Jean-Claude Junqua

The influence of acoustics on speech production: a noise-induced stress phenomenon known as the Lombard reflex

Abstract Recently, a number of researchers reported quantitative results about the acoustic changes between normal and Lombard speech. These results highlighted that the nature of the Lombard reflex is highly speaker-dependent. In this paper, after briefly discussing the influence of acoustics on speech production, we summarize some important characteristics of the Lombard reflex. Then, we review some experimental results showing how the Lombard reflex varies with the speaker gender, the language, and the environment (type of noise). Finally, we briefly discuss the use of relational features as a way to reduce the influence of the Lombard reflex on automatic speech recognizers.

Recent advances in biometric person authentication

Biometrics is an emerging topic in the field of signal processing. While technologies (e.g. audio, video) for biometrics have mostly been studied separately, ultimately, biometric technologies could find their strongest role as interwined and complementary pieces of a multi-modal authentication system. In this paper, a short overview of voice, fingerprint, and face authentication algorithms is provided.

Unsupervised hierarchical adaptation using reliable selection of cluster-dependent parameters

Adaptation of speaker-independent hidden Markov models (HMMs) to a new speaker using speaker-specific data is an effective approach to improve speech recognition performance for the enrolled speaker. Practically, it is desirable to flexibly perform the adaptation without any prior knowledge or limitation on the enrolled adaptation data (e.g. data transcription, length and content). However, the inevitable transcription errors may cause unreliability in the model adaptation (or transformation). The variable length and content of adaptation data usually make it necessary to dynamically control the degree of sharing in transformation-based adaptation. This paper presents an unsupervised hierarchical adaptation algorithm for flexible speaker adaptation. We build a tree structure of HMMs such that the control of transformation sharing can be achieved. To perform the unsupervised learning, we apply Bayesian theory to estimate the transformation parameters and data transcription. To select the parameters for hierarchical model transformation, we developed a new algorithm based on the maximum confidence measure (MCM) and minimum description length (MDL) criteria. Experimental comparisons on unsupervised speaker adaptation show that the hybrid adaptation scheme based on MCM and MDL criteria achieves the best recognition results for any lengths of enrollment data.

α-Jacobian environmental adaptation

The robustness of automatic speech recognition systems to noise is still a problem, especially for small footprint systems. This paper addresses the problem of noise robustness using model compensation methods. Such algorithms are already available, but their complexity is usually high. An often-referenced method for achieving noise robustness is parallel model combination (PMC). Several algorithms have been proposed to develop more computationally efficient methods than PMC. For example, Jacobian adaptation approximates PMC with a linear transformation function in the cepstral domain. However, the Jacobian approximation is valid only for test environments that are close to the training conditions whereas, in real test conditions, the mismatch between the test and training environments is usually large. In this paper, we propose two methods, respectively called static and dynamic α-Jacobian adaptation (or α-JAC), to compute new linear approximations of PMC for realistic test environments. We further extend both algorithms to compensate for additive and convolutional noise and we derive the corresponding non-linear algorithm that is approximated. All these algorithms are experimentally compared in important mismatch conditions. As compared to Jacobian adaptation, improvements are observed with both static and dynamic α-Jacobian adaptation.

Blind channel estimation based on speech correlation structure

Cepstral mean normalization is the standard technique for channel robustness. Despite its good performance, the effectiveness of cepstral mean normalization (CMN) for short sentences is argued. CMN underlying hypothesis that the speech cepstral mean is constant is not valid for short processing windows. This implies the removal of some phonetic information. In this paper we show that the speech correlation structure may be used to estimate the communication channel and we propose an efficient algorithm to compute this estimate. We argue that the resulting channel estimate is more accurate because the underlying hypothesis is better verified than the original CMN hypothesis. Results for the Kai-Fu Lee phone recognition task on NTIMIT, with acoustic models trained on TIMIT (mismatch conditions), show that our method provides an 8% relative error rate reduction as compared to CMN.

Editorial of Special issue on adaptation methods for speech recognition

This special issue of the Arab World English Journal focuses on Computer-Assisted Language Learning: innovative uses of ICT, impact and modes of integration in the classroom. Referring to acronyms such as ICT or CALL, or generic terms such as “technology” or “computers”, seems to point to a well-identified area of research and practice in education. Yet, this terminology may well be the cause of an ontological illusion leading readers of the scientific literature to think that CALL is a unified field and that the results of research may therefore be valid for any situation.

Latitudes: Where speech met the beach the time of a workshop

This invention is a composition of matter and method of preparing the same. More specifically the present invention is a medicated vapor candle consisting of a homogenous mixture having approximately three parts petroleum jelly containing the active ingredients of camphor, menthol and eucalyptus, melted and mixed with four parts candle wax. The active ingredients, by volume, of approximately 5% camphor, 2.5% menthol and 1.2% eucalyptus oil.

A knowledge engineering approach to speech processing using a blackboard model: Application to automatic labelling

Abstract In many applications related to speech processing various types of knowledge are often combined to try to solve a particular problem. This comes mainly from the fact that our knowledge about speech in one particular domain is very limited. This paper focuses on the automatic segmentation and labelling of speech with a method which uses several knowledge sources. We propose a new segmentation and labelling approach which relies on the co-operation between multi-knowledge sources (heuristic, phonetic, suprasegmental) through a blackboard model controlled by a hierarchical strategy. The blackboard model allowed us to develop the system in a modular way in order to facilitate its adaptation to different applications as the automatic segmentation of a speech database or automatic labelling in the case of speech recognition. The blackboard is composed of instances of speech objects hierarchically organized into levels of analysis. A knowledge consistency verification module has been developed to deal with hypothetical cases and data dependency.