Authors:Hill, Adam J.; Hawksford, Malcolm O. J.; Newell, Philip
Affiliation:Department of Engineering, University of Derby, United Kingdom; School of Computer Science and Electronic Engineering, University of Essex, United Kingdom; Acoustic consultant, Moaña Spain
The current strategies for the low-frequency calibration of cinema sound systems are based on a flawed premise of low-frequency acoustics and psychoacoustics. This research shows that there is virtually no benefit in terms of spatiotemporal variance reduction: pre- and post-calibrated systems will exhibit equally position-dependent listening experience differences. For modern cinemas, the typical focus on room-modes when designing a low frequency calibration system is not necessary because the dimensions of the space coupled with low reverberation time results in Schroeder frequencies around 35 Hz. Above this value, effects of room-modes are not perceptible. Comb-filtering between sources and low-order reflections is the primary cause of high spatial variance. Furthermore, there is no evidence that spatial averaging techniques used for measurement and equalization are subjectively beneficial. A new approach needs to be invented.
Download: PDF (HIGH Res) (5.2MB)
Download: PDF (LOW Res) (773KB)
Authors:Stitt, Peter; Bertet, Stéphanie; van Walstijn, Maarten
Affiliation:LIMSI-CNRS, Orsay, France; Queen's University Belfast, Belfast, UK
Spatial audio techniques, such as Ambisonics and Wave Field Synthesis, aim to reproduce the sound field in the limited area of the sweet spot, which is approximately equidistant from all loudspeakers. This research extends the energy vector technique to improved localization at off-centered positions. In determining the source direction, a perceptual weight is assigned to each loudspeaker gain that takes into account the relative arrival times, levels, and directions of the loudspeaker signals. Conversely, uncompensated differences in arrival time can trigger the precedence effect. The proposed model was evaluated and compared to the original energy vector model and two binaural models. The extended energy vector version was at least 50% more accurate than the second best predictor with an average error of about 4°.
Download: PDF (HIGH Res) (1.5MB)
Download: PDF (LOW Res) (361KB)
Authors:Kolbrek, Bjørn; Evensen, Karen Brastad; Svensson, U. Peter
Affiliation:Department of Electronics and Telecommunications, NTNU – Norwegian University of Science and Technology, Trondheim, Norway
Because Finite Element and Boundary Element methods used to simulate cone loudspeaker directivity are slow and demanding on memory, this research explores the use of mode matching methods. A mode matching method for simulating concave, axisymmetric geometries offers a viable alternative. This method is based on a method presented by Pagneux et al. for simulation of horns, but has been extended so that vibrating walls can be taken into account. With the previous method it was possible to find impedance, pressure, and volume velocity by working step by step from one end of the structure to the other. With the new method, pressure and volume velocity at all points in the structure are found simultaneously by solving a large but sparse linear system of equations. Two examples of loudspeaker diaphragm shapes demonstrate how the method could be used to optimize diaphragm shape.
Download: PDF (HIGH Res) (1.4MB)
Download: PDF (LOW Res) (507KB)
Authors:G, Christian; Herrera, ; Donoso-Garcia, Pedro F.; Medeiros, Eduardo B.
Affiliation:Department of Mechatronic Engineering, CEFET-MG, Divinopolis, Minas Gerais, Brazil; Department of Electronic Engineering, UFMG, Belo Horizonte, Minas Gerais, Brazil; Department of Mechanical Engineering, UFMG, Belo Horizonte, Minas Gerais, Brazil
When using the automobile’s audio system with cell phones, the intelligibility of speech communications inside the vehicle is important because it influences the inhabitant’s comfort and safety. The articulation index (AI), which combines the S/N ratio of numerous spectral bands, is one objective proxy for the subjective experience. By analyzing the acoustic conditions inside an automobile while in use, it was shown that intelligibility depends both on the characteristics of car audio system and on the intensity of various noise sources. The main noise sources in two different test vehicles were examined. For example, the road noise of a vehicle moving a 100 km/h on asphalt pavement showed the lowest AI, while the noise produced by a vehicle on polyhedral pavement produced more user discomfort. In addition, air conditioner noise of a stationary vehicle also has a significant impact on the AI. For electronic communications, such as cell phones, equalization could lead to an improved AI.
Download: PDF (HIGH Res) (7.0MB)
Download: PDF (LOW Res) (717KB)
Authors:Aguilera, Emanuel; Lopez, Jose J.; Cooperstock, Jeremy R.
Affiliation:Institute of Telecommunications and Multimedia Applic., Universitat Politécnica de Valéncia, Valencia, Spain; Centre for Intelligent Machines, McGill University, Montreal, Canada
Audioconferencing systems are becoming increasingly sophisticated, seeking to improve immersion, intelligibility, and sense of presence. Various opportunities exist to make mobile multiparty conference calls more realistic, interactive, and immersive. These include spatial audio, interactive manipulation of avatar positions, and whisper mode. The authors analyze the utility of avatar movement and sidebar whisper-mode functionality within which a subset of participants can engage in an ad-hoc sidebar conversation privately from the remaining participants. Participants found that avatar movement was useful, but approximately half the participants only used this feature for initial positioning of the avatars in a desired configuration in the first task. The ability to maintain sidebar conversations was used frequently and rated highly in negotiation activity. However, participants made almost no use of this feature in the conversation activity for which negotiation was not involved.
Download: PDF (HIGH Res) (3.6MB)
Download: PDF (LOW Res) (314KB)
At the 61st Conference, held recently in London, topics of papers ranged from audio content management and user interfaces, through spatial rendering, to synthesis and sound design. Easy-to-reuse sonic content is likely to be more widely available in the future as a result of the Audio Commons project. We also find that drivers can be helped to drive more safely and economically with the assistance of a game that can modify behavior. Procedural audio shows promise as a means of synthesizing weapon sounds and it may also be possible to emulate mammalian voices successfully. Finally, it may be possible to simulate spatial motion using crossfaded binaural responses from a limited number of locations, and there is the possibility to source generic HRTFs from an ear shape-based averaging process.
Download: PDF (535KB)