Patrick A. Naylor

Eric Grinstein, Vincent W. Neo, Patrick A. Naylor

In many signal processing applications, metadata may be advantageously used in conjunction with a high dimensional signal to produce a desired output. In the case of classical Sound Source Localization (SSL) algorithms, information from a high dimensional, multichannel audio signals received by many distributed microphones is combined with information describing acoustic properties of the scene, such as the microphones' coordinates in space, to estimate the position of a sound source. We introduce Dual Input Neural Networks (DI-NNs) as a simple and effective way to model these two data types in a neural network. We train and evaluate our proposed DI-NN on scenarios of varying difficulty and realism and compare it against an alternative architecture, a classical Least-Squares (LS) method as well as a classical Convolutional Recurrent Neural Network (CRNN). Our results show that the DI-NN significantly outperforms the baselines, achieving a five times lower localization error than the LS method and two times lower than the CRNN in a test dataset of real recordings.

Dushyant Sharma, Rong Gong, James Fosburgh et al.

We present a novel multi-channel front-end based on channel shortening with theWeighted Prediction Error (WPE) method followed by a fixed MVDR beamformer used in combination with a recently proposed self-attention-based channel combination (SACC) scheme, for tackling the distant ASR problem. We show that the proposed system used as part of a ContextNet based end-to-end (E2E) ASR system outperforms leading ASR systems as demonstrated by a 21.6% reduction in relative WER on a multi-channel LibriSpeech playback dataset. We also show how dereverberation prior to beamforming is beneficial and compare the WPE method with a modified neural channel shortening approach. An analysis of the non-intrusive estimate of the signal C50 confirms that the 8 channel WPE method provides significant dereverberation of the signals (13.6 dB improvement). We also show how the weights of the SACC system allow the extraction of accurate spatial information which can be beneficial for other speech processing applications like diarization.

Francesco Nespoli, Jule Pohlhausen, Patrick A. Naylor et al.

The analysis of conversations recorded in everyday life requires privacy protection. In this contribution, we explore a privacy-preserving feature extraction method based on input feature dimension reduction, spectral smoothing and the low-cost speaker anonymization technique based on McAdams coefficient. We assess the utility of the feature extraction methods with a voice activity detection and a speaker diarization system, while privacy protection is determined with a speech recognition and a speaker verification model. We show that the combination of McAdams coefficient and spectral smoothing maintains the utility while improving privacy.

Costas Yiallourides, Patrick A. Naylor

Objective: In this work the potential of non-invasive detection of knee osteoarthritis is investigated using the sounds generated by the knee joint during walking. Methods: The information contained in the time-frequency domain of these signals and its compressed representations is exploited and their discriminant properties are studied. Their efficacy for the task of normal vs abnormal signal classification is evaluated using a comprehensive experimental framework. Based on this, the impact of the feature extraction parameters on the classification performance is investigated using Classification and Regression Trees (CART), Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM) classifiers. Results: It is shown that classification is successful with an area under the Receiver Operating Characteristic (ROC) curve of 0.92. Conclusion: The analysis indicates improvements in classification performance when using non-uniform frequency scaling and identifies specific frequency bands that contain discriminative features. Significance: Contrary to other studies that focus on sit-to-stand movements and knee flexion/extension, this study used knee sounds obtained during walking. The analysis of such signals leads to non-invasive detection of knee osteoarthritis with high accuracy and could potentially extend the range of available tools for the assessment of the disease as a more practical and cost effective method without requiring clinical setups.

Constantinos Papayiannis, Christine Evers, Patrick A. Naylor

The materials of surfaces in a room play an important room in shaping the auditory experience within them. Different materials absorb energy at different levels. The level of absorption also varies across frequencies. This paper investigates how cues from a measured impulse response in the room can be exploited by machines to detect the materials present. With this motivation, this paper proposes a method for estimating the probability of presence of 10 material categories, based on their frequency-dependent absorption characteristics. The method is based on a CNN-RNN, trained as a multi-task classifier. The network is trained using a priori knowledge about the absorption characteristics of materials from the literature. In the experiments shown, the network is tested on over 5,00 impulse responses and 167 materials. The F1 score of the detections was 98%, with an even precision and recall. The method finds direct applications in architectural acoustics and in creating more parsimonious models for acoustic reflections.

Constantinos Papayiannis, Christine Evers, Patrick A. Naylor

The classification of acoustic environments allows for machines to better understand the auditory world around them. The use of deep learning in order to teach machines to discriminate between different rooms is a new area of research. Similarly to other learning tasks, this task suffers from the high-dimensionality and the limited availability of training data. Data augmentation methods have proven useful in addressing this issue in the tasks of sound event detection and scene classification. This paper proposes a method for data augmentation for the task of room classification from reverberant speech. Generative Adversarial Networks (GANs) are trained that generate artificial data as if they were measured in real rooms. This provides additional training examples to the classifiers without the need for any additional data collection, which is time-consuming and often impractical. A representation of acoustic environments is proposed, which is used to train the GANs. The representation is based on a sparse model for the early reflections, a stochastic model for the reverberant tail and a mixing mechanism between the two. In the experiments shown, the proposed data augmentation method increases the test accuracy of a CNN-RNN room classifier from 89.4% to 95.5%.

Constantinos Papayiannis, Christine Evers, Patrick A. Naylor

Reverberation is present in our workplaces, our homes, concert halls and theatres. This paper investigates how deep learning can use the effect of reverberation on speech to classify a recording in terms of the room in which it was recorded. Existing approaches in the literature rely on domain expertise to manually select acoustic parameters as inputs to classifiers. Estimation of these parameters from reverberant speech is adversely affected by estimation errors, impacting the classification accuracy. In order to overcome the limitations of previously proposed methods, this paper shows how DNNs can perform the classification by operating directly on reverberant speech spectra and a CRNN with an attention-mechanism is proposed for the task. The relationship is investigated between the reverberant speech representations learned by the DNNs and acoustic parameters. For evaluation, AIRs are used from the ACE-challenge dataset that were measured in 7 real rooms. The classification accuracy of the CRNN classifier in the experiments is 78% when using 5 hours of training data and 90% when using 10 hours.

Heinrich W. Loellmann, Christine Evers, Alexander Schmidt et al.

Algorithms for acoustic source localization and tracking provide estimates of the positional information about active sound sources in acoustic environments and are essential for a wide range of applications such as personal assistants, smart homes, tele-conferencing systems, hearing aids, or autonomous systems. The aim of the IEEE-AASP Challenge on sound source localization and tracking (LOCATA) was to objectively benchmark state-of-the-art localization and tracking algorithms using an open-access data corpus of recordings for scenarios typically encountered in audio and acoustic signal processing applications. The challenge tasks ranged from the localization of a single source with a static microphone array to the tracking of multiple moving sources with a moving microphone array.

James Eaton, Nikolay D. Gaubitch, Alastair H. Moore et al.

This document provides the results of the tests of acoustic parameter estimation algorithms on the Acoustic Characterization of Environments (ACE) Challenge Evaluation dataset which were subsequently submitted and written up into papers for the Proceedings of the ACE Challenge. This document is supporting material for a forthcoming journal paper on the ACE Challenge which will provide further analysis of the results.

James Eaton, Patrick A. Naylor

Direct-to-Reverberant Ratio (DRR) is an important measure for characterizing the properties of a room. The recently proposed DRR Estimation using a Null-Steered Beamformer (DENBE) algorithm was originally tested on simulated data where noise was artificially added to the speech after convolution with impulse responses simulated using the image-source method. This paper evaluates the performance of this algorithm on speech convolved with measured impulse responses and noise using the Acoustic Characterization of Environments (ACE) Evaluation corpus. The fullband DRR estimation performance of the DENBE algorithm exceeds that of the baselines in all Signal-to-Noise Ratios (SNRs) and noise types. In addition, estimation of the DRR in one third-octave ISO frequency bands is demonstrated.

Pablo Peso Parada, Dushyant Sharma, Toon van Waterschoot et al.

We present a single channel data driven method for non-intrusive estimation of full-band reverberation time and full-band direct-to-reverberant ratio. The method extracts a number of features from reverberant speech and builds a model using a recurrent neural network to estimate the reverberant acoustic parameters. We explore three configurations by including different data and also by combining the recurrent neural network estimates using a support vector machine. Our best method to estimate DRR provides a Root Mean Square Deviation (RMSD) of 3.84 dB and a RMSD of 43.19 % for T60 estimation.

James Eaton, Patrick A. Naylor

Reverberation Time (T60) is an important measure for characterizing the properties of a room. The author's T60 estimation algorithm was previously tested on simulated data where the noise is artificially added to the speech after convolution with a impulse responses simulated using the image method. We test the algorithm on speech convolved with real recorded impulse responses and noise from the same rooms from the Acoustic Characterization of Environments (ACE) corpus and achieve results comparable results to those using simulated data.

James Eaton, Nikolay D. Gaubitch, Alastair H. Moore et al.

Several established parameters and metrics have been used to characterize the acoustics of a room. The most important are the Direct-To-Reverberant Ratio (DRR), the Reverberation Time (T60) and the reflection coefficient. The acoustic characteristics of a room based on such parameters can be used to predict the quality and intelligibility of speech signals in that room. Recently, several important methods in speech enhancement and speech recognition have been developed that show an increase in performance compared to the predecessors but do require knowledge of one or more fundamental acoustical parameters such as the T60. Traditionally, these parameters have been estimated using carefully measured Acoustic Impulse Responses (AIRs). However, in most applications it is not practical or even possible to measure the acoustic impulse response. Consequently, there is increasing research activity in the estimation of such parameters directly from speech and audio signals. The aim of this challenge was to evaluate state-of-the-art algorithms for blind acoustic parameter estimation from speech and to promote the emerging area of research in this field. Participants evaluated their algorithms for T60 and DRR estimation against the 'ground truth' values provided with the data-sets and presented the results in a paper describing the method used.