• MALSORI
• /
• no.43
• /
• pp.137-150
• /
• 2002

Speech segmentation at phoneme level is important for corpus-based text-to-speech synthesis. In this paper, we examine acoustic modeling methods to improve the performance of automatic speech segmentation system based on Hidden Markov Model (HMM). We compare monophone and triphone models, and evaluate several model training approaches. In addition, we employ an energy-based postprocessing scheme to make correction of frequent boundary location errors between silence and speech sounds. Experimental results show that our system provides 71.3% and 84.2% correct boundary locations given tolerance of 10 ms and 20 ms, respectively.

• Proceedings of the IEEK Conference
• /
• 2003.07e
• /
• pp.2216-2219
• /
• 2003

Accurate speech region detection and automatic syllable segmentation is important part of speech recognition system. In automatic speech recognition system, they are needed for the purpose of accurate recognition and less computational complexity, In this paper, we Propose improved syllable segmentation method using ramp edge detection method and residual signal Peak energy. These methods were used to ensure accuracy and robustness for endpoint detection and syllable segmentation system. They have almost invariant response to various background noise levels. As experimental results, we obtained the rate of 90.7％ accuracy in syllable segmentation in a condition of accurate endpoint detection environments.

• MALSORI
• /
• no.58
• /
• pp.67-81
• /
• 2006

For a large corpus of time-aligned data, HMM based approaches are most widely used for automatic segmentation, providing a consistent and accurate phone labeling scheme. There are two methods for training in HMM. Flat starting method has a property that human interference is minimized but it has low accuracy. Bootstrap method has a high accuracy, but it has a defect that manual segmentation is required In this paper, a new algorithm is proposed to minimize manual work and to improve the performance of automatic segmentation. At first phase, voiced, unvoiced and silence classification is performed for each speech data frame. At second phase, the phoneme sequence is aligned dynamically to the voiced/unvoiced/silence sequence according to the acoustic phonetic rules. Finally, using these segmented speech data as a bootstrap, phoneme model parameters based on HMM are trained. For the performance test, hand labeled ETRI speech DB was used. The experiment results showed that our algorithm achieved 10% improvement of segmentation accuracy within 20 ms tolerable error range. Especially for the unvoiced consonants, it showed 30% improvement.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.5
• /
• pp.213-221
• /
• 2006

Phone segmentation of speech waveform is especially important for concatenative text to speech synthesis which uses segmented corpora for the construction of synthetic units. because the quality of synthesized speech depends critically on the accuracy of the segmentation. In the beginning. the phone segmentation was manually performed. but it brings the huge effort and the large time delay. HMM-based approaches adopted from automatic speech recognition are most widely used for automatic segmentation in speech synthesis, providing a consistent and accurate phone labeling scheme. Even the HMM-based approach has been successful, it may locate a phone boundary at a different position than expected. In this paper. we categorized adjacent phoneme pairs and analyzed the mismatches between hand-labeled transcriptions and HMM-based labels. Then we described the dominant error patterns that must be improved for the speech synthesis. For the experiment. hand labeled standard Korean speech DB from ETRI was used as a reference DB. Time difference larger than 20ms between hand-labeled phoneme boundary and auto-aligned boundary is treated as an automatic segmentation error. Our experimental results from female speaker revealed that plosive-vowel, affricate-vowel and vowel-liquid pairs showed high accuracies, 99%, 99.5% and 99% respectively. But stop-nasal, stop-liquid and nasal-liquid pairs showed very low accuracies, 45%, 50% and 55%. And these from male speaker revealed similar tendency.

• The Journal of the Acoustical Society of Korea
• /
• v.24 no.4E
• /
• pp.128-131
• /
• 2005

A new algorithm for speech signal segmentation is proposed. This algorithm is based on finding successive similar frames belonging to a segment and represents it by an average spectrum. The speech signal is a slowly time varying signal in the sense that, when examined over a sufficiently short period of time (between 10 and 100 ms), its characteristics are fairly stationary. Generally this approach is based on finding these fairly stationary periods. Advantages of the. algorithm are accurate border decision of segments and simple computation. The automatic segmentations using frame averaging show as much as $82.20\%$ coincided with manually verified segmentation of CMU ARCTIC corpus within time range 16 ms. More than $90\%$ segment boundaries are coincided within a range of 32 ms. Also it can be combined with many types of automatic segmentations (HMM based, acoustic cues or feature based etc.).

• Speech Sciences
• /
• v.9 no.4
• /
• pp.265-274
• /
• 2002

In this paper we propose an automatic segmentation system that outputs the time alignment information of phoneme boundary using Viterbi search with HMM (Hidden Markov Model) and corrects these results by an UVS (unvoiced/voiced/silence) classification algorithm. We selecte a set of 39 monophones and a set of 647 extended phones for HMM models. For the UVS classification we use the feature parameters such as ZCR (Zero Crossing Rate), log energy, spectral distribution. The result of forced alignment using the extended phone set is 11% better than that of the monophone set. The UVS classification algorithm shows high performance to correct the segmentation results.

• MALSORI
• /
• no.35_36
• /
• pp.175-188
• /
• 1998

Database segmented and labeled up to phoneme level plays an important role in phonetic research and speech engineering. However, it usually requires manual segmentation and labeling, which is time-consuming and may also lead to inconsistent consequences. Automatic segmentation and labeling can be introduced to solve these problems. In this paper, we investigate a method to improve the performance of automatic segmentation and labeling system, where Spectral Variation Function(SVF), modification of silence model, and use of energy variations in postprocessing stage are considered. In this paper, SVF is applied in three ways: (1) addition to feature parameters, (2) postprocessing of phoneme boundaries, (3) restricting the Viterbi path so that the resulting phoneme boundaries may be located in frames around SVF peaks. In the postprocessing stage, positions with greatest energy variation during transitional period between silence and other phonemes were used to modify boundaries. In order to evaluate the performance of the system, we used 452 phonetically balanced word(PBW) database for training phoneme models and phonetically balanced sentence(PBS) database for testing. According to our experiments, 83.1% (6.2% improved) and 95.8% (0.9% improved) of phoneme boundaries were within 20ms and 40ms of the manually segmented boundaries, respectively.

• Proceedings of the KSPS conference
• /
• 2003.10a
• /
• pp.3-16
• /
• 2003

ㆍAcoustic characteristics of stops in speech with contextual variability ㆍPosibility of stop recognition by post processing technique ㆍFurther work - Speech database - Modification of decoder - automatic segmentation of acoustic parameters

• The Journal of the Acoustical Society of Korea
• /
• v.20 no.8
• /
• pp.24-30
• /
• 2001

This article is concerned with automatic segmentation for Korean speech signals. All kinds of transition cases of phonetic units are classified into 3 types and different strategies for each type are applied. The type 1 is the discrimination of silence, voiced-speech and unvoiced-speech. The histogram analysis of each indicators which consists of wavelet coefficients and SVF (Spectral Variation Function) in wavelet coefficients are used for type 1 segmentation. The type 2 is the discrimination of adjacent vowels. The vowel transition cases can be characterized by spectrogram. Given phonetic transcription and transition pattern spectrogram, the speech signal, having consecutive vowels, are automatically segmented by the template matching. The type 3 is the discrimination of vowel and voiced-consonants. The smoothed short-time RMS energy of Wavelet low pass component and SVF in cepstral coefficients are adopted for type 3 segmentation. The experiment is performed for 342 words utterance set. The speech data are gathered from 6 speakers. The result shows the validity of the method.

• The Journal of the Acoustical Society of Korea
• /
• v.16 no.5
• /
• pp.50-59
• /
• 1997

In this paper, we implement an automatic speech segmentation and labeling system which marks phone boundaries automatically for constructing the Korean speech database. We specify and implement the system based on conventional speech segmentation and labeling techniques, and also develop the graphic user interface(GUI) on Hangul $Motif^{TM}$ environment for the users to examine the automatic alignment boundaries and to refine them easily. The developed system is applied to 16kHz sampled speech, and the labeling unit is composed of 46 phoneme-like units(PLUs) and silence. The system uses both of the phonetic and orthographic transcription as input methods of linguistic information. For pattern-matching method, hidden Markov models(HMM) is employed. Each phoneme model is trained using the manually segmented 445 phonetically balanced word (PBW) database. In order to evaluate the performance of the system, we test it using another database consisting of sentence-type speech. According to our experiment, 74.7% of phoneme boundaries are within 20ms of the true boundary and 92.8% are within 40ms.