Comparing the different sound source localization techniques, proposed in the literature during the last decade, represents a relevant topic in order to establish advantages and disadvantages of a given approach in a real-time implementation. Traditionally, algorithms for sound source localization rely on an estimation of time difference of arrival (TDOA) at microphone pairs through GCC-PHAT When several microphone pairs are available the source position can be estimated as the point in space that best fits the set of TDOA measurements by applying global coherence field (GCF), also known as SRP-PHAT, or oriented global coherence field (OGCF). A first interesting analysis compares the performance of GCF and OGCF to a suboptimal LS search method. In a second step, Adaptive Eigenvalue Decomposition is implemented as an alternative to GCC-PHAT in TDOA estimation. Comparative experiments are conducted on signals acquired by a linear array during WOZ experiments in an interactive-TV scenario. Changes in performance according to different SNR levels are reported.
This paper addresses the problem of hands-free speech recognition in a noisy office environment. An array of six omnidirectional microphones and a corresponding time delay compensation module are used to provide a beamformed signal as input to a HMM-based recognizer. Training of HMMs is performed either using a clean speech database or using a filtered version of the same database. Filtering consists in a convolution with the acoustic impulse response between the speaker and microphone, to reproduce the reverberation effect. Background noise is summed to provide the desired SNR. The paper shows that the new models trained on these data perform better than the baseline ones. Furthermore, the paper investigates on maximum likelihood linear regression (MLLR) adaptation of the new models. It is shown that a further performance improvement is obtained, allowing to reach a 98.7% WRR in a connected digit recognition task, when the talker is at 1.5 m distance from the array.
Acoustic maps created on the basis of the signals acquired by distributed networks of microphones allow to identify position and orientation of an active talker in an enclosure. In adverse situations of high background noise, high reverberation or unavailability of direct paths to the microphones, localization may fail. This paper proposes a novel approach to talker localization and estimation of head orientation based on the classification of global coherence field (GCF) or oriented GCF maps. Preliminary experiments with data obtained by simulated propagation as well as with data acquired in a real room show that the match with precalculated map models provides a robust behavior in adverse conditions.
A new approach to the permutation problem for Blind Source Separation (BSS) in the frequency domain is presented. The independence of the separation across the frequencies, and thus the probability that a permutation may occur, is minimized by a recursive linking of the ICA stage. A recursive adaptive estimation of smooth demixing matrices is used to initialize the Independent Component Analysis (ICA) in order to force it to converge with a coherent permutation across the whole spectrum. Since no information about non stationarity of the signals is exploited, the proposed method works also for short utterances (0.5-1 s) and in highly reverberant environments (T60sime 700 ms). Furthermore it is shown that the recursive initialization increases the accuracy of the ICA when a small amount of data observations is available.
This paper refers to a distant-talking interaction system in the car environment being developed at our labs. The system shares many aspects with the European project VICO (Virtual Intelligent CO-driver), to which ITC-irst and Bosch GmbH contributes. Target of the project is to allow a real natural dialogue with a user while driving a car, enabling him/her to get driving assistance, hotel and restaurant reservation and tourist information. The system is based on a HMM speech recognizer for the Italian language, a Dialog Manager module that handles the interaction with the user and a module that retrieves data from databases, called Car Wide Web. In this paper we focus on the structure of the Car Wide Web module, the databases containing the information being used in the project, the specific API developed to access the databases and the corpus of spontaneous speech interactions collected using the Wizard-of-Oz method.