The aim of this paper is to investigate the feasibility of an audio-based context recognition system. Here, context recognition refers to the automatic classification of the context or an environment around a device. A system is developed and compared to the accuracy of human listeners in the same task. Particular emphasis is placed on the computational complexity of the methods, since the application is of particular interest in resource-constrained portable devices. Simplistic low-dimensional feature vectors are evaluated against more standard spectral features. Using discriminative training, competitive recognition accuracies are achieved with very low-order hidden Markov models (1-3 Gaussian components). Slight improvement in recognition accuracy is observed when linear data-driven feature transformations are applied to mel-cepstral features. The recognition rate of the system as a function of the test sequence length appears to converge only after about 30 to 60 s. Some degree of accuracy can be achieved even with less than 1-s test sequence lengths. The average reaction time of the human listeners was 14 s, i.e., somewhat smaller, but of the same order as that of the system. The average recognition accuracy of the system was 58% against 69%, obtained in the listening tests in recognizing between 24 everyday contexts. The accuracies in recognizing six high-level classes were 82% for the system and 88% for the subjects.

We present the virtual acoustics modeling framework that is a part of the MPEG-4 standard. A scene description language called the binary format for scenes (BIFS) is defined within MPEG-4 for making multimedia presentations that include various types of audio and visual data. BIFS also provides means for creating three-dimensional (3-D) virtual worlds or scenes, where visual and sound objects can be positioned and given temporal behavior. Local interaction between the user and the scene can be added to MPEG-4 applications. Typically the user can navigate in a 3-D scene so that it is viewed from different positions. In case that there are sound source objects in a scene, the sounds may be spatialized so that they are heard coming from the positions defined for them. A subset of BIFS, called the advanced AudioBIFS aims at enhanced modeling of 3-D sound environments. In this framework, sounds can be given positions, and also the virtual environment where they appear can be associated with acoustic properties that allow modeling of phenomena such as air absorption, Doppler effect, sound reflections, and reverberation. These features can be used for adding room acoustic effects to sound in the MPEG-4 terminal, and for creating immersive 3-D audiovisual scenes.

The paper concerns the development of a system for the recognition of a context or an environment based on acoustic information only. Our system uses Mel-frequency cepstral coefficients and their derivatives as features, and continuous density hidden Markov models (HMM) as acoustic models. We evaluate different model topologies and training methods for HMMs and show that discriminative training can yield a 10% reduction in error rate compared to maximum-likelihood training. A listening test is made to study the human accuracy in the task and to obtain a baseline for the assessment of the performance of the system. Direct comparison to human performance indicates that the system performs somewhat worse than human subjects do in the recognition of 18 everyday contexts and almost comparably in recognizing six higher level categories.

In this paper, we address the problem of computational auditory scene recognition and describe methods to classify auditory scenes into predefined classes. By auditory scene recognition we mean recognition of an environment using audio information only. The auditory scenes comprised tens of everyday outside and inside environments, such as streets, restaurants, offices, family homes, and cars. Two completely different but almost equally effective classification systems were used: band-energy ratio features with 1-NN classifier and Mel-frequency cepstral coefficients with Gaussian mixture models. The best obtained recognition rate for 17 different scenes out of 26 and for an analysis duration of 30 seconds was 68.4%. For comparison, the recognition accuracy of humans was 70% for 25 different scenes and the average response time was around 20 seconds. The efficiency of different acoustic features and the effect of test sequence length were studied.

The goal of this paper is to give an overview of real-time room acoustic rendering. The approach is based on the source-medium-receiver model, in which we model sound sources, room acoustics, and a listener. The basic techniques for each of these are presented, but the main emphasis is on the room acoustic modeling and interactive auralization. As a case study we present the structure of the DIVA auralization system developed at the Helsinki University of Technology. In addition, we describe subjective evaluations made to our system. Finally, a discussion of some applications of virtual acoustics and their computational needs are given.

We present an overview of the AudioBIFS system, part of the Binary Format for Scene Description (BIFS) tool in the MPEG-4 International Standard. AudioBIFS is the tool that integrates the synthetic and natural sound coding functions in MPEG-4. It allows the flexible construction of soundtracks and sound scenes using compressed sound, sound synthesis, streaming audio, interactive and terminal-dependent presentation, three-dimensional (3-D) spatialization, environmental auralization, and dynamic download of custom signal-processing effects algorithms. MPEG-4 sound scenes are based on a model that is a superset of the model in VRML 2.0, and we describe how MPEG-4 is built upon VRML and the new capabilities provided by MPEG-4. We discuss the use of structured audio orchestra language, the MPEG-4 SAOL, for writing downloadable effects, present an example sound scene built with AudioBIFS, and describe the current state of implementations of the standard.

We propose a novel technique for the design of low-order infinite impulse response (IIR) filter models of head-related transfer functions (HRTFs) that uses balanced model truncation. We design tenth-order IIR filters, suitable for efficient real-time implementation, from preprocessed HRTF impulse responses that are of significantly superior quality to current IIR models derived with the Prony and Yule-Walker methods.