• The Journal of the Acoustical Society of Korea
• /
• v.30 no.6
• /
• pp.324-329
• /
• 2011

In this paper, we propose a novel approach to improve the performance of a statistical model-based voice activity detection (VAD) which is based on the conditional maximum a posteriori (CMAP) with deviation. In our approach, the VAD decision rule is expressed as the geometric mean of likelihood ratios (LRs) based on adapted threshold according to the speech presence probability conditioned on both the speech activity decisions and spectral deviation in the pervious frame. Experimental results show that the proposed approach yields better results compared to the CMAP-based VAD using the LR test.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.11C
• /
• pp.1161-1168
• /
• 2003

In this paper, we propose an improved statistical model-based voice activity detection algorithm and threshold update method. We first improve signal-to-noise ratio by using speech enhancement preprocessing algorithm combined power subtraction method and matched filter, then apply it to LLR test optimum decision rule for improving the performance even in low SNR conditions. And we propose an adaptive threshold update method that was not concerned in any papers. We also perform extensive computer simulations to demonstrate the performance improvement of the proposed VAD algorithm employing the proposed speech enhancement preprocessing algorithm and adaptive threshold update method under various background noise environments. Finally we verify our results by comparing ITU-T G.729 Annex B.

• The Journal of the Acoustical Society of Korea
• /
• v.32 no.2
• /
• pp.131-137
• /
• 2013

In this paper we propose an improved voice activity detection (VAD) algorithm using statistical models in the signal subspace domain. A uncorrelated signal subspace is generated using embedded prewhitening technique and the statistical characteristics of the noisy speech and noise are investigated in this domain. According to the characteristics of the signals in the signal subspace, a new statistical VAD method using GGD (Generalized Gaussian Distribution) is proposed. Experimental results show that the proposed GGD-based approach outperforms the Gaussian-based signal subspace method at 0-15 dB SNR simulation conditions.

• ETRI Journal
• /
• v.33 no.1
• /
• pp.99-109
• /
• 2011

This paper concerns a robust real-time voice activity detection (VAD) approach which is easy to understand and implement. The proposed approach employs several short-term speech/nonspeech discriminating features in a voting paradigm to achieve a reliable performance in different environments. This paper mainly focuses on the performance improvement of a recently proposed approach which uses spectral peak valley difference (SPVD) as a feature for silence detection. The main issue of this paper is to apply a set of features with SPVD to improve the VAD robustness. The proposed approach uses a weighted voting scheme in order to take the discriminative power of the employed feature set into account. The experiments show that the proposed approach is more robust than the baseline approach from different points of view, including channel distortion and threshold selection. The proposed approach is also compared with some other VAD techniques for better confirmation of its achievements. Using the proposed weighted voting approach, the average VAD performance is increased to 89.29% for 5 different noise types and 8 SNR levels. The resulting performance is 13.79% higher than the approach based only on SPVD and even 2.25% higher than the not-weighted voting scheme.

• Speech Sciences
• /
• v.15 no.2
• /
• pp.55-65
• /
• 2008

Recently, the multi-model based speech recognizer has been used quite successfully for noisy speech recognition. For the selection of the reference HMM (hidden Markov model) which best matches the noise type and SNR (signal to noise ratio) of the input testing speech, the estimation of the SNR value using the VAD (voice activity detection) algorithm and the classification of the noise type based on the GMM (Gaussian mixture model) have been done separately in the multi-model framework. As the SNR estimation process is vulnerable to errors, we propose an efficient method which can classify simultaneously the SNR values and noise types. The KL (Kullback-Leibler) distance between the single Gaussian distributions for the noise signal during the training and testing is utilized for the classification. The recognition experiments have been done on the Aurora 2 database showing the usefulness of the model compensation method in the multi-model based speech recognizer. We could also see that further performance improvement was achievable by combining the probability density function of the MCT (multi-condition training) with that of the reference HMM compensated by the D-JA (data-driven Jacobian adaptation) in the multi-model based speech recognizer.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.4
• /
• pp.286-292
• /
• 2003

Voice activity detection (VAD) is a process that separates the noise region from silence or noise region of input speech signal. Since unvoiced consonant signals have very similar characteristics to those of noise signals, it may result in serious distortion of unvoiced consonants, or in erroneous noise estimation to can out VAD without paying special attention on unvoiced consonants. In this paper, we propose a method to extract in an explicit way the boundaries between unvoiced consonant and noise in fluent speech so that more exact VAD could be performed. The proposed method is based on histogram in frequency domain which was successfully used by Hirsch for noise estimation, and a1so on similarity measure of frequency components between adjacent frames, To evaluate the performance of the proposed method, experiments on unvoiced consonant boundary extraction was performed on seven kinds of noisy speech signals of 10 ㏈ and 15 ㏈ SNR respectively.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.48 no.1
• /
• pp.101-107
• /
• 2011

The background noises and distortions by channel are major factors that disturb the practical use of speech recognition. Usually, noise reduce the performance of speech recognition system DSR(Distributed Speech Recognition) based speech recognition also bas difficulty of improving performance for this reason. Therefore, to improve DSR-based speech recognition under noisy environment, this paper proposes a method which detects accurate speech region to extract accurate features. The proposed method distinguish speech and noise by using entropy and detection of spectral energy of speech. The speech detection by the spectral energy of speech shows good performance under relatively high SNR(SNR 15dB). But when the noise environment varies, the threshold between speech and noise also varies, and speech detection performance reduces under low SNR(SNR 0dB) environment. The proposed method uses the spectral entropy and harmonics of speech for better speech detection. Also, the performance of AFE is increased by precise speech detections. According to the result of experiment, the proposed method shows better recognition performance under noise environment.

• Journal of Korea Multimedia Society
• /
• v.4 no.1
• /
• pp.56-66
• /
• 2001

The spectral subtraction algorithm using a criterion based on the human perception has been recently developed. The speech processed with Virag's algorithm sounds more pleasant to a human listener than those obtained by the classical methods. However, Virag's algorithm requires a robust voice activity detector (VAD). In the ESS (extended spectral subtraction) algorithm without VAD, the residual noise becomes more noticeable as the SNR decrease. In this paper we propose a new speech enhancement method, the combination of Wiener filter and spectral subtraction based on noise masking characteristics in the human auditory system. There is no need of VAD because the noise can be successively updated even during speech activity using Wiener filter. The adjustment of the subtraction parameter based on the masking threshold makes the residual noise inaudible. The proposed method has been compared with conventional spectral subtraction algorithms. Objective and subjective evaluation of the proposed system is performed with several noise types having different time-frequency distributions. The application of objective measures, the study of the speech spectrograms, as well as subjective listening tests, confirm that the enhanced speech with proposed algorithm is more pleasant to a human listener.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.10 no.8
• /
• pp.901-906
• /
• 2015

We presented a method for Vad(Voice Activity Detection) using Bi-level HMM. Conventional methods need to do an additional post processing or set rule-based delayed frames. To cope with the problem, we applied to VAD a Bi-level HMM that has an inserted state layer into a typical HMM. And we used posterior ratio of voice states to detect voice period. Considering MFCCs(: Mel-Frequency Cepstral Coefficients) as observation vectors, we performed some experiments with voice data of different SNRs and achieved satisfactory results compared with well-known methods.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.49 no.3
• /
• pp.89-97
• /
• 2012

In this paper, we propose a novel approach to noise power estimation for speech enhancement in noisy environments. The method based on IMCRA (improved minima controlled recursive averaging) which is widely used in speech enhancement utilizes a rough VAD (voice activity detection) algorithm which excludes speech components during speech periods in order to improves the performance of the noise power estimation by reducing the speech distortion caused by the conventional algorithm based on the minimum power spectrum derived from the noisy speech. However, since the VAD algorithm is not sufficient to distinguish speech from noise at non-stationary noise and low SNRs (signal-to-noise ratios), the speech distortion resulted from the minimum tracking during speech periods still remained. In the proposed method, minimum power estimate obtained by IMCRA is modified by SMT (spectral minima tracking) to reduce the speech distortion derived from the bias of the estimated minimum power. In addition, in order to effectively estimate minimum power by considering the distribution characteristic of the speech and noise spectrum, the presented method combines the minimum estimates provided by IMCRA and SMT depending on the weighting factor based on the subband. Performance of the proposed algorithm is evaluated by subjective and objective quality tests under various environments and better results compared with the conventional method are obtained.