The theoretically minimum length of a signal for fundamental frequency estimation in a noisy environment is discussed. Assuming that the noise is additive white Gaussian, it is known that a Cramér-Rao lower bound (CRLB) is given by the length and other parameters of the signal. In this paper, we define the minimum length as the length whose CRLB is less than or equal to the specific variance for any parameters of the signal. The specific variance is allowable variance of the estimate within an application of fundamental frequency estimation. By reformulating the CRLB with respect to the initial phase of the signal, the algorithms for determining the minimum length are proposed. In addition, we develop the methods of deciding the specific variance for general fundamental frequency estimation and pitch estimation. Simulation results in terms of both the fundamental frequency estimation and the pitch estimation show the validity of our approach.

Non-negative matrix factorization (NMF) is widely used for monaural musical sound source separation because of its efficiency and good performance. However, an additional clustering process is required because the musical sound mixture is separated into more signals than the number of musical tracks during NMF separation. In the conventional method, manual clustering or training-based clustering is performed with an additional learning process. Recently, a clustering algorithm based on the mel-frequency cepstrum coefficient (MFCC) was proposed for unsupervised clustering. However, MFCC clustering supplies limited information for clustering. In this paper, we propose various timbre features for unsupervised clustering and a clustering algorithm with these features. Simulation experiments are carried out using various musical sound mixtures. The results indicate that the proposed method improves clustering performance, as compared to conventional MFCC-based clustering.

Dai-Yuan (DY) conjugate gradient method is an effective method for solving large-scale unconstrained optimization problems. In this paper, a new DY method, possessing a spectral conjugate parameter βk, is presented. An attractive property of the proposed method is that the search direction generated at each iteration is descent, which is independent of the line search. Global convergence of the proposed method is also established when strong Wolfe conditions are employed. Finally, comparison experiments on impulse noise removal are reported to demonstrate the effectiveness of the proposed method.

A stacked structure consisting of ∼ 1 nm-thick MgO and ∼ 4 nm-thick HfO2 was formed on thermally grown SiO2/Si(100) by MOCVD using dipivaloymethanato (DPM) precursors, and the influences of N2 anneal on interfacial reaction and defect state density in this stacked structure were examined. The chemical bonding features of Mg atom were evaluated by using an Auger parameter independently of positive charge-up during XPS measurements. With Mg incorporation into HfO2, a slight decrease in the oxidation number of Mg was detectable. The result suggests that Mg atoms are incorporated preferentially near oxygen vacancies in the HfO2, which can be responsible for a reduction of the flat band voltage shifts observed from C-V characteristics.

A novel delta-sigma modulator that employs a non-uniform quantizer whose spacing is adjusted by reference to the statistical properties of the input signal is proposed. The proposed delta-sigma modulator has less quantization noise compared to the one that uses a uniform quantizer with the same number of output values. With respect to the quantizer on its own, Lloyd proposed a non-uniform quantizer that is best for minimizing the average quantization noise power. The applicable condition of the method is that the statistical properties of the input signal, the probability density, are given. However, the procedure cannot be directly applied to the quantizer in the delta-sigma modulator because it jeopardizes the modulator's stability. In this paper, a procedure is proposed that determine the spacing of the quantizer with avoiding instability. Simulation results show that the proposed method reduces quantization noise by up to 3.8 dB and 2.8 dB with the input signal having a PAPR of 16 dB and 12 dB, respectively, compared to the one employing a uniform quantizer. Two alternative types of probability density function (PDF) are used in the proposed method for the calculation of the output values. One is the PDF of the input signal to the delta-sigma modulator and the other is an approximated PDF of the input signal to the quantizer inside the delta-sigma modulator. Both approaches are evaluated to find that the latter gives lower quantization noise.

The multicast ATM switch has been developed for the purpose of the point-to-multipoint cell transmission. The basic structure of conventional multicast ATM switches is mainly based on a T. T. Lee's multicast switch, since it has a very simple and expandable structure. However, in spite of these benefits, it requires excessive hardware for the loss-free cell transmission, since it employs the blocking network, i.e., the banyan network as a copy network and a routing network. In this paper, we propose a new network for the multicast ATM switching. In proposed copy network, we adopt a new structure, the parallel broadcast banyan network with bypass links between switch planes, to offer the maximum cell transmission capacity and the fault tolerance. All conflict cells, which are blocked during the cell routing process, are bypassed to the next switch plane through bypass links and try to be routed. And to support the highly efficient cell transmission, we propose Alternate Path Scheme (APS) and copy-number (CN) comparator in the proposed copy network. APS is a kind of cell transmission schemes and guarantees multicasting capability to achieve a high performance. To estimate the performance of a proposed copy network, we provide several simulation results.

A method to recognize planar objects undergoing affine transformation is proposed in this paper. The method is based upon wavelet multiscale features and Hopfield neural networks. The feature vector consists of the multiscale wavelet transformed extremal evolution. The evolution contains the information of the contour primitives in a multiscale manner, which can be used to discriminate dominant points, hence a good initial state of the Hopfield network can be obtained. Such good initiation enables the network to converge more efficiently. A wavelet normalization scheme was applied to make our method scale invariant and to reduce the distortion resulting from normalizing the object contours. The Hopfield neural network was employed as a global processing mechanism for feature matching and made our method suitable to recognize planar objects whose shape distortion arising from an affine transformation. The Hopfield network was improved to guarantee unique and more stable matching results. A new matching evaluation scheme, which is computationally efficient, was proposed to evaluate the goodness of matching. Two sets of images, noiseless and noisy industrial tools, undergoing affine transformation were used to test the performance of the proposed method. Experimental results showed that our method is not only effective and robust under affine transformation but also can limit the effect of noises.

In this letter, we propose a sample and hold circuit (S/H circuit) with the clock boost technique and the input signal tracking technique. The proposed circuit block generates the clock with the amplitude of VDD + vin, and the clock is used to control the MOS switch. By applying this circuit to a S/H circuit, we can deal with the rail-to-rail signal with maintaining low distortion. Furthermore, the hold error caused by the charge injection and the clock feedthrough can be also reduced by using the dummy switch. The Star-HSPICE simulation results are reported in this letter.