Chinese Syllables Research Articles

The Left and Right Context of a Word

Since a Chinese syllable can correspond to many characters (homophones), the syllable-to-character conversion task is quite challenging for Chinese phonetic input methods (CPIM). There are usually two stages in a CPIM: 1. segment the syllable sequence into syllable words, and 2. select the most likely character words for each syllable word. A CPIM usually assumes that the input is a complete sentence, and evaluates the performance based on a well-formed corpus. However, in practice, most Pinyin users prefer progressive text entry in several short chunks, mainly in one or two words each (most Chinese words consist of two or more characters). Short chunks do not provide enough contexts to perform the best possible syllable-to-character conversion, especially when a chunk consists of overlapping syllable words. In such cases, a conversion system often selects the boundary of a word with the highest frequency. Short chunk input is even more popular on platforms with limited computing power, such as mobile phones. Based on the observation that the relative strength of a word can be quite different when calculated leftwards or rightwards, we propose a simple division of the word context into the left context and the right context. Furthermore, we design a double ranking strategy for each word to reduce the number of errors in Step 1. Our strategy is modeled as the minimum feedback arc set problem on bipartite tournament with approximate solutions derived from genetic algorithm. Experiments show that, compared to the frequency-based method (FBM) (low memory and fast) and the conditional random fields (CRF) model (larger memory and slower), our double ranking strategy has the benefits of less memory and low power requirement with competitive performance. We believe a similar strategy could also be adopted to disambiguate conflicting linguistic patterns effectively.

Modeling pitch contour of Chinese Mandarin sentences with the PENTA model

In continuous speech, the pitch contour of the same syllable may vary much due to its contextual information. The Parallel Encoding and Target Approximation (PENTA) model is applied here to Mandarin speech synthesis with a method to predict pitch contours for Chinese syllables with different contexts by combining the Classification And Regression Tree (CART) with the PENTA model to improve its prediction accuracy. CART was first used to cluster the syllables' normalized pitch contours according to the syllables contextual information and the distances between pitch contours. The average pitch contour was used to train the PENTA model with the average contour for each cluster. The initial pitch is required with the PENTA model to predict a continuous pitch contour. A Pitch Discontinuity Model (PDM) was used to predict the initial pitches at positions with voiceless consonants and prosodic boundaries. Initial tests on a Chinese four-syllable word corpus containing 2048 words were extended to tests with a continuous speech corpus containing 5445 sentences. The results are satisfactory in terms of the Root Mean Square Error (RMSE) comparing the predicted pitch contour with the original contour. This method can model pitch contours for Mandarin sentences with any text for speech synthesis.

Aphasia Assessment in Chinese Speakers

Aphasia Assessment in Chinese Speakers

Learning a Tonal Language by Attending to the Tone: An In Vivo Experiment

Learning the Chinese tone system is a major challenge to students of Chinese as a second or foreign language. Part of the problem is that the spoken Chinese syllable presents a complex perceptual input that overlaps tone with segments. This complexity can be addressed through directing attention to the critical features of a component (tone in this case) within a complex perceptual input stimulus. We tested hypotheses based on this feature‐focusing assumption in an in vivo classroom setting. First‐year students in a Chinese language program at a U.S. university were trained to identify the tones of 228 syllables learned across eight lessons in the first semester. Three learning conditions were designed to support tone learning by presenting (a) visual pitch contours that depict the acoustic shape of the tones, together with pinyin spelling of the spoken syllables (Contour + Pinyin condition); (b) numbers that represent the tones in traditional computer interface, together with pinyin spelling of the spoken syllables (Number + Pinyin condition); and (c) visual pitch contours without pinyin spelling (Contour Only condition). Analyses of student activity logs (learning curves) and pretests and posttests showed significant effects of learning condition. The results suggested that the Contour + Pinyin condition had more error reduction in tone recognition over the activity log than the Contour Only condition and greater improvement from the pretest to posttest than the Number + Pinyin condition. These findings point at the value of separate support for the two major components (tone and segments) of a tonal language.

Cognitive mechanism of writing to dictation of logographic characters

ABSTRACTThe cognitive mechanisms for writing to dictation of Chinese syllables by healthy adults were investigated using large-sample multiple regression analyses. In the experiment, subjects wrote down a corresponding character upon hearing a syllable. We mainly examined the effects of three types of attributes (i.e., lexical, semantic, and phonology to orthography conversion [POC] ones) in predicting the production probability of specific characters out of the homophone families for target syllables. We observed significant effects for all three types of attributes, as well as interactions between POC and the lexical attributes, and between POC and the semantic attributes. We further found that the semantic effects vanished for the writing stimuli without homophones. A feedback procedure (i.e., phonetic radical transparency) was also observed to influence Chinese writing performances. Our results support the hypothesis that the extent of semantic involvement in writing (spelling) to dictation is influenced by the effectiveness of POC procedure in a certain language and/or word set. The existence of an interaction between the lexical semantic route and the POC route in writing is further consolidated.

Dynamic Chinese viseme model based on phones and control function

The aspirate and pronunciation of a Chinese syllable is similar to an `olivary nucleus` and syllables are mildly connected by coarticulation. Accordingly a dynamic viseme model is presented. Inner-syllable visemes are described by initial-final phones, and inter-syllable visemes are determined by a hierarchical control function. Experimental results show improvement in naturalness of lip animation, as compared to the visual speech synthesis system based on triphones.

Rules for Connecting the Tones of Two-Syllable Words in Beijing Mandarin

When two Chinese syllables are uttered sequentially, the tone of each syllable has its own pitch and contour. Two adjacent tones in spontaneous speech must overcome the contradictions in pitch and contour in order to form a tonally smooth unit, and there must be some rules governing the process. Taking the tonal pattern of two-syllable words in Beijing Mandarin as the object of study and based on analyses of experimental phonetics, this paper reveals three rules in connecting two tones in actual speech, namely connection at the same pitch level, connection with pitch adjustment, and connection with highlighting.

Rules for Connecting the Tones of Two-Syllable Words in Beijing Mandarin

When two Chinese syllables are uttered sequentially, the tone of each syllable has its own pitch and contour. Two adjacent tones in spontaneous speech must overcome the contradictions in pitch and contour in order to form a tonally smooth unit, and there must be some rules governing the process. Taking the tonal pattern of two-syllable words in Beijing Mandarin as the object of study and based on analyses of experimental phonetics, this paper reveals three rules in connecting two tones in actual speech, namely connection at the same pitch level, connection with pitch adjustment, and connection with highlighting.

Effects of language background on tonal perception for young school‐aged children.

Given that Chinese is a tonal language while English is not, the present study investigates how language background affects tonal perception for young children. Tonal identification and discrimination are measured on for three groups of school‐aged children (6–7 years old): Chinese‐monolingual (CM), English‐monolingual (EM), and English‐Chinese bilingual (ECB) children. Children’s task is to identify and discriminate tone 1 (level), tone 2 (rising), and tone 4 (falling) of a Chinese syllable /ma/, for which the fundamental frequency contour is systematically manipulated from tone 1 to tone 2 and from tone 1 to tone 4. CM and ECB children show typical categorical perception in both identification and discrimination, while EM children cannot identify and discriminate the three tones. These results suggest that learning and exposure experience of the tonal language is critical for children to perceive tonal changes in speech sounds.

Neural correlates of speech and non-speech processing: Role of language experience in brain activation

In adults, neural responses to the acoustic properties of native and non-native speech sounds differ. Recent magnetoencephalography (MEG) studies indicate that two regions of the brain, the superior temporal area and inferior parietal area (Broca's), in both the left and right hemispheres, may contribute to how the adult brain responds to the acoustic properties of speech. The aim of the present study was to investigate neural activation in adults when listening to native and non-native speech, and to non-speech analogs of speech. We used MEG to examine activation of the superior temporal and inferior parietal areas. Native speakers of Finnish were presented with Finnish (/pa/ and /ta/) and Chinese (/t "looped c" "superscript h" i/ and /"looped c" i/) syllables, and sine-wave replicas of /pa/ and /ta/, in a passive oddball paradigm. Our analysis examines the properties of localized neural activities to determine how activation of sensory and motor speech areas differs for speech and non-speech, and the role of language experience in brain activation.

Processing of Segmental and Tonal Information of Chinese Syllables

In the phonological system of Chinese, a unique Chinese syllable usually comprises three components: onset, rime, and tone. The present study, with 30 participants, examined how native Chinese listeners process these subsyllabic components of Chinese by a syllable comparison task. Analysis showed that the cognitive processing of these subsyllabic components was different after the encoding process of the Chinese syllable. These results are discussed in light of the perception of Chinese language.

Enhancement of Chinese speech based on nonlinear dynamics

Based on recently observed nonlinear dynamic features of human speech, the local projection (LP) method, originally developed for noisy chaotic time series, is generalized and adapted to the enhancement of Chinese speech. The analysis of minimum embedding dimensions estimated by the false nearest neighbor algorithm shows that all the basic phonemes and syllables in Chinese can be faithfully embedded in some low-dimensional phase space. Over-embedding is applied to reconstruct the dynamics of continuous speech in some extended phase space of higher dimension, thus solving the problem of nonstationarity in continuous speech. A generalization of the LP method, named the local subspace method, is presented for speech enhancement in the phase space. It is demonstrated that, the local subspace method is essentially an extension of the well-known linear subspace technique in the local phase space, and the LP method is the least square case of this generalization. Noise reduction is then carried out in the local phase space. Results show that the LP method, with 2 or 3 iterations, achieves better performances than the local subspace method. For both isolated and continuous speech with additive white noise, experiments show the superiority of the LP method over two other popular algorithms.

Tone sandhi effect on Chinese speech perception

Mandarin Chinese has four tones with various sandhi processes. The third tone sandhi process is when a third tone (T3) becomes a second tone (T2) before another third tone (T3). Previous studies on third tone sandhi used connected speech in their experiments. This study examines the third tone sandhi effect on Chinese listeners in a nonconnected speech condition. Full‐tone Chinese syllables were recorded in isolation. Ten Chinese listeners participated in tonal discrimination tasks which consisted of combinations of T2 and T3 (i.e., T22, T23, T32, T33). Three interstimulus intervals (ISIs) (i.e., 50, 250, 500 ms) and three types of pair conditions (minimal‐, rhyme‐, mixed‐pairs) were introduced in the experiments. The findings of this study confirm the phonological hypothesis that the third tone sandhi effect has a great impact across all the pair conditions. When discriminating T23, Chinese listeners had longer response times and more errors than other tone pairs. The phonological phenomenon of third tone sandhi persists regardless of connected or nonconnected speech conditions in Chinese spoken language. Differences in processing the tone sandhi pairs (T23) and the nontone sandhi pairs (T32) reveal the psychological reality of the third tone sandhi phonological rule.

Orthographic input and phonological representations in learners of Chinese as a foreign language

This paper provides evidence that the second language orthographic input affects the mental representations of L2 phonology in instructed beginner L2 learners. Previous research has shown that orthographic representations affect monolinguals’ performance in phonological awareness tasks; in instructed L2 learners such representations could also affect pronunciation. This study looked at the phonological representations of Chinese rimes in beginner learners of Chinese as a foreign language, using a phoneme counting task and a phoneme segmentation task. Results show that learners do not count or segment the main vowel in those syllables where it is not represented in the pinyin (romanisation) orthographic representations. It appears that the pinyin orthographic input is reinterpreted according to L1 phonology–orthography correspondences, and interacts with the phonological input in shaping the phonological representations of Chinese syllables in beginner learners. This explains previous findings that learners of Chinese do not pronounce the main vowel in these syllables.

Auditory and speech processing and reading development in Chinese school children: behavioural and ERP evidence

By measuring behavioural performance and event-related potentials (ERPs) this study investigated the extent to which Chinese school children's reading development is influenced by their skills in auditory, speech, and temporal processing. In Experiment 1, 102 normal school children's performance in pure tone temporal order judgment, tone frequency discrimination, temporal interval discrimination and composite tone pattern discrimination was measured. Results showed that children's auditory processing skills correlated significantly with their reading fluency, phonological awareness, word naming latency, and the number of Chinese characters learned. Regression analyses found that tone temporal order judgment, temporal interval discrimination and composite tone pattern discrimination could account for 32% of variance in phonological awareness. Controlling for the effect of phonological awareness, auditory processing measures still contributed significantly to variance in reading fluency and character naming. In Experiment 2, mismatch negativities (MMN) in event-related brain potentials were recorded from dyslexic children and the matched normal children, while these children listened passively to Chinese syllables and auditory stimuli composed of pure tones. The two groups of children did not differ in MMN to stimuli deviated in pure tone frequency and Chinese lexical tones. But dyslexic children showed smaller MMN to stimuli deviated in initial consonants or vowels of Chinese syllables and to stimuli deviated in temporal information of composite tone patterns. These results suggested that Chinese dyslexic children have deficits in auditory temporal processing as well as in linguistic processing and that auditory and temporal processing is possibly as important to reading development of children in a logographic writing system as in an alphabetic system.

Intra- and inter-speaker variations of formant pattern for lateral syllables in Standard Chinese

Speech variation of speakers is a crucial issue for speaker recognition and identification, especially for forensic practice. Greater intra-speaker variation is one main reason for incorrect speaker identification in real forensic situations. Understanding the stability of acoustic parameters and their variation in speech is therefore significant for the evaluation of effective parameters for speaker identification. In this paper, all vowels in Standard Chinese including five monophthongs, eight diphthongs and four triphthongs were combined with lateral /l/. Finally, 15 lateral syllables with different tones for 10 speakers were selected and acoustically analyzed. Central frequencies of the first four formants for each syllable were measured for quantitative comparison of intra- and inter-speaker variation in order to provide a general idea of speaker variation in Standard Chinese, and finally serving for forensic application. Results show that the overall intra-speaker variation is less than the inter-speaker variation in great extent under laboratory condition though occasionally they are contrary. This supports the basis for forensic speaker identification, that is, intra-speaker variation should be less than inter-speaker variation in many acoustic features, and further validates the probability and reliability of forensic speaker identification.

Analysis and synthesis of fundamental frequency contours of Standard Chinese using the command–response model

While the tonal characteristics of Chinese syllables have been qualitatively described in traditional phonetics, quantitative analysis requires a mathematical model. This paper presents such a model for the fundamental frequency contours of Standard Chinese, based on an extension of a model that has already been proved to be applicable to non-tone languages including Japanese, English, and others. The model allows one to interpret a given fundamental frequency contour in terms of tone commands and phrase commands, and to analyze various tonal phenomena in quantitative terms. The paper then describes the results of analysis of fundamental frequency contours of a number of utterances, revealing systematic relationships between the timing of the tone commands and the final of each syllable. The results are used to derive constraints for tone and phrase command generation in speech synthesis. The validity of the rules is confirmed by evaluating the naturalness of prosody of synthetic speech. The validity of introducing these constraints in speech synthesis of Standard Chinese is confirmed by perceptual tests on naturalness of prosody as well as on intelligibility of tones, using speech synthesized with and without these constraints.

Large vocabulary Mandarin Final recognition based on Two-Level Time-Delay Neural Networks (TLTDNN)

A Two-Level Time-Delay Neural Network (TLTDNN) technique has been developed to recognize all Mandarin Finals of the entire Chinese syllables. The first level discriminates the vowel-group based on (a, e, i, o, u, v) and the nasal-group based on nasal ending (-n, -ng, -others). The nasal-group discriminator is used to further split the large /a/ subgroup produced by the vowel-group discriminator. The two groupings in the first level produce 8 subgroups in the second level. Further discrimination in the second level enables the identification of all 35 Mandarin Finals. The technique was thoroughly tested using 8 sets of 1265 isolated Hanyu Pinyin syllables, with 6 sets used for training and 2 sets used for testing. The overall result shows that a high recognition rate of 99.4% on the training datasets and 95.6% on the test datasets, is achievable. The top 4 recognition rate attained on the test datasets is as high as 99.1%.

Prosody or Pharyngealization in Old Chinese? The Origin of the Distinction between Type A and Type B Syllables

Some comments are offered on a recent paper in these pages by Jerry Norman, in which be proposed an alternative to the A.'s theory of the origin of the distinction between Type A and Type B syllables in Old and Middle Chinese, that is, the distinction marked by the absence or presence of medial yod in Karlgren's reconstructions

Internal temporal structure of Mandarin syllables

The durational interactions among different combinations of onsets, nuclei, and codas in Mandarin Chinese syllables are examined, based on data from six speakers at normal and fast rates of speech. This report extends and refines preliminary results, based on four speakers at a normal tempo, that were reported earlier [A. Bell and M. Liu, J. Acoust. Soc. Am. Suppl. 1 87, S67 (1990)]. Effects of different onsets (/n t shtsh/), vowels (/a i u/), diphthongs (/ai au ou ia iau/), and nasal codas (/n sh/) were considered. Consistent effects of onsets, vowel height, complex nuclei, and nasal codas on both syllable and rhyme durations were found. These effects, however, were relatively small, of the order of 15% for rhyme durations, and from about 5% to 10% for syllables, apparently consistent with a hypothesis that Mandarin rhymes and syllables are perceptually invariant. On the other hand, the nature of segment compensations was more consistent with segment-local and universal mechanisms than with language-particular global timing units of rhyme or syllable. [Work supported by the University of Colorado Council on Research and Creative Work.]