Research Interests:
ABSTRACT In this paper, we review a blind musical-noise-free speech extraction method using a microphone array that can be applied to nonstationary noise. In our previous study, it was found that optimized iterative spectral subtraction... more
ABSTRACT In this paper, we review a blind musical-noise-free speech extraction method using a microphone array that can be applied to nonstationary noise. In our previous study, it was found that optimized iterative spectral subtraction (SS) results in speech enhancement with almost no musical noise generation, but this method is valid only for stationary noise. The proposed method consists of iterative blind dynamic noise estimation by, e.g., ICA or multichannel Wiener filtering, and musical-noise-free speech extraction by modified iterative SS, where multiple iterative SS is applied to each channel while maintaining the multichannel property reused for the dynamic noise estimators. Also, related to the proposed method, we discuss the justification of applying ICA to such signals nonlinearly distorted by SS. From objective and subjective evaluations simulating real-world hands-free speech communication system, we reveal that the proposed method outperforms the conventional methods.
Research Interests:
Research Interests:
Research Interests:
An echo canceller removes undesired echo in fullduplex speech communication. The cancellation is done by modeling the echo path impulse response with an adaptive finite impulse response filter and subtracting the echo estimate from the... more
An echo canceller removes undesired echo in fullduplex speech communication. The cancellation is done by modeling the echo path impulse response with an adaptive finite impulse response filter and subtracting the echo estimate from the received signal. A ...
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Abstract The impact of auditory masking when obtaining sparse representations for audio signals is investigated by integrating a masking model into a Locally Competitive Algorithm. The masking model considers both temporal and frequency... more
Abstract The impact of auditory masking when obtaining sparse representations for audio signals is investigated by integrating a masking model into a Locally Competitive Algorithm. The masking model considers both temporal and frequency domain masking effects. The performance of the new algorithm is verified against the original Locally Competitive Algorithm for different types of sound files. Results show that the new algorithm allows for larger residual error between an input signal and its reconstructed version while ...
Research Interests:
Abstract A novel framework based on graph theory for structure discovery is applied to audio to find new types of audio objects which enable the compression of an input signal. It converts the sparse time-frequency representation of an... more
Abstract A novel framework based on graph theory for structure discovery is applied to audio to find new types of audio objects which enable the compression of an input signal. It converts the sparse time-frequency representation of an audio signal into a graph by representing each data point as a vertex and the relationship between two vertices as an edge. Each edge is labelled based on a clustering algorithm which preserves a quality guarantee on the clusters. Frequent subgraphs are then extracted from this graph, via a ...
Research Interests:
Research Interests:
ABSTRACT
Research Interests:
Research Interests:
Research Interests: Algorithms and Sensors
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
ABSTRACT
Research Interests:
ABSTRACT
Research Interests:
In this paper, we present new versions of the fast RLS adaptive algorithm. These versions are based on a combination of the block filtering technique and a use of a scalar accelerator parameter in each block. These mixing techniques lead... more
In this paper, we present new versions of the fast RLS adaptive algorithm. These versions are based on a combination of the block filtering technique and a use of a scalar accelerator parameter in each block. These mixing techniques lead a faster convergence of the fast RLS algorithm and behave better with time-varying acoustic systems. The proposed versions have approximately
Research Interests:
Research Interests:
Research Interests:
... gilbert.soulodre@crc.ca ABSTRACT Inverse filtering methods commonly use techniques such as regularization and/or smoothing to reduce artifacts created by the ... In [1] the frequency-dependent regularization term was based on the 1/3... more
... gilbert.soulodre@crc.ca ABSTRACT Inverse filtering methods commonly use techniques such as regularization and/or smoothing to reduce artifacts created by the ... In [1] the frequency-dependent regularization term was based on the 1/3 octave spectrum of the transfer function C ...
ABSTRACT A new algorithm for blind signal separation of speech signals that does not require pre-whitening is proposed in this paper. The algorithm is based on second order optimization using Riemannian geometry. The algorithm employs... more
ABSTRACT A new algorithm for blind signal separation of speech signals that does not require pre-whitening is proposed in this paper. The algorithm is based on second order optimization using Riemannian geometry. The algorithm employs several practical approximations to the Hessian matrix of the maximum-likelihood blind separation cost function, to produce a computationally efficient algorithm that is capable of working on-line. Simulation results show the improved performance of the proposed algorithm with different mixing data
Research Interests: Information Technology, Information Theory, Computational Modeling, Convergence, Blind Source Separation, and 9 moreSpeech Processing, Speech, Geometry, Maximum Likelihood, Riemannian Geometry, Efficient Algorithm for ECG Coding, Maximum Likelihood Estimation, Cost Function, and Blind Signal Separation
Research Interests:
ABSTRACT A classical adaptive filtering view of the problem of instantaneous blind signal separation is presented. This classical form enables an easy understanding of the natural gradient algorithm. A new RLS-based algorithm is developed... more
ABSTRACT A classical adaptive filtering view of the problem of instantaneous blind signal separation is presented. This classical form enables an easy understanding of the natural gradient algorithm. A new RLS-based algorithm is developed using this classical interpretation. The algorithm provides improved on-line separation speed under the same steady state error compared to the natural gradient algorithm without requiring pre-whitening
Research Interests:
ABSTRACT A new algorithm for blind signal separation that does not require pre-whitening is proposed in this paper. The algorithm is based on an iterative inversion of the mixing matrix. The algorithm is capable of working on-line and... more
ABSTRACT A new algorithm for blind signal separation that does not require pre-whitening is proposed in this paper. The algorithm is based on an iterative inversion of the mixing matrix. The algorithm is capable of working on-line and provides improved convergence speed and steady state error compared to the popular natural gradient algorithm, with a low additional computational cost