Automatic speech recognition based on cepstral coefficients and a mel-based discrete energy operator

Abstract

In this paper, a novel feature vector based on both Mel Frequency Cepstral Coefficients (MFCCs) and a Mel-based nonlinear Discrete-time Energy Operator (MDEO) is proposed to be used as the input of an HMM-based Automatic Continuous Speech Recognition (ACSR) system. Our goal is to improve the performance of such a recognizer using the new feature vector. Experiments show that the use of the new feature vector increases the recognition rate of the ACSR system. The HTK Hidden Markov Model Toolkit was used throughout. Experiments were done on both the TIMIT and NTIMIT databases. For the TIMIT database, when the MDEO was included in the feature vector to test a multi-speaker ACSR system, we found that the error rate decreased by about 9:51. On the other hand, for NTIMIT, the MDEO deteriorates the performance of the recognizer. That is, the new feature vector is useful for clean speech but not for telephone speech.

Time-frequency distributions for automatic speech recognition

“…The energy operator has been applied successfully to demodulation and has many attractive features such as simplicity, efficiency, and adaptability to instantaneous signal variations [3]. The attractive physical interpretation of the energy operator has led to its use as an ASR feature extractor in various forms, see for example [12], [13].…”

“…In general, using (5) the sum of any quadratic operator output (e.g., see [4], [1]) can be expressed as (12) where are arbitrary constants. For narrowband signals , can be assumed constant around and the short-time average of can be expressed as (13) i.e., the difference between the log of any time-frequency distribution produced by the generalized ASR front-end in Fig.…”

“…Given the similarity between the time-frequency distributions of quadratic operators it is expected that ASR performance will also be similar for various front-ends that use short-time averages of quadratic operators as features. However, as the size of the short-time window decreases and/or the bandwidth of the filter increases the differences among are no longer time-invariant, i.e., and significant ASR performance differences may arise between various front-ends (see for example [12] where the energy operator is applied to the unfiltered signal). The equivalence between , and as features (in the cepstrum domain) for ASR is experimentally shown in Section IV.…”

Auditory Teager energy cepstrum coefficients for robust speech recognition

“…For resonance signals, the Teager-Kaiser Energy and the nonlinear energy operator Ψ provide a good estimation of the "real" source energy. Recently, Teager energy has been used for speech recognition in [10,12]. In this paper, we extend this work and design a front-end that combines an auditorymotivated filterbank with the Teager energy estimation method.…”

A wavelet-based parameterization for speech/music discrimination

“…The Teager energy is a noise robust parameter for speech recognition because the effect of additive noise is attenuated: good results are obtained in presence of car engine noise [20]. The Instantaneous energy reflects only the amplitude of the signal whereas the Teager energy operator reflects the variations in both amplitude and frequency of the signal [45]. Figure 4 is an example of two spectrograms: one based on wavelet coefficients (Coiflet, 5 bands, Teager energy) and the other based on STFT coefficients for the same signal.…”