ETSI Advanced Front-End Research Articles

An investigation of the impact of MVA normalization on the advanced front-end features

Feature normalization is a key objective in speech related applications. In this paper, we study the effects of the Mean subtraction, Variance normalization, and Autoregressive Moving Average (ARMA) filtering (MVA) normalization method on the ETSI Advanced Front-End (AFE) features. A series of experiments, on the Aurora-2 task, was conducted to show the impact of MVA normalization for different subsets of AFE feature components. Compared to the AFE baseline system, recognition results show performance improvement when only the logarithmic energy coefficient is normalized. However, the performance is degraded through the normalization of the rest of AFE coefficients. To investigate this degradation, other experiments were performed by eliminating the AFE implemented blind equalization post-processing block. It has shown that one part of this degradation can plausibly be interpreted as the effect of over-normalization caused by the MVA post-processing to the AFE original features. Furthermore, by analyzing the statistical distributions of AFE features we found that the effectiveness of MVA could also be affected by the high intra-frame variability of AFE features.

An experimental framework for Arabic digits speech recognition in noisy environments

In this paper we present an experimental framework for Arabic isolated digits speech recognition named ARADIGITS-2. This framework provides a performance evaluation of Modern Standard Arabic devoted to a Distributed Speech Recognition system, under noisy environments at various Signal-to-Noise Ratio (SNR) levels. The data preparation and the evaluation scripts are designed by deploying a similar methodology to that followed in AURORA-2 database. The original speech data contains a total of 2704 clean utterances, spoken by 112 (56 male and 56 female) Algerian native speakers, down-sampled at 8 kHz. The feature vectors, which consist of a set of Mel Frequency Cepstral Coefficients and log energy, are extracted from speech samples using ETSI Advanced Front-End (ETSI-AFE) standard; whereas, the Hidden Markov Models (HMMs) Toolkit is used for building the speech recognition engine. The recognition task is conducted in speaker-independent mode by considering both word and syllable as acoustic units. Therefore, an optimal fitting of HMM parameters, as well as the temporal derivatives window, is carried out through a series of experiments performed on the two training modes: clean and multi-condition. Better results are obtained by exploiting the polysyllabic nature of Arabic digits. These results show the effectiveness of syllable-like unit in building Arabic digits recognition system, which exceeds word-like unit by an overall Word Accuracy Rate of 0.44 and 0.58% for clean and multi-condition training modes, respectively.

Improved ETSI advanced front-end for automatic speech recognition using quadrature mirror filter synthesis based on robust complex speech analysis

An automatic speech recognition (ASR) is commonly used in these days. Current ASR systems perform well in ideal environment; however, it does not perform well in realistic noisy environment. As a robust ASR, ETSI has standardized Advanced Front-End (AFE) that adopts two-stage of iterative Wiener filter (IWF) to realize a speech enhancement as the front-end of ASR. In the ETSI AFE, 16 kHz speech is divided uniformly by QMF (Quadrature Mirror Filter) into lower-band and higher-band signal and only lower-band signal is emphasized by the IWF and MFCC (Mel Frequency Cepstral Coefficient) is extracted from both the second-stage of emphasized lower-band signal and non-emphasized higher-band signal. FFT is used to estimate speech spectrum that designs the Wiener filter. On the other hand, we have already proposed robust complex speech analysis for an analytic signal. It can estimate more robust and more accurate speech spectrum due to the introduced robust criterion and nature of analytic signal. This paper proposes an improved AFE using wide-band robust ELS (Extended Least Square) complex analysis and real-valued analysis instead of FFT. Moreover, the QMF synthesis is also introduced for second-stage of wide-band analysis. The experimental results using the CENSREC-2 speech database demonstrate that the performance is improved.

Power-Normalized Cepstral Coefficients (PNCC) for Robust Speech Recognition

This paper presents a new feature extraction algorithm called power normalized Cepstral coefficients (PNCC) that is motivated by auditory processing. Major new features of PNCC include the use of a power-law nonlinearity that replaces the traditional log nonlinearity used in MFCC coefficients, a noise-suppression algorithm based on asymmetric filtering that suppresses background excitation, and a module that accomplishes temporal masking. We also propose the use of medium-time power analysis in which environmental parameters are estimated over a longer duration than is commonly used for speech, as well as frequency smoothing. Experimental results demonstrate that PNCC provides substantial improvements in recognition accuracy compared to MFCC and PLP for speech in the presence of various types of additive noise and in reverberant environments, with only slightly greater computational cost than conventional MFCC processing, and without degrading the recognition accuracy that is observed while training and testing using clean speech. PNCC also provides better recognition accuracy in noisy environments than techniques such as vector Taylor series (VTS) and the ETSI advanced front end (AFE) while requiring much less computation. We describe an implementation of PNCC using online processing that does not require future knowledge of the input.

An efficient low bit-rate compression scheme of acoustic features for distributed speech recognition

Due to the limited network bandwidth, a noise robust low bit-rate compression scheme of Mel frequency cepstral coefficients (MFCCs) is desired for distributed speech recognition (DSR) services. In this paper, we present an efficient MFCCs compression method based on weighted least squares (W-LS) polynomial approximation through the exploitation of the high correlation across consecutive MFCC frames. Polynomial coefficients are quantized by designing a tree structured vector quantization (TSVQ) based scheme. Recognition experiments are conducted on the noisy Aurora-2 database, under both clean and multi-condition training modes. The results show that the proposed W-LS encoder slightly exceeds the ETSI advanced front-end (ETSI-AFE) baseline system for bit-rates ranging from 1400 bps to 1925 bps under clean training mode. However, a negligible degradation is observed in case of multi-condition training mode (around 0.6% and 0.2% at 1400 bps and 1925 bps, respectively). Furthermore, the obtained performance generally outperforms the ETSI-AFE source encoder at 4400 bps under clean training and provides similar performance, at 1925 bps, under multi-condition training.

Robust feature extraction based on an asymmetric level-dependent auditory filterbank and a subband spectrum enhancement technique

In this paper we introduce a robust feature extractor, dubbed as robust compressive gammachirp filterbank cepstral coefficients (RCGCC), based on an asymmetric and level-dependent compressive gammachirp filterbank and a sigmoid shape weighting rule for the enhancement of speech spectra in the auditory domain. The goal of this work is to improve the robustness of speech recognition systems in additive noise and real-time reverberant environments. As a post processing scheme we employ a short-time feature normalization technique called short-time cepstral mean and scale normalization (STCMSN), which, by adjusting the scale and mean of cepstral features, reduces the difference of cepstra between the training and test environments. For performance evaluation, in the context of speech recognition, of the proposed feature extractor we use the standard noisy AURORA-2 connected digit corpus, the meeting recorder digits (MRDs) subset of the AURORA-5 corpus, and the AURORA-4 LVCSR corpus, which represent additive noise, reverberant acoustic conditions and additive noise as well as different microphone channel conditions, respectively. The ETSI advanced front-end (ETSI-AFE), the recently proposed power normalized cepstral coefficients (PNCC), conventional MFCC and PLP features are used for comparison purposes. Experimental speech recognition results demonstrate that the proposed method is robust against both additive and reverberant environments. The proposed method provides comparable results to that of the ETSI-AFE and PNCC on the AURORA-2 as well as AURORA-4 corpora and provides considerable improvements with respect to the other feature extractors on the AURORA-5 corpus.

Utilizing Compressibility in Reconstructing Spectrographic Data, With Applications to Noise Robust ASR

In this letter, we propose a novel algorithm for reconstructing unreliable spectrographic data, a method applicable to missing feature-based automatic speech recognition (ASR). We provide quantitative analysis illustrating the high compressibility of spectrographic speech data. The existence of sparse representations for spectrographic data motivates the spectral reconstruction solution to be posed as an optimization problem minimizing the lscr1-norm. When applied to the Aurora-2 database, the proposed missing feature estimation algorithm is shown to provide significant improvements in recognition accuracy relative to the baseline MFCC system. Even without an oracle mask, performance approaches that of the ETSI advanced front end (AFE) , with less complexity.

A robust scheme for distributed speech recognition over loss-prone packet channels

In this paper, we propose a whole recovery scheme designed to improve robustness against packet losses in distributed speech recognition systems. This scheme integrates two sender-driven techniques, namely, media-specific forward error correction (FEC) and frame interleaving, along with a receiver-based error concealment (EC) technique, the weighted Viterbi algorithm (WVA). Although these techniques have been already tested separately, providing a significant increase of performance in clean acoustic environments, in this paper they are jointly applied and their performance in adverse acoustic conditions is evaluated. In particular, a noisy speech database and the ETSI Advanced Front-end are used, while the dynamic features, which play an important role in adverse acoustic environments, and their confidences for the WVA algorithm are examined. In order to solve the issue of mixing two sender-driven techniques (both causing a delay) whose direct composition causes an increase of the global latency, we propose a double stream scheme which limits the latency to the maximum delay of both techniques. As a result, with very few overhead bits and a very limited delay, the integrated scheme achieves a significant improvement in the performance of a DSR system over a degraded transmission channel, both in clean and noisy acoustic conditions.

Robust Speech Recognition Using a Cepstral Minimum-Mean-Square-Error-Motivated Noise Suppressor

We present an efficient and effective nonlinear feature-domain noise suppression algorithm, motivated by the minimum-mean-square-error (MMSE) optimization criterion, for noise-robust speech recognition. Distinguishing from the log-MMSE spectral amplitude noise suppressor proposed by Ephraim and Malah (E&M), our new algorithm is aimed to minimize the error expressed explicitly for the Mel-frequency cepstra instead of discrete Fourier transform (DFT) spectra, and it operates on the Mel-frequency filter bank's output. As a consequence, the statistics used to estimate the suppression factor become vastly different from those used in the E&M log-MMSE suppressor. Our algorithm is significantly more efficient than the E&M's log-MMSE suppressor since the number of the channels in the Mel-frequency filter bank is much smaller (23 in our case) than the number of bins (256) in DFT. We have conducted extensive speech recognition experiments on the standard Aurora-3 task. The experimental results demonstrate a reduction of the recognition word error rate by 48% over the standard ICSLP02 baseline, 26% over the cepstral mean normalization baseline, and 13% over the popular E&M's log-MMSE noise suppressor. The experiments also show that our new algorithm performs slightly better than the ETSI advanced front end (AFE) on the well-matched and mid-mismatched settings, and has 8% and 10% fewer errors than our earlier SPLICE (stereo-based piecewise linear compensation for environments) system on these settings, respectively.

Robust In-Car Speech Recognition Based on Nonlinear Multiple Regressions

We address issues for improving handsfree speech recognition performance in different car environments using a single distant microphone. In this paper, we propose a nonlinear multiple-regression-based enhancement method for in-car speech recognition. In order to develop a data-driven in-car recognition system, we develop an effective algorithm for adapting the regression parameters to different driving conditions. We also devise the model compensation scheme by synthesizing the training data using the optimal regression parameters and by selecting the optimal HMM for the test speech. Based on isolated word recognition experiments conducted in 15 real car environments, the proposed adaptive regression approach shows an advantage in average relative word error rate (WER) reductions of 52.5% and 14.8%, compared to original noisy speech and ETSI advanced front end, respectively.