Background Chinese tones are considered important in Chinese discrimination. However, the relevant reports on auditory central mechanisms concerning Chinese tones are limited. In this study, mismatch negativity （MMN...Background Chinese tones are considered important in Chinese discrimination. However, the relevant reports on auditory central mechanisms concerning Chinese tones are limited. In this study, mismatch negativity （MMN）, one of the event related potentials （ERP）, was used to investigate pre-attentive processing of Chinese tones, and the differences between the function of oddball MMN and that of control MMN are discussed. Methods Ten subjects （six men and four women） with normal hearing participated in the study. A sequence was presented to these subjects through a loudspeaker, the sequence included four blocks, a control block and three oddball blocks. The control block was made up of five components （one pure tone and four Chinese tones） with equiprobability. The oddball blocks were made up of two components, one was a standard stimulus （tone 1） and the other was a deviant stimulus （tone 2 or tone 3 or tone 4）. Electroencephalogram （EEG） data were recorded when the sequence was presented and MMNs were obtained from the analysis of the EEG data. Results Two kinds of MMNs were obtained, oddball MMN and control MMN. Oddball MMN was obtained by subtracting the ERP elicited by standard stimulation （tone 1） from that elicited by deviant stimulation （tone 2 or tone 3 or tone 4） in the oddball block; control MMN was obtained by subtracting the ERP elicited by the tone in control block, which was the same tone as the deviant stimulation in the oddball block, from the ERP elicited by deviant stimulation （tone 2 or tone 3 or tone 4） in the oddball block. There were two negative waves in oddball MMN, one appeared around 150 ms （oddball MMN 1）, the other around 300 ms （oddball MMN 2）. Only one negative wave appeared around 300 ms in control MMN, which was corresponding to the oddball MMN 2. We performed the statistical analyses in each paradigm for latencies and amplitudes for oddball MMN 2 in discriminating the three Chinese tones and reported no significant differences. But the latencies and amplitudes for control MMN in discriminating the three tones all were significantly different. Conclusions There are evident waveforms for oddball and control MMN obtained in normal hearing persons, indicating that the change of Chinese tones could be detected in the pre-attentive stage. Because control MMN can eliminate reflects for physical characteristics of sound, it is the genuine memory-based pre-attentive processing. Since control MMN can reflect the differences of Chinese tones processing better than oddball MMN, it is more sensitive in evaluating pre-attentive processing in Chinese tones discrimination.展开更多
Objective To assess the utility of low- and high-frequency tympanometry in the diagnosis of middle ear dysfunction in Chinese infants. Methods Tympanograms were obtained with 226 Hz, 678 Hz and 1000 Hz probe tones fro...Objective To assess the utility of low- and high-frequency tympanometry in the diagnosis of middle ear dysfunction in Chinese infants. Methods Tympanograms were obtained with 226 Hz, 678 Hz and 1000 Hz probe tones from infants aged 5-25 weeks with normal auditory brainstem responses(ABRs)(15 infants, 30 ears) and with prolonged wave I latencies(17 infants, 20 ears), suggesting middle ear dysfunction, using a GSI Tympstar middle ear analyzer Version II. Results The single-peaked tympanogram was the most characteristic type in both groups and seen in 25 ears (83.3%) in the normal ABR group and in 18 ears (90%) in the delayed wave I group, respectively. The peak pressure, peak compensated static acoustic admittance and gradient of 226 Hz tympanometry were of no significant differences between the two groups. The 678 Hz tympanograms of admittance, susceptance and conductance demonstrated non-peak, single-, double- and tri-peaked patterns in both groups. The agreement between ABRs and 678 Hz tympanograms of admittance,susceptance and conductance were 70.0%, 58.0% and 64.0%(kappa=0.324, 0.234 and 0.118) respectively. For 1000 Hz probe tone, admittance, susceptance and conductance tympanograms showed single peaked patterns in 28(93.3%), 25 (83.3%) and 26(86.7%) of the 30 normal ears. Admittance, susceptance and conductance tympanograms using the 1000 Hz probe tone were flat in 15 (75%), 17(85%) and 13 (65%) of the ears in infants with prolonged wave I latencies. For 1000 Hz admittance, susceptance and conductance Tympanograms, the agreement between tympanometry and ABR results were 90.0%, 92.0% and 86.0% with kappa at 0.783, 0.831 and 0.690, respectively. Conclusion 1000 Hz probe tone tympanometry is a promising middle ear function test for infants of 1-6 months age, while 226 Hz and 678 Hz probe tones are less efficient in detecting middle ear dysfunction in infants.展开更多
A perceptual experiment was conducted in which stimuli sliced from 12 StandardChinese syllables spoken in isolation were identified by 10 native speakers of Beijing dialect.The results of the experiment indicate that ...A perceptual experiment was conducted in which stimuli sliced from 12 StandardChinese syllables spoken in isolation were identified by 10 native speakers of Beijing dialect.The results of the experiment indicate that when the pre-onset or post-offest of a tone is extended to a large extent, the initial voiced consonant, the medial, the nasal coda, and the vocalic ending are not related to the tonal pitch; even when the post-offset is extended to a short extent, the nasal coda and the vocalic cnding are neither related to the tona1 pitch. Thus, with regard to the tonal domain of a Standard Chinese monosyllable in isolation, not only the initial voiced consonant and the medial, but also the nasal coda and the vocalic ending are excludedi in other words, the basic contour of Standard Chincse tone in isolation is mainly related to the syllabic vowel and its adjacent transition. Because F0 in the pre-onset and post- onset are not related to the tonal pitch, there is no exact match between F0 contour and tonal pitch.展开更多
The fundamental frequency F0 patterns of Chinese were compared with that of English, Japanese, and German at both macro (passage) and micro (fundamental period) levels by using a Laryngograph in this paper. 14 Chinese...The fundamental frequency F0 patterns of Chinese were compared with that of English, Japanese, and German at both macro (passage) and micro (fundamental period) levels by using a Laryngograph in this paper. 14 Chinese and 14 foreigners were selected as speakers. The results show that: 1 .The 90% pitch range of Chinese is somewhat larger than that of other languages studied in Hz but it is near the same for all languages in relative scale logF0, 2. The average signed jitter factor in connected speech is neariy the same, about 3.2%, for both male and female speakers and for dif ferent languages studied; 3. The (-)jitter facior is highcr than (+) jitter factor in connected speech; 4. The rate of F0 fluctuation (peak and valley) for Chinese is higher than that for English; 5.The speech rated in syllables per second of Chinese speakers is lower than that of foreign speakers.展开更多
The Chinese foreign policy with principled,theoretical and innovative presentation by the documents of the CPC Eighteenth National Congress reflects the continuity and innovation of China’s diplomacy,on the one hand,...The Chinese foreign policy with principled,theoretical and innovative presentation by the documents of the CPC Eighteenth National Congress reflects the continuity and innovation of China’s diplomacy,on the one hand,is a farsighted summary of China’s diplomacy in the past 10 years,and is also the response to challenges posed by profound changes in the world situation on the other.展开更多
Several methods were developed to improve grapheme-to-phoneme （G2P） conversion models for Chinese text-to-speech （TTS） systems. The critical problem of data sparsity was handled by combining approaches. First, a t...Several methods were developed to improve grapheme-to-phoneme （G2P） conversion models for Chinese text-to-speech （TTS） systems. The critical problem of data sparsity was handled by combining approaches. First, a text-selection method was designed to cover as many G2P text corpus contexts as possible. Then, various data-driven modeling methods were used with comparisons to select the best method for each polyphonic word. Finally, independent models were used for some neutral tone words in addition to the normal G2P models to achieve more compact and flexible G2P models. Tests show that these methods reduce the relative errors by 50% for both normal polyphonic words and Chinese neutral tones.展开更多
文摘Background Chinese tones are considered important in Chinese discrimination. However, the relevant reports on auditory central mechanisms concerning Chinese tones are limited. In this study, mismatch negativity （MMN）, one of the event related potentials （ERP）, was used to investigate pre-attentive processing of Chinese tones, and the differences between the function of oddball MMN and that of control MMN are discussed. Methods Ten subjects （six men and four women） with normal hearing participated in the study. A sequence was presented to these subjects through a loudspeaker, the sequence included four blocks, a control block and three oddball blocks. The control block was made up of five components （one pure tone and four Chinese tones） with equiprobability. The oddball blocks were made up of two components, one was a standard stimulus （tone 1） and the other was a deviant stimulus （tone 2 or tone 3 or tone 4）. Electroencephalogram （EEG） data were recorded when the sequence was presented and MMNs were obtained from the analysis of the EEG data. Results Two kinds of MMNs were obtained, oddball MMN and control MMN. Oddball MMN was obtained by subtracting the ERP elicited by standard stimulation （tone 1） from that elicited by deviant stimulation （tone 2 or tone 3 or tone 4） in the oddball block; control MMN was obtained by subtracting the ERP elicited by the tone in control block, which was the same tone as the deviant stimulation in the oddball block, from the ERP elicited by deviant stimulation （tone 2 or tone 3 or tone 4） in the oddball block. There were two negative waves in oddball MMN, one appeared around 150 ms （oddball MMN 1）, the other around 300 ms （oddball MMN 2）. Only one negative wave appeared around 300 ms in control MMN, which was corresponding to the oddball MMN 2. We performed the statistical analyses in each paradigm for latencies and amplitudes for oddball MMN 2 in discriminating the three Chinese tones and reported no significant differences. But the latencies and amplitudes for control MMN in discriminating the three tones all were significantly different. Conclusions There are evident waveforms for oddball and control MMN obtained in normal hearing persons, indicating that the change of Chinese tones could be detected in the pre-attentive stage. Because control MMN can eliminate reflects for physical characteristics of sound, it is the genuine memory-based pre-attentive processing. Since control MMN can reflect the differences of Chinese tones processing better than oddball MMN, it is more sensitive in evaluating pre-attentive processing in Chinese tones discrimination.
文摘Objective To assess the utility of low- and high-frequency tympanometry in the diagnosis of middle ear dysfunction in Chinese infants. Methods Tympanograms were obtained with 226 Hz, 678 Hz and 1000 Hz probe tones from infants aged 5-25 weeks with normal auditory brainstem responses(ABRs)(15 infants, 30 ears) and with prolonged wave I latencies(17 infants, 20 ears), suggesting middle ear dysfunction, using a GSI Tympstar middle ear analyzer Version II. Results The single-peaked tympanogram was the most characteristic type in both groups and seen in 25 ears (83.3%) in the normal ABR group and in 18 ears (90%) in the delayed wave I group, respectively. The peak pressure, peak compensated static acoustic admittance and gradient of 226 Hz tympanometry were of no significant differences between the two groups. The 678 Hz tympanograms of admittance, susceptance and conductance demonstrated non-peak, single-, double- and tri-peaked patterns in both groups. The agreement between ABRs and 678 Hz tympanograms of admittance,susceptance and conductance were 70.0%, 58.0% and 64.0%(kappa=0.324, 0.234 and 0.118) respectively. For 1000 Hz probe tone, admittance, susceptance and conductance tympanograms showed single peaked patterns in 28(93.3%), 25 (83.3%) and 26(86.7%) of the 30 normal ears. Admittance, susceptance and conductance tympanograms using the 1000 Hz probe tone were flat in 15 (75%), 17(85%) and 13 (65%) of the ears in infants with prolonged wave I latencies. For 1000 Hz admittance, susceptance and conductance Tympanograms, the agreement between tympanometry and ABR results were 90.0%, 92.0% and 86.0% with kappa at 0.783, 0.831 and 0.690, respectively. Conclusion 1000 Hz probe tone tympanometry is a promising middle ear function test for infants of 1-6 months age, while 226 Hz and 678 Hz probe tones are less efficient in detecting middle ear dysfunction in infants.
文摘A perceptual experiment was conducted in which stimuli sliced from 12 StandardChinese syllables spoken in isolation were identified by 10 native speakers of Beijing dialect.The results of the experiment indicate that when the pre-onset or post-offest of a tone is extended to a large extent, the initial voiced consonant, the medial, the nasal coda, and the vocalic ending are not related to the tonal pitch; even when the post-offset is extended to a short extent, the nasal coda and the vocalic cnding are neither related to the tona1 pitch. Thus, with regard to the tonal domain of a Standard Chinese monosyllable in isolation, not only the initial voiced consonant and the medial, but also the nasal coda and the vocalic ending are excludedi in other words, the basic contour of Standard Chincse tone in isolation is mainly related to the syllabic vowel and its adjacent transition. Because F0 in the pre-onset and post- onset are not related to the tonal pitch, there is no exact match between F0 contour and tonal pitch.
文摘The fundamental frequency F0 patterns of Chinese were compared with that of English, Japanese, and German at both macro (passage) and micro (fundamental period) levels by using a Laryngograph in this paper. 14 Chinese and 14 foreigners were selected as speakers. The results show that: 1 .The 90% pitch range of Chinese is somewhat larger than that of other languages studied in Hz but it is near the same for all languages in relative scale logF0, 2. The average signed jitter factor in connected speech is neariy the same, about 3.2%, for both male and female speakers and for dif ferent languages studied; 3. The (-)jitter facior is highcr than (+) jitter factor in connected speech; 4. The rate of F0 fluctuation (peak and valley) for Chinese is higher than that for English; 5.The speech rated in syllables per second of Chinese speakers is lower than that of foreign speakers.
文摘The Chinese foreign policy with principled,theoretical and innovative presentation by the documents of the CPC Eighteenth National Congress reflects the continuity and innovation of China’s diplomacy,on the one hand,is a farsighted summary of China’s diplomacy in the past 10 years,and is also the response to challenges posed by profound changes in the world situation on the other.
文摘Several methods were developed to improve grapheme-to-phoneme （G2P） conversion models for Chinese text-to-speech （TTS） systems. The critical problem of data sparsity was handled by combining approaches. First, a text-selection method was designed to cover as many G2P text corpus contexts as possible. Then, various data-driven modeling methods were used with comparisons to select the best method for each polyphonic word. Finally, independent models were used for some neutral tone words in addition to the normal G2P models to achieve more compact and flexible G2P models. Tests show that these methods reduce the relative errors by 50% for both normal polyphonic words and Chinese neutral tones.
