In this paper, we introduce a newly developed sound-field-reproducing and -sharing system. The system consists of an 80-channel fullerene-shaped microphone array and a 96-channel loudspeaker array mounted in an enclosure called a sound cask, so named because of its shape. The cask has two functions. First, it functions as a precise sound field reproduction system. The sound signals acquired from a microphone array in any sound field can be reproduced in the sound cask after passing through filters that modify the amplitude and phase on the basis of the boundary surface control principle. The large number of loudspeakers result in the precise orientation and depth of sound images. Second, it functions as a platform for a sound-field-sharing system. Several casks located remotely can appear to exist in the same sound field for subjects inside a cask. In addition, the cask is large enough for one to be able to play a musical instrument inside it. The musical sound or voices produced by subjects can be shared by subjects in a distant cask after convoluting the impulse responses of the original sound field. The concept of the system is explained in detail.
Principal component and spectral-based feature sets were applied to the recognition of gamelan instrument sounds using support vector machines (SVMs). The principal components were calculated on the basis of a segmented scalogram from the first harmonic frequency of the gamelan recordings. The segmented scalogram is assumed as a ``facial image'' of the gamelan instrument sound in a frontal pose, neutral expression, and normal lighting. The scalogram was computed from the gamelan sound signal using a continuous wavelet transform (CWT). The performance and contribution of the principal component and spectral-based features were compared using an F-measure. For the training phase, the feature sets were extracted from isolated tones that were recorded over the entire frequency range of four gamelan instruments (demung, saron, peking, and bonang families). Using 90%/10% splits between the training and validating data sets, model classifiers were constructed from the radial basis function (RBF) kernel SVM. The classifiers are composed of 28 separate One-Against-One multiclass classifiers. The experiment showed that the spectral-based feature set shows an average F-measure of 74.05% and the appearance-based feature yields 71.87%. For saron-only note tracking, the spectral-based feature set had an F-measure of 83.79%, higher than the demung-only note tracking, which yielded 63.89%.
Wind turbine noise has become a serious environmental noise issue. To investigate this problem, a research project has been conducted over the three years from fiscal year 2010. In this project, experimental studies on human auditory sensation of wind turbine noise have been conducted together with field measurements and social surveys. Since wind turbine noise is often argued to be a low-frequency noise issue, the experiments were conducted with emphasis on the low-frequency components of wind turbine noise. Among them, an auditory experiment was performed to investigate noise indicators suitable for the assessment of wind turbine noise. In this experiment, not only wind turbine noise but also various environmental noises that were recorded so as to include low-frequency components down to infrasound were reproduced down to 4 Hz and loudness tests were performed. The experimental results were evaluated using the A- and C-weighted sound pressure levels, Zwicker loudness level, and Moore-Glasberg loudness level. It has been found that the A-weighted sound pressure level is a simple and appropriate indicator for the loudness assessment of general environmental noises. In addition, the relationships between the noise indicators are discussed on the basis of the results of numerical investigations.