Synthetic speech is occasionally required to be enhanced by a fixed amount before being presented to the listener. However, this approach neglects the diverse types and levels of noise, potentially resulting in either unintelligible or unpleasant speech. This paper proposes a dynamic speech enhancement approach, which generates speech while accounting for the presence of noise. Our methodology extends the WaveRNN vocoder by conditioning not only the speech mel-spectrograms but also the spectrogram from the background noise. In training, the target speech is tilted according to the listeners’ preference and we further introduce an objective metric that aims at maximizing the spectro-temporal regions where target energy exceeds the masker energy. The generated speech was tested in speech-shaped noise at various noise levels. Our evaluation results showed that (a) in more adverse conditions, both a fixed-amount post-enhanced baseline system and the suggested dynamically-enhanced system performed equally well in terms of intelligibility and preference and (b) in the least noisy conditions, where the intelligibility scores of all models were nearly identical, listeners preferred the dynamically-enhanced speech. Our findings demonstrate and reinforce the benefits of using dynamic speech enhancement techniques in noisy environments.

This paper describes a system for discriminating innocent from pathologic systolic heart murmurs in children based on auscultation recordings. For sound signal analysis the use of reassigned spectrogram is suggested. Both dimensions and noise of the time-frequency representation were significantly reduced using higher order singular value decomposition. Optimal dimensions were selected through cross-validation experiments on a database of auscultation recordings with systolic murmurs from the University Hospital of Heraklion. The database only consisted with recordings of high misclassification rate by general practitioners. Using support vector machines for classification, the suggested approach achieved an Equal Error Rate of 6.71 ± 1.18% and an Area Under the Curve score of 0.9758 ± 0.0053 (95% confidence intervals). The performance of the suggested classification system is comparable to the reported accuracy of experienced pediatric cardiologists on the same database, while it outperforms alternative signal representations based on simple STFT schemes.

In this paper, we explore the information provided by a joint acoustic and modulation frequency representation, referred to as modulation spectrum, for detection and discrimination of voice disorders. The initial representation is first transformed to a lower dimensional domain using higher order singular value decomposition (HOSVD). From this dimension-reduced representation a feature selection process is suggested using an information-theoretic criterion based on the mutual information between voice classes (i.e., normophonic/dysphonic) and features. To evaluate the suggested approach and representation, we conducted cross-validation experiments on a database of sustained vowel recordings from healthy and pathological voices, using support vector machines (SVMs) for classification. For voice pathology detection, the suggested approach achieved a classification accuracy of 94.1±0.28% (95% confidence interval), which is comparable to the accuracy achieved using cepstral-based features. However, for voice pathology classification the suggested approach significantly outperformed the performance of cepstral-based features.

In this paper, we combine modulation spectral features with mel-frequency cepstral coefficients for automatic detection of dysphonia. For classification purposes, dimensions of the original modulation spectra are reduced using higher order singular value decomposition (HOSVD). Most relevant features are selected based on their mutual information to discrimination between normophonic and dysphonic speakers made by experts. Features that highly correlate with voice alterations are associated then with a support vector machine (SVM) classifier to provide an automatic decision. Recognition experiments using two different databases suggest that the system provides complementary information to the standard mel-cepstral features.

In this paper, we employ normalized modulation spectral anal-ysis for voice pathology detection. Such normalization is im-portant when there is a mismatch between training and testing conditions, or in other words, employing the detection system in real (testing) conditions. Modulation spectra usually produce a high-dimensionality space. For classification purposes, the size of the original space is reduced using Higher Order Singular Value Decomposition (SVD). Further, we select most relevant features based on the mutual information between subjective voice quality and computed features, which leads to an adap-tive to the classification task modulation spectra representation. For voice pathology detection, the adaptive modulation spectra is combined with an SVM classifier. To simulate the real test-ing conditions; one for training and the other for testing. We address the difference of signal characteristics between training and testing data through subband normalization of modulation spectral features. Simulations show that feature normalization enables the cross-database detection of pathological voices even when training and test data are different. Index Terms: pathologic voice detection, modulation spec-trum, feature normalization, mutual information, SVD.

We describe a dimensionality reduction method for modulation spectral features, which keeps the time-varying information of interest to the classification task. Due to the varying degrees of redundancy and discriminative power of the acoustic and mod-ulation frequency subspaces, we first employ a generalization of SVD to tensors (Higher Order SVD) to reduce dimensions. Projection of modulation spectral features on the principal axes with the higher energy in each subspace results in a compact feature set. We further estimate the relevance of these projec-tions to speech discrimination based on mutual information to the target class. Reconstruction of modulation spectrograms from the “best ” 22 features back to the initial dimensions, shows that modulation spectral features close to syllable and phoneme rates as well as pitch values of speakers are preserved. Index Terms: modulation spectrum, multilinear algebra, fea-ture selection, mutual information, speech discrimination

In this work, we adopt an information theoretic approach- the Information Bottleneck method- to extract the relevant spectro-temporal modulations for the task of speech / non-speech dis-crimination- non-speech events include music, noise and an-imal vocalizations. A compact representation (a “cluster pro-totype”) is built for each class consisting of the maximally in-formative features with respect to the classification task. We assess the similarity of a sound to each representative cluster using the spectro-temporal modulation index (STMI) adapted to handle the contribution of different frequency bands. A sim-ple threshold check is then used for discriminating speech from non-speech events. Conducted experiments have shown that the proposed method has low complexity and high accuracy of dis-crimination in low SNR conditions compared to recently pro-posed methods for the same task. Index Terms: audio classification, speech discrimination, au-ditory model

In next-generation IP networks, there will be a mixture of traffic classes with different QoS requirements. Differentiated Services (DiffServ) has evolved from the IETF's earlier Integrated Services (IntServ) standard as a solution to the service quality challenges. In this paper, a novel buffer management scheme called CBQ priority buffer management (CPBM) is presented for IP routers, using the class based queueing (CBQ) scheduler, in a combined IntServ and DiffServ architecture. The objectives of the CPBM scheme are to cope with the various QoS requirements of the IP layer service classes, to utilize the buffer space as efficiently as possible and to successfully coordinate with the CBQ scheduling mechanism in order to improve packet loss performance. The proposed CPBM scheme has been compared with conventional buffer management schemes as complete partitioning (CP), push-out (PO) and push-out with threshold (POT), using OPNET release 5.1.D as a simulation platform.

The design of a buffer management algorithm, with controllable quality-of-service (QoS) commitments in terms of packet loss, is addressed for IP routers in a differentiated services architecture. The objectives of the proposed policy are to utilize the buffer space efficiently to coordinate with the class-based queuing (CBQ) scheduling algorithm adopted by the IP routers, and mainly to be able to adjust the quality spacing between the different QoS traffic classes. Using OPNET Modeler 6.0.L as a simulation platform, the proposed scheme is compared along with conventional buffer management policies, such as complete partitioning (CP), push-out (PO) and push-out with threshold (POT). Numerical results demonstrate the controllable relative differentiation between classes achieved by the proposed scheme.

In this short paper an approach to the ATM hierarchical (PNNI) routing issue is presented based on a simple algorithm that chooses a route for a connection request out of a predefined list of possible routes between a specific source-destination pair in a probabilistic way. A route fitness (cost) function is used to assign selection probabilities to every candidate path and arriving connection requests are routed independently according to these path probabilities. This cost function has been chosen in a way to achieve a successful trade-off between the use of minimum-hop routes and the application of the load-balancing concept. Moreover we show that the performance of the proposed algorithm and other probabilistic routing algorithms can be enhanced with the application of scaling techniques that are used in genetic algorithms (GAs) maximizing the network revenue. Simulation results over a wide range of uniform, time-varying and skewed loading conditions show the effectiveness of the proposed routing algorithm.

The evolving broadband integrated services digital network is reinforcing the demand for high-speed and high-performance multiple access networks. The number of channels available to support the isochronous traffic in these networks is limited by technology, due to implementation costs. We introduce a method using channel sharing/reusing in an effort to provide efficient management of isochronous traffic under this limitation. The proposed method is based on a hybrid genetic algorithm and aims to accomplish the establishment of a maximal number of connections with the minimal number of isochronous channels. Experimental results are provided and they are compared with those of a deterministic graph coloring algorithm. The performance of the proposed algorithm in all simulation runs reveals the robustness, the flexibility and the efficiency of using evolutionary approaches to complex real-world problems.

Protocols such as the real-time transport protocol (RTP), are used to provide suitable voice/video communication channels over the Internet. RTP transports real-time media data along with synchronization information over a datagram protocol such as the user datagram protocol (UDP). The Casner-Jacobson (see Internet-Draft, 1997) RTP/UDP/IP header compression algorithm can significantly reduce the overhead and increase the throughput of a connection. We evaluate the behavior of header compression in noisy links. It is shown that error indication and error correction can improve the resilience of header compression over noisy links.