Voice extraction by on-line signal separation and recovery

Abstract

Abstract-The paper presents a formulation and an implementation of a system for voice output extraction (VOX) in real-time and near-real-time realistic real-world applications. A key component includes voice-signal separation and recovery from a mixture in practical environments. The signal separation and extraction component includes several algorithmic modules with a variety of sophistication levels, which include dynamic processing neural networks in tandem with (dynamic) adaptive methods. These adaptive methods make use of optimization theory subject to the dynamic network constraints to enable practical algorithms. The underlying technology platforms used in the compiled VOX software can significantly facilitate the embedding of speech recognition into many environments. Two demonstrations are described: one is PC-based and is near-realtime, the second is digital signal processing based and is real time. Sample results are described to quantify the performance of the overall systems.Index Terms-Adaptive networks, audio signal processing, DSP, gradient descent, independent component analysis, neural networks, nonlinear networks and systems, optimization, speech processing, state-space models, statistical independence criteria.

A unified balanced approach to multichannel blind deconvolution

“…Similar to the general continuous time MBD algorithm [11], the optimization problem can be solved using the Lagrange multiplier method as follows.…”

“…In this section, we first summarize the concept of the delta operator as presented in [15], then we apply it to the MBD problem such that the general state space MBD theories presented in [10] and [11] are rearranged into a unified framework.…”

“…First we will compare the benchmark algorithm, the proposed algorithm with the continuous time algorithm presented in [11] and the information backpropagation algorithm [7]. We run each algorithm ten times with different initial values.…”

“…6.3 we will examine the benchmark algorithm and the proposed algorithm using a set of artificially mixed speech observations. In particular, we will compare the performance of the two algorithms with the general continuous time state space algorithm [11] and the information backpropagation algorithm [7]; then we will examine the proposed algorithm's ability of estimating the number of states of the mixer. In Sect.…”

“…In [10], a discrete time MBD problem is considered. Correspondingly, in [11], a continuous time version of MBD problem is studied. In both cases, the mixer is modelled as a LTI dynamical system using state space approach; 2 Lagrange multiplier method is used to tackle the resulting optimization problem.…”