Statistical voice activity detection using a multiple observation likelihood ratio test

Abstract

Currently, there are technology barriers inhibiting speech processing systems that work in extremely noisy conditions from meeting the demands of modern applications. This letter presents a new voice activity detector (VAD) for improving speech detection robustness in noisy environments and the performance of speech recognition systems. The algorithm defines an optimum likelihood ratio test (LRT) involving multiple and independent observations. The so-defined decision rule reports significant improvements in speech/nonspeech discrimination accuracy over existing VAD methods that are defined on a single observation and need empirically tuned hangover mechanisms. The algorithm has an inherent delay that, for several applications, including robust speech recognition, does not represent a serious implementation obstacle. An analysis of the overlap between the distributions of the decision variable shows the improved robustness of the proposed approach by means of a clear reduction of the classification error as the number of observations is increased. The proposed strategy is also compared to different VAD methods, including the G.729, AMR, and AFE standards, as well as recently reported algorithms showing a sustained advantage in speech/nonspeech detection accuracy and speech recognition performance.Index Terms-Multiple observation likelihood ratio test (MO-LRT), robust speech recognition, voice activity detection.

Deep Belief Networks Based Voice Activity Detection

“…It is mainly for the detection efficiency. We apply the multiple observation technique [14], [15], [25] to the DFT and MFCC features. This technique has two advantages.…”

“…It fuses the likelihood ratio tests of different statistical models in a linear weighted combination way with the weights optimized by the gradient descent algorithm. Yu and Hansen [14] inherited the advantages of the statistical-model-based multiple observation techniques [25], [27], and proposed to fuse the likelihood ratio tests of multiple observations by the discriminative weight training. Inspired by Kang's VAD and Yu's VAD [14], Suh and Kim [18] further proposed to conduct the linear weighted combination of multiple acoustic models and multiple observations together with all weights optimized by a generalized probabilistic descent algorithm.…”

Voice Activity Detection. Fundamentals and Speech Recognition System Robustness

“…An improvement over the LRT proposed by Sohn (Sohn et al, 1999) is the multiple observation LRT (MO-LRT) proposed by Ramírez (Ramírez et al, 2005b). The performance of the decision rule was improved by incorporating more observations to the statistical test.…”

“…The non-speech hit rate (HR0) and the false alarm rate (FAR0= 100-HR1) were determined in each noise condition being the actual speech frames and actual speech pauses determined by handlabeling the database on the close-talking microphone. Figure 9 shows the ROC curves of the MO-LRT VAD (Ramírez et al, 2005b) and other frequently referred algorithms for recordings from the distant microphone in quiet and high noisy conditions. The working points of the G.729, AMR, and AFE VADs are also included.…”

Fuzzy Logic Speech/Non-speech Discrimination for Noise Robust Speech Processing

“…The classification task is not as trivial as it appears, and most of the VAD algorithms often fail in high noise conditions. During the last decade, numerous researchers have developed different strategies for detecting speech in a noisy signal [6,7,8,9] and have evaluated the influence of the VAD effectiveness on the performance of speech processing systems [10,11,12,13].…”