Browsing by Author "Wang, Yih-Ru"

Now showing 1 - 20 of 42

A New Approach of Speaking Rate Modeling for Mandarin Speech Prosody
(2012)
Hsieh, Chiao-Hua
;
Chiang, Chen-Yu
;
Wang, Yih-Ru
;
Yu, Hsiu-Min
;
Chen, Sin-Horng
A new approach of Mandarin-speech prosody modeling to consider the effects of speaking rate is proposed. The approach is a modification of our previous prosody labeling and modeling method to take speaking rate as a continuous independent variable and let prosodic-acoustic features and some parameters of prosodic models depend on it in order to count its influences. A speaking rate-dependent hierarchical prosodic model is hence constructed from four speech corpora of a single female speaker with fast, normal, medium and slow speaking rates. An analysis of the effects of speaking rate on the model parameters showed that they agreed well with our prior knowledge. So, the proposed approach provides a systematic and effective way to quantify the effects of speaking rate on Mandarin-speech prosody.
A New Model-based Mandarin-speech Coding System
(2011-01-01)
Chiang, Chen-Yu
;
Yang, Jyh-Her
;
Liu, Ming-Chieh
;
Wang, Yih-Ru
;
Liao, Yuan-Fu
In this paper, a new model-based Mandarin-speech coding system is proposed. It employs a prosody-enriched ASR with a hierarchical prosodic model (HPM) to generate from the input speech enriched transcriptions, including linguistic features, prosodic tags and spectral parameters in the encoder. By sending these features to the decoder, we can first reconstruct the prosodic-acoustic features of syllable pitch contour, syllable duration, syllable energy level, and inter-syllable pause duration by HPM using the linguistic features and prosodic tags; and then combined with spectral parameters to reconstruct the input speech signal by an HMM-based speech synthesizer. Experimental results show that the reconstructed speech has good quality at a low data rate of 543 bits/s.
A New Model-based Prosody Coder for Mandarin Speech
(2013)
Chiang, Chen-Yu
;
Hung, Yu-Ping
;
Chen, Sin-Horng
;
Wang, Yih-Ru
;
電機工程學系
In this paper, a novel parametric prosody coding approach for Mandarin speech is proposed. It employs a hierarchical prosodic model (HPM) as a prosody generating model in the encoder to analyze the speech prosody of the input utterance to obtain a parametric representation of four prosodic-acoustic features of syllable pitch contour, syllable duration, syllable energy level, and syllable-juncture pause duration for encoding. In the decoder, the four prosodic-acoustic features are reconstructed by a synthesis operation using the decoded HPM parameters. The reconstructed prosodic features are lastly used in an HMM-based speech synthesizer to help to generate the reconstructed speech. Experimental results show that the reconstructed speech has good quality at low data rates of 114.9 bits/s for a speaker-dependent task. An informal listening test confirmed decoded speeches sounded very fluently.
A New Prosody-Assisted Mandarin ASR System
(2012-08-01)
Chen, Sin-Horng
;
Yang, Jyh-Her
;
Chiang, Chen-Yu
;
Liu, Ming-Chieh
;
Wang, Yih-Ru
This paper presents a new prosody-assisted automatic speech recognition (ASR) system for Mandarin speech. It differs from the conventional approach of using simple prosodic cues on employing a sophisticated prosody modeling approach based on a four-layer prosody-hierarchy structure to automatically generate 12 prosodic models from a large unlabeled speech database by the joint prosody labeling and modeling (PLM) algorithm proposed previously. By incorporating these 12 prosodic models into a two-stage ASR system to rescore the word lattice generated in the first stage by the conventional hidden Markov model (HMM) recognizer, we can obtain a better recognized word string. Besides, some other information can also be decoded, including part of speech (POS), punctuation mark (PM), and two types of prosodic tags which can be used to construct the prosody-hierarchy structure of the testing speech. Experimental results on the TCC300 database, which consists of long paragraphic utterances, showed that the proposed system significantly outperformed the baseline scheme using an HMM recognizer with a factored language model which models word, POS, and PM. Performances of 20.7%, 14.4%, and 9.6% in word, character, and base-syllable error rates were obtained. They corresponded to 3.7%, 3.7%, and 2.4% absolute (or 15.2%, 20.4%, and 20% relative) error reductions. By an error analysis, we found that many word segmentation errors and tone recognition errors were corrected.
A NEW SIMILARITY MEASURE BETWEEN HMMS
(2008)
Wang, Yih-Ru
;
電信工程研究所
;
Institute of Communications Engineering
In this paper, a new similarity measure between HMM models which extended the well-known Kullback-Leibler distance was proposed. The Kullback-Leibler distance was defined as the mean of log-likelihood ratio (LLR) in a hypotheses test and the Kullback-Leibler distance was frequently used as a similarity measure for HMM models. Here, the standard deviation of LLR between HMM models was deviated first. Besides, the ratio of mean and standard variation of LLR was used as a new similarity measure between HMM models. Experiments were done in a Mandarin speech database, TCC-300, in order to check the effectiveness of the proposed similarity measure. The accuracy of the standard deviation of LLR estimated from the syllable HMM models was checked by comparison with the standard deviation of LLR of top-10 candidates found from HMM decoder. And, the confusion sets of 411 syllables were also found by using both the KL distance and the proposed similarity measure. Comparing to the top-10 confusion models, 94.9% and 95.3% inclusion rates can be achieved by using KL distance and the proposed similarity measure of HMM models.
A Probabilistic Interpretation for Artificial Neural Network-based Voice Conversion
(2015)
Hwang, Hsin-Te
;
Tsao, Yu
;
Wang, Hsin-Min
;
Wang, Yih-Ru
;
Chen, Sin-Horng
Voice conversion (VC) using artificial neural networks (ANNs) has shown its capability to produce better sound quality of the converted speech than that using Gaussian mixture model (GMM). Although ANN-based VC works reasonably well, there is still room for further improvement. One of the promising ways is to adopt the successful techniques in statistical model-based parameter generation (SMPG), such as trajectory-based mapping approaches that are originally designed for GMM-based VC and hidden Markov model (HMM)-based speech synthesis. This study presents a probabilistic interpretation for ANN-based VC. In this way, ANN-based VC can easily incorporate the successful techniques in SMPG. Experimental results demonstrate that the performance of ANN-based VC can be effectively improved by two trajectory-based mapping techniques (maximum likelihood parameter generation (MLPG) algorithm and maximum likelihood-based trajectory mapping considering global variance (referred to as MLGV)), compared to the conventional ANN-based VC with frame-based mapping and the GMM-based VC with the MLPG algorithm. Moreover, ANN-based VC with the trajectory-based mapping techniques can achieve comparable performance when compared to the state-of-the-art GMM-based VC with the MLGV algorithm.
A SPEAKING RATE-CONTROLLED MANDARIN TTS SYSTEM
(2013)
Hsieh, Chiao-Hua
;
Wang, Yih-Ru
;
Chiang, Chen-Yu
;
Chen, Sin-Horng
;
電子工程學系及電子研究所
In this paper, a new speaking rate-controlled Mandarin TTS system based on a speaking rate-dependent hierarchical prosodic model (SR-HPM) [6] is proposed. In the training phase, a data-driven approach is employed to automatically build the SR-HPM directly from a large prosody-unlabeled speech database containing utterances of various speaking rates. The SR-HPM comprises 15 sub-models designed to describe various relationships among 3 types of prosodic-acoustic features of speech utterances, two types of prosodic tags specifying a 4-layer prosody hierarchy, linguistic features of various levels of the associated texts, and the speaking rates. In the test phase, the SR-HPM is employed to generate 4 prosodic-acoustic features, including syllable pitch contours, syllable durations, syllable energy levels, and syllable juncture pause durations. Combining these prosodic features with the spectral features generated by the HTS synthesizer, the system can generate natural speech for any speaking rate in a wide range of 0.15-0.3 seconds/syllable. A distinct feature of the system to control the occurrence frequencies of breaks of various types as well as their pause durations according to the given speaking rate was demonstrated. A subjective test showed that MOS scores of 3.35, 3.44 and 3.28 were achieved respectively for fast (SR= 0.17 sec/syllable), medium (SR=0.2 sec/syllable) and slow (SR= 0.25 sec/syllable) synthetic speeches.
A Study of Mutual Information for GMM-Based Spectral Conversion
(2012)
Hwang, Hsin-Te
;
Tsao, Yu
;
Wang, Hsin-Min
;
Wang, Yih-Ru
;
Chen, Sin-Horng
The Gaussian mixture model (GMM)-based method has dominated the field of voice conversion (VC) for last decade. However, the converted spectra are excessively smoothed and thus produce muffled converted sound. In this study, we improve the speech quality by enhancing the dependency between the source (natural sound) and converted feature vectors (converted sound). It is believed that enhancing this dependency can make the converted sound closer to the natural sound. To this end, we propose an integrated maximum a posteriori and mutual information (MAPMI) criterion for parameter generation on spectral conversion. Experimental results demonstrate that the quality of converted speech by the proposed MAPMI method outperforms that by the conventional method in terms of formal listening test.
A STUDY ON CROSS-LANGUAGE KNOWLEDGE INTEGRATION IN MANDARIN LVCSR
(2012)
Chiang, Chen-Yu
;
Siniscalchi, Sabato Marco
;
Wang, Yih-Ru
;
Chen, Sin-Horng
;
Lee, Chin-Hui
We present a cross-language knowledge integration framework to improve the performance in large vocabulary continuous speech recognition. Two types of knowledge sources, manner attribute and prosodic structure, are incorporated. For manner of articulation, cross-lingual attribute detectors trained with an American English corpus (WSJ0) are utilized to verify and rescore hypothesized Mandarin syllables in word lattices obtained with state-of-the-art systems. For the prosodic structure, models trained with an unsupervised joint prosody labeling and modeling technique using a Mandarin corpus (TCC300) are used in lattice rescoring. Experimental results on Mandarin syllable, character and word recognition with the TCC300 corpus show that the proposed approach significantly outperforms the baseline system that does not use articulatory and prosodic information. It also demonstrates a potential of utilizing results from cross-lingual attribute detectors as a language-universal frontend for automatic speech recognition.
A SYLLABLE-BASED PROSODY MODELING FOR L1 AND L2 ENGLISH SPEECHES
(2012)
Chen, Wei-Fan
;
Kuo, Chin-Kuan
;
Wang, Yih-Ru
;
Chen, Sin-Horng
;
電機工程學系
In this paper, a statistical prosody modeling approach for L1 and L2 English speeches is proposed. The study focuses on the modeling of two prosodic-acoustic features: syllable duration and log-pitch contour. Several major affecting factors (AFs) that influence the variations of these two features are considered. They include lexical stress, word length, nearby break type, phonemic constituent of syllable, and prosodic state. A sequential optimization procedure is adopted to automatically train the two models from the TWNAESOP corpus recorded in Taiwan. Experimental results showed that most AFs estimated agreed well with our prior linguistic knowledge. The differences in the prosody of L1 and L2 speeches were also explored.
A Two-Stage Sample-based Phone Boundary Detector using Segmental Similarity Features
(2011-01-01)
Wang, Yih-Ru
;
傳播研究所
;
Institute of Communication Studies
In this paper, a two-stage sample-based phone boundary detection algorithm is proposed. In the first stage, some local sample-based acoustic parameters are used to pre-select some phone boundary candidates. Then, in the second stage, some high-order statistics of the log-likelihood differences of two adjacent speech segments around each boundary candidate are calculated to serve as similarity measure for candidate verification. Experimental results on the TIMIT speech corpus showed that ERRS of 8.6% and 7.6% were achieved for one-stage and two-stage sample-based phone boundary detections, respectively. Moreover, for the two-stage system, 42.1% and 81.9% of boundaries detected were within 5- and 15-sample error tolerance from manual labeling results.
Advanced Unsupervised Joint Prosody Labeling and Modeling for Mandarin Speech and Its Application to Prosody Generation for TTS
(2009)
Chiang, Chen-Yu
;
Chen, Sin-Horng
;
Wang, Yih-Ru
;
電信工程研究所
;
Institute of Communications Engineering
Motivated by the success of the unsupervised joint prosody labeling and modeling (UJPLM) method for Mandarin speech on modeling of syllable pitch contour in our previous study, in this paper, the advanced UJPLM (A-UJPLM) method is proposed based on UJPLM to jointly label prosodic tags and model syllable pitch contour, duration and energy level. Experimental results on the Sinica Treebank corpus showed that most prosodic tags labeled were linguistically meaningful and the model parameters estimated were interpretable and generally agreed with other previous study. In virtue of the functions given by the model parameters, an application of A-UJPLM to the prosody generation for Mandarin TTS is proposed. Experimental results showed that the proposed method performed well. Most predicted prosodic features matched well to their original counterparts. This also reconfirmed the effectiveness of the A-UJPLM method.
Alleviating the Over-Smoothing Problem in GMM-Based Voice Conversion with Discriminative Training
(2013-01-01)
Hwang, Hsin-Te
;
Tsao, Yu
;
Wang, Hsin-Min
;
Wang, Yih-Ru
;
Chen, Sin-Horng
In this paper, we propose a discriminative training (DT) method to alleviate the muffled sound effect caused by over smoothing in the Gaussian mixture model (GMM)-based voice conversion (VC). For the conventional GMM-based VC, we often observed a large degree of ambiguities among acoustic classes (generative classes), determined by the source feature vectors for generating the converted feature vectors, causing the "muffled sound" effect on the converted voice. The proposed DT method is applied to refine the parameters in the maximum likelihood (ML)-trained joint density GMM (JDGMM) in the training stage to reduce the ambiguities among acoustic classes (generative classes) to alleviate the muffled sound effect. Experimental results demonstrate that the DT method significantly enhances the discriminative power between acoustic classes (generative classes) in the objective evaluation and effectively alleviates the muffled sound effect in the subjective evaluation.
An Automatic Prosody Labeling Method for Mandarin Speech
(2007)
Chiang, Chen-Yu
;
Yu, Hsiu-Min
;
Wang, Yih-Ru
;
Chen, Sin-Horng
;
電信工程研究所
A new model-based automatic prosody labeling method for Mandarin speech is proposed. It first introduces four models to describe the relationships of the prosody tags to be labeled, the prosodic features of the speech signals, and the linguistic features of the associated texts. It then employs a sequential optimization procedure to estimate parameters of these four models and find all prosody tags. Experimental results on the Sinica Tree-Bank corpus showed that most prosody tags labeled were meaningful and the estimated parameters of these four models matched well with our a priori knowledge about Mandarin prosody.
An Exploration of Local Speaking Rate Variations in Mandarin Read Speech
(2018-01-01)
Liou, Guan-Tin
;
Chiang, Chen-Yu
;
Wang, Yih-Ru
;
Chen, Sin-Horng
;
電機工程學系
This paper explores speaking rate variation in Mandarin read speech. In contrast to assuming that each utterance is generated in a constant or global speaking rate, this study seeks to estimate local speaking rate for each prosodic unit in an utterance. The exploration is based on the existing speaking rate-dependent hierarchical prosodic model (SR-HPM). The main idea is to first use the SR-HPM to explore the prosodic structures of utterances and extract the prosodic units. Then, local speaking rate is estimated for each prosodic unit (prosodic phrase in this study). Some major influence factors including tone, base syllable type, prosodic structure, and speaking rate of the higher prosodic units (utterance and BG/PG) are compensated in the local SR estimation. A syntactic-local SR model is constructed and use in the prosody generation of Mandarin TTS. Experimental results on a large read speech corpus generated by a professional female announcer showed that the generated prosody with local speaking rate variations is proved to be more vivid than the one with a constant speaking rate.
An Investigation on the Mandarin Prosody of a Parallel Multi-Speaking Rate Speech Corpus
(2009)
Chiang, Chen-Yu
;
Tang, Cheng-Chang
;
Yu, Hsiu-Min
;
Wang, Yih-Ru
;
Chen, Sin-Horng
In this paper, the prosody of a parallel multi-speaking rate Mandarin read speech corpus is investigated. The corpus contains four parallel speech datasets uttered by a female professional announcer with various speech rates (SRs) of 4.40 (fast), 3.82 (normal), 2.97 (median) and 2.45 (slow) syllables/second. By using the unsupervised joint prosody labeling and modeling (PLM) method proposed previously, the relationship between SR and various prosodic features, including pause duration, patterns of three high-level prosodic constituents, and the break labels, are investigated. The analyses reported in this study could be very informative in developing prosody generation mechanism for text-to-speech and prosody modeling for automatic speech recognition in various SRs.
Distributed speech recognition of mandarin digits string
(2006)
Wang, Yih-Ru
;
Lu, Bo-Xuan
;
Liao, Yuan-Fu
;
Chen, Sin-Horng
;
交大名義發表
In this paper, the performance of the pitch detection algorithm in ETSI ES-202-212 XAFE standard is evaluated on. a Mandarin digit string recognition task. Experimental results showed that the performance of the pitch detection algorithm degraded seriously when the SNR of speech signal was lower than 10dB. This makes the recognizer using pitch information perform inferior to the original recognizer without using pitch information in low SNR environments. A modification of the pitch detection algorithm is therefore proposed to improve the performance of pitch detection in low SNR environments. The recognition performance can be improved for most SNR levels by integrating the recognizers with and without using pitch information. Overall recognition rates of 82.1% and 86.8% were achieved for clean and multi-condition training cases.
ENRICHING MANDARIN SPEECH RECOGNITION BY INCORPORATING A HIERARCHICAL PROSODY MODEL
(2011)
Yang, Jyh-Her
;
Liu, Ming-Chieh
;
Chang, Hao-Hsiang
;
Chiang, Chen-Yu
;
Wang, Yih-Ru
This paper presents a new probabilistic framework of Mandarin speech recognition by incorporating a sophisticated hierarchical prosody model into the conventional HMM-based system. The prosody model describes the relations of linguistic cues of various levels, break types and prosodic states which represent the prosody hierarchical structure, and prosody-related acoustic features. Aside from producing the recognized word sequences, the system also decodes other information including word's part-of-speech, punctuation marks, inter-syllable break types, and prosodic states of syllables. Experimental results on the TCC300 corpus, which consists of paragraphic utterances, showed that the proposed system significantly outperformed the baseline system. The word and character error rates decreased from 24.4% and 18.1% to 20.7% and 14.4% (or 15.2% and 20.4% relative improvements), respectively.
Evaluation of Weighted Graph and Neural Network Models on Predicting the Valence-Arousal Ratings of Chinese Words
(2016-01-01)
Chou, Wei-Chieh
;
Lin, Chin-Kui
;
Wang, Yih-Ru
;
Liao, Yuan-Fu
;
電機工程學系
Automatic sentiment information extraction of social network articles has many essential applications. Following the valence-arousal space framework, in this paper, two approaches including (1) a weighted graph (WG) and (2) a neural network (NN) model that could predict the valence-arousal ratings of words are evaluated on the Chinese valence-arousal words (CVAW) database provide by the IALP-2016 shared task. According the official evaluation results, our NN systems achieved (0.621,1.165) MAEs and (0.853,0.631) PCCs for valence and arousal predictions. Compared with the results of other participants, the performance of our systems are quite nice.
Exploration of high-level prosodic patterns for continuous Mandarin speech
(2008)
Chiang, Chen-Yu
;
Yu, Hsiu-Min
;
Wang, Yih-Ru
;
Chen, Sin-Horng
;
電信工程研究所
In this paper, the high-level prosodic patterns of prosodic word (PW), prosodic phrase (PPh) and breath group/prosodic phrase group (BG/PG) for syllable pitch-level and duration are explored using an automatic joint prosody labeling and modeling method. Experimental results on a treebank speech corpus showed that the explored high-level prosodic patterns not only matched well with our a priori knowledge about Mandarin prosody, but also conformed well to other previous studies. They can therefore be integrated to form a meaningful Mandarin prosody hierarchy.