No abstract

This paper proposes a discriminative layered nonnegative matrix factorization (DL-NMF) for monaural speech separation. The standard NMF conducts the parts-based representation using a single-layer of bases which was recently upgraded to the layered NMF (L-NMF) where a tree of bases was estimated for multi-level or multi-aspect decomposition of a complex mixed signal. In this study, we develop the DL-NMF by extending the generative bases in L-NMF to the discriminative bases which are estimated according to a discriminative criterion. The discriminative criterion is conducted by optimizing the recovery of the mixed spectra from the separated spectra and minimizing the reconstruction errors between separated spectra and original source spectra. The experiments on single-channel speech separation show the superiority of DL-NMF to NMF and L-NMF in terms of the SDR, SIR and SAR measures.

SUMMARY This paper proposes a voice activity detection (VAD) algorithm based on an energy related feature of the frequency modulation of harmonics. A multi-resolution spectro-temporal analysis framework, which was developed to extract texture features of the audio signal from its Fourier spectrogram, is used to extract frequency modulation features of the speech signal. The proposed algorithm labels the voice active segments of the speech signal by comparing the energy related feature of the frequency modulation of harmonics with a threshold. Then, the proposed VAD is implemented on one of Texas Instruments (TI) digital signal processor (DSP) platforms for real-time operation. Simulations conducted on the DSP platform demonstrate the proposed VAD performs significantly better than three standard VADs, ITU-T G.729B, ETSI AMR1 and AMR2, in non-stationary noise in terms of the receiver operating characteristic (ROC) curves and the recognition rates from a practical distributed speech recognition (DSR) system.

Spectro-temporal modulations of speech encode speech structures and speaker characteristics. An algorithm which distinguishes speech from non-speech based on spectro-temporal modulation energies is proposed and evaluated in robust text-independent closed-set speaker identification simulations using the TIMIT and GRID corpora. Simulation results show the proposed method produces much higher speaker identification rates in all signal-to-noise ratio (SNR) conditions than the baseline system using mel-frequency cepstral coefficients. In addition, the proposed method also outperforms the system, which uses auditory-based nonnegative tensor cepstral coefficients [Q. Wu and L. Zhang, "Auditory sparse representation for robust speaker recognition based on tensor structure," EURASIP J. Audio, Speech, Music Process. 2008, 578612 (2008)], in low SNR (≤ 10 dB) conditions.

No abstract