Norihiro Takamune

Sparse sound field decomposition with multichannel extension of complex NMF

A sparse sound field decomposition method using prior information on source signals in the time-frequency domain is proposed. Sparse sound field decomposition has been proved to be effective for various acoustic signal processing applications. Current methods for sparse decomposition are based only on the spatial sparsity of the source distribution. However, it can be assumed that possible source signals to be decomposed are approximately known in advance. To exploit this prior information, we incorporated the complex nonnegative factorization model into sparse sound field decomposition. Since the magnitude spectrum of the possible source signals can be trained in advance, accuracy of the sparse decomposition can be improved even when the source signals are highly correlated and the sources are in a highly noisy environment. In addition, the proposed decomposition algorithm is derived using the auxiliary function method. Numerical experiments indicated that the sparse decomposition performance was significantly improved using the proposed method.

Music signal separation using supervised NMF with all-pole-model-based discriminative basis deformation

In this paper, we address the music signal separation problem and propose a new supervised nonnegative matrix factorization (SNMF) algorithm employing the deformation of a spectral supervision basis trained in advance. Conventional SNMF has a problem that the separation accuracy is degraded by a mismatch between the trained basis and the spectrogram of the actual target sound in open data. To reduce the mismatch problem, we propose a new method with two features. First, we introduce a deformation with an all-pole model that is optimized to make the trained basis fit the spectrogram of the target signal, even if the true target component is hidden in the observed mixture. Next, to avoid an excess deformation, we limit the degree of freedom in the deformation by performing discriminative training. Our experimental evaluation reveals that the proposed method outperforms conventional SNMFs.

Maximum reconstruction probability training of Restricted Boltzmann machines with auxiliary function approach

Restricted Boltzmann machines (RBMs) are stochastic neural networks that can be used to learn features from raw data. They have attracted particular attention recently after being proposed as building blocks for deep belief network (DBN) and have been applied with notable success in a range of problems including speech recognition and object recognition. The success of these models raises the issue of how best to train them. At present, the most popular training algorithm for RBMs is the Contrastive Divergence (CD) learning algorithm. We propose deriving a new training algorithm based on an auxiliary function approach for RBMs using the reconstruction probability of observations as the optimization criterion. Through an experiment on parameter training of an RBM, we confirmed that the present algorithm outperformed the CD algorithm in terms of the convergence speed and the reconstruction error when used as an autoencoder.

Training Restricted Boltzmann Machines with auxiliary function approach

Restricted Boltzmann Machines (RBMs) are neural network models for unsupervised learning, but have recently found a wide range of applications as feature extractors for supervised learning algorithms. They have also received a lot of attention recently after being proposed as building blocks for deep belief networks. The success of these models raises the issue of how best to train them. At present, the most popular training algorithm for RBMs is the Contrastive Divergence (CD) learning algorithm. The aim of this paper is to seek for a new optimization algorithm tailored for training RBMs in the hope of obtaining a faster algorithm than the CD algorithm. We propose deriving a new training algorithm for RBMs based on an auxiliary function approach. Through an experiment on parameter training of an RBM, we confirmed that the present algorithm converged faster and to a better solution than the CD algorithm.

Spatio-temporal sparse sound field decomposition considering acoustic source signal characteristics

We propose a sound field decomposition method that takes into consideration spatio-temporal sparsity. It has been proved that sparse representation of a sound field is effective in reducing errors originating from spatial aliasing artifacts compared with conventional plane wave decomposition. In most current methods of sparse sound field decomposition, the spatial sparsity of the sound source distribution is only assumed. However, it is known that the temporal structure of the source signal to be decomposed can also be sparse in the time-frequency domain. We formulate an objective function for sparse sound field decomposition by using the ℓp,q-norm to simultaneously induce sparsity in the space and time domains. An optimization algorithm on the auxiliary function method is derived to solve it. Numerical simulations of acoustic holography indicate that the reconstruction accuracy can be improved by controlling the parameter of temporal sparsity. We also demonstrate that a statistical measure of the source signals can be used as an indicator to determine a nearly optimal parameter.

Audio signal separation using supervised NMF with time-variant all-pole-model-based basis deformation

We address a novel nonnegative matrix factorization (NMF) with a new basis deformation method to handle various music sounds. Conventional supervised NMF has a critical problem that a mismatch between bases trained in advance and an actual target sound reduces the accuracy of separation. To solve this problem, we proposed an advanced supervised NMF that applies a single time-invariant filter to the bases for making them fit into the target sound. However, this method suffers from limitations on basis deformation ability, especially for transient instrumental sounds. In this paper, we propose a new time-variant all-pole-model-based basis deformation method. Our proposed deformation method consists of two types of filter that individually deforms attack and sustain parts in one note. Each of the all-pole models can be automatically selected and adapted to the open data via a statistical signal sampling approach. Experimental results show that the proposed method outperforms conventional methods in many types of instrumental sound.

Sparse sound field decomposition with parametric dictionary learning for super-resolution recording and reproduction

A method for sparse sound field decomposition with parametric dictionary learning is proposed. Sound field decomposition forms the foundation of various acoustic signal processing applications. Our main focus is sound field recording and reproduction for high-fidelity audio systems. To improve the reproduction accuracy above the spatial Nyquist frequency, determined by the intervals between array elements, i.e., super-resolution in recording and reproduction, We have proposed a method based on sparse sound field decomposition. For more accurate decomposition, we propose a method for parametric dictionary learning to adaptively optimize the dictionary parameters using input signals. The proposed method is based on Newtons method including pruning of the columns of a dictionary matrix. Numerical simulation results indicate that more accurate sparse decomposition is achieved by the proposed method. The reproduction accuracy is also improved above the spatial Nyquist frequency.