• Phonetics and Speech Sciences
• /
• v.10 no.2
• /
• pp.7-14
• /
• 2018

The paper investigates the acoustic characteristics of English fricatives in the TIMIT corpus, with a special focus on the role of voicing in rendering fricatives in American English. The TIMIT database includes 630 talkers and 2,342 different sentences, and comprises more than five hours of speech. Acoustic analyses are conducted in the domain of spectral and temporal properties by treating gender, voicing, and place of articulation as independent factors. The results of the acoustic analyses revealed that acoustic signals interact in a complex way to signal the gender, place, and voicing of fricatives. Classification experiments using a multiclass support vector machine (SVM) revealed that 78.7% of fricatives are correctly classified. The majority of errors stem from the misclassification of /θ/ as [f] and /ʒ/ as [z]. The average accuracy of gender classification is 78.7%. Most errors result from the classification of female speakers as male speakers. The paper contributes to the understanding of the effects of voicing and gender on fricatives in a large-scale speech corpus.

• Phonetics and Speech Sciences
• /
• v.7 no.1
• /
• pp.117-124
• /
• 2015

This paper investigates the acoustic characteristics of English fricatives in the TIMIT corpus, with a special focus on the role of gender in rendering fricatives in American English. The TIMIT database includes 630 talkers and 2342 different sentences, comprising over five hours of speech. Acoustic analyses are conducted in the domain of spectral and temporal properties by treating gender as an independent factor. The results of acoustic analyses revealed that the most acoustic properties of voiceless sibilants turned out to be different between male and female speakers, but those of voiceless non-sibilants did not show differences. A classification experiment using linear discriminant analysis (LDA) revealed that 85.73% of voiceless fricatives are correctly classified. The sibilants are 88.61% correctly classified, whereas the non-sibilants are only 57.91% correctly classified. The majority of the errors are from the misclassification of /ɵ/ as [f]. The average accuracy of gender classification is 77.67%. Most of the inaccuracy results are from the classification of female speakers in non-sibilants. The results are accounted for by resorting to biological differences as well as macro-social factors. The paper contributes to the understanding of the role of gender in a large-scale speech corpus.

• The Journal of the Acoustical Society of Korea
• /
• v.21 no.8
• /
• pp.727-734
• /
• 2002

A simple algorithm for discriminating voiced sounds in a speech is proposed in this paper. In addition to low-frequency energy and zero-crossing rate (ZCR), both of which have been widely used in the past for identifying voiced sounds, the proposed algorithm incorporates pitch variation to improve the discrimination rate. Based on TIMIT corpus, evaluation result shows an improvement of 13% in the discrimination of voiced phonemes over that of the traditional algorithm using only energy and ZCR.

• Speech Sciences
• /
• v.9 no.3
• /
• pp.35-45
• /
• 2002

A simple algorithm for discriminating voiced sounds in a speech is proposed. In addition to low-frequency energy and zero-crossing rate (ZCR), both of which have been widely used in the past for identifying voiced sounds, the proposed algorithm incorporates pitch variation to improve the discrimination rate. Based on TIMIT corpus, evaluation result shows an improvement of 13% in the discrimination of voiced phonemes over that of the traditional algorithm using only energy and ZCR.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.154.2-154
• /
• 2001

This paper presented a adaptation approach based on maximum model distance (MMD) method. This method shares the same framework as they are used for training speech recognizers with abundant training data. The MMD method could adapt to all the models with or without adaptation data. If large amount of adaptation data is available, these methods could gradually approximate the speaker-dependent ones. The approach is evaluated through the phoneme recognition task on the TIMIT corpus. On the speaker adaptation experiments, up to 65.55% phoneme error reduction is achieved. The MMD could reduce phoneme error by 16.91% even when ...

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.99.1-99
• /
• 2001

This paper presented an adaptation approach based on maximum model distance (MMD) method. This method shares the same framework as they are used for training speech recognizers with abundant training data. The MMD method could adapt to all the models with or without adaptation data. If large amount of adaptation data is available, these methods could gradually approximate the speaker-dependent ones. The approach is evaluated through the phoneme recognition task on the TIMIT corpus. On the speaker adaptation experiments, up to 65.55% phoneme error reduction is achieved. The MMD could reduce phoneme error by 16.91% even when only one adaptation utterance is used.

• Journal of Advanced Marine Engineering and Technology
• /
• v.38 no.2
• /
• pp.182-187
• /
• 2014

We propose nonlinear GMM-based transformation functions in an attempt to deal with the over-smoothing effects of linear transformation for voice processing. The proposed methods adopt RBF networks as a local transformation function to overcome the drawbacks of global nonlinear transformation functions. In order to obtain high-quality modifications of speech signals, our voice conversion is implemented using the Harmonic plus Noise Model analysis/synthesis framework. Experimental results are reported on the English corpus, MOCHA-TIMIT.

• The KIPS Transactions:PartB
• /
• v.12B no.4 s.100
• /
• pp.437-442
• /
• 2005

For the purpose of speaker normalization in speaker independent speech recognition systems, experiments are conducted on a method based on Gaussian mixture model(GMM). The method, which is an improvement of the previous study based on vector quantizer, consists of modeling the probability distribution of canonical feature vectors by a GMM with an appropriate number of clusters, and of estimating the warp factor of a test speaker by making use of the obtained probabilistic model. The purpose of this study is twofold: improving the existing ML based methods, and comparing the performance of what is called 'soft decision' method with that of the previous study based on vector quantizer. The effectiveness of the proposed method is investigated by recognition experiments on the TIMIT corpus. The experimental results showed that a little improvement could be obtained tv adjusting the number of clusters in GMM appropriately.

• The Journal of the Acoustical Society of Korea
• /
• v.23 no.8
• /
• pp.583-589
• /
• 2004

Proposed is a speaker normalization method based on vector quantizer for continuous speech recognition (CSR) system in which no acoustic information is made use of. The proposed method, which is an improvement of the previously reported speaker normalization scheme for a simple digit recognizer, builds up a canonical codebook by iteratively training the codebook while the size of codebook is increased after each iteration from a relatively small initial size. Once the codebook established, the warp factors of speakers are estimated by comparing exhaustively the warped versions of each speaker's utterance with the codebook. Two sets of phones are used to estimate the warp factors: one, a set of vowels only. and the other, a set composed of all the Phonemes. A Piecewise linear warping function which corresponds to the estimated warp factor is adopted to warp the power spectrum of the utterance. Then the warped feature vectors are extracted to be used to train and to test the speech recognizer. The effectiveness of the proposed method is investigated by a set of recognition experiments using the TIMIT corpus and HTK speech recognition tool kit. The experimental results showed comparable recognition rate improvement with the formant based warping method.

• The Journal of the Acoustical Society of Korea
• /
• v.39 no.5
• /
• pp.476-482
• /
• 2020

Sequence-to-sequence deep neural networks with attention mechanisms have shown superior performance across various domains, where the sizes of the input and the output sequences may differ. However, if the input sequences are much longer than the output sequences, and the characteristic of the input sequence changes within a single output token, the conventional attention mechanisms are inappropriate, because only a single context vector is used for each output token. In this paper, we propose a double-attention mechanism to handle this problem by using two context vectors that cover the left and the right parts of the input focus separately. The effectiveness of the proposed method is evaluated using speech recognition experiments on the TIMIT corpus.