events” are to be re-assigned for new wireless mobile telephony and Internet connec - Session P12 Friday, May 8 tivity services: the FCC recently decided to per - 13:30 – 15:00 K4 Foyer mit so-called White Space Devices to operate in sections of UHF spectrum previously reserved POSTERS: LOUDSPEAKERS for shared use by analog TV and wireless micro - phones. This paper examines the key perfor - mance aspects of low bit-rate audio codecs for 13:30 the next generation of bandwidth-efficient digital P12-1 Reduction of Distortion in Conical Horn wireless microphone systems that meet the Loudspeakers at High Levels— Sverre Holm ,1 future needs of live events. Rune Skramstad 2 Convention Paper 7714 1University of Oslo, Oslo, Norway 2Paragon Arrays, Drammen, Norway 15:30 Many horns have audible distortion at high lev - els. We measured a horn consisting of 6 conical P11-6 Krasner’s Audio Coder Revisited— Jamie sections with a 10-inch element at 99 dB SPL. A Angus, Chris Ball, Thomas Peeters, Rowan closed back gave maximum 2.4 percent second Williams , University of Salford, Salford, Greater harmonic and 3.4 percent third harmonic distor - Manchester, UK tion in the 100–1000 Hz range, while an open An audio compression encoder and decoder construction had 1.25 percent and 0.6 percent. A system based on Krasner’s work was imple - new semi-permeable back chamber reduced this mented. An improved Quadrature Mirror Filter to 0.7 percent and 0.35 percent. We hypothesize tree, which more closely approximates modern that the distortion is partly due to the nonlinear critical band measurements, splits the input sig - compliance of air in the back chamber, and part - nal into sub bands that are encoded using both ly is due to the element’s interaction with the adaptive quantization and entropy coding. The front and back loading of the horn, and that the uniform adaptive quantization scheme devel - new construction loads the element in a more oped by Jayant was implemented and enhanced optimal way. through the addition of non-uniform quantization Convention Paper 7717 steps and look ahead. The complete codecs are evaluated using the perceptual audio evaluation algorithm PEAQ and their performance com - 13:30 pared to equivalent MPEG-1 Layer III files. Ini - P12-2 Comparison of Different Methods for the tial, limited, tests reveal that the proposed Subjective Sound Quality Evaluation of codecs score Objective Difference Grades close Compression Drivers— José Martínez ,1 to or even better than MPEG-1 Layer III files Joan Croañes ,2 Jorge Francés Monllor ,3 encoded at a similar bit rate. Jaime Ramis 3 Convention Paper 7715 1Acustica Beyma S.L., Valencia, Spain 2Escola Politecnica Superior de Gandia, 16:00 Valencia, Spain 3Universidad de Alicante, Alicante, Spain P11-7 Inter-Channel Prediction to Prevent Unmasking of Quantization Noise in In this paper an approach to the problem of Beamforming— Mauri Väänänen , Nokia sound quality evaluation of radiating systems is Research Center, Tampere, Finland considered, applying a perceptual model. One of the objectives is to use the parameter pro - This paper studies the use of inter-channel pre - posed by Moore to test if it provides satisfactory diction for the purpose of preventing or reducing results when it is applied to the quality evalua - the risk of noise unmasking when beamforming tion of indirect radiation loudspeakers. Three type of processing is applied to quantized micro - compression drives have been used for these phone array signals. The envisaged application proposals. Recordings with different test sig - is the re-use and postprocessing of user-created nals at different input voltages have been done. content. Simulations with an AAC coder and Using this experimental base, an approach to real-world recordings using two microphones are the problem from different points of view is performed to study the suitability of two existing done: Taking in consideration classic sound coding tools for this purpose: M/S stereo coding quality parameters such as roughness, sharp - and the AAC Long Term Predictor (LTP) tool ness, and tonality. Applying the parameter adapted for inter-channel prediction. The results suggested by Moore obtained from the applica - indicate that LTP adapted for inter-channel pre - tion of a perceptual model. Moreover, a psy - ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 19 Technical Progra m choacoustic experiment has been made on a party devices, transport of control data and Ethernet, population of 25 people. The results, although software control, and firmware upgrades. preliminary and strongly dependant on the ref - erence signal used to obtain Rnonlin, show a Student Event/Career Development good correlation with the Rnonlin values. RECORDING COMPETITION—STEREO Convention Paper 7718 Friday, May 8, 13:30 – 17:00 Room K2 13:30 The Student Recording Competition is a highlight at P12-3 Membrane Modes in Transducers with the each Convention. A distinguished panel of judges par - Direct D/A Conversion— Libor Husník , Czech ticipates in critiquing finalists in each of the categories Technical University in Prague, Prague, Czech in an interactive presentation during the Convention. Republic Student members can submit stereo and surround Operating principle of systems with the direct recordings in the categories classical, jazz, folk/world acoustic D/A conversion, which are sometimes music, and pop/rock. Meritorious awards will be pre - called digital loudspeakers, brings new features sented at the closing Student Delegate Assembly Meet - to the field of transducer design. There are many ing on Saturday. design possibilities to these systems, using dif - Judges include: Jim Anderson, Akira Fukada, Andres ferent transduction principles and spatial Mayo, Ronald Prent, and Darcy Proper. arrangement of constituting parts. This paper deals with the single-acting condenser transduc - Session P13 Friday, May 8 er, suitable for micromachining applications, in 14:00 – 18:30 Room K3 which the membrane is driven by a partitioned back electrode. While in conventional transduc - SPATIAL AUDIO AND SPATIAL PERCEPTION ers the electric force between the back electrode and the membrane is evenly distributed, in digi - Chair: Tapio Lokki , Helsinki University of tal transducers it is no longer the case. Conse - Technology, Espoo, Finland quences to membrane vibrations for some cases of excitation by various distributions of forces 14:00 representing given binary combinations from the dynamic level are presented. P13-1 Evaluation of Equalization Methods for Convention Paper 7719 Binaural Signals— Zora Schärer, Alexander Lindau , TU Berlin, Berlin, Germany 13:30 The most demanding test criterion for the quality P12-4 Increasing Active Radiating Factor of High- of binaural simulations of acoustical environ - Frequency Horns by Using Staggered ments is whether they can be perceptually distin - Arrangement in Loudspeaker Line Array— guished from a real sound field or not. If the Kang An, Yong Shen, Aiping Zhang , Nanjing simulation provides a natural interaction and suf - University, Nanjing, China ficient spatial resolution, differences are predom - inantly perceived in terms of spectral distortions Active Radiating Factor (ARF) is an important due to a non-perfect equalization of the transfer parameter to analyze the loudspeaker line array functions of the recording and reproduction sys - when considering the gap between each two tems (dummy head microphones, headphones). transducers, especially for high-frequency horns. In order to evaluate different compensation As ARF is desired to be as high as possible, the methods, several headphone transfer functions staggered arrangement of horns is introduced in were measured on a dummy head. Based upon this paper. The responses in vertical direction these measurements, the performance of differ - and horizontal direction are analyzed. Compared ent inverse filtering techniques re-implemented with the conventional arrangement, the negative from literature was evaluated using auditory effects of gaps are reduced and responses are measures for spectral differences. Additionally, improved in simulation. an ABC/HR listening test was conducted, using Convention Paper 7720 two different headphones and two different audio stimuli (pink noise, acoustical guitar). In the lis - tening test, a real loudspeaker was directly com - pared to a binaural simulation with high spatial Exhibit Seminar Friday, May 8 resolution, which was compensated using seven 13:00 – 14:00 Exhibit Floor Booth 2303 different equalization methods. ES4 OPTOCORE –C Convention Paper 7721 OPTOCORE IN PRACTICE 14:30 Presenters: Martin Barbour , OPTOCORE Support P13-2 Crosstalk Cancellation between Phantom Engineer Sources — Florian Völk, Thomas Musialik, Andreas Kaspar , OPTOCORE Support Hugo Fastl , Technical University of München, Engineer München, Germany Thorsten Schulze , OPTOCORE Support and Product Manager This paper presents an approach using phantom sources (resulting from the so-called summing This seminar will cover practical exercises with OPTO - localization of two loudspeakers) as sources for CORE system/devices. Content will include practical set - crosstalk cancellation (CTC). The phantom up of a network, remote pre-amp control through third sources can be rotated synchronously with the

20 Audio Engineering Society 126th Convention Program, 2009 Spring listener’s head, thus demanding significantly less reflections need not be lateral, and electroa - processing power than traditional approaches coustic enhancement of late reverberation may using fixed CTC loudspeakers, as an online be vital in small halls. re-computation of the CTC filters is (under cer - Convention Paper 7724 tain circumstances) not necessary. First results of localization experiments show the general 16:00 applicability of this procedure. Convention Paper 7722 P13-5 Frequency-Domain Interpolation of Empirical HRTF Data— Brian Carty, Victor Lazzarini , National University of Ireland, Maynooth, Ireland 15:00 This paper discusses Head Related Transfer P13-3 Preliminary Evaluation of Sweet Spot Size in Function (HRTF)-based artificial spatialization Virtual Sound Reproduction Using Dipoles— of audio. Two alternatives to the minimum Yesenia Lacouture Parodi, Per Rubak , Aalborg phase method of HRTF interpolation are University, Aalborg, Denmark suggested, offering novel approaches to the In a previous study, three crosstalk cancellation challenge of phase interpolation. A phase trun - techniques were evaluated and compared cation, magnitude interpolation technique aims under different conditions. Least square approxi - to avoid complex preparation, manipulation or mations in frequency and time domain were transformation of empirical HRTF data, and any evaluated along with a method based on mini - inaccuracies that may be introduced by these mum-phase approximation and a frequency operations. A second technique adds low fre - independent delay. In general, the least square quency nonlinear frequency scaling to a func - methods outperformed the method based on tionally based phase model. This approach minimum-phase approximation. However, the aims to provide a low frequency spectrum more evaluation was only done for the best-case sce - closely aligned to the empirical HRTF data. nario, where the transfer functions used to Test results indicate favorable performance of design the filters correspond to the listener’s the new techniques. transfer functions and his/her location and orien - Convention Paper 7725 tation relative to the loudspeakers. In this paper we present a follow-up evaluation of the perfor - mance of the three inversion techniques when 16:30 these conditions are violated. A setup to mea - P13-6 Analysis and Implementation of a sure the sweet spot of different loudspeaker Stereophonic Play Back System for arrangements is described. Preliminary mea - Adjusting the “Sweet Spot” to the Listener’s surement results are presented for loudspeakers Position— Sebastian Merchel, Stephan Groth , placed at the horizontal plane and an elevated Dresden University of Technology, Dresden, position, where a typical 60-degree stereo setup Germany is compared with two closely spaced loudspeak - ers. Additionally, two- and four-channel arrange - This paper focuses on a stereophonic play back ments are evaluated. system designed to adjust the “sweet spot” to Convention Paper 7723 the listener’s position. The system includes an optical face tracker that provides information 15:30 about the listener’s x-y position. Accordingly, the loudspeaker signals are manipulated in real-time P13-4 The Importance of the Direct to Reverberant in order to move the “sweet spot.” The stereo - Ratio in the Perception of Distance, phonic perception with an adjusted “sweet spot” Localization, Clarity, and Envelopment— is theoretically investigated on the basis of sev - David Griesinger , Consultant, Cambridge, MA, eral models of binaural hearing. The results indi - USA cate that an adjustment of signals corresponding to the center of the listener’s head does improve The Direct to Reverberant ratio (D/R)—the ratio the localization over the whole listening area. of the energy in the first wave front to the reflect - Although some localization error remains due to ed sound energy—is absent from most discus - asymmetric signal paths for off-center listening sions of room acoustics. Yet only the direct positions, which can be estimated and compen - sound (DS) provides information about the local - sated for. ization and distance of a sound source. This Convention Paper 7726 paper discusses how the perception of DS in a reverberant field depends on the D/R and the time delay between the DS and the reverberant 17:00 energy. Threshold data for DS perception will be P13-7 Issues on Dummy-Head HRTFs presented, and the implications for listening Measurements— Daniela Toledo, Henrik Møller , rooms, hall design, and electronic enhancement Aalborg University, Aalborg, Denmark will be discussed. We find that both clarity and envelopment depend on DS detection. In listen - The dimensions of a person are small compared ing rooms the direct sound must be at least to the wavelength at low frequencies. Therefore, equal to the total reflected energy for accurate at these frequencies head-related transfer func - imaging. As the room becomes larger (and the tions (HRTFs) should decrease asymptotically time delay increases) the threshold goes down. until they reach 0 dB—i.e., unity gain—at DC. Some conclusions: typical listening rooms bene - This is not the case in measured HRTFs: the fit from directional loudspeakers, small concert limitations of the equipment used result in a halls should not have a shoe-box shape, early wrong—and random—value at DC and the effect ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 21 Technical Progra m can be seen well within the audio frequencies. Chair: Jörg Wuttke, Schoeps, Technical Director We have measured HRTFs on a commercially Emeritus available dummy-head Neumann KU-100 and analyzed issues associated to calibration, DC Presenters: Ulrich Apel , Microtech Gefell GmbH correction, and low-frequency response. Infor - Sean Davies , S.W. Davies mal listening tests suggest that the ripples seen Stephan Peus , Neumann GmbH in HRTFs with a wrong DC value affect the This tutorial will be presented in 3 parts. sound quality in binaural synthesis. Stephan Peus ’ presentation, “35 Years of Microphone Convention Paper 7727 Development at Neumann—What Touched Us, What Moved Us,” gives an insight to specific development top - 17:30 ics and to some very special test procedures including: microphone’s transient response: insights beyond fre - P13-8 Binaural Processing Algorithms: Importance quency response or polar pattern; RF susceptibility: of Clustering Analysis for Preference Tests— already a topic before the era of mobile phones; capsule Andreas Silzle, Bernhard Neugebauer, Sunish distortion measurement: difficult procedure giving a lot of George, Jan Plogsties , Fraunhofer Institute for interesting results; dynamic range and self noise level of Integrated Circuits IIS, Erlangen, Germany studio microphones: a remarkable development within The acceptability of a newly proposed technolo - the 35 years in question. gy for commercial application is often assumed if Ulrich Apel will report on “The Importance of Vacuum the sound quality reached in a listening test for Condenser Microphones.” He will speak on such top - surpasses a certain target threshold. As an ics as: the electron-tube was and is still an important example, it is a well-established procedure for step in the development of condenser microphones; the decisions on the deployment of audio codecs to construction of special-made tubes for use in mics such run a listening test comparing the coded/ as RE084k, Hiller MSC2, Telefunken AC701k, EF804, decoded signal with the uncoded reference sig - Valvo EF86, 6072, etc.; and special measuring capabili - nal. For other technologies, e.g., upmix or binau - ties to select tubes regarding noise, stability. and sound” ral processing, however, the unprocessed signal Sean Davies ’s presentation is "Microphone History: only can act as a "comparison signal." Here, the The Why, The How, and The Who.” The developments in goal is to achieve a significant preference of the microphone technology are reviewed from the earliest processed over the comparison signal. For such telephone based type through the decades as far as the preference listening tests, we underline the 1970s. The “Why” section looks at the reasons behind importance of clustering the test results to obtain the different designs, e.g., directional characteristics, out - additional valuable information, as opposed to put signal levels, diffraction effects, frequency range. The using the standard statistic metrics like mean “How” examines the solutions proposed for the “Why” and confidence interval. This approach allows section, and the “Who” identifies the landmark determining the size of the user group that sig - designs and the designers behind them. nificantly prefers to use the proposed algorithm when it would be available in a consumer Friday, May 8 14:00 Room D120 device. As an example, listening test data for Technical Committee Meeting on Network Audio binaural processing algorithms are analyzed in Systems this investigation. Convention Paper 7728 Friday, May 8 14:00 Room D104 Standards Committee Meeting on SC-04-03 18:00 Loudspeaker Modeling and Measurement P13-9 Perception of Head-Position-Dependent Variations in Interaural Cross-Correlation Coefficient— Russell Mason, Chungeun Kim, Workshop 9 Friday, May 8 Tim Brookes , University of Surrey, Guildford, 15:00 – 16:30 Room K1 Surrey, UK MIXING SPORTS IN 5.1, PART 1 Experiments were undertaken to elicit the perceived effects of head-position-dependent Chair: Gerhard Stoll , IRT Munich variations in the interaural cross-correlation coef - ficient of a range of signals. A graphical elicita - Panelists: Dennis Baxter , Sound for the Olympics, USA tion experiment showed that the variations in the Beat Joss , tpc AG, Switzerland IACC strongly affected the perceived width and Ales Koman , Slovenia TV, Slovenia depth of the reverberant environment, as well as the perceived width and distance of the source. Sports has proven to be a major driving force for the in - A verbal experiment gave similar results and troduction of HDTV into the market. Events like the Foot - also indicated that the head-position-dependent ball World Championships 2006, the Football European IACC variations caused changes in the per - Championships in 2008, and the 2008 Bejing Olympics ceived spaciousness and envelopment of the provide an excellent opportunity to showcase the stimuli. strengths of high resolution pictures. Teaming up with Convention Paper 7729 HD picture is High Definition Surround Sound and a gen - erally more elaborate and creative sound design. The challenges are manifold: Tutorial 8 Friday, May 8 • from capturing a roaring stadium crowd of 80,000 to 14:00 – 16:00 Room D123 hearing the details of every ball kick, the so-called “close ball”; MICROPHONE HISTORY • from producing a surround mix to still serving your

22 Audio Engineering Society 126th Convention Program, 2009 Spring stereo-viewers with an appropriate downmix and an intel - 16:30 ligible commentator; P14-1 Further EBU Tests of Multichannel Audio • from getting your signal to and through your broad - 1 2 cast center unharmed to arriving at the consumer in sync Codecs— David Marston , Franc Kozamernik , Gerhard Stoll ,3 Gerhard Spikofski 3 with the picture; 1 • from making meaningful use of LFE to using the BBC R&D, Tadworth, Surrey, UK 2EBU, Geneva, Switzerland center channel not only for commentary but for the 3 "sound of sports" as well. IRT, Munich, Germany A panel of experienced protagonists will discuss these The European Broadcasting Union technical issues and other challenges. Examples will give the audi - group D/MAE has been assessing the quality of ence the chance to judge the effectiveness themselves. multichannel audio codecs in a series of subjec - In Part 2 of this workshop a multitrack recording of one tive tests. The two most recent tests and results of the finals of the Swiss Ice-Hockey Championship 2009 are described in this paper. The first set of tests will be mixed live for the audience to witness the covered 5.1 multichannel audio emission codecs approach to creating a compelling surround sound field, at a range of bit-rates from 128 kbit/s to 448 which includes the atmosphere of thousands of bawling, kbit/s. The second set of tests covered cascaded booing, and applauding fans and the “sounds of the contribution codecs, followed by the most promi - match,” which you see on your screen. nent emission codecs. Codecs under test in - clude offerings from Dolby, DTS, MPEG, APT, and Linear Acoustics. The conclusions observe Friday, May 8 15:00 Room D120 that while high quality is achievable at lower bit- Technical Committee Meeting on Human Factors rates, there are still precautions to be aware of. in Audio Systems The results from cascading of codecs have shown that the emission codec is usually the Exhibitor Seminar Friday, May 8 bottleneck of quality. 15:00 – 16:00 Exhibit Floor Booth 2303 Convention Paper 7730 ES5 OPTOCORE –A

TECHNICAL INTRODUCTION TO OPTOCORE 17:00 Presenters: Martin Barbour , OPTOCORE Support P14-2 Spatial Parameter Decision by Least Squared Engineer Error in Parametric Stereo Coding and MPEG Andreas Kaspar , OPTOCORE Support Surround — Chi-Min Liu, Han-Wen Hsu, Yung- Engineer Hsuan Kao, Wen-Chieh Lee , National Chiao Thorsten Schulze , OPTOCORE Support Tung University, Hsinchu, Taiwan and Product Manager Parametric stereo coding (PS) and MPEG Sur - The OPTOCORE Synchronous Fibre Network is intro - round (MPS) are used to reconstruct stereo or duced. Content will include an Introduction/Overview, multichannel signals from down-mixed signals design of the network, synchronicity, latency, topology, with a few spatial parameters. For extracting spa - redundancy, protocol, and devices. tial parameters, the first issue is to decide a time- frequency (T-F) tile that controls the resolution of reconstructed spatial scenes and highly deter - Friday, May 8 16:00 Room D120 mines the amount of consumed bits. On the oth - Technical Committee Meeting on Automotive Audio er hand, according to the standard syntax, the up-mixing matrices for time slots not on time bor - Exhibitor Seminar Friday, May 8 ders are reconstructed by interpolation in the 16:00 – 17:00 Exhibit Floor Booth 2303 decoder. Therefore, the second issue is to ES6 OPTOCORE –B decide the transmitted parameter values on the time borders for confirming the minimum recon - struction error of matrices. For both PS and MPS, BASICS OF FIBRE OPTICS based on the criterion of least squared error, this Presenters: Martin Barbour , OPTOCORE Support paper proposes a generic dynamic programming Engineer method for deciding the two issues under the Andreas Kaspar , OPTOCORE Support tradeoff of audio quality and limited bits. Engineer Convention Paper 7731 Thorsten Schulze , OPTOCORE Support and Product Manager 17:30 Technical, economical, safety aspects, and advantages P14-3 The Potential of High Performance of fibre optics is the topic of this seminar. Content will Computing in Audio Engineering— David include fibre advantages, reliability, types of fibre, para - Moore, Jonathan Wakefield , University meters of fibres (attenuation, bandwidth length product, of Huddersfield, West Yorkshire, UK latency), connector types, cleaning, and testing. High Performance Computing (HPC) resources are fast becoming more readily available. HPC Session P14 Friday, May 8 hardware now exists for use in conjunction with 16:30 – 18:30 Room K4 standard desktop computers. This paper looks at what impact this could have on the audio engi - MULTICHANNEL CODING neering industry. Several potential applications of HPC within audio engineering research are Chair: Ville Pulkki , Helsinki University of Technology, discussed. A case study is also presented that Espoo, Finland highlights the benefits of using the Single ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 23 Technical Progra m Instruction, Multiple Data (SIMD) architecture parameters mentioned. The survey has been when employing a search algorithm to produce designed to be answered by the workers that are surround sound decoders for the standard exposed to the noise, so that conclusions could 5-speaker surround sound layout. be derived about the feelings and annoyance Convention Paper 7732 that the noise can cause. Convention Paper 7734 18:00 16:30 P14-4 Efficient Methods for High Quality Merging of Spatial Audio Streams in Directional Audio P15-2 On the Design of Automatic Sound Coding— Giovanni Del Galdo ,1 Ville Pulkki ,2 Classification Systems for Digital Hearing Fabian Kuech ,1 Mikko-Ville Laitinen ,2 Richard Aids— Enrique Alexandre, Lorena Álvarez- Schultz-Amling ,1 Markus Kallinger 1 Perez, Roberto Gil-Pita, Raúl Vicen-Bueno, 1Fraunhofer Institute for Integrated Circuits IIS, Lucas Cuadra , University of Alcalá, Alcalá de Erlangen, Germany Henares, Spain 2Helsinki University of Technology, Espoo, Finland The design of digital hearing aids able to carry out advanced functionalities (such as, for Directional Audio Coding (DirAC) is an efficient instance, classify the acoustic environment and technique to capture and reproduce spatial automatically select the best amplification pro - sound. The analysis step outputs a mono DirAC gram for the user’s comfort) exhibits a great diffi - stream, comprising an omnidirectional micro - culty. Since hearing aids have to work at very phone pressure signal and side information, i.e., low clock frequency in order to minimize power direction of arrival and diffuseness of the sound consumption and maximize life battery, the num - field expressed in time-frequency domain. This ber of available instructions per second is actual - paper proposes efficient methods to merge mul - ly very small. This enforces the design of tiple mono DirAC streams to allow a joint play - efficient algorithms with a reduced number of back at the reproduction side. The problem of instructions. In particular, this paper will focus on merging two or more streams arises in applica - three extremely related topics: (1) the design of tions such as immersive spatial audio teleconfer - low-complexity features; (2) the use of automatic encing, virtual reality, and online gaming. feature selection algorithms to optimize the per - Compared to a trivial direct merging of the formance of the classifier; and (3) the critical decoder outputs, the proposed methods are analysis of a variety of different classification more efficient as they do not require the synthe - algorithms, basically based on their complexity sis step. From this it follows the benefit that the and performance and determining whether or loudspeaker setup at the reproduction side does not they are feasible to be implemented. not have to be known in advance. Simulations Convention Paper 7735 and listening tests confirm the absence of any artifacts and that the proposed methods are 16:30 practically indistinguishable from the ideal P15-3 Pruning Algorithms for Multilayer merging. Perceptrons Tailored for Speech/Non-Speech Convention Paper 7733 Classification in Digital Hearing Aids— Lorena Álvarez, Enrique Alexandre, Manuel Rosa-Zurera , University of Alcalá, Alcalá de Session P15 Friday, May 8 Henares, Spain 16:30 – 18:00 K4 Foyer This paper explores the feasibility of using differ - ent pruning algorithms for multilayer perceptrons POSTERS: HEARING (MLPs) applied to the problem of speech/non- speech classification in digital hearing aids. A 16:30 classifier based on MLPs is considered the best option in spite of its presumably high computa - P15-1 Psychoacoustics and Noise Perception tional cost. Nevertheless, its implementation has Survey in Workers of the Construction Sector been proven to be feasible: it requires some —Marcos D. Fernández, Bálder Vitón, José trade-offs involving a balance between reducing Antonio Ballesteros, Samuel Quintana, Isabel the computational demands (that is, the number González , Escuela Universitaria Politécnica de of neurons) and the quality perceived by the Cuenca, Universidad de Castilla-La Mancha, user. In this respect, this paper will focus on the Cuenca, Spain design of three novel pruning algorithms for MLPs, which attempt to converge to the mini - Noise levels are not enough to assess complete - mum complexity network (that is, the lowest ly the influence of the noise. Therefore, psychoa - number of neurons in the hidden layer) without coustics and perception surveys should be taken degrading the performance of it. The results into account. The noise that the construction obtained with the proposed algorithms will be workers produce in their tasks is recorded with a compared with those obtained when using anoth - HATS. Later, those recordings are processed to er pruning algorithm proposed in the literature. derive different parameters: spectrum, weighted Convention Paper 7736 equivalent levels, and the main psychoacoustics parameters. After that, a specific survey has 16:30 been developed to assess the perception of such activity noises during the working time to P15-4 Evolutionary Optimization for Hearing Aids correlate adjectives of perception with those of Computational Auditory Scene Analysis—

24 Audio Engineering Society 126th Convention Program, 2009 Spring Anton Schlesinger, Marinus M. Boone , Technical aids, the acoustic behavior with insert head - University of Delft, Delft, The Netherlands phones is not as well known. Our research focused on the effects of outer ear physical Computational auditory scene analysis (CASA) dimensions, particularly on sound pressure at provides an excellent means to improve speech the eardrum. The main parameter was the intelligibility in adverse acoustical situations. In length of the canal, but eardrum’s damping of order to utilize algorithms of CASA in hearing resonances was also studied. Ear canal simula - aids, sets of algorithmic parameters need to be tors and a dummy head were constructed. Mea - adjusted to the individual auditory performance surements were also performed from human ear of the listener and the acoustic scene in which canals. The study was carried out both with un - they are employed. Performed manually, the blocked ear canals and when the canal entrance optimization is an expensive procedure. We was blocked with an insert earphone. Special therefore developed a framework in which algo - insert earphones with in-ear microphones were rithms of CASA are automatically optimized by constructed for this purpose. Physics-based the principles of evolution, i.e., by a genetic algo - computational models were finally used to vali - rithm. By using the speech transmission index date the approach. (STI) as an objective function, the presented Convention Paper 7739 framework presents a holistic routine that is solely based on psychoacoustical and physiolog - ical models to improve and to assess speech Workshop 10 Friday, May 8 intelligibility. The initial listening test revealed a 16:30 – 18:30 Room D123 discrepancy between the objective and subjec - tive assessment of speech intelligibility, which suggests a review of the objective function. AUDIO NETWORK CONTROL PROTOCOLS Once the objective function is in accordance with the individual perception of speech intelligibility, Chair: Richard Foss , Rhodes University, the presented framework could be applied in the Grahamstown, South Africa optimization of all complex speech processors and therewith accelerate their assessment and Panelists: John Grant , Nine Tiles application. Robby Gurdan , UMAN Convention Paper 7737 Rick Kreifeldt , Harman Professional Philip Nye , Engineering Arts 16:30 With the advent of digital audio networking, there has P15-5 Enhanced Control of On-Screen Faders with been a need to manage the connection of devices on a Computer Mouse— Michael Hlatky, 1 Kristian networks and also to control and access various parame - Gohlke, 1 David Black ,1 Jörn Loviscach 2 ters of the devices. A number of standard and proprietary 1Hochschule Bremen (University of Applied protocols have been developed. In this workshop a panel Sciences), Bremen, Germany of experts who have helped develop some of these pro - 2Fachhochschule Bielefeld (University of Applied tocols will discuss their approaches and attempt to define Sciences), Bielefeld, Germany a way forward, whereby devices with differing protocol implementations can communicate. Input devices of the audio studio that formerly were physical have mostly been converted into virtual controls on the computer screen. Where - Friday, May 8 16:30 Room D104 as this transition saves space and cost, it has Standards Committee Meeting on SC-04-04 reduced the performance of these controls, as Microphone Measurement and Characterization virtual controls adjusted using the computer mouse do not exhibit the accuracy and accessi - bility of their physical counterparts. Previous Workshop 11 Friday, May 8 studies show that interaction with scrollable 17:00 – 18:30 Room K1 timelines can be enhanced by an intelligent interpretation of the mouse movement. We apply MIXING SPORTS IN 5.1, PART 2 similar techniques to virtual faders as used for audio control, leveraging such approaches as Chair: Gerhard Stoll , IRT Munich controllable zoom levels and pseudo-haptic interaction. Tests conducted on five such meth - Panelists: Dennis Baxter , Sound for the Olympics, USA ods provide insight into how to decouple the Beat Joss , tpc AG, Switzerland fader from the mouse movement to improve Ales Koman , Slovenia TV, Slovenia accuracy without impairing the speed of the Sports has proven to be a major driving force for the interaction. introduction of HDTV into the market. Events like the Convention Paper 7738 Football World Championships 2006, the Football Euro - pean Championships in 2008, and the 2008 Bejing 16:30 Olympics provide an excellent opportunity to showcase the strengths of high resolution pictures. Teaming up P15-6 Modeling of External Ear Acoustics for Insert with HD picture is High Definition Surround Sound and a Headphone Usage— Marko Hiipakka, Miikka generally more elaborate and creative sound design. The Tikander, Matti Karjalainen , Helsinki University challenges are manifold: of Technology, Espoo, Finland • From capturing a roaring stadium crowd of 80,000 to Although acoustics of the external ear has been hearing the details of every ball kick, the so-called “close studied extensively for auralization and hearing ball”; ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 25 Technical Progra m • From producing a surround mix to still serving your This year the Banquet will take place in a small old rail - stereo-viewers with an appropriate downmix and an intel - way station, above the valley of the River Isar. The ligible commentator; railway opened in 1891 and steam trains took people • From getting your signal to and through your broad - from the city to many beautiful places in the south of cast center unharmed to arriving at the consumer in sync Munich. Today the steam trains have been replaced with the picture; and the line is now part of the S-Bahn, so the old • From making meaningful use of LFE to using the station is not needed anymore and has been turned center channel not only for commentary but for the into a traditional Bavarian style restaurant with its own “sound of sports” as well. micro-brewery. What could be more natural than A panel of experienced protagonists will discuss these making this location a pleasant place for a “get issues and other challenges. Examples will give the audi - together” in a lovely atmosphere? ence the chance to judge the effectiveness themselves. The welcome beer from the micro brewery and other In Part 2 of this workshop a multitrack recording of one drinks will be followed by a fine buffet with Bavarian deli - of the finals of the Swiss Ice-Hockey Championship 2009 cacies. At the end of a long day at the Convention, these will be mixed live for the audience to witness the Schmankerl will be a good way to relax and enjoy the approach to creating a compelling surround sound field, evening with old and new friends and colleagues. Come which includes the atmosphere of thousands of bawling, and savour Munich’s lifestyle. The ticket price includes booing, and applauding fans and the “sounds of the all food and drinks and the bus to the restaurant and match,” which you see on your screen. back. 55 Euros for AES members; 65 Euros for nonmembers Exhibitor Seminar Friday, May 8 Tickets will be available at the Special Events desk. 17:00 – 18:00 Exhibit Floor Booth 2303 ES7 OPTOCORE –C Session P16 Saturday, May 9 OPTOCORE IN PRACTICE 09:00 – 11:30 Room K4 Presenters: Martin Barbour , OPTOCORE Support Engineer SPATIAL RENDERING, PART 1 Andreas Kaspar , OPTOCORE Support Engineer Chair: Andreas Silzle , Fraunhofer Institute for Thorsten Schulze , OPTOCORE Support Integrated Circuits IIS, Erlangen, Germany and Product Manager 09:00 This seminar will cover practical exercises with OPTO - CORE system/devices. Content will include practical set - P16-1 An Alternative Ambisonics Formulation: up of a network, remote pre-amp control through third Modal Source Strength Matching and the party devices, transport of control data and Ethernet, Effect of Spatial Aliasing— Franz Zotter, 1 software control, and firmware upgrades. Hannes Pomberger ,1 Matthias Frank 2 1University of Music and Dramatic Arts, Graz, Austria Tutorial 9 Friday, May 8 2Graz University of Technology, Graz, Austria 17:30 – 18:30 Room K1 Ambisonics synthesizes sound fields as a sum TECHNIQUES OF AUDIO LOCALIZATION over angular (spherical/cylindrical harmonic) FOR GAMES modes, resulting in the definition of an isotropi - cally smooth angular resolution. This means, vir - Presenters: Fabio Minazzi , Binari Sonori Srl tual sources are synthesized with outstanding Francesco Zambon , Binari Sonori Srl smoothness across all angles of incidence, using discrete loudspeakers that uniformly cover Videogaming is a new form of communication, which a spherical or circular surface around the listen - heavily relies on audio. The production of game sound - ing area. The classical Ambisonics approach tracks shares many technologies and techniques with models the fields of these discrete loudspeakers the production of audio for other media. At the same in terms of a sampled continuum of plane- time, videogames are software products, with nonlinear waves. More accurately, the contemporary con - and dynamic behaviors, and these features of the media cept of Ambisonics uses a continuous angular affect the way audio is produced and localized. This pre - distribution of point-sources at finite distance sentation reviews a set of specific audio production tech - instead, which is considerably easier to imagine. niques that are applied in games localization, like the This also improves the accuracy of holophonic pre- and postproduction methods for localizing voices, sound field synthesis and the analytic descrip - the A/V asset management, the team that attend the tion of the sweet spot. The sweet spot is a limit - local recording sessions, as well as the international ed area of faultless synthesis emerging from quality assurance processes. Some theoretical aspects angular harmonics truncation. Additionally, play - are also described, like the typical constraints that need back with loudspeakers causes spatial aliasing. to be accounted for during the speech voice recording In this sense, it allows for a successive consider - phase, in order to allow the software to seamlessly load, ation of the major shortcomings of Ambisonics: mix, and combine audio assets coming from various the limited sweet spot size and spatial aliasing. countries, recorded in different times. To elaborate on this concept this paper starts with the solution of the nonhomogeneous wave Special Event equation for a spherical point-source distribution, BANQUET and ends with a novel study on spatial aliasing in Friday, May 8, 20:00 – 23:30 Ambisonics. Isar Brau, Munchen Pullach Convention Paper 7740

26 Audio Engineering Society 126th Convention Program, 2009 Spring 09:30 Single User, Work Space, and Work Group, are introduced and analyzed. P16-2 Sound Field Reproduction Employing Convention Paper 7743 Non-Omnidirectional Loudspeakers— Jens Ahrens, Sascha Spors , Deutsche Telekom Laboratories, Techniche Universität Berlin, 11:00 Berlin, Germany P16-5 Spatial Sampling Artifacts of Wave Field In this paper we treat sound field reproduction Synthesis for the Reproduction of Virtual via circular distributions of loudspeakers. The Point Sources— Sascha Spors, Jens Ahrens , general formulation of the approach has been Deutsche Telekom Laboratories, Techniche recently published by the authors. In this paper Universität Berlin, Berlin, Germany we concentrate on the employment of secondary Spatial sound reproduction systems with a large sources (i.e., loudspeakers) whose spatio-tem - number of loudspeakers are increasingly being poral transfer function is not omnidirectional. The used. Wave field synthesis is a reproduction presented approach allows us to treat each spa - technique using a large number of densely tial mode of the secondary source’s spatio-tem - placed loudspeakers (loudspeaker array). The poral transfer function individually. We finally underlying theory, however, assumes a continu - outline the general process of incorporating spa - ous distribution of loudspeakers. Individual loud - tio-temporal transfer functions obtained from speakers placed at discrete positions constitute microphone array measurements. a spatial sampling process that may lead to Convention Paper 7741 sampling artifacts. These may degrade the per - ceived reproduction quality and will limit the 10:00 application of active control techniques like active room compensation. The sampling arti - P16-3 Alterations of the Temporal Spectrum in facts for the reproduction of plane waves have High-Resolution Sound Field Reproduction already been discussed in previous papers. This of Different Spatial Bandwidths— Jens paper derives the spatial sampling artifacts and Ahrens, Sascha Spors , Deutsche Telekom anti-aliasing conditions for the reproduction of Laboratories, Techniche Universität Berlin, virtual point sources on linear loudspeaker Berlin, Germany arrays using wave field synthesis techniques. We present simulations of the wave field Convention Paper 7744 reproduced by a discrete circular distribution of loudspeakers. The loudspeaker distribution is driven either with signals of infinite spatial Session P17 Saturday, May 9 bandwidth (as it happens in wave field synthe - 09:00 – 11:00 Room K3 sis), or the loudspeaker distribution is driven with signals of finite spatial bandwidth (as it is ROOM ACOUSTICS AND LOUDSPEAKER the case in near-field compensated higher INTERACTION order Ambisonics). The different spatial band - widths lead to different accuracies of the Chair: Eddy B. Brixen , EBB-consult, Smørum, desired component of the reproduced wave Denmark field and to spatial aliasing artifacts with essen - tially different properties. Our investigation 09:00 focuses on the potential consequences of the artifacts on human perception. P17-1 Effects of Loudspeaker Directivity on Convention Paper 7742 Perceived Sound Quality—A Review of Existing Studies— William Evans ,1 Jakob Dyreby ,2 Søren Bech ,2 10:30 Slawomir Zielinski ,1 Francis Rumsey 1 P16-4 Cooperative Spatial Audio Authoring: 1University of Surrey, Guildford, Surrey, UK Systems Approach and Analysis of Use 2Bang & Olufsen A/S, Struer, Denmark 1 Cases— Jens-Oliver Fischer , Francis The directivity of a loudspeaker system is often Gropengiesser ,2 Sandra Brix 1 1 regarded as a prominent factor in the overall Fraunhofer Institute for Digital Media subjective quality of the reproduced sound expe - Technology IDMT, Ilmenau, Germany 2 rience. Much literature is available on the topic, TU Ilmenau, Ilmenau, Germany and currently a broad field of opinion exists Today’s audio production process is highly par - among designers. This paper provides an allel and segregated. This is especially the case overview of the available literature, as well as an in the field of audio postproduction for motion extended investigation into listener-based pictures. The introduction of spatial audio sys - research. Results indicate that for such a widely tems like 5.1, 22.2 or Wave Field Synthesis debated topic, conclusive measurement data results in even more production steps, namely with regard to human listeners is limited, and, the spatial authoring, to accomplish a rich expe - therefore, a proposal for more informative listen - rience for the audience. This paper proposes a ing tests is presented. system that enables the audio engineers to work Convention Paper 7745 together on the same project. The proposed sys - tem is planned to be implemented for an existing 09:30 spatial authoring software but can be utilized by any other application that organizes its data in a P17-2 Subjective Validity of Figures of Merit for tree structured way. Three major use cases, i.e., Room Aspect Ratio Design— Matthew ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 27 Technical Progra m Wankling, Bruno Fazenda , University of Workshop 12 Saturday, May 9 Huddersfield, Huddersfield, West Yorkshire, UK 09:00 – 11:00 Room K1

Attempts have long been made to classify a REALITY IS NOT A RECORDING / A RECORDING IS room’s low frequency audio reproduction capabili - NOT REALITY ty with regard to its aspect ratio. Common metrics used have relied on the homogeneous distribution Presenter: Jim Anderson , New York University, New of modal frequencies and from these a number of York, NY, USA “optimal” aspect ratios have emerged. However, most of these metrics ignore the source and re - The former New York Times film critic, Vincent Canby, ceiver coupling to the mode shapes—only a few wrote “all of us have different thresholds at which we account for this in the derivation of a figure of suspend disbelief, and then gladly follow fictions to con - merit. The subjective validity of these attempts is clusions that we find logical." Any recording is a “fiction,” tested and discussed. Examples are given of sup - a falsity, even in its most pure form. It is the responsibili - posedly good room ratios with bad performance ty, if not the duty, of the recording engineer and produc - and vice versa. Subjective assessment of various er, to create a universe so compelling and transparent room scenarios is undertaken and a ranking order that the listener isn't aware of any manipulation. Using has been obtained to correlate with a proposed basic recording techniques, and standard manipulation figure of merit. of audio, a recording is made, giving the listener an Convention Paper 7746 experience that is not merely logical but better than reali - ty. How does this occur? What techniques can be applied? How does an engineer create a convincing 10:00 loudspeaker illusion that a listener will perceive as a P17-3 A Study of Low-Frequency Near- and plausible reality? Recordings will be played. Far-Field Loudspeaker Behavior— John Vanderkooy ,1,2 Martial Rousseau 2 Student Event/Career Development 1University of Waterloo, Waterloo, Ontario, Canada EDUCATION FORUM PANEL 2B&W Group Ltd., Steyning, West Sussex, UK Saturday, May 9, 09:00 – 11:00 Room K2 Low-frequency loudspeaker measurements are difficult. Room reflections, mediocre anechoic Moderator: John Krivit , New England Institute of Art, chambers, and random noise play havoc with Brookline, MA, USA the quest. Diffraction is different in nearfield Teaching Quality and farfield. This paper covers a range of topics that bear on these problems, such as boundary John Krivit invites all educators to join him for this open element diffraction simulations, an approximate discussion on how we teach students about sound quality theory for low frequencies, methods to shorten and high production standards. In an age when many the impulse response, and nearfield character - graduates will not have the chance to work with the finest istics. A few points are illustrated with measure - equipment in the finest rooms, what are the responsibili - ments. An earlier simplified diffraction theory of ties of the faculty and what are the expectations for the fa - Kessel is checked for axisymmetric cylindrical cilities at institutions of higher learning? Come share your and rectangular boxes by boundary-element approaches and learn from others as we all seek to teach simulations, in an attempt to pin down the quality. An open forum for educators, students are also diffractive 4pi to 2pi transition. It turns out to welcome to attend and participate. have a strong connection to the acoustic center of a loudspeaker. Some measurements are Saturday, May 9 09:00 Room D120 made under various conditions. Shortening Technical Committee Meeting on Audio Recording methods are used to minimize the deleterious and Mastering Systems effect of truncating room reflections from the impulse response. Saturday, May 9 09:00 Room D104 Convention Paper 7747 Standards Committee Meeting on SC-05-05 Grounding and EMC Practices 10:30 Saturday, May 9 10:00 Room D120 P17-4 Subwoofers in Symmetrical and Technical Committee Meeting on High Resolution Asymmetrical Rooms— Juha Backman , Audio Nokia Corporation, Espoo, Finland A theoretical study of behavior of single and mul - Session P18 Saturday, May 9 tiple subwoofers, taking also geometrical and 10:30 – 12:00 K4 Foyer acoustical asymmetry of practical listening envi - ronments into account, is presented. The results indicate that configurations aimed at precise POSTERS: ASSESSMENT, EVALUATION cancellation of individual modes have a high sensitivity to deviations from the ideal. However, 10:30 with multiple subwoofers it is possible to find ro - P18-1 WhisPER—A New Tool for Performing bust placements that both reduce the spatial Listening Tests— Simon Ciba, André variation of the sound field and the frequency Wlodarski, Hans-Joachim Maempel , Technical variation of the response. This, however, University of Berlin, Berlin, Germany requires loudspeaker placements where also the height of the source from the floor is varied. A software tool is presented for performing ex - Convention Paper 7748 periments in the field of perceptual audio evalua -

28 Audio Engineering Society 126th Convention Program, 2009 Spring tion and psychoacoustic measurement, control - Live Sound Seminar 1 Saturday, May 9 ling the interaction with both the subject and the 10:30 – 13:00 Atrium 2 playback environment. For this purpose a reper - toire of test procedures has been implemented, NEUMANN & MULLER AND D & B including popular qualitative and quantitative approaches. By using OpenSound Control com - Presenters: Stefan Goertz mands, not only traditional multichannel repro - Michael Kennedy duction is supported, but also advanced spatial Sound System Design and Commissioning audio reproduction such as dynamic binaural in Critical Acoustic Environments synthesis or wavefield synthesis. WhisPER has been written in MATLAB to facilitate its further The interaction of sound systems with special attention to development within the scientific community. As the excitation of diffuse sound will be examined in theory opposed to existing libraries it provides a coher - and practical demonstrations using speech intelligibility ent graphical user interface system allowing measurements as an indicator. Software-aided line and easier access and configuration also for users subwoofer array designs will be discussed followed by a without advanced programming experience. live demonstration of the tuning and alignment process. Convention Paper 7749 [Paper presented by Hans-Joachim Maempel] Exhibitor Seminar Saturday, May 9 10:30 – 11:30 Room D124 10:30 ES8 MINNETONKA AUDIO SOFTWARE (GERMANY) P18-2 Psychoacoustic Assessment of the Noise Emitted by the Machines. The Case of the SURCODE ENCODER FOR DTS-HD WORKFLOW ON Grinders— Marcos D. Fernández, José Antonio APPLE FINAL CUT PRO – COMPRESSOR Ballesteros, Iván Suárez, Samuel Quintana, Isabel González , Escuela Universitaria Presenters: Markus Hintz , Minnetonka Audio Software Politécnica de Cuenca, Universidad de Achim Scherner , DTS Castilla-La Mancha, Cuenca, Spain Jayson Tomlin , Minnetonka Audio Software Sound quality is used as the suitability of the DTS HD has become a preferred audio format for Blu- sound emitted by a machine, depending on the ray production. DTS and Minnetonka will explain how characteristics of that sound and the perceptual authoring houses can increase workflow efficiency and sensation received that reflects the degree of make best use of available disc capacity for both Blu-ray acceptance of the machine by the user. In order and DVD-Video production. to evaluate the sound quality of the grinders under study, binaural recordings, are required of Saturday, May 9 11:00 Room D120 the emitted sounds to determine the objective Technical Committee Meeting on Electronic psychoacoustic parameters, and then, to make Distribution of Audio subjective tests to a representative number of people about the impression made by that Saturday, May 9 11:00 Room D104 particular sound. Standards Committee Meeting on SC-05-02 Audio Convention Paper 7750 Connectors

10:30 Exhibitor Seminar Saturday, May 9 11:00 – 12:00 Exhibit Floor Booth 2303 P18-3 Investigations of the Effects of Nonlinear ES9 OPTOCORE –A Distortions on Psychoacoustical Measures —Stephan Herzog , Technical University TECHNICAL INTRODUCTION TO OPTOCORE Kaiserslautern, Kaiserslautern, Germany Presenters: Martin Barbour , OPTOCORE Support The perception of nonlinear distortions of audio Engineer devices, in particular the perception of nonlinear Andreas Kaspar , OPTOCORE Support distortions of digital audio, is only insufficiently Engineer described by typical measures like THD. To pro - Thorsten Schulze , OPTOCORE Support vide a better insight into perceptual effects of and Product Manager nonlinear distortions, their audibility, and the impact on the psychoacoustical measures like The OPTOCORE Synchronous Fibre Network is intro - loudness and sharpness is examined. For this duced. Content will include an Introduction/Overview, purpose a test method has been developed. The design of the network, synchronicity, latency, topology, first step in the test is the measurement of the redundancy, protocol, and devices. frequency response of the device under test with an efficient method to enable the separation of Workshop 13 Saturday, May 9 linear and nonlinear processing. The second 11:30 – 13:00 Room K1 step of the test consists of the computation of the psychoacoustical measures and the thresh - AES 42 DIGITAL MICROPHONES olds for the audibility of nonlinear distortions. Both computations are based on the same psy - Chair: Gregor Zielinsky , Sennheiser choacoustical model to obtain consistent results. Results for several types of distortion obtained Panelists: Malgorzata Albinska-Frank with simulations and measurements on analog Stephan Flock , Direct Out circuits are presented. Stephan Peus , Neumann Convention Paper 7751 Helmut Wittek , Schoeps ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 29 Technical Progra m The panel will discuss daily work with digital micro - Saturday, May 9 13:00 Room D120 phones and their peripheral devices. Technical Committee Meeting on Audio Forensics While the first digital microphones were launched a few years ago, their use is still within an exclusive com - Saturday, May 9 13:00 Room D104 munity of users. Standards Committee Meeting on SC-02-01 Digital During the last two years, prices of digital microphones Audio Measurement Techniques have dropped, while the choice of mics—as well as choic - es of interfaces—have risen very strongly. More and Exhibitor Seminar Saturday, May 9 more companies are incorporating the AES 42 standard. 13:00 – 14:00 Exhibit Floor Booth 2303 In this panel manufacturers and users discuss possibil - ES12 OPTOCORE –C ities and workflows of digital microphones. Experienced users will give their view on the AES 42 standard. OPTOCORE IN PRACTICE

Student Event/Career Development Presenters: Martin Barbour , OPTOCORE Support RECORDING COMPETITION—SURROUND Engineer Saturday, May 9, 11:30 – 13:30 Andreas Kaspar , OPTOCORE Support Room K2 Engineer Thorsten Schulze , OPTOCORE Support The Student Recording Competition is a highlight at each and Product Manager Convention. A distinguished panel of judges participates in critiquing finalists in each of the categories in an inter - This seminar will cover practical exercises with OPTO - active presentation during the Convention. Student mem - CORE system/devices. Content will include practical set - bers can submit stereo and surround recordings in the up of a network, remote pre-amp control through third categories classical, jazz, folk/world music, and pop/rock. party devices, transport of control data and Ethernet, Meritorious awards will be presented at the closing Stu - software control, and firmware upgrades. dent Delegate Assembly Meeting later on Saturday. Judges include: Jim Anderson, Akira Fukada, Andres Session P19 Saturday, May 9 Mayo, Ronald Prent, and Darcy Proper. 13:30 – 15:00 Room K4

Exhibitor Seminar Saturday, May 9 EVENT, STAGE, AND SOUND REINFORCEMENT 11:30 – 12:30 Room D124 ES10 KRK (USA) Chair: Francis Rumsey , University of Surrey, Guildford, Surrey, UK ROOM OPTIMIZATION WITH ERGO 13:30 AND STUDIO MONITOR DESIGN P19-1 Comparative Evaluation of Howling Detection Presenter: Nils Karsten Criteria in Notch-Filter-Based Howling Every room has his own characteristics. Especially these Suppression— Toon van Waterschoot, Marc days it is important, to optimize the acoustics of the lis - Moonen , Katholieke Universiteit Leuven, tening environment. An acoustically well treated environ - Leuven, Belgium ment is needed to ensure a well translating mix. ERGO Notch-filter-based howling suppression (NHS) is makes any monitor system more accurate by reducing one of the most popular methods for acoustic “bad” room sound. Besides the room acoustic also the feedback control in public address and hands- speaker design has a fundamental impact on the radia - free communication systems. The NHS method tion of the sound. The KRK design is known to built bullet consists of two stages: howling detection and proof tools for the recording and mastering engineer to notch filter design. While the design of notch fil - ensure to receive a well translating mix. ters is based on well-established filter design techniques, there is little agreement in the NHS Saturday, May 9 12:00 Room D120 literature on how the howling detection subprob - Technical Committee Meeting on Studio Practices lem should be tackled. Moreover, since the NHS and Production literature mainly consists of patents, only few ex - perimental results have been reported. The aim of this paper is to describe a unifying framework Exhibitor Seminar Saturday, May 9 for howling detection and to provide a compara - 12:00 – 13:00 Exhibit Floor Booth 2303 tive evaluation of existing and novel howling ES11 OPTOCORE –B detection criteria. Convention Paper 7752 BASICS OF FIBRE OPTICS 14:00 Presenters: Martin Barbour , OPTOCORE Support Engineer P19-2 Professional Wireless Microphone Systems: Andreas Kaspar , OPTOCORE Support Current Situation and Upcoming Changes in Engineer Regulatory Issues in Europe and USA— Thorsten Schulze , OPTOCORE Support Frank Ernst , Beyerdynamic GmbH & Co. KG, and Product Manager Heilbronn, Germany Technical, economical, safety aspects, and advantages Professional wireless microphones have been in of fibre optics is the topic of this seminar. Content will use for almost 50 years now. The operation is include fibre advantages, reliability, types of fibre, para - based on a frequency sharing with TV broadcast meters of fibres (attenuation, bandwidth length product, transmitters. With the transition to digital TV, this latency), connector types, cleaning, and testing. situation changes. Digital TV is more spectrum 30 Audio Engineering Society 126th Convention Program, 2009 Spring efficient. After the transition is completed, areas loss rate in the simulated system is presented. of the spectrum will be cleared from TV broad - Convention Paper 7755 cast and will be available for new services. These cleared spectrum areas are referred to as 13:30 “digital dividend” or white spaces. By the alloca - tion of these bands to other services, valuable P20-2 A Joint Approach to Extract Multiple resources for the operation of professional wire - Fundamental Frequency in Polyphonic less microphones will be lost. This paper will Signals Minimizing Gaussian Spectral give an overview on the current situation for pro - Distance— Francisco J. Cañadas-Quesada, fessional wireless microphones and the upcom - Pedro Vera-Candeas, Nicolás Ruiz-Reyes, ing changes with the transition to digital TV. Julio Jose Carabias-Orti, D. Martínez-Muñoz , Convention Paper 7753 University of Jaén, Jaén, Spain This paper presents a joint estimation approach 14:30 to extract multiple fundamental frequency (F0) in P19-3 Sound Field Reconstruction: An Improved monaural polyphonic music signals. In a frame- Approach for Wave Field Synthesis— Mihailo based analysis, we generate a spectral envelope Kolundzija, 1 Christof Faller ,1 Martin Vetterli 1,2 for each combination of F0 candidates, from 1Ecole Polytechnique Fédérale de Lausanne, non-overlapped partials, under assumption that Lausanne, Switzerland; a harmonic sound is characterized by a Gauss - 2University of California at Berkeley, Berkeley, ian mixture model (GMM). The optimal F0 CA, USA candidates combination minimizes a spectral Euclidean distance measure between the origi - Wave field synthesis (WFS) is a prevalent nal spectrum and Gaussian spectral models. approach to multiple-loudspeaker sound repro - Evaluation was carried out using several piano duction for an extended listening area. Although recordings. Evaluation shows promising results. powerful as a theoretical concept, its deployment Convention Paper 7756 is hampered by practical limitations due to dif - fraction, aliasing, and the effects of the listening 13:30 room. Reconstructing the desired sound field in the listening area, accounting for the medium P20-3 A Mixture-of-Experts Approach for Note propagation characteristic, is another approach Onset Detection— Norberto Degara, 1 Antonio termed as sound field reproduction (SFR). It is Pena, 1 Manuel Sobreira-Seoane ,1 Soledad based on the essential band-limitedness of the Torres-Gijarro 2 sound field, which enables a continuous match - 1Universidade de Vigo, Vigo, Spain ing of the reconstructed and the desired sound 2Laboratorio Oficial de Metroloxía de Galicia field by their matching on a discrete set of points (LOMG), Tecnópole, Ourense, Spain spaced below the Nyquist distance. We compare the two approaches in a common single-source Finding the starting time of events (onsets) is free-field setup, and show that SFR provides useful in a number of applications for audio sig - improved sound field reproduction compared to nals. The goal of this paper is to present a com - WFS in a wide listening area around a defined bination of techniques for automatic detection of reference line. events in audio signals. The proposed system Convention Paper 7754 uses a supervised classification algorithm to combine a set of features extracted from the audio signal and reduce the original signal to a Session P20 Saturday, May 9 robust detection function. Onsets are obtained 13:30 – 15:00 K4 Foyer by using a simple peak-picking algorithm. This paper describes the analysis system used to POSTERS: SIGNAL PROCESSING extract the features and the details of the neural network algorithm used to combine them. We 13:30 conclude by comparing the performance of the proposed algorithm with the system that P20-1 Subjective Audio Quality with Multiple obtained the first place in the 2005 Music Infor - Description Coding in a WLAN 802.11-Based mation Retrieval Evaluation eXchange. Multicast Distribution Network— Marcus Convention Paper 7757 Purat, Tom Ritter , TFH Berlin, Berlin, Germany Paper presented by Soledad Torres-Gijarro This paper presents a number of results of a study of different methods to mitigate the 13:30 impact of packet loss on the subjective quality P20-4 Automatic Adjustment of Off-the-Shelf in a wireless distribution network of com - Reverberation Effects— Sebastian Heise, 1 pressed high fidelity audio. The system was Michael Hlatky ,1 Jörn Loviscach 2 simulated in Matlab based on parameters of 1Hochschule Bremen (University of Applied an 802.11a WLAN in multicast-mode and the Sciences), Bremen, Germany Vorbis codec. The performance of multiple 2Fachhochschule Bielefeld (University of Applied description coding in terms of expected Sciences), Bielefeld, Germany subjective audio quality and its impact on the data rate is quantified and compared to other Virtually all effect units that process digital receiver-based packet loss concealment meth - audio—software plug-ins as well as dedicated ods. The optimum set of parameters for a hardware—can be controlled digitally. This multiple description coder that achieves the allows subjecting their settings to optimization best subjective audio quality for a given packet processes. We demonstrate the automatic adap - ¯ Audio Engineering Society 126th Convention Program, 2009 Spring 31 Technical Progra m tation of reverberation plug-ins to given room However, this configuration is subject to Com - impulse responses. This facilitates replacing mon-Mode voltage Induced Distortion—CMID. computationally expensive convolution reverber - The susceptibility is further increased if the ation units with standard ones, which also are amplifier source impedance is not perfectly amenable to easier parameter tweaking after matched. This is illustrated by tests on popular their overall setting has been adjusted through audio op amps. Advanced transformer input our method. We propose optimization strategies stage topology proposed in this paper complete - for this multi-dimensional nonlinear problem that ly prevents this kind of distortion. Noise perfor - need no adaptation to the particularities of each mance remains unaffected, yet listening tests in effect unit, are sped up using multicore proces - practical application confirm the sound to be sors and networked computers. The optimization more pleasant. process evaluates the difference between the Convention Paper 7760 actual response and the targeted response on the basis of psychoacoustic features. An acoustic comparison with effect parameter set - Workshop 14 Saturday, May 9 tings crafted by professional human operators 13:30 – 15:30 Room K1 indicates that the computationally optimized set - tings yield comparable or better results. Convention Paper 7758 5.1 HIGH PROFILE MIXING

Cochairs: Akira Fukada , Senior Engineer, NHK, 13:30 Tokyo, Japan P20-5 Improvements on Automatic Parametric Ulrike Schwarz , Engineer, Bavarian Radio, Equalization and Cross-Over Alignment Munich, Germany of Audio Systems— German Ramos, Pedro Tomas , Technical University of Valencia, Panelists: Jean-Marie Geijsen , Director & Balance Valencia, Spain Engineer, Polyhymnia, Baarn, The Netherlands The idea and algorithm implementation of an Sascha Paeth , Owner/Engineer, Gate automatic parametric equalizer and cross-over Studios, Wolfsburg, Germany alignment of audio systems was proposed previ - Ronald Prent , Residential Surround ously by one of the authors with proven success. Engineer, Galaxy Studios, Mol, Belgium This method designed Infinite Impulse Response (IIR) equalization and cross-over filters directly in It is very interesting how the perception of music can be a series of second-order-sections (SOS) altered by a mixing engineer. A conductor or a musician employing peak filters, and pre-initialized high- changes the figure of written music, the composer's pass and low-pass filters, defined by its parame - work, by his or her interpretation and expression. For the ters (frequency, gains, and Q). The method sup - recording of music it is rather important what the music ported the inclusion of constraints (maximum conveys to the engineer. In the process of recording and and minimum parameter values) and designed mixing the engineer will approach and embrace the the SOS in order of importance in the equaliza - music like an artist or musician. tion, providing thus a scalable filter implementa - In this workshop engineers who have various cultural tion. In order to lower the order of the needed and musical backgrounds present their different mixing filter, and looking also for an automatic decision results. What did each engineer consider and what did on the selection of filter types and initialization, they aim at? We believe that considering the result is a several improvements are presented. It is now very important element for understanding music and art. possible for the algorithm to select, configure, and use shelving filters in the SOS chain for Student Event/Career Development equalization. Also, the decision and initialization EDUCATION FAIR of the needed high-pass and low-pass SOS fil - Saturday, May 9, 13:30 – 15:00 ters could be automatic, helping in the cross- K4 Foyer over design stage for active audio systems. Convention Paper 7759 Institutions offering studies in audio (from short cours - es to graduate degrees) will be represented in a “table top” session. Information on each school’s respective 13:30 programs will be made available through displays and P20-6 Low Noise Transformer Input Preamp academic guidance. There is no charge for schools to Design—A Solution that Eliminates CMID participate. Admission is free and open to all Conven - —Milan Kovinic ,1 Dragan Drincic ,2 Sasha tion attendees. Jankovic 3 1MMK Instruments, Belgrade, Serbia 2Advanced School for Electrical & Computer Workshop 15 Saturday, May 9 Engineering, Belgrade, Serbia 14:00 – 15:00 Room K2 3OXYGEN-Digital, Parkgate Studio, Sussex, UK WAR ON MUSIC This paper examines more closely the advan - tages of input transformer-op amplifier configura - Chair: Thomas Lund , tc electronic tions, especially those implemented in low-noise designs. Usual transformer input stage topology Music is currently under fire from two sides: Hyper-level works in non-inverting architecture, since it and data reduction. allows the transformer to work with optimum Using theory as well as listening examples, the ses - loading, to maximize the signal-to-noise ratio. sion will demonstrate how this is a lethal combination. 16

32 Audio Engineering Society 126th Convention Program, 2009 Spring bit audio recorded 20 years ago using less perfect con - fashion. Because the inverse BEM technique version and processing equipment than today easily end can deal with arbitrary shaped source or bound - up sounding better than modern recordings. ing surfaces, one can simultaneously consider Based on years of research, some of the significant the effect of irregular radiation surface and "quality bottlenecks" of today's production and delivery reflection boundaries having impedances such system are identified, and ways of improving on the situ - as walls, floor, and ceiling. To examine the ation are suggested. applicability, a field rendering example was test - ed to control the relative spatial distribution of Saturday, May 9 14:00 Room D120 sound pressure in the enclosed field. Technical Committee Meeting on Perception Convention Paper 7762 and Subjective Evaluation of Audio 16:00 Exhibitor Seminar Saturday, May 9 P21-3 Validation of a Loudspeaker-Based Room 14:30 – 15:30 Room D124 Auralization System Using Speech ES13 MINNETONKA AUDIO SOFTWARE Intelligibility Measures— Sylvain Favrot, (GERMANY) Jörg M. Buchholz , Technical University of Denmark, Lyngby, Denmark FILE- AND TAPE-BASED WORKFLOWS USING SURCODE FOR DOLBY E A novel loudspeaker-based room auralization (LoRA) system has been proposed to generate Presenters: Markus Hintz , Minnetonka Audio Software versatile and realistic virtual auditory environ - Jayson Tomlin , Minnetonka Audio Software ments (VAEs) for investigating human auditory TBA , Dolby Laboratories perception. This system efficiently combines SurCode for Dolby E is quickly becoming the standard modern room acoustic models with loudspeaker for Dolby E software-based encoding and decoding. auralization using either single loudspeaker or Minnetonka Audio Software will present how SurCode for high-order Ambisonics (HOA) auralization. The Dolby E is the best choice for both file and tape based LoRA signal processing of the direct sound and workflows on a variety of workstation platforms. the early reflections was investigated by measur - ing the speech intelligibility enhancement by ear - ly reflections in diffuse background noise. Dan - Session P21 Saturday, May 9 ish sentences were simulated in a classroom 15:00 – 18:30 Room K4 and the direct sound and each early reflection were either auralized with a single loudspeaker, SPATIAL RENDERING, PART 2 HOA or first-order Ambisonics. Results indicated that (i) absolute intelligibility scores are signifi - Chair: Sascha Spors , Technische Universität cantly dependent on the reproduced technique Berlin, Berlin, Germany and that (ii) early reflections reproduced with HOA provide a similar benefit on intelligibility as 15:00 when reproduced with a single loudspeaker. It is concluded that speech intelligibility experiments P21-1 Score File Generators for Boids-Based can be carried out with the LoRA system either Granular Synthesis in Csound— Enda Bates, with the single loudspeaker or HOA technique. Dermot Furlong , Trinity College, Dublin, Ireland Convention Paper 7763 In this paper we present a set of score file gener - ators and granular synthesis instruments for the 16:30 Csound language. The applications use spatial data generated by the Boids flocking algorithm P21-4 Low Complexity Directional Sound Sources along with various user-defined values to gener - for Finite Difference Time Domain Room ate score files for grainlet additive synthesis, Acoustic Models— Alexander Southern, granulation, and glisson synthesis instruments. Damian Murphy , University of York, York, UK Spatialization is accomplished using Higher The demand for more natural and realistic aural - Order Ambisonics and distance effects are mod - ization has resulted in a number of approaches eled using the Doppler Effect, early reflections, to the time domain implementation of directional and global reverberation. The sonic quality of sound sources in wave-based acoustic modeling each synthesis method is assessed and an origi - schemes such as the Finite Difference Time nal composition by the author is presented. Domain (FDTD) method and the Digital Wave - Convention Paper 7761 guide Mesh (DWM). This paper discusses an approach for implementing simple regular direc - 15:30 tive sound sources using multiple monopole excitations with distributed spatial positioning. P21-2 Acoustical Rendering of an Interior Space These arrangements are tested along with a dis - Using the Holographically Designed Sound cussion of the characteristic limitations for each Array— Wan-Ho Cho, Jeong-Guon Ih , KAIST, setup scenario. Daejeon, Korea Convention Paper 7764 It was reported that the filter for the acoustic array can be inversely designed in a holographic 17:00 way, which was demonstrated in a free-field. In this study the same method using the boundary P21-5 Binaural Reverberation Using a Modified Jot element method (BEM) was employed to render Reverberator with Frequency-Dependent and the interior sound field in an acoustically desired Interaural Coherence Matching— Fritz Menzer, ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 33 Technical Progra m Christof Faller , Ecole Polytechnique Fédérale Session P22 Saturday, May 9 de Lausanne, Lausanne, Switzerland 15:00 – 18:30 Room K3 An extension of the Jot reverberator is present - MICROPHONES AND HEADPHONES ed, producing binaural late reverberation where the interaural coherence can be controlled as a function of frequency such that it matches the Chair: William Evans , University of Surrey, Guildford, frequency-dependent interaural coherence of a Surrey, UK reference binaural room impulse response (BRIR). The control of the interaural coherence 15:00 is implemented using linear filters outside the P22-1 Frequency Response Adaptation in Binaural reverberator’s recursive loop. In the absence of Hearing— David Griesinger , Consultant, a reference BRIR, these filters can be calculated Cambridge, MA, USA from an HRTF set. Convention Paper 7765 The pinna and ear canals act as listening trum - pets to concentrate sound pressure on the eardrum. This concentration is strongly frequen - cy dependent, typically showing a rise in pres - 17:30 sure of 20 dB at 3000 Hz. In addition, diffraction and reflections from the pinna substantially alter P21-6 Design and Limitations of Non-Coincidence the frequency response of the eardrum pressure Correction Filters for Soundfield as a function of the direction of a sound source. 1 Microphones— Christof Faller , Mihailo In spite of these large departures from flat 2 Kolundzija response, listeners usually report that a uniform 1 Illusonic LLC, Lausanne, Switzerland pink power spectrum sounds frequency bal - 2 Ecole Polytechnique Fédérale de Lausanne, anced, and loudspeakers are manufactured to Lausanne, Switzerland this standard. But on close listening frontal pink The tetrahedral microphone capsule arrange - noise does not sound uniform. The ear clearly ment in a soundfield microphone captures a uses adaptive correction of timbre to achieve so-called A-format signal, which is then convert - these results. This paper discusses and demon - ed to a corresponding B-format signal. The strates the properties and limits of this adapta - phase differences between the A-format signal tion. The results are important for our experience channels due to non-coincidence of the micro - of live music in halls and in reproduction of phone capsules cause coloration and errors in music through loudspeakers and headphones. the corresponding B-format signals and linear Convention Paper 7768 combinations thereof. Various strategies for designing B-format non-coincidence correction 15:30 filters are compared and limitations are discussed. P22-2 Concha Headphones and Their Coupling to Convention Paper 7766 the Ear— Lola Blanchard ,1 Finn T. Agerkvist 2 1Bang & Olufsen ICEpower s/a, Lyngby, Denmark 2Technical University of Denmark, Lyngby, 18:00 Denmark P21-7 Generalized Multiple Sweep Measurement— The purpose of the study is to obtain a better Stefan Weinzieri, Andre Giese, Alexander understanding of concha headphones. Concha Lindau , TU Berlin, Berlin, Germany headphones are the small types of earpiece that are placed in the concha. They are not sealed to A system identification by impulse response the ear and therefore, there is a leak between measurements with multiple sound source the earpiece and the ear. This leak is the reason configurations can benefit greatly from time- why there is a significant lack of bass when efficient measurement procedures. An opti - using such headphones. This paper investigates mized method by interleaving and overlapping the coupling between the headphone and the of multiple exponential sweeps (MESM) used ear, by means of measurement in artificial ears as excitation signals was presented by Majdak and models. The influence of the back volume is et al. (2007). For single system identifications, taken into account. however, much higher signal-to-noise ratios Convention Paper 7769 (SNR) can be reached with sweeps whose Paper presented by Finn Agerkvist magnitude spectra are adapted to the back - ground noise spectrum of the acoustical environment, as proposed by Müller & 16:00 Massarani (2001). We investigated on which P22-3 Subjective Evaluation of Headphone Target conditions and to what extent the efficiency of Frequency Responses— Gaëtan Lorho , Nokia multiple sweep measurements can be Corporation, Finland increased by using arbitrary, spectrally adapt - ed sweeps. An extension of the MESM ap - The effect of headphone frequency response proach toward generalized sweep spectra is equalization on listeners’ preference was studied presented, along with a recommended mea - for music and speech reproduction. The high- surement procedure and a prediction of the quality circum-aural headphones selected for efficiency of multiple sweep measurements this listening experiment were first equalized to depending on typical measurement conditions. produce a flat frequency response. Then, a set Convention Paper 7767 of filters was created based on two parameters

34 Audio Engineering Society 126th Convention Program, 2009 Spring defining the amplitude and center frequency of motivates the design of high-order differential the main peak found around 3 kHz in the free- microphone arrays using the aforementioned field and diffuse-field equalization curves. Two spatio-temporal gradient analysis. different listening tests were carried out to evalu - Convention Paper 7772 ate these equalization candidates using a differ - ent methodology and a total of 80 listeners. The 17:30 results of this study indicate that a target fre - quency response with a 3 kHz peak of lower P22-6 The Analog Microphone Interface and its amplitude than in the diffuse-field response is pre - History— Jörg Wuttke , Jörg Wuttke Consultancy, ferred by listeners for both music and speech . Pfinztal, Germany, Schoeps GmbH, Karlsruhe, Convention Paper 7770 Germany The interface between microphones and micro - 16:30 phone inputs has special characteristics and requires special attention. The low output levels P22-4 Study and Consideration on Symmetrical of microphones and the possible need for long KEMAR HATS Conforming to IEC60959— cables have made it necessary to think about Kiyofumi Inanaga, 1 Homare Kon ,1 Gunnar noise and interference of all kinds. A microphone Rasmussen, 2 Per Rasmussen ,2 Yasuhiro Riko 3 input is also the electrical load for a microphone 1Sony Corporation, Tokyo, Japan and can have an adverse influence on the its 2G.R.A.S. Sound & Vibration A/S, Holte, Denmark performance. Condenser microphones contain 3Riko Associates, Yokohama, Japan active circuitry that required some form of pow - KEMAR is widely recognized as a leading model ering. With the introduction of transistorized cir - of head and torso simulators (HATS) for different cuitry in the 1960s, it became practical for this types of acoustic measurements meeting require - powering to be incorporated into microphone ments of a global industrial standard, ANSI inputs. Various methods appeared in the begin - S3.36/ASA58-1985 and IEC 60959:1990. One of ning; 48-Volt phantom powering is now the dom - the KEMAR HATS pinna models has a reputation inant standard, but this standard method is still for good reproducibility of measured results in ex - not always implemented correctly. amining headphones and earphones. However, it Convention Paper 7773 requires free field compensation in order to con - duct the measurements; thus, the head-related transfer function (HRTF) of HATS fitted with the 18:00 pinna model must be corrected. Because head - P22-7 Handling Noise Analysis in Large Cavity phones and earphones are usually designed Microphone Windshields—Improved Solution symmetrically, we developed a prototype of Sym - —Philippe Chenevez , CINELA, Paris, France metrical KEMAR HATS based on the original KE - MAR mounted with the pinna model with good Pressure gradient microphones are well known reproducibility. We measured and evaluated a to be highly sensitive to vibrations. Respectable set of HRTFs from the sound source to both suspensions are made to create the best isola - ears. Our study concluded that the HATS we tion possible, but when the microphone is placed developed carries symmetrical characteristics inside a large cavity windshield, the external skin and is also suitable to be utilized as a tool to behaves as a drum excited by the vibrations of measure the qualities of variety of acoustic the support (boom or stand). As a consequence devices along with the conventional KEMAR and structure-borne noise is also transmitted acousti - it can serve as a new common platform for differ - cally to the microphone, due to its hard proximity ent types of electroacoustic measurements. effect. Some theoretical aspects and practical Convention Paper 7771 measurements are presented, in conjunction with a proposed improved solution. 17:00 Convention Paper 7774 P22-5 Spatio-Temporal Gradient Analysis of Differential Microphone Arrays— Mihailo 1 1 1,2 Kolundzija, Christof Faller , Martin Vetterli Live Sound Seminar 2 Saturday, May 9 1 Ecole Polytechnique Fédérale de Lausanne, 15:00 – 18:30 Atrium 2 Lausanne, Switzerland 2University of California at Berkeley, Berkeley, CA, USA SENNHEISER & YAMAHA—LIVE SOUND WORKSHOP The literature on gradient and differential micro - phone arrays makes a distinction between the Presenters: Svenja Dunkel two, and nevertheless shows how both types can Wayne "Heights" Gittens be used to obtain the same response. A more Oliver Voges theoretically sound rationale for using delays in Jürgen Wilhelm differential microphone arrays has not yet been Gregor Zielinsky given. This paper presents a gradient analysis of the sound field viewed as a spatio-temporal phe - The workshop will cover two important aspects of PA: nomenon, and gives a theoretical interpretation of (1) Digital mixing consoles for PA and IT basic require - the working principles of gradient and differential ments, e.g., switches, cables, fiber optics, connectors. microphone arrays. It shows that both types of (2) A complete soundcheck of a live rockband on microphone arrays can be viewed as devices for stage. The visitors will experience how high-end profes - approximately measuring the spatio-temporal sionals do their soundcheck, including FOH, monitors, ¯ derivatives of the sound field. Furthermore, it also and wireless techniques.

Audio Engineering Society 126th Convention Program, 2009 Spring 35 Technical Progra m The special task will be to get the tricky sound of the and professionals alike, are welcome to come talk with location under control. The hall Atrium 2 is very reverber - representatives from the companies and find out more ant and is not an optimized condition. This is what pros about job and internship opportunities in the audio indus - have to work with every day. The band "Rauschenberg - try. Bring your resume! er" is a new upcoming group from Hannover around singer and leader Rauschenberger, who has a splendid Workshop 17 Saturday, May 9 and very characteristic voice. After the soundcheck, 16:00 – 18:30 Room K1 there will be a half-hour concert. SOUND DESIGN FOR DIE FALSCHER Saturday, May 9 15:00 Room D120 Technical Committee Meeting on Electro Magnetic Presenters: Tobi Fleig , Dubbing Mixer Compatibility Rainer Heesch , Sound Designer Tatjana Jakob , Sound Designer Saturday, May 9 15:00 Room D104 Olaf Mierau , Postproduction Sound Standards Committee Meeting on SC-03-06 Digital Supervisor Library and Archive Systems Die Fälscher (The Counterfeiters ) was awarded the Oscar in 2008 for the best non-English-speaking film. Saturday, May 9 15:00 Room D105 Key members of the sound team will discuss the con - Standards Committee Meeting on SC-04-01 cepts of sound designing this movie, communication with Acoustics and Sound Source Modeling the director, freedom and constraints, sound as a strong aesthetic means for the narrative, and other issues. Sev - Exhibitor Seminar Saturday, May 9 eral examples from the film will be shown, sometimes in 15:00 – 16:00 Exhibit Floor Booth 2303 different versions and stages of the mix. ES14 OPTOCORE –A Saturday, May 9 16:00 Room D120 TECHNICAL INTRODUCTION TO OPTOCORE Technical Committee Meeting on Transmission Presenters: Martin Barbour , OPTOCORE Support and Broadcasting Engineer Andreas Kaspar , OPTOCORE Support Exhibitor Seminar Saturday, May 9 Engineer 16:00 – 17:00 Exhibit Floor Booth 2303 Thorsten Schulze , OPTOCORE Support ES15 OPTOCORE –B and Product Manager BASICS OF FIBRE OPTICS The OPTOCORE Synchronous Fibre Network is intro - duced. Content will include an Introduction/Overview, Presenters: Martin Barbour , OPTOCORE Support design of the network, synchronicity, latency, topology, Engineer redundancy, protocol, and devices. Andreas Kaspar , OPTOCORE Support Engineer Thorsten Schulze , OPTOCORE Support Workshop 16 Saturday, May 9 and Product Manager 15:30 – 18:30 Room K2 Technical, economical, safety aspects, and advantages LOUDSPEAKER FE/BE MODELING of fibre optics is the topic of this seminar. Content will include fibre advantages, reliability, types of fibre, para - Chair: Steve Hutt , Equity Sound Investments meters of fibres (attenuation, bandwidth length product, latency), connector types, cleaning, and testing. Panelists: Wolfgang Klippel , Klippel GmbH Hermann Landes , SIMetris GmbH Session P23 Saturday, May 9 Rich Little , Tymphany Inc. 16:30 – 18:00 K4 Foyer Peter Larsen , Loudsoft Ltd . Patrick Macey , PACSYS Limited POSTERS: PSYCHOACOUSTICS AND PERCEPTION Joerg Panzer , R and D Team GmbH Pierre Thieffry , ANSYS Inc. 16:30 This workshop will explore FE/BE modeling of loud - P23-1 Influence of the Listening Room in the speaker drivers. A case study of an existing loudspeaker Perception of a Musical Work— Nelia driver will be modeled by each panelist to benchmark the Valverde, Marcos D. Fernández, José Antonio capabilities of their modeling software package. A loud - Ballesteros, Leticia Martínez, Samuel Quintana, speaker driver's dimensions and material properties will Isabel González , Escuela Universitaria be provided to panelists in advance of the convention so Politénica de Cuenca, Cuenca, Spain that they may develop a thorough model for presentation at the workshop. Results will be discussed along with The listening of the same musical composition gen - measurement analysis of the real loudspeaker driver. erates a unique perception for every listener but, simultaneously, the specific acoustic conditions of Student Event/Career Development the chosen room have a decisive influence on the CAREER/JOB FAIR perception. In order to evaluate such differences Saturday, May 9, 15:30 – 17:00 depending on the listening room, a musical work K4 Foyer for choir has been composed and recorded with a HATS in an anechoic room, in a reverberant room, The Career Fair will feature several companies from the and in a normal room. With those records, surveys exhibit floor. All attendees of the Convention, students to professional musicians and non-expert listeners 36 Audio Engineering Society 126th Convention Program, 2009 Spring have been carried out, once they have previously of listener responses to paired comparisons of heard the recording with headphones, and finally, stimuli generated by the morpher, showed the answers obtained have been evaluated in order movement in three perceptually-orthogonal to determine the influence of the listening room in directions. These directions were labeled in a the perception of the musical work. subsequent verbal elicitation experiment that Convention Paper 7775 found that the effects of the brightness and soft - ness controls were perceived as intended. A 16:30 Timbre Morpher, adjusting additional timbral attributes with perceptually-meaningful controls, P23-2 Comparison of Methods for Measuring can now be considered for further work. Sound Quality through HATS and Binaural Convention Paper 7778 Microphones— José Antonio Ballesteros, Marcos D. Fernández, Samuel Quintana, Isabel González, Laura Rodríguez , Escuela 16:30 Universitaria Politécnica de Cuenca, Universidad P23-5 Resolution of Spatial Distribution Perception de Castilla-La Mancha, Cuenca, Spain with Distributed Sound Source in Anechoic Sound quality techniques are currently becoming Conditions — Olli Santala, Ville Pulkki , Helsinki more important as they take into account the University of Technology, Espoo, Finland human perception of sound. By now, there is no The resolution of directional perception of spatially well established international standards for mea - distributed sound sources was investigated with a suring sound quality and no well recognizable ref - listening test in an anechoic chamber using vari - erence index for its assessment. Then, a HATS or ous sound source distributions. Fifteen loudspeak - a pair of binaural microphones can be used for ers were used to produce test cases that included measuring the typical sound quality parameters. A sound sources with varying widths and wide sound set of measurements, under the same condition, sources with gaps in the distribution. The subjects has been carried out using both devices for were asked to distinguish which loudspeakers assessing the differences and the possible varia - emitted sound according to their own perception. tion in the results. As a consequence of all of this, Results show that small gaps in the sound source guidance is given for choosing the device that best were not perceived accurately and wide sound fits depending on each measurement context. sources were perceived narrower than they actual - Convention Paper 7776 ly were. The results also indicate that the resolu - tion for fine spatial details was worse than 15 16:30 degrees when the sound source was wide. Convention Paper 7779 P23-3 Improving Perceived Tempo Estimation by Statistical Modeling of Higher-Level Musical Descriptors— Ching-Wei Chen, Markus Cremer, 16:30 Kyogu Lee, Peter DiMaria, Ho-Hsiang Wu , Gracenote, Inc., Emeryville, CA, USA P23-6 Perceived Roughness—A Recent Psychoacoustic Measurement— Robert Conventional tempo estimation algorithms gen - Mores, Thorsten Smit, Jana-Marie Wiese , erally work by detecting significant audio events University of Applied Science, Hamburg, and finding periodicities of repetitive patterns in Germany an audio signal. However, human perception of tempo is subjective and relies on a far richer set This paper relates to an investigation on per - of information, causing many tempo estimation ceived roughness from Aures in 1984 where algorithms to suffer from octave errors, or “dou - findings are based on psychoacoustic tests with ble/half-time” confusion. In this paper we pro - synthetic sounds and a small group of people. pose a system that uses higher-level musical The related results have repeatedly been used descriptors such as mood to train a statistical for modeling roughness perception since then, model of perceived tempo classes, which can for instance in the context of noise perception. then be used to correct the estimate from a con - Roughness is again an issue when investigating ventional tempo estimation algorithm. Our exper - the perceived quality or timbre of musical imental results show reliable classification of sounds. In this context roughness is one among perceived tempo class, as well as a significant some ten mid-level features to be extracted. reduction of octave errors when applied to an Here, perceived roughness is measured again, array of available tempo estimation algorithms. but on a wider basis than in the earlier investiga - Convention Paper 7777 tion. This paper outlines the psychoacoustic investigation, basically following the method of Aures, but modifying some of the issues under 16:30 question. The results are reasonable and differ P23-4 Perceptually-Motivated Audio Morphing: from the earlier findings in various aspects. Softness— Duncan Williams, Tim Brookes , Convention Paper 7780 University of Surrey, Guildford, Surrey, UK A system for morphing the softness and bright - 16:30 ness of two sounds independently from their oth - P23-7 A Physiological Auditory Model— Václav er perceptual or acoustic attributes was coded. Vencovsky , Czech Technical University in The system is an extension of a previous one Prague, Prague, Czech Republic that morphed brightness only, that was based on the Spectral Modeling Synthesis additive/resid - A physiological auditory model is described. The ual model. A Multidimensional Scaling analysis, model simulates a processing of a sound by an ¯ Audio Engineering Society 126th Convention Program, 2009 Spring 37 Technical Progra m outer, middle, and inner ear. A nonlinear inner ear Both dinner and party will take place at the saf at Zerwirk , model comprises the cochlear frequency selectivi - which is restaurant in downtown Munich. The dinner is ty model and the inner hair cells model proposed scheduled for 8 pm, the party will start at 10 pm. The according to mammalian physiological data. A restaurant is not open to the public on that evening, so it capability of the auditory model to simulate is a private event. The restaurant is located in the Leder - human psychophysical masking data is verified. erstrasse only a two minute walk from Marienplatz, Mu - Convention Paper 7781 nich’s central square and an important transportation hub. The food: Saf is a vegan restaurant. There will either be Saturday, May 9 17:00 Room D104 a buffet or a menu for approximately EUR 20 apiece. Standards Committee Meeting on SC-03-07 Audio There will be fifty tickets for the dinner that will be sold Metadata after SDA-1 and at the Student Science Spot.

Saturday, May 9 17:00 Room D105 The party: After the dinner there will be the chance to get Standards Committee Meeting on SC-03-12 Forensic together and have a good time at the Student Party. The Audio Graz student section will take care of the music. There won't be an entrance fee for the party, but of course Exhibitor Seminar Saturday, May 9 everybody will have to pay for their own drinks. 17:00 – 18:00 Exhibit Floor Booth 2303 ES16 OPTOCORE –C Special Event ORGAN RECITAL BY GRAHAM BLYTH OPTOCORE IN PRACTICE Saturday, May 9, 20:30 – 21:30 Cathedral Church of Our Lady (also called the Presenters: Martin Barbour , OPTOCORE Support Liebfrauendom) Engineer Frauenplatz 1, Munich Andreas Kaspar , OPTOCORE Support Engineer Graham Blyth’s traditional organ recital will be given on Thorsten Schulze , OPTOCORE Support the organ of the Münchner Dom, the “Frauenkirche.” The and Product Manager 99 and 100 meter high respectively Twin Towers of the Frauenkirche with the “Welch” caps are high up in the This seminar will cover practical exercises with OPTO - sky of Munich. This is the oldest landmark of the city. CORE system/devices. Content will include practical set - Built between 1468 and 1488 by Jörg von Halback (nick - up of a network, remote pre-amp control through third named “Ganghofer”), this cathedral is the biggest late- party devices, transport of control data and Ethernet, gothic hall church of Europe. Here you will find the software control, and firmware upgrades. splendid tomb of the German Kaiser (emperor) Ludwig the Bavarian. Student Event/Career Development The featured works are by Bach, plus Mendelssohn’s STUDENT DELEGATE ASSEMBLY MEETING— 1st Sonata, and Vierne’s 1st Symphony. PART 2 Graham Blyth was born in 1948, began playing the Saturday, May 9, 17:30 – 18:30 piano aged 4 and received his early musical training as a Room D123 Junior Exhibitioner at Trinity College of Music in London, At this meeting the SDA will elect a new vice chair. One England. Subsequently, at Bristol University, he took up vote will be cast by the designated representative from each conducting, performing Bach’s St. Matthew Passion recognized AES student section in the European and Inter - before he was 21. He holds diplomas in Organ Perfor - national Regions. Judges’ comments and awards will be mance from the Royal College of Organists, The Royal presented for the Recording Competitions. Plans for future College of Music and Trinity College of Music. In the late student activities at local, regional, and international levels 1980s he renewed his studies with Sulemita Aronowsky will be summarized. for piano and Robert Munns for organ. He gives numer - ous concerts each year, principally as organist and pianist, but also as a conductor and harpsichord player. Special Event He made his international debut with an organ recital at LIVE CONCERT St. Thomas Church, New York in 1993 and since then Saturday, May 9, 18:30 – 19:00 has played in San Francisco (Grace Cathedral), Los Atrium 2 Angeles (Cathedral of Our Lady of Los Angeles), Ams - The band featured in the Live Sound Workshop LS2, terdam, Copenhagen, Munich (Liebfrauendom), Paris Rauschenberger, will continue to play after the workshop (Madeleine and St. Etienne du Mont) and Berlin. He has finishes, in a concert open to all attendees. lived in Wantage, Oxfordshire, since 1984 where he is currently Artistic Director of Wantage Chamber Concerts Special Event and Director of the Wantage Festival of Arts. MIXER PARTY He divides his time between being a designer of pro - Saturday, May 9, 18:30 – 19:30 fessional audio equipment (he is a co-founder and Tech - The Bistro (between Atriums 3 and 4) nical Director of Soundcraft) and organ related activities. In 2006 he was elected a Fellow of the Royal Society of A Mixer Party will be held on Saturday evening to enable Arts in recognition of his work in product design relating Convention attendees to meet in a social atmosphere to to the performing arts. catch up with friends and colleagues from the world of audio. There will be a cash bar. Session P24 Sunday, May 10 09:00 – 12:30 Room K3 Student Event/Career Development STUDENT PARTY ASSESSMENT AND EVALUATION Saturday, May 7, 20:00 –22:00 saf at Zerwirk Chair: Gaëtan Lorho , Nokia Corporation, Finland

38 Audio Engineering Society 126th Convention Program, 2009 Spring 09:00 ditions are usually included in subjective tests. In P24-1 Influence of Level Setting on Loudspeaker the field of quality of transmitted speech, refer - Preference Ratings— Vincent Koehl, Mathieu ence conditions correspond to a speech sample Paquier , Université de Brest, Plouzané, France impaired by a known amount of degradation. In this paper several kinds of reference conditions The perceived audio quality of a sound-repro - and the process used for their production are duction device such as a loudspeaker is hard to presented. Examples of the corresponding nor - evaluate. Industrial and academic researchers malization procedure of each kind of reference are still focusing on the design of reliable are given. assessment procedures to measure this subjec - Convention Paper 7784 tive character. One of the main issues of listen - ing tests is about their validity in regard to real 10:30 comparison situations (Hi-Fi magazine evalua - tions, audiophile, sound engineer, customer, P24-4 The Influence of Sound Processing on etc.). Are the conclusions of laboratory tests Listeners’ Program Choice in Radio consistent with these almost informal compar - Broadcasting— Hans-Joachim Maempel, isons? As an example, one of the main differ - Fabian Gawlik , Technische Universität Berlin, ences between listening tests and real-life Berlin, Germany comparisons is about the loudness matching. This paper is aimed at comparing paired-com - Many opinions on broadcast sound processing parison tests that are commonly accomplished are founded on tacit assumptions about certain under laboratory conditions with a procedure effects on listeners. These, however, have lacked assumed to be closer to real-life conditions. It support by internally and ecologically valid empir - shows that differences in the test procedures led ical data so far. Thus, under largely realistic con - to differences in the subjective assessments. ditions it has been experimentally investigated to Convention Paper 7782 what extent broadcast sound processing influ - ences listeners’ program choice. Technical fea - tures of stimuli, socio-demographic data of the 09:30 test persons, and data of listening conditions P24-2 Comparing Three Methods for Sound Quality have been additionally collected. In the main Evaluation with Respect to Speed and experiment, subjects were asked to choose one Accuracy— Florian Wickelmaier ,1 Nora out of six audio stimuli varied in content and Umbach ,1 Konstantin Sering ,1 Sylvain Choisel 2 sound processing. The varied sound processing 1University of Tübingen, Tübingen, Germany caused marginal and statistically not significant 2Bang & Olufsen A/S, Struer, Denmark, differences in frequencies of program choice. By now at Philips Consumer Lifestyle, Leuven, Belgium contrast, a subsequent experiment enabling a direct comparison of different sound processings The goal of the present study was to compare of the same audio content yielded distinct prefer - three response-collection methods that may be ences for certain sound processings. used in sound quality evaluation. To this end, 52 Convention Paper 7785 listeners took part in an experiment where they assessed the audio quality of musical excerpts 11:00 and six processed versions thereof. For different types of program material, participants per - P24-5 Free Choice Profiling and Natural Grouping formed (a) a direct ranking of the seven sound as Methods for the Assessment of Emotions samples, (b) pairwise comparisons, and (c) a in Musical Audio Signals— Sebastian novel procedure, called ranking by elimination. Schneider ,1 Florian Raschke ,1 Gabriel The latter requires subjects on each trial to elimi - Gatzsche ,2 Dominik Strohmeier 1 nate the least preferred sound; the elimination 1Ilmenau University of Technology, Ilmenau, continues until only the sample with the highest Germany audio quality is left. The methods are compared 2Fraunhofer Institute for Digital Media with respect to the resulting ranking/scaling and Technology, IDMT, Ilmenau, Germany; the time required to obtain the results. Convention Paper 7783 To measure the perceived emotions caused by musical audio signals we propose to use “Free Choice Profiling” (FCP) combined with “Natural 10:00 Grouping” (NG). FCP/NG—originally derived P24-3 Reference Units for the Comparison of from food research and new to the research of Speech Quality Test Results— Nicolas Côté ,1,2 music perception—allow participants to evaluate Vincent Koehl ,3 Valérie Gautier-Turbin ,4 stimuli using their own vocabulary. To evaluate Alexander Raake, 2 Sebastian Möller 2 the proposed methods, we conducted an 1Ecole Nationale d’Ingénieurs de Brest, experiment where 16 participants had to assess Plouzané, France major-major and minor-minor chord pairs. Unlike 2Deutsche Telekom Laboratories, Berlin, Germany one could expect, allowing participants to 3Université de Brest, Plouzané, France express themselves freely does not lead to a 4France Telecom R&D, Lannion, France degeneration of the quality of the data. Instead, clearly interpretable results consistent with Subjective tests are carried out to assess the music theory and emotional psychology were quality of an entity as perceived by a user. How - obtained. These results encourage further inves - ever, several characteristics inherent to the sub - tigations, which could lead to a general method ject or to the test methodology might influence for assessing emotions in music. the users’ judgments. As a result, reference con - Convention Paper 7786 ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 39 Technical Progra m 11:30 enhanced dynamic range and lower crest factor with no unwanted artifacts, compared to the tra - P24-6 Subjective Quality Evaluation of Audio ditionally mixed signals. The overall headroom is Streaming Applications on Absolute and improved, while clarity and tonal balance are Paired Rating Scales— Bernhard Feiten, preserved. Alexander Raake, Marie-Neige Garcia, Ulf Convention Paper 7789 Wüstenhagen, Jens Kroll , Deutsche Telekom Laboratories, Berlin, Germany 09:30 In the context of the development of a paramet - ric model for the quality assessment of audiovi - P25-2 Source-Filter Modeling in Sinusoid Domain— sual IP-based multimedia applications, audio Wen Xue, Mark Sandler , Queen Mary, University tests have been carried out. The test method of London, London, UK used for the subjective audio tests was aligned This paper presents the theory and implementa - to the method used for video tests. Hence, the tion of source-filter modeling in sinusoid domain Absolute Category Ranking (ACR) method was and its applications on timbre processing. The applied. To prove the usability of ACR tests for technique decomposes the instantaneous ampli - this purpose MUSHRA and ACR were applied in tude in a sinusoid model into a source part and a parallel listening tests. The MPEG audio codecs filter part, each capturing a different aspect of the AAC, HE-AAC, MP2, and MP3 at different timbral property. We show that the sinusoid bitrates and different packet loss conditions were domain source-filter modeling is approximately evaluated. The test results show that the ACR equivalent to its time or frequency domain counter - method also reveals the quality differences for parts. Two methods are proposed for the evalua - higher qualities, even though MUSHRA has tion of the source and filter, including a superior resolution. least-square method based on the assumption of Convention Paper 7787 slow variation of source and filter in time, and a filter bank method that models the global spectral 12:00 envelope in the filter. Tests show the effectiveness of the algorithms for isolation frequency-driven P24-7 Assessor Selection Process for Multisensory amplitude variations. Example applications are Applications— Søren Vase Legarth, Nick given to demonstrate the use of the technique for Zacharov , DELTA SenseLab, Hørsholm, Denmark timbre processing. Assessor panels are used to perform perceptual Convention Paper 7790 evaluation tasks in the form of listening and viewing tests. In order to ensure the quality of 10:00 collected data it is vital that the selected asses - sors have the desired qualities in terms of dis - P25-3 Analysis of a Modified Boss DS-1 Distortion crimination aptitude as well as consistent rating Pedal — Matthew Schneiderman, Mark Sarisky , ability. This work extends existing procedures in University of Texas at Austin, Austin, TX, USA this field to provide a statistically robust and eff - cient manner for assessing and evaluating the Guitar players are increasingly modifying (or paying performance of assessors for listening and view - someone else to modify) inexpensive mass-pro - ing tasks. duced guitar pedals into boutique units. The Keeley Convention Paper 7788 modification of the Boss DS-1 is a prime example. In this paper we compare the measured and perceived performance of a Boss DS-1 before and after applying the Keeley All-Seeing-Eye and Ultra Session P25 Sunday, May 10 mods. This paper sheds light on psychoacoustics, 09:00 – 12:30 Room K4 signal processing, and guitar recording techniques in relation to low fidelity guitar distortion pedals. SOUND DESIGN AND PROCESSING Convention Paper 7791

Chair: Michael Hlatky , Hochschule Bremen 10:30 (University of Applied Sciences), Bremen, Germany P25-4 Phase and Amplitude Distortion Methods for Digital Synthesis of Classic Analog Waveforms— Joseph Timoney, 1 Victor 09:00 Lazzarini, 1 Brian Carty ,1 Jussi Pekonen 2 P25-1 Hierarchical Perceptual Mixing— Alexandros 1NUI Maynooth, Maynooth, Ireland Tsilfidis, Charalambos Papadakos, John 2Helsinki University of Technology, Espoo, Mourjopoulos , University of Patras, Patras, Finland Greece An essential component of digital emulations of A novel technique of perceptually-motivated sig - subtractive synthesizer systems are the algo - nal-dependent audio mixing is presented. The rithms used to generate the classic oscillator proposed Hierarchical Perceptual Mixing (HPM) waveforms of sawtooth, square, and triangle method is implemented in the spectro-temporal waves. Not only should these be perceived to be domain; its principle is to combine only the authentic sonically, but they should also exhibit perceptually relevant components of the audio minimal aliasing distortions and be computational - signals, derived after the calculation of the mini - ly efficient to implement. This paper examines a mum masking threshold, which is introduced in set of novel techniques for the production of the the mixing stage. Objective measures are pre - classic oscillator waveforms of analog subtractive sented indicating that the resulting signals have synthesis that are derived from using amplitude or 40 Audio Engineering Society 126th Convention Program, 2009 Spring phase distortion of a mono-component input designer in music information retrieval (MIR) and waveform. Expressions for the outputs of these sound categorization applications. Initial testing distortion methods are given that allow parameter regarding the grouping of individual sounds into control to ensure proper bandlimited behavior. groups based on similarity has shown that Additionally, their implementation is demonstrably participants tended to group certain sounds efficient. Last, the results presented illustrate their together, often reflecting the groupings our team equivalence to their original analog counterparts. constructed. Convention Paper 7792 Convention Paper 7795

11:00 Workshop 18 Sunday, May 10 P25-5 Soundscape Attribute Identification— Martin 09:00 – 11:30 Room K1 Ljungdahl Eriksson, Jan Berg , Luleå University of Technology, Luleå, Sweden BLU-RAY—THE NEXT (AND ONLY?) CHANCE FOR HIGH RESOLUTION MUSIC MEDIA In soundscape research, the field’s methods can be employed in combination with approaches Presenters: Johannes Müller , msm mastering involving sound quality attributes in order to cre - Munich ate a deeper understanding of sound images Ronald Prent , Freelance Engineer, Galaxy and soundscapes and how these may be Studios described and designed. The integration of four methods are outlined, two from the soundscape The concept of utilizing Blu-ray as a pure audio format domain and two from the sound engineering will be explained, and Blu-ray will be positioned as suc - domain. cessor to both SACD and DVD-A. The operational func - Convention Paper 7793 tionality and a double concept of making it usable both with and without screen will be demonstrated by showing a few products that are already on the market. 11:30

P25-6 SonoSketch: Querying Sound Effect Workshop 19 Sunday, May 10 Databases through Painting— Michael 09:00 – 11:30 Room K2 Battermann, 1 Sebastian Heise, 1 Jörn 2 Loviscach AUDIO OVER IP 1Hochschule Bremen (University of Applied Sciences), Bremen, Germany Chair Heinz-Peter Reykers , WDR, Germany 2Fachhochschule Bielefeld (University of Applied Sciences), Bielefeld, Germany Panelists: Joost Bloemen , Technica Del Arte BV Jeremy R. Cooperstock , McGill University, Numerous techniques support finding sounds Montreal, Canada that are acoustically similar to a given one. It is Gerald List , TransTel Communications GmbH hard, however, to find a sound to start the similar - Peter Stevens , BBC R&D ity search with. Inspired by systems for image search that allow drawing the shape to be found, In radio broadcasting most transmissions and communi - we address quick input for audio retrieval. In our cation between remote broadcast venues and the radio system, the user literally sketches a sound house are based on ISDN lines. Within TV the communi - effect, placing curved strokes on a canvas. Each cation lines are also based on ISDN circuits. But the of these represents one sound from a collection technology of ISDN networks is gradually being of basic sounds. The audio feedback is interac - exchanged by IP networks and packet- based transmis - tive, as is the continuous update of the list of sion. What are the consequences of this migration for the retrieval results. The retrieval is based on broadcasters? In this workshop the focus will be on prac - symbol sequences formed from MFCC data com - tical demonstrations. The topics range from SIP-Servers pared with the help of a neural net using an edit - and infrastructure issues to audio examples regarding ing distance to allow small temporal changes. latency and other related parameters. Quality of Service Convention Paper 7794 and current problems as well as possible solutions will be examined. 12:00 P25-7 Generic Sound Effects to Aid in Audio 1 1 Sunday, May 10 09:00 Room D104 Retrieval— David Black, Sebastian Heise, Standards Committee Meeting: AESSC Plenary Jörn Loviscach 2 1Hochschule Bremen (University of Applied Sciences), Bremen, Germany Session P26 Sunday, May 10 2Fachhochschule Bielefeld (University of Applied 10:30 – 12:00 K4 Foyer Sciences), Bielefeld, Germany POSTERS: ROOM ACOUSTICS Sound design applications are often hampered AND LOUDSPEAKER INTERACTION because the sound engineer must either pro - duce new sounds using physical objects, or search through a database of sounds to find a 10:30 suitable sample. We created a set of basic P26-1 Acoustic Design of Classrooms— Suthikshn sounds to mimic these physical sound-producing Kumar , PESIT, Bangalore, India objects, leveraging the mind's onomatopoetic clustering capabilities. These sounds, grouped Acoustic principles when used effectively in into onomatopoetic categories, aid the sound classroom design can improve the audibility of ¯ Audio Engineering Society 126th Convention Program, 2009 Spring 41 Technical Progra m the professor in dramatic way. The cost-effective Increasing demands on flexibility, scalability, and effi - way to enhance the acoustics serves several ciency in sound reinforcement applications encouraged purposes: Less speaking effort on the part of the one loudspeaker development configurable for point and lecturer; students can easily hear the lecturer line source applications by an easy mechanical modifica - more clearly; improved communication and, tion. Several implemented design technologies will be hence, improved learning experience. Several discussed before the listening demonstration of the per - improvements can be done to the classroom formance of the system under critical acoustic conditions architecture to enhance the signal-to-noise ratio, comparing Q-Series line arrays. reduce reverberation, and background noise. We propose an innovative way of providing par - Yamaha abolic reflectors near the platform for amplifying Modern IT-Compatible Audio Networks the lecturer’s voice. This paper focuses on the cost-effective, energy efficient acoustic design of IT-compatible audio networks and standards: Cobranet classrooms. and Ethersound. Both formats will be discussed regard - Convention Paper 7796 ing their advantages and limitations. Advanced network strategies like VLAN programming offers high channel 10:30 counts and a wide range of additional services via Giga - bit audio networks. Now video, intercom, remote control, P26-2 Epidaurus: Comments on the Acoustics of and DMX services may be included in a modern network the Legendary Ancient Greek Theater— infrastructure. Of course, such a network has to be as Christos Goussios, Christos Sevastiadis, Kalliopi safe and stable as possible, so the redundancy concepts Chourmouziadou, George Kalliris , Aristotle developed by the IT industry like link aggregation / trunk - University of Thessaloniki, Thessaloniki, Greece ing / spanning tree will be discussed. The ancient Greek theaters and especially the well preserved theater of Epidaurus are of great Exhibitor Seminar Sunday, May 10 interest because of their legendary acoustic 11:00 – 12:00 Exhibit Floor Booth 2303 characteristics. In the present paper the history ES17 OPTOCORE –A and the construction characteristics of the spe - cific theater are presented. The differences TECHNICAL INTRODUCTION TO OPTOCORE between the ancient and modern use of it are explained. Important acoustic parameters calcu - Presenters: Martin Barbour , OPTOCORE Support lated using in situ measurements are presented. Engineer The conclusions show the relation between its Andreas Kaspar , OPTOCORE Support excellent acoustic performance and the obtained Engineer results. Thorsten Schulze , OPTOCORE Support Convention Paper 7797 and Product Manager The OPTOCORE Synchronous Fibre Network is intro - 10:30 duced. Content will include an Introduction/Overview, design of the network, synchronicity, latency, topology, P26-3 A Matlab Toolbox for the Analysis of Ando’s redundancy, protocol, and devices. Factors— Dario D'Orazio, Paolo Guidorzi, Mas - simo Garai , University of Bologna, Bologna, Italy Workshop 20 Sunday, May 10 The autocorrelation and crosscorrelation 12:00 – 14:00 Room K2 functions analysis, as well-known in litera - ture, obtains remarkable results in different scientific fields. The autocorrelation function STANDARDS-BASED AUDIO NETWORKS (ACF) and the interaural crosscorrelation USING IEEE 802.1 AVB function (IACF) analysis in architectural acoustics is known thanks to Y. Ando's work. Presenters : Ed Clarke , XMOS Semiconductor, The Toolbox presented in this work has been Bristol, UK developed in order to compute Ando's signifi - Rick Kreifeldt , Harman International, UT, cant and spatial factors (as the factors obtained USA from ACF and IACF are called), to subjective Recent work by IEEE 802 working groups will allow preference functions and to investigate furthe r vendors to build a standards-based network with the applications. appropriate quality of service for high quality audio per - Convention Paper 7798 formance and production. This new set of standards, de - veloped by the IEEE 802.1 Audio Video Bridging Task Group, provides three major enhancements for 802- Live Sound Seminar 3 Sunday, May 10 based networks: (1) Precise timing to support low-jitter 10:30 – 12:30 Atrium 2 media clocks and accurate synchronization of multiple streams; (2) A simple reservation protocol that allows an YAMAHA AND D & B endpoint device to notify the various network elements in a path so that they can reserve the resources necessary to support a particular stream; and (3) Queuing and for - Presenters: Stefan Goertz warding rules that ensure that such a stream will pass Arthur Koll through the network within the delay specified by the d & b reservation. These enhancements require no changes to the Ethernet lower layers and are compatible with all the Line Source to Point Source Transformation other functions of a standard Ethernet switch (a device —Technical Concept of the d & b T-Series that follows the IEEE 802.1Q bridge specification). As a

42 Audio Engineering Society 122nd Convention Program, 2007 Spring result, all of the rest of the Ethernet ecosystem is avail - 13:00 able to developers, in particular, the various high speed physical layers (up to 10 gigabit/sec in current standards, P27-1 On the Myth of Pulse Width Modulated Spectrum in Theory and Practice— Arnold even higher speeds are in development), security fea - 1,2 2 tures (encryption and authorization), and advanced man - Knott , Gerhard Pfaffinger , Michael A. E. Andersen 1 agement (remote testing and configuration) features can 1 be used. This workshop will outline the basic protocols Technical University of Denmark, Lyngby, Denmark and capabilities of AVB networks, describe how such a 2 network can be used, and provide some simple demon - Harman/Becker Automotive Systems GmbH, strations of network operation (including a live compari - Straubing, Germany son with a legacy Ethernet network). Switch-mode audio power amplifiers are common - ly used in sound reproduction. Their well-known Workshop 21 Sunday, May 10 drawback is the radiation of high frequent energy, 12:00 – 14:00 Room K1 which can disturb radio and TV receivers. The designer of switch-mode audio equipment there - fore need to make arrangements to prevent this MULTIMODAL INTEGRATION: coupling, which would otherwise result in bad AUDIO-VISUAL-HAPTIC-TACTILE audio performance. A deep understanding of the pulse width modulated (PWM) signal is therefore Cochairs: Durand Begault , NASA Ames Research essential, which resulted in different mythic models Center as pulse, trapezoidal, or Double Fourier Series Ellen Haas , US Army Research Lab, MD, (DFS) representations in the past. This paper will USA clarify these theoretical approaches by comparing them with reality from both the time and the Panelists: Ercan Altinsoy , Technical University of frequency domain perspective. For validation a Dresden, Dresden, Germany switch-mode audio power amplifier was built, deliv - Jeremy Cooperstock , McGill University, ering the contents material with less than 0.06 Montreal, Quebec, Canada percent distortion across the audio band at 50 W. Jürgen Hellbrück , KU Eichstätt, Eichstätt, The switch-mode signals have been evaluated Germany very precisely in time and spectral domain to Gerhard Mauter , Audi AG, Germany enlighten the assumptions about the PWM spectra Alexander S. Treiber , ACC Akustik, Germany and decrypt this myth. Convention Paper 7799 To design the best human interfaces for complex sys - tems such as automobiles, mixing consoles, shared real - ity spaces, games, military, and aerospace, an increas - 13:30 ing number of designers, psychologists, and human P27-2 Design Approaches for Psychoacoustical factors specialists are reconsidering multimodal displays In-Band Noise Shaping Filters— Jochen Hahn , of information that includes haptic and tactile feedback in University of Kaiserslautern, Kaiserslautern, addition to auditory-visual communication. This work - Germany shop considers in particular the way in which various modalities of information can be best integrated for a Noise shaping is a state-of-the-art technique to particular application, and identifies future requirements preserve the perceived quality of audio signals for such research. when requantization happens. Noise shaping fil - ters are special filters because of the nonlinear characteristics in hearing. They have to be taken Exhibitor Seminar Sunday, May 10 into account when designing these special digital 12:00 – 13:00 Exhibit Floor Booth 2303 audio filters. The design approaches presented ES18 OPTOCORE –B in this contribution meet these requirements. They minimize or limit the filter magnitude, the BASICS OF FIBRE OPTICS unweighted noise amplification, and the group delay characteristics of the filter. Presenters: Martin Barbour , OPTOCORE Support Convention Paper 7800 Engineer Andreas Kaspar , OPTOCORE Support 14:00 Engineer Thorsten Schulze , OPTOCORE Support P27-3 A New Analog Input Topology for Extreme and Product Manager Dynamic Range Analog to Digital Conversion —Jamie Angus , University of Salford, Salford, Technical, economical, safety aspects, and advantages Greater Manchester, UK of fibre optics is the topic of this seminar. Content will include fibre advantages, reliability, types of fibre, para - The purpose of this paper is to introduce a novel meters of fibres (attenuation, bandwidth length product, form of input topology for the analog inputs of over - latency), connector types, cleaning, and testing. sampled analog to digital converters. This new topology, when used with existing components, can achieve a dynamic range of 28 linear bits but Session P27 Sunday, May 10 has the potential to achieve even more if suitable 13:00 – 17:00 Room K3 technology can be developed. The paper analyzes the current limitations of existing topologies, pre - SIGNAL PROCESSING sents the new topology, and shows how it can achieve much higher dynamic ranges. The optimal Chair: Günther Thiele application of the topology and means of extending ¯

Audio Engineering Society 122nd Convention Program, 2007 Spring 43 Technical Progra m it for higher dynamic ranges is also discussed. maximizing the information flow through several Convention Paper 7801 nonlinear neurons. The algorithm requires that for every system audio output there be a corre - 14:30 sponding microphone for effective feedback sig - nal separation/ cancellation. The desired talker P27-4 Automatic Equalization of Flat TV signal from the algorithm outputs is recognized Loudspeakers Using Parametric IIR Filters— and transmitted while the identified feedback sig - Herwig Behrends ,1 Adrian von dem Knesebeck ,2 nals are discarded. Werner Bradinal, 1 Peter Neumann ,1 Udo Zölzer 2 Convention Paper 7804 1NXP Semiconductors, Hamburg, Germany 2Helmut Schimidt University, University of the 16:00 Federal Armed Forces, Hamburg, Germany P27-7 Implementing Audio Algorithms and Small loudspeakers used in today’s flat television Integrating Processor-Specific Code Using set cabinets and the requirement for “invisible Model Based Design— Arvind Ananthan ,1 sound” lead to a frequency response that is influ - Mark Corless ,2 Marco Roggero 3 enced in a very disadvantageous way by the 1The MathWorks, Natick, MA, USA physical design constraints. Loudspeakers are 2The MathWorks, Novi, MI, USA deeply embedded within the cabinet—the sound 3The MathWorks, Ismaning, Germany is thus forced through narrow vents, funnels or other waveguides. Down- or backfiring place - This paper explores the final stages in the mod - ments of the loudspeakers are also common el-based design and implementation of an audio practice, in order to minimize their visibility as algorithm on a fixed-point embedded processor. much as possible. Generally, this leads to a Once the algorithm, a 3-band parametric equal - non-flat frequency response with a strong izer in this example, is designed and simulated coloration of the sound. We present an approach to using a combination of scripting and graphical compensate these effects by means of simple sec - modeling tools, real-time C-code is automatically ond order equalizer sections (biquads), where cen - generated from this model. This paper illustrates ter frequency, gain, and bandwidth of the equalizer how algorithmic C-code generated from such a sections are automatically calculated from a mea - model in Simulink can be integrated into a stand- sured frequency response. The tool is usable in a alone embedded project as a library and imple - ® laboratory environment, with relatively inexpensive mented on an Analog Devices Blackfin 537 standard PC sound cards and microphones. processor. It also elaborates how processor spe - Convention Paper 7802 cific C-callable assembly code can then be inte - grated into the model for both simulation and code generation to improve its execution perfor - 15:00 mance on this processor. P27-5 Audio n-Genie: Domain Specific Language Convention Paper 7805 for Audio Processing— Tiziano Leidi ,1 Thierry Heeb ,2 Marco Colla ,1 Jean-Philippe Thiran 3 16:30 1 Institute for Applied Computer Science and P27-8 Subjective and Objective Evaluation of the Industrial Technologies of Southern Audio Vacuum-Tube Amplifiers— Andrzej Switzerland (ICIMSI), Manno, Switzerland Dobrucki ,1 Stanislaw Maleczek ,2 Maurycy Kin 1 2 ANAGRAM Technologies SA, Préverenges, 1Wroclaw University of Technology, Wroclaw, Switzerland Poland 3 Ecole Polytechnique Federale de Lausanne 2Military Institute of Engineering Technology, (EPFL), Lausanne, Switzerland Wroclaw, Poland Specialized development environments represent The subjective and objective evaluation of 5 today an important added value for domain high-quality vacuum-tube audio amplifiers is pre - specific system providers suffering the lack of a sented in this paper. As the reference the pro - dedicated, ergonomic, efficient, and affordable tool fessional transistor amplifier has been used. The able to boost their core business. This paper de - subjective evaluation has been done by the scribes Audio n-genie, a domain-specific language team of judges as well as with the computer- and its associated development environment sup - based psychoacoustic model according with porting the automatic production, by means of PAQM protocol. The amplifiers’ sound quality component-based model-driven generative pro - assessed by the listeners is consistent with the gramming, of digital audio processing applications. one evaluated with the use of the psychoa - Convention Paper 7803 coustic model. It was found that the best sound quality is obtained by vacuum-tube amplifiers, 15:30 the worst—by the reference amplifier. The P27-6 Acoustic Echo Cancellation Using MIMO results of subjective evaluation are inconsistent Blind Deconvolution— Ephraim Gower, with quality assessed by measurement of objec - Malcolm Hawksford , University of Essex, tive parameters: all amplifiers have comparable Colchester, UK quality, but the best is the transistor amplifier because of lowest level of THD+N level. A new multiple-input multiple-output frame-pro - Convention Paper 7806 cessing algorithm is introduced that exploits blind deconvolution for acoustic echo cancella - Session P28 Sunday, May 10 tion. The channel deconvolution filters, which 13:00 – 16:30 Room K4 can be blindly estimated as either finite impulse or infinite impulse responses, are optimized by PSYCOACOUSTICS AND PERCEPTION

44 Audio Engineering Society 126th Convention Program, 2009 Spring Chair: Florian Wickelmaier , University of Tübingen, 1Gdansk University of Technology, Gdansk, Tübingen, Germany Poland 2Excellence Center PROKSIM Institute of 13:00 Physiology and Pathology of Hearing, Warsaw, P28-1 Localization of Consecutive Sound Events in Poland Reverberant Environment— Marko Takanen, A new way of assessment of noise-induced Antti Jylhä, Tapani Pihlajamäki, Juha Holm, Ilkka harmful effects on the human hearing system is Huhtakallio, Ville Pulkki , Helsinki University presented in this paper. The method takes into of Technology, Espoo, Finland consideration properties of the selected physiologi - A listening test was conducted to assess the cal human hearing system. On the basis of the localization of consecutive sound events in simulat - hearing examinations and noise measurements ed reverberant conditions. The stimuli consisted of results and psychoacoustical noise dosimeter per - two sound events, which were reverberant wide - formance the new indicators of the noise harmful - band harmonic sounds reproduced in a multichan - ness were proposed. The evaluation of the pro - nel anechoic chamber. Localization threshold for posed indicators were conducted on the basis of the latter sound event was measured as the direct- hearing examinations in the real noise exposure to-reverberant sound level ratio with an adaptive situations and also on the basis of the simulation transformed up-down method. The studied factors results using standard test signals (such as white, affecting the localization threshold were the time in - pink, and brown noise). The performed analysis terval and pitch difference between the two sound and obtained results confirmed the practical useful - events and the time gap between the direct sound ness and correctness of the proposed indicators. and reverberation. The results indicate that all fac - Convention Paper 7813 tors have a significant effect on localization. Paper presented by Bozena Kostek Convention Paper 7807 15:00

13:30 P28-5 Time-Expanded Speech to Improve the Intelligibility of Consonants for the Hard-of- P28-2 The Contrasting and Conflicting Definitions Hearing— N H Shobha ,1 T G Thomas ,2 K of Envelopment— Jan Berg , Luleå University Subbarao 1 of Technology, Luleå, Sweden 1Osmania University, Hyderabad, India 2BITS-Pilani, Dubai, UAE In spatial audio, the term envelopment is not unambiguously defined and the different de facto Our investigations addressed the efficacy of definitions both overlap and contradict one temporal processing to the clear speech advan - another. This unclarity may pose a problem tage. A case for synthetic clear speech in the where the sensation of being surrounded by context of hearing impairment was constituted. sound is subject for investigation and analysis. Stop–vowels of English language were used as This paper reviews the different concepts of test stimuli, which were subjected to non-uniform envelopment in order to point to where possible time-expansion. Burst duration, voice onset time, problems may occur. A tentative suggestion for and formant transition duration segments were a terminology that can serve the different con - selectively time-expanded by 50 to 100 percent texts of enveloping sounds is also given. of their original duration. Hearing impairment Convention Paper 7808 was simulated at three SNR levels. The percep - tual analysis was accomplished in terms of infor - 14:00 mation transmission analysis measure and statistical measures. Burst duration expansion P28-3 Apparent Source Width in ITU Surround— by 50 percent was reported to be beneficial at a Jorge Medina Victoria, Thomas Görne , Hamburg lower SNR level; while VOT and FTD lengthen - University of Applied Sciences, Hamburg, ing did not contribute to improved intelligibility. Germany Convention Paper 7810 Apparent Source Widths (ASW) of phantom im - Paper not presented but available for purchase ages in a ITU-R BS.775-1 standard surround loud - speaker configuration have been investigated for 15:30 different signals by means of a randomized blind P28-6 Octave-Band Analysis on ITU-R Listening test. Test signals were generated from anechoic Test Data— Ian M. Dash , Australian recordings by amplitude panning between adjacent Broadcasting Corporation, Sydney, NSW, channels. The listening test showed that an Australia increase of Apparent Source Width coincides with the increase of localization uncertainty at the side Listening test data collected in 2003 on 49 audio and back areas of the ITU setup. Largest ASW val - program samples were used to formulate the ues were found between RS and LS channels. ITU-R BS.1770 program loudness prediction Convention Paper 7809 algorithm. The validity of this data at low fre - quencies was unproven. Octave-band analysis 14:30 has therefore been performed on the test sam - ples to test for audibility in each band. Results P28-4 A New Methodological Approach to the Noise suggest that further listening tests may be need - Threat Evaluation Based on the Selected ed to obtain reliable low-frequency data. A multi - Physiological Properties of the Human Hear - ple regression analysis was also performed on ing System— Jozef Kotus ,1 Bozena Kostek ,1,2 the octave-band data to obtain a least-squares Andrzej Czyzewski 1 weighting curve for comparison with the ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 45 Technical Progra m BS.1770/RLB2 weighting curve. Results suggest Session P29 Sunday, May 10 that while the BS.1770 curve performs well, 13:30 – 15:00 K4 Foyer there is still room for improvement. Convention Paper 7811 POSTERS: SIGNAL ANALYSIS, MEASUREMENTS, RESTORATION 16:00 13:30 P28-7 Windowed Sine Bursts: In Search of Optimal Test Signals for Detecting the Threshold of P29-1 Evaluation and Comparison of Audio Chroma Audibility of Temporal Decays— Andrew Feature Extraction Methods— Michael Stein ,1 Goldberg , Helsinki University of Technology, Benjamin M. Schubert ,1 Matthias Gruhne ,2 Espoo, Finland Gabriel Gatzsche ,2 Markus Mehnert 1 1Ilmenau University of Technology, Ilmenau, A slow decay in an audio signal is perceived as Germany ringing and is commonly caused by room modes. 2 Fraunhofer Institute for Digital Media This affects the perception of intelligibility, clarity, Technology IDMT, Ilmenau, Germany definition, and spatial rendering. A method has previously been devised to find the threshold of This paper analyzes and compares different audibility of the decay in low-frequency narrow- methods for digital audio chroma feature extrac - band signals. One of the test signals in the tion. The chroma feature is a descriptor, which large-scale listening test will be a low-frequency represents the tonal content of a musical audio sine burst, but spectral spreading at the start signal in a condensed form. Therefore chroma and end of the test signal acts as an additional features can be considered as an important non-modal cue. This effect is removed by prerequisite for high-level semantic analysis, like windowing, for example a half Hann. The aim of chord recognition or harmonic similarity estima - this paper is to determine the window length tion. A better quality of the extracted chroma required (threshold) to render the end of the test feature enables much better results in these signal free from audible spectral spreading. The high-level tasks. In order to discover the quality Parameter Estimation by Sequential Testing of chroma features, seven different state-of-the- (PEST) method and calibrated headphones (to art chroma feature extraction methods have remove factors associated with the listening been implemented. Based on an audio data - environment) are used in subjective listening base, containing 55 variations of triads, the out - tests. The window length threshold is found to put of these algorithms is critically evaluated. be constant above 200 Hz but rises exponential - The best results were obtained with the ly toward low frequencies, and is replay level Enhanced Pitch Class Profile. dependent. Threshold may be related to the Convention Paper 7814 absolute threshold of hearing, masking curves and/or auditory filter bandwidth. 13:30 Convention Paper 7812 P29-2 Measuring Transient Structure-Borne Sound in Musical Instruments—Proposal and First Live Sound Seminar 4 Sunday, May 10 Results from a Laser Intensity Measurement 13:00 – 15:00 Atrium 2 Setup— Robert Mores ,1 Marcel thor Straten ,2 Andreas Selk 3 SENNHEISER 1Hamburg University of Applied Sciences, Hamburg, Germany Chair: Matthias Fehr , DKE (German Comission 2Consultant, Seevetal, Germany for Electronics and IT-Techn.) 3Consultant, Hamburg, Germany Panelists: Hubert Eckart , DTHG (German Society The proposal for this new measurement setup of Theatre, Events, etc.) is motivated by curiosity in transients propagat - Bruno Marx , APWTP (Association ing across arched tops of violins. Understand - of Professional Wireless Production ing the impact of edge construction on transient Technologies) wave reflection back to the top of a violin or on Walter Möller , BLM, TKLM (Technical conduction into the rib requires single-shot Conference of State Media Authority) recordings possibly without statistical process - Florian von Hofen , VPLT (Society of Private ing. Signal-to-noise ratio should be high al - Light and Soundtechique) though mechanical amplitudes at distinct loca - tions on the structure surface are in the range Loss of Spectrum for Wireless Production Tools of a few micrometers only. In the proposed set - The WRC 07, World Radio Conference 2007, has identified up, the intensity of a laser beam is directly the Spectrum 790 – 862MHz for IMT, International Mobile measured after passing a screen attached to Telecom. In Germany more than 630,000 users operate in the device under test. The signal-to-noise ratio that spectrum: they need a new home for their activities. achieved for one micrometer transients in single- The discussion will focus on the alternative spectrum shot recordings is significantly more than 60 dB. that has to be opened and the transition time that can be Convention Paper 7815 expected. For how long can the quality of wireless pro - duction tools be guaranteed? 13:30 Experts of the Federal Network Agency, the IRT, and the APWPT user organization and from the Manufactur - P29-3 Evaluating Ground Truth for ADRess as a ers will give Spectrum users an outlook on expected Preprocess for Automatic Musical Instrument changes. Identification — Joseph McKay, Mikel Gainza,

46 Audio Engineering Society 126th Convention Program, 2009 Spring Dan Barry , Dublin Institute of Technology, cation of the system with the use of a monitoring Dublin, Ireland station is shown. The plan of further experiments is presented and the conclusions are derived. Most research in musical instrument identification Convention Paper 7818 has focused on labeling isolated samples or solo phrases. A robust instrument identification system 13:30 capable of dealing with polytimbral recordings of instruments remains a necessity in music informa - P29-6 Estimating Instrument Spectral Envelopes for tion retrieval. Experiments are described that eval - Polyphonic Music Transcription in a Music uate the ground truth of ADRess as a sound Scene-Adaptive Approach— Julio J. Carabias- source separation technique used as a preprocess Orti, Pedro Vera-Candeas, Nicolas Ruiz-Reyes, to automatic musical instrument identification. The Francisco J. Cañadas-Quesada, Pablo ground truth experiments are based on a number Cabañas-Molero , University of Jaén, Linares, of basic acoustic features, while using a Gaussian Spain Mixture Model as the classification algorithm. We propose a method for estimating the spectral Using all 44 acoustic feature dimensions, success - envelope pattern of musical instruments in a ful identification rates are achieved. musical scene-adaptive scheme, without having Convention Paper 7816 any prior knowledge about the real transcription. A musical note is defined as stable when varia - 13:30 tions between its harmonic amplitudes are held constant during a certain period of time. A densi - P29-4 Improving Rhythmic Pattern Features Based ty-based clustering algorithm is used with the sta - on Logarithmic Preprocessing— Matthias ble notes in order to separate different envelope Gruhne, Christian Dittmar , Fraunhofer Institute models for each note. Music scene-adaptive en - for Digital Media Technology IDMT, Ilmenau, velope patterns are finally obtained from similarity Germany and continuity of the different note models. Our In the area of Music Information Retrieval, the approach has been tested in a polyphonic music rhythmic analysis of music plays an important role. transcription scheme with synthesized and real In order to derive rhythmic information from music music recordings obtaining very promising results. signals, several feature extraction algorithms have Convention Paper 7819 been described in the literature. Most of them extract the rhythmic information by auto-correlating the temporal envelope derived from different Exhibitor Seminar Sunday, May 10 frequency bands of the music signal. Using the 13:00 – 14:00 Exhibit Floor Booth 2303 auto-correlated envelopes directly as an audio- ES19 OPTOCORE –C feature is afflicted with the disadvantage of tempo OPTOCORE IN PRACTICE dependency. To circumvent this problem, further postprocessing via higher-order statistics has been Presenters: Martin Barbour , OPTOCORE Support proposed. However, the resulting statistical fea - Engineer tures are still tempo dependent to a certain extent. Andreas Kaspar , OPTOCORE Support This paper describes a novel method, which loga - Engineer rithmizes the lag-axis of the auto-correlated enve - Thorsten Schulze , OPTOCORE Support lope and discards the tempo-dependent part. This and Product Manager approach leads to tempo-invariant rhythmic features. A quantitative comparison of the original This seminar will cover practical exercises with OPTO - methods versus the proposed procedure is CORE system/devices. Content will include practical set - described and discussed in this paper. up of a network, remote pre-amp control through third Convention Paper 7817 party devices, transport of control data and Ethernet, software control, and firmware upgrades. 13:30 Workshop 22 Sunday, May 10 P29-5 Further Developments of Parameterization 14:30 – 16:30 Room K2 Methods of Audio Stream Analysis for Security Purposes— Pawel Zwan, Andrzej MPEG SAOC: INTERACTIVE AUDIO Czyzewski , Gdansk University of Technology, AND BROADCASTING, MUSIC 2.0, NEXT Gdansk, Poland GENERATION TELECOMMUNICATION The paper presents an automatic sound recogni - Chair: Oliver Hellmuth , Fraunhofer IIS, Erlangen, tion algorithm intended for application in an Germany audiovisual security monitoring system. A distrib - uted character of security systems does not Panelists: Jonas Engdegård , Dolby, Stockholm, allow for simultaneous observation of multiple mul - Sweden timedia streams, thus an automatic recognition al - Christof Faller , Illusonic LLC, Lausanne, gorithm must be introduced. In the paper a module Switzerland for the parameterization and automatic detection Jürgen Herre , Fraunhofer IIS, Erlangen, of audio events is described. The spectral analysis Germany of sounds of a broken window, gunshot, and Werner Oomen , Philips Applied scream are performed and parameterization meth - Technologies, Eindhoven, The Netherlands ods are proposed and discussed. Moreover, a sound classification system based on the Support Recently the ISO/MPEG standardization group launched Vector Machines (SVM) algorithm is presented an activity for bit rate-efficient and backward compatible and its accuracy is discussed. The practical appli - coding of multiple sound objects that heavily exploits the ¯

Audio Engineering Society 126th Convention Program, 2009 Spring 47 Technical Progra m human perception of spatial sound. On the receiving power line treatments such as technical power, balanced side, such a "Spatial Audio Object Coding" (SAOC) power, isolation transformers, and surge suppressors. system renders the transmitted objects interactively into a sound scene on any desired reproduction. Based on Exhibitor Seminar Sunday, May 10 the SAOC technology elegant solutions for Interactive 15:00 – 16:00 Exhibit Floor Booth 2303 Audio and Broadcasting, Music 2.0, Next Generation ES20 OPTOCORE –A Telecommunication become feasible. The workshop reviews the ideas and principles behind Spatial Audio TECHNICAL INTRODUCTION TO OPTOCORE Object Coding, especially highlighting its possibilities and benefits for those new types of applications. Additionally, Presenters: Martin Barbour , OPTOCORE Support the potential of SAOC is illustrated by means of real-time Engineer demonstrations. Andreas Kaspar , OPTOCORE Support Engineer Thorsten Schulze , OPTOCORE Support Tutorial 10 Sunday, May 10 and Product Manager 14:30 – 17:00 Room K1 The OPTOCORE Synchronous Fibre Network is intro - AUDIO SYSTEM GROUNDING & INTERFACING— duced. Content will include an Introduction/Overview, AN OVERVIEW design of the network, synchronicity, latency, topology, redundancy, protocol, and devices. Presenter: Bill Whitlock , Jensen Transformers, Inc., Chatsworth, CA, USA Exhibitor Seminar Sunday, May 10 Although the subject has a black art reputation, this tutorial 16:00 – 17:00 Exhibit Floor Booth 2303 replaces myth and hype with insight and knowledge, reveal - ES18 OPTOCORE –B ing the true causes of system noise and ground loops. Although safety must be the top priority, some widely used BASICS OF FIBRE OPTICS cures are both illegal and deadly. Both balanced and unbal - Presenters: Martin Barbour , OPTOCORE Support anced interfaces are vulnerable to noise coupling, but the Engineer unbalanced interface is exquisitely so due to an intrinsic Andreas Kaspar , OPTOCORE Support problem. Because balanced interfaces are widely misunder - Engineer stood, their theoretically perfect noise rejection is severely Thorsten Schulze , OPTOCORE Support degraded in most real-world systems. Some equipment, and Product Manager because of an innocent design error, has a built-in noise problem. A simple, no-test-equipment, troubleshooting Technical, economical, safety aspects, and advantages method can pinpoint the location and cause of system of fibre optics is the topic of this seminar. Content will noise. Ground isolators in the signal path solve the funda - include fibre advantages, reliability, types of fibre, para - mental noise coupling problems. Also discussed are meters of fibres (attenuation, bandwidth length product, unbalanced to balanced connections, RF interference, and latency), connector types, cleaning, and testing.

th Presen 2 ted at 009 Ma the 12 126 Convention Papers and CD-ROM y 7 – M 6th Co ay 10 nven • Mun tion ich, Ge rmany

Convention Papers of many of the presentations given at the 126th Convention and a CD-ROM containing the 126th Convention Papers are available from the Audio Engineering Society. Prices follow:

126 th CONVENTION PAPERS (single copy) Member: $ 5. (USA) Nonmember: $ 20. (USA)

126th Convention th 126 CD-ROM (168 papers) 2009 May 7– May 10 Member : $160. (USA) Munich, Germany Nonmember : $200. (USA)

48 Audio Engineering Society 126th Convention Program, 2009 Spring