150 likes | 161 Views
This study focuses on enhancing binaural noise reduction algorithms using ITD and ILD cues, with a focus on speech preservation and noise reduction. Experimental results and binaural auditory cues are discussed.
E N D
Extension of the multi-channel Wiener filter with localization cues for binaural noise reduction Simon Doclo1,2, Rong Dong2, Thomas J. Klasen1,3, Jan Wouters3 , Simon Haykin2, Marc Moonen1 1Dept. of Electrical Engineering (ESAT-SCD), KU Leuven, Belgium 2Adaptive Systems Laboratory, McMaster University, Canada 3 Laboratory for Exp. ORL, KU Leuven, Belgium WASPAA-2005, Oct 17 2005
Overview • Binaural hearing aids: noise reduction and preservation of binaural cues • Overview of binaural noise reduction algorithms • Binaural multi-channel Wiener filter: • Estimate of speech component at both hearing aids • Speech cues are preserved – noise cues may be distorted • Preservation of binaural noise cues: • Partial estimation of noise component • Extension with ITD and ILD cost function • Experimental results • Audio demonstration
Binaural auditory cues: • Interaural Time Difference (ITD) – Interaural Level Difference (ILD) • Binaural cues, in addition to spectral and temporal cues, play an important role in binaural noise reduction and sound localization t PR PL Problem statement • Hearing impairment reduction of speech intelligibility in background noise • Signal processing to selectively enhance useful speech signal • Many hearing impaired are fitted with hearing aid at both ears • Problem statement • Binaural noise reduction • Multi-channel Wiener filter • Preservation of binaural cues • Experiments • Conclusions
Problem statement • Bilateral system: • Independent processing of left and right hearing aid • Binaural cues may be distorted (cf. poster Tim Van den Bogaert) • Binaural system: • Cooperation between left and right hearing aid (e.g. wireless link) • Assumption: all microphone signals are available at the same time • Problem statement • Binaural noise reduction • Multi-channel Wiener filter • Preservation of binaural cues • Experiments • Conclusions • Objectives/requirements for binaural algorithm: • SNR improvement: noise reduction, limit speech distortion • Preservation of binaural cues (speech/noise) to exploit binaural hearing advantage • No assumption about position of speech source and microphones
[Desloge, 1997] [Lotter, 2004] • Adaptive beamforming: based on GSC-structure • Divide frequency spectrum: low-pass portion unaltered to preserve ITD cues, high-pass portion processed using GSC low-pass: no noise reduction, high-pass: no cue preservation • TF-GSC: minimize output energy, constraint: output speech component = speech component in reference microphone signal binaural noise cues may be distorted [Welker, 1997] [Gannot, 2001] • CASA-based techniques • Computation and application of (real-valued) binaural mask based on binaural and temporal/spectral cues mostly for 2 microphones, “spectral-subtraction”-like problems [Kollmeier, Peissig, Wittkop, Dong, Haykin] Binaural noise reduction techniques • Fixed beamforming: spatial selectivity + binaural speech cues • Maximize directivity index while restricting ITD error • Superdirective beamformer using HRTFS broadside array, limited performance, known geometry • Problem statement • Binaural noise reduction • Multi-channel Wiener filter • Preservation of binaural cues • Experiments • Conclusions
Extension of MWF: preservation of binaural speech and noise cues without comprimising noise reduction performance Binaural noise reduction techniques • Binaural multi-channel Wiener filter • MMSE estimate of speech component in reference microphone signal at both ears speech cues are preserved, no assumptions about position of speech source and microphones noise cues may be distorted • Problem statement • Binaural noise reduction • Multi-channel Wiener filter • Preservation of binaural cues • Experiments • Conclusions [Doclo, Klasen, Wouters, Moonen]
Use all available microphone signals to compute output signal at both ears computation of filters W0 and W1 Binaural multi-channel Wiener filter • Configuration: microphone array at left and right hearing aid • Problem statement • Binaural noise reduction • Multi-channel Wiener filter • Preservation of binaural cues • Experiments • Conclusions
Binaural multi-channel Wiener filter • SDW-MWF: estimate speech component in reference mic signal at both ears + trade-off noise reduction and speech distortion • Problem statement • Binaural noise reduction • Multi-channel Wiener filter • Preservation of binaural cues • Experiments • Conclusions
Partial estimation of noise component • Estimate of sum of speech component and scaled noise component considerable reduction of noise reduction performance [Klasen, 2005] • Extension of SDW-MWF with binaural cues • Add term related to ITD and ILD cue to SDW cost function • Link computation of filters W0 and W1 • Weight factors can be frequency-dependent • Task: perceptually relevant expressions for binaural cues Preservation of binaural cues • SDW-MWF: binaural speech cues are generally preserved, binaural noise cues may be distorted • Problem statement • Binaural noise reduction • Multi-channel Wiener filter • Preservation of binaural cues • Experiments • Conclusions
ILD: power ratio of two signals • Output: input: • Other possibility: specifiy desired angle v and use HRTFs • Estimate noise cross-correlation/power during noise segments • No closed-form expression iterative optimization techniques ITD-ILD cost functions • ITD: phase of cross-correlation between two signals • Output: input: • Cost function: cosine of phase difference between cross-correlations • Problem statement • Binaural noise reduction • Multi-channel Wiener filter • Preservation of binaural cues • Experiments • Conclusions
Reverberation time= 125 msec Noise 1 HAL HAR Experimental results • Binaural recordings on KEMAR, 2 microphones on each hearing aid (d=1cm) • Speech source in front, multi-talker babble noise at 45 • SNR=0 dB, fs=16 kHz, FFT-size N=256, 0= 1=1 • Performance measures: • SNR improvement (left/right) • Mean ITD/ILD cost function (speech/noise component) • Problem statement • Binaural noise reduction • Multi-channel Wiener filter • Preservation of binaural cues • Experiments • Conclusions
Experimental results • Partial estimation of noise component • ITD and ILD cost function of speech and noise components decrease • SNR improvement is significantly reduced • Problem statement • Binaural noise reduction • Multi-channel Wiener filter • Preservation of binaural cues • Experiments • Conclusions
Experimental results • Extension with ITD-ILD cost function • ITD and ILD cost function of noise component decrease • SNR improvement is practically not reduced • ITD and ILD cost function of speech component not degraded • Problem statement • Binaural noise reduction • Multi-channel Wiener filter • Preservation of binaural cues • Experiments • Conclusions
Audio demo • PS: Other demos available (=5, 2 noise sources, T60=250 ms)
Conclusions and future work • Speech enhancement for binaural hearing aids: • SNR improvement • Preservation of binaural speech and noise cues • No assumptions about position speech source and microphones • Suitable noise reduction algorithms: • Multi-channel Wiener filter (but also e.g. Transfer Function GSC) speech cues are preserved noise cues may be distorted • Preservation of binaural noise cues: • Extension with ITD and ILD cost functions • Other possible extensions: Interaural Transfer Function (ITF) • Future work: • Multiple noise sources ? • Better perceptual cost functions/performance measures • Tuning of frequency-dependent weight factors • Problem statement • Binaural noise reduction • Multi-channel Wiener filter • Preservation of binaural cues • Experiments • Conclusions