Speech coding and synthesis using parametric curves

Formant trajectories in the vicinity of consonants in continuous speech are observed to discriminate the place of articulation of the consonants. The coarticulation model [S. Chiba, Proc. IEEE ICASSP‐86, Paper 49‐10 (1986)] with a target formant frequency for consonants is developed based on the observation of the formant trajetories in natural VCV utterances. The coarticulation model proposed here is as follows: Fb − Ft = a(Fp − Ft) + b(Fe − Ft) + c(Ff − Ft) + E, where Fb, Fp, Fe, and Ff are the second formant frequencies at the centers and the consonant boundaries of the initial and final vowel segments in a VCV sequence, Ft is the target second formant frequency of the consonant, and E is the estimation error. The parameter values of a, b, c, and Ft are estimated by the method of least squares using actual values. Three coarticulation models for labials, alveolars, and velars are estimated and these models are applied to the discrimination of place of articulation of consonants in continuous speech. Ex...

GPU-based rendering of curved polygons using simplicial coverings

In this paper, we describe BYBLOS, the BBN continuous speech recognition system. The system, designed for large vocabulary applications, integrates acoustic, phonetic, lexical, and linguistic knowledge sources to achieve high recognition performance. The basic approach, as described in previous papers [1, 2], makes extensive use of robust context-dependent models of phonetic coarticulation using Hidden Markov Models (HMM). We describe the components of the BYBLOS system, including: signal processing frontend, dictionary, phonetic model training system, word model generator, grammar and decoder. In recognition experiments, we demonstrate consistently high word recognition performance on continuous speech across: speakers, task domains, and grammars of varying complexity. In speaker-dependent mode, where 15 minutes of speech is required for training to a speaker, 98.5% word accuracy has been achieved in continuous speech for a 350-word task, using grammars with perplexity ranging from 30 to 60. With only 15 seconds of training speech we demonstrate performance of 97% using a grammar.

Approximation of helices has been studied by using in many ways. In this study, it has been examined how a circular helix can be written as Bezier curve and written the 3th degree, 5th degree, and the 7th degree Maclaurin series expansions of helices for the polynomial forms. Hence, they can be written cubic, 5th order, and 7th order Bezier curves, based on the control points with matrix form we have already given in E3. Further we have given the control points of the Bezier curve based on the coefficients of the Maclaurin series expansion of the circular helix.

Speech articulation has been modeled to be a kind of compromise between ease of articulation for the speaker and speech signal clarity that is needed by the listener. This compromise is most clearly reflected in Lindblom's Hyper- and Hypoarticulation theory of speech, in which speech realization is the result of speaker- and listener-oriented constraints. Coarticulation means the articulation of different sounds at the same time; another term being coproduction. Several coarticulation models have been developed since the sixties of the 20th century. One model which is still influential is Keating's window model. Categorical perception has been a widely discussed phenomenon since its introduction by Liberman, Harris, Hoffman and Griffith and its official labeling as such by Eimas. The basis was the interesting "relation" between classification and discrimination of speech sounds on a continuum of stimuli. Several claims have been made regarding categorical perception. One of those claims is that categorical perception is limited to human speech.

Speech segmentation is an important sub problem of automatic speech recognition. This research is concerned with the development of a continuous speech segmentation system using Bangla Language. This paper presents a dynamic thresholding algorithm to segment the continuous Bngla speech sentences into words/sub-words. The research uses Otsu’s method for dynamic thresholding and introduces a new approach, named blocking black area method to identify the voiced regions of the continuous speech in speech segmentation. The developed system has been justified with continuously spoken several Bangla sentences. To test the performance of the system, 100 Bangla sentences have been recorded from 5 (five) male speakers of different ages and 656 words have been presented in the 100 Bangla sentences. So, the speech database contains 500 Bangla sentences with 3280 words. All the algorithms and methods used in this research are implemented in MATLAB and the proposed system has been achieved the average segmentation accuracy of 90.58%.

In this paper we present two path planning algorithms based on Bezier curves for autonomous vehicles with waypoints and corridor constraints. Bezier curves have useful properties for the path generation problem. The paper describes how the algorithms apply these properties to generate the reference trajectory for vehicles to satisfy the path constraints. Both algorithms join cubic Bezier curve segments smoothly to generate the path. Additionally, we discuss the constrained optimization problem that optimizes the resulting path for a user-defined cost function. The simulation shows the generation of successful routes for autonomous vehicles using these algorithms as well as control results for a simple kinematic vehicle. Extensions of these algorithms towards navigating through an unstructured environment with limited sensor range are discussed.

We report on some recent improvements to an HMM-based, continuous speech recognition system which is being developed at AT&T Bell Laboratories. These advances, which include the incorporation of inter-word, context-dependent units and an improved feature analysis, lead to a recognition system which achieve better than 95% word accuracy for speaker independent recognition of the 1000-word, DARPA resource management task using the standard word-pair grammar (with a perplexity of about 60). It will be shown that the incorporation of inter-word units into training results in better acoustic models of word juncture coarticulation and gives a 20% reduction in error rate. The effect of an improved set of spectral and log energy features is to further reduce word error rate by about 30%. We also found that the spectral vectors, corresponding to the same speech unit, behave differently statistically, depending on whether they are at word boundaries or within a word. The results suggest that intra-word and inter-word units should be modeled independently, even when they appear in the same context. Using a set of sub-word units which included variants for intra-word and inter-word, context-dependent phones, an additional decrease of about 10% in word error rate resulted.

An ongoing study is reported of all sixteen of the American English vowels using subsets of the DARPA acoustic-phonetic database. Formants are obtained and normalized for each talker's formant range based on one sentence. The resulting formant tracks are smoothed using splines and sampled at nine equally spaced points in time within vowel-centered triphone regions. Triphones with semivowels in them are clustered separately. These formant values are k-means clustered using subsets of the sampled formant values. The additional supervised training is done using other parameters, including duration. The resulting clusters are used as a classifier on the basis of the modified Euclidean distance from the cluster centers. This results in approximately 80% first choice vowel recognition of the outer edges of the vowel quadrilateral. Stressed vowels were found to have spectra which statistically were no more stable than unstressed vowels. >

Continuous speech is far more natural and efficient than isolated speech for communication. However, for current state-of-the-art automatic speech recognition systems, isolated speech recognition (ISR) is far more accurate than continuous speech recognition (CSR). It is common practice in the speech research community to build CSR systems using only CS data. However, slowing of the speaking rate is a natural reaction for a user faced with the high error rates of current CSR systems. Ironically, CSR systems typically have a much higher word error rate when speakers slow down since the acoustic models are usually derived exclusively from continuous speech corpora. In this paper, we summarize our efforts to improve the robustness of our speaker-independent CSR system against speaking styles, without suffering a recognition accuracy penalty. In particular the multi-style trained system described in this paper attains a 7.0% word error rate for a test set consisting of both isolated and continuous speech, in contrast to the 10.9% word error rate achieved by the same system trained only on continuous speech.

The Hungarian spontaneous speech recording and annotation subproject is being carried out by our Computational Linguistics research group and my PhD work at the University of Debrecen and is a part of a comprehensive multimodal human-machine interaction development project and multimodal (audio and video) database collection. The efficiency of speech recognition systems can be increased by proper acoustic preprocessing and by investigation of the suprasegmental characteristics of spontaneous speech. The research aims to contribute to the exact knowledge of prosody through the examination of spontaneous speech, with special regard to syntactic embeddings, insertions, iterations, hesitations and restarts, various kinds of emotions and discourse markers regarding Hungarian, the lack of a prosodically labelled, representative spontaneous speech database makes the development more difficult. The spontaneous multimodal database is being recorded via guided formal and informal conversations. During the conversation, several points are to be discussed in order to provoke longer monologues, including those phenomena of spontaneous speech, which are to be examined within our research. Designing a continuous spontaneous speech recognition system that is speaker-independent and is able to contribute to our theoretical assumptions, requires the construction of a speech database for which we need to take several personnel and technical aspects into account. The visual channel also needs to be annotated, which will enable us to examine and implement multimodal features as well.Keywordsdatabase planningspontaneous speechprosody researchmultimodality

Automatic and accurate detection of the closure-burst transition events of stops and affricates serves many applications in speech processing. A temporal measure named the plosion index is proposed to detect such events, which are characterized by an abrupt increase in energy. Using the maxima of the pitch-synchronous normalized cross correlation as an additional temporal feature, a rule-based algorithm is designed that aims at selecting only those events associated with the closure-burst transitions of stops and affricates. The performance of the algorithm, characterized by receiver operating characteristic curves and temporal accuracy, is evaluated using the labeled closure-burst transitions of stops and affricates of the entire TIMIT test and training databases. The robustness of the algorithm is studied with respect to global white and babble noise as well as local noise using the TIMIT test set and on telephone quality speech using the NTIMIT test set. For these experiments, the proposed algorithm, which does not require explicit statistical training and is based on two one-dimensional temporal measures, gives a performance comparable to or better than the state-of-the-art methods. In addition, to test the scalability, the algorithm is applied on the Buckeye conversational speech corpus and databases of two Indian languages.

The development and evaluation of large vocabulary, speaker-independent continuous speech recognition systems are mainly done for the American English language. In this paper we present the work done to date in the development of an hybrid large vocabulary, speaker-independent continuous speech recognition system for the European Portuguese language. Due to the lack of a large appropriate speech and text database to be used in the development of that system we started collecting a large database and at the same time began developing a baseline system based on a smaller database. On this baseline system we applied techniques for automatic segmentation and labeling, in parallel with the development of a basic lexicon and language model for Portuguese. In the last part of this paper we also present the rst steps of our work over the new database.

This work deals with the packet loss problem in a Distributed Speech Recognition architecture. A packet loss simulation model is first proposed in order to simulate different channel degradation conditions. In these conditions, the performance of our continuous French speech recognition system is evaluated for packets containing different numbers of speech feature vectors. Several reconstruction strategies, to recover lost information, are proposed and evaluated. The results first confirm the intuitive fact that the word error rate obviously increases with the size of the lost packets and with the channel degradation level. However, it is shown that simple reconstruction strategies allow to recover acceptable performance. The most efficient ones are those using interleaving technique to distribute the speech information among packets, combined with interpolation methods to estimate lost acoustic features.

Emotion recognition or affect detection from speech is an old and challenging problem in the field of artificial intelligence. Many significant research works have been done on emotion recognition. In this paper, the recent works on affect detection using speech and different issues related to affect detection has been presented. The primary challenges of emotion recognition are choosing the emotion recognition corpora (speech database), identification of different features related to speech and an appropriate choice of a classification model. Different types of methods to collect emotional speech data and issues related to them are covered by this presentation along with the previous works review. Literature survey on different features used for recognizing emotion from human speech has been discussed. The significance of various classification models has been presented along with some recent research works review. A detailed description of a prime feature extraction technique named Mel Frequency Cepstral Coefficient (MFCC) and brief description of the working principle of some classification models are also discussed here. In this paper terms like affect detection and emotion recognition are used interchangeably.