AES San Francisco 2010
Paper Session P11
Friday, November 5, 2:30 pm — 6:30 pm (Room 220)
Paper Session: P11 - Acoustical and Physical Modeling
Julius O. Smith
P11-1 Virtual Acoustic Prototyping—Practical Applications for Loudspeaker Development—Alex Salvatti, JBL Professional - Northridge, CA, USA
Acoustic simulations using finite elements have been used in loudspeaker development for over 20 years, with complexity and accuracy accelerating in tandem with the increases in computing power generally available on the engineering desktop. Using user-friendly, modern FEA software, the author presents an overview of methods to build virtual prototypes of both horns and loudspeaker drivers that allows a significant reduction in the number of physical prototypes, as well as reduced development time. A comparison of simulated vs. measured data proves the validity of the methods.
Convention Paper 8213 (Purchase now)
P11-2 Simulation of Horn Driver Response by Combination of Matrix Analysis and FEA—Alex Voishvillo, JBL Professional - CA, USA
To access performance of a horn driver (compression driver loaded by a horn), measurement of frequency response on-axis and off-axis must be carried out. The measurement process is time-consuming especially if the entire 3-dimensional “balloon” of responses is to be measured. Prediction of directional responses of the horn only (without compression driver) can be performed by the FEA (Finite Elements Analysis) or BEA (Boundary Elements Analysis). However, FEA or BEA of horn only provides relative directional properties of the horn. The SPL responses of horn driver at different angles remain unknown because these responses depend on interaction of electrical, mechanical, and acoustical parameters of the compression driver and the acoustical parameters of the horn. New methods based on a combination of FEA and matrix analysis makes it possible to predict the response of a combination of various compression drivers and horns without actually measuring each combination and even without physically building horns. This method was verified during the development of a new AM series of JBL professional loudspeaker systems and showed high accuracy.
Convention Paper 8214 (Purchase now)
P11-3 Dynamic Motion of the Corrugated Ribbon In a Ribbon Microphone—Daniel Moses Schlessinger, Sennheiser DSP Research Laboratory - Palo Alto, CA, USA; Jonathan S. Abel, Stanford University - Stanford, CA, USA
Ribbon microphones are known for their warm sonics, owing in part to the unique ribbon motion induced by the sound field. Here the motion of the corrugated ribbon element in a sound field is considered, and a physical model of the ribbon motion is presented. The model separately computes propagating torsional disturbances and coupled transverse and longitudinal disturbances. Each propagation mode is implemented as a mass-spring model where a mass is identified with a ribbon corrugation fold. The model is parametrized using ribbon material and geometric properties. Laser vibrometer measurements are presented, revealing stiffness in the transverse and longitudinal propagation and showing close agreement between measured and modeled ribbon motion.
Convention Paper 8215 (Purchase now)
P11-4 Modeling of Leaky Acoustic Tube for Narrow-Angle Directional Microphone—Kazuho Ono, Takehiro Sugimoto, Akio Ando, Kimio Hamasaki, NHK Science and Technology Research Laboratories - Kinuta Setagaya-ku, Tokyo, Japan; Takeshi Ishii, Yutaka Chiba, Keishi Imanaga, Sanken Microphone Co. Ltd. - Suginami-ku, Tokyo, Japan
Line microphones have been popular as narrow directional microphones for a long time. Their structure adopts a leaky acoustical tube with many slits to suppress off-axis sensitivity, together with a directional capsule attached to this tube. Although many microphones of this type are on the market, we have no quantitative theory to explain its behavior, which is very important for effectively designing directivity. We thus modeled the leaky acoustical tube using a distributed equivalent circuit and combined it with the directional capsule’s equivalent circuit model. The analysis showed that the model agreed well with the measurement results, particularly at the directional characteristics, while an ordinary model of acoustical tube using delay and sum modeling did not.
Convention Paper 8216 (Purchase now)
P11-5 Modeling Viscoelasticity of Loudspeaker Suspensions Using Retardation Spectra—Tobias Ritter, Finn Agerkvist, Technical University of Denmark - Kgs. Lyngby, Denmark
It is well known that, due to viscoelastic effects in the suspension, the displacement of the loudspeaker increases with decreasing frequency below the resonance. Present creep models are either not precise enough or purely empirical and not derived from the basis of physics. In this investigation, the viscoelastic retardation spectrum, which provides a more fundamental description of the suspension viscoelasticity, is first used to explain the accuracy of the empirical LOG creep model (Knudsen et al.). Then, two extensions to the LOG model are proposed that include the low and high frequency limit of the compliance, not accounted for in the original LOG model. The new creep models are verified by measurements on two 5.5 loudspeakers with different surrounds.
Convention Paper 8217 (Purchase now)
P11-6 Physical Modeling and Synthesis of Motor Noise for Replication of a Sound Effects Library—Simon Hendry, Josh Reiss, Queen Mary University of London - London, UK
This paper presents the results of objective tests exploring the concept of using a small number of physical models to create and replicate a large number of samples from a traditional sound effects library. The design of a DC motor model is presented and this model is used to create both a household drill and a small boat engine. The harmonic characteristics, as well as the spectral centroid were compared with the original samples, and all the features agree to within 6.1%. The results of the tests are discussed with a heavy emphasis on realism and perceived accuracy, and the parameters that have to be improved in order to humanize a model are explored.
Convention Paper 8218 (Purchase now)
P11-7 Measures and Parameter Estimation of Triodes for the Real-Time Simulation of a Multi-Stage Guitar Preamplifier—Ivan Cohen, Ircam - Paris, France, Orosys R&D, Montpellier, France; Thomas Hélie, Ircam - Paris, France
This paper deals with the real-time simulation of a multi-stage guitar preamplifier. Dynamic triode models based on Norman Koren’s model, and "secondary phenomena" as grid rectification effect and parasitic capacitances are considered. Then, the circuit is modeled by a nonlinear differential algebraic system, with extended state-space representations. Standard numerical schemes yield efficient stable simulations of the circuit and are implemented as VST plug-ins. Measures of real triodes have been realized, to develop new triode models, and to characterize the capabilities of aged and new triodes. The results are compared for all the models, using lookup tables generated with the measures and Norman Koren’s model with its parameters estimated from the measures.
Convention Paper 8219 (Purchase now)
P11-8 ZFIT: A MATLAB Tool for Thiele-Small Parameter Fitting and Optimization—Christopher Struck, CJS Labs - San Francisco, CA, USA
Over the years, many approaches to the calculation of the Thiele-Small parameters have been presented. Most current methods rely upon curve-fitting to the impedance magnitude data for a specific lumped parameter model. A flexible Matlab least-mean-squares optimization tool for complex loudspeaker impedance data is described. Magnitude and phase data are fit to a user-selected lumped parameter model of variable complexity. Appropriate constraints on the optimization help identify if the selected model is of sufficient order or overly complex for the given data. Examples are shown for impedance data from several different loudspeaker drivers.
Convention Paper 8220 (Purchase now)