Experiments in syllable-based recognition of continuous speech

Abstract

An exploratory implementation of a syllable-based recognizer is described. Continuous speech is first divided into syllabic unite, and the units are then matched against syllable templates using a dynamic programming algorithm. A hierarchical transition network is used to limit the syllable search to possible continuations of the current partial sentence hypotheses. Competing hypotheses are pruned by a 'beam search'.Experiments are reported on automatic recognition of English sentences with a 70-word vocabulary and restricted syntax produced by one male speaker. 85% of the sentences were correctly recognized.Comparable results were obtained for a similar task in French using a female speaker. The method is computationally efficient: real-time performance on dedicated hardware should bs obtainable without difficulty. A method of scaling the distance measures used in the syllable matching is described. This scaling takes into account variability in syllable production, both as a function of position within the syllable and as a function of the various spectral parameters being used. 1, INTRODUCTIONAt some level in a continuous speech recognition system there is an interface between acoustically derived information about the speech to be recognized and syntactic constraints on the interpretation of that information. For maximum efficiency in the application of the constraints, the interface level should be as low as possible. At the lowest levels, however, the interpretation of acoustic information is highly context dependent. The syllable level is attractive because it is low enough to make the inventory of units to be recognized reasonably small while being high enough to make the phonetic interpretation of each unit reasonably context independent.In addition, syllable boundaries cam be determined independently of the task of identifying the syllable. This property, which does not hold for words or for phonemes, makes for efficient comparison of syllables with stored reference forms and for efficient evaluation of competing sentence hypotheses. The theoreticsl advantages of the syllable-level representation are discussed by Msrmslstein (1) and Fujimura (2). Davis and Mermelstein (3) report results on syllable-based recognition assisted by lexical information but using CH1559-418010000-0880$O0.75 © 1980 IEEE 880 hand-segmented speech data. Zwicker, Terhardt and Psulus (4) use a. similar syllable-based psychoacoustics-oriented scheme for the recognition of German city names.The experiments reported pertain to speakerdependsnt recognition of pra-specified sentences produced with a limited vocabulary and restricted syntax. One male speaker was used to train and test a 70-word vocabulary English system intended to recognize 'date and time' sentences, e.g., "January fourth at eight forty-five p.m.". A female speaker was used to test and train a conpsrabls French systen. The French vocabulary comprised 43 words and an additional 25 word variants corresponding to sequences of two or more words in which syllables had been ...

Graphical models for integrating syllabic information

“…The experiments in Zwicker et al [1979] automatically segmented speech by half syllables by detecting both syllable boundaries and peaks of nuclei. Speech was automatically segmented into whole syllables using onset estimations in Hunt et al [1980]. All of these methods made hard decisions about syllable locations.…”

“…This work proposed that speech should be characterized in terms of syllable onset, nucleus, and coda along with sub-classes within each of these sub-syllable units. This was actually deployed in a dynamic template matching system in Zwicker et al [1979] where template matching was used to recognize half-syllable units, and in Hunt et al [1980] where a similar approach was used on whole-syllable units. Syllables were employed as the basic recognition unit in an HMM based system in Green et al [1993].…”

Use of syllable nuclei locations to improve ASR

“…Note that since this is based only on a count, there is flexibility in the exact placement in time of the syllable centers. Thus, the location information is used, but it does not need to be as precise as methods that segment the utterance based on syllable onsets [6,8]. The motivation for this is that the exact placement of the onset may not be well defined due to coarticulation [19].…”

“…The syllable was proposed as a basic unit of recognition as early as 1975 [5]. In [6], the utterances were segmented via syllable onset estimations as a precursor to template matching, and in [7] syllables were employed as the basic recognition unit in an HMM . The most closely related method to this paper was presented by Wu in 1997 [8,9].…”

Operações de consulta por similaridade em grandes bases de dados complexos

“…Dentre os exemplos de características extraídas a partir do resultado da aplicação desses métodos estão as características baseadas na transformada STFT, descritas em [Wold et al, 1996, Tzanetakis, 2002, e os Mel Frequency Cepstral Coefficients (MFCC) [Hunt et al, 1980]. Esse tipo de características pode ser utilizado para capturar aspectos relacionados ao timbre de um sinal deáudio e pode ser empregado tanto em aplicações de análise da fala como em aplicações de análise de músicas [Tzanetakis, 2002].…”