In this paper, we present a novel way to design biorthogonal and paraunitary linear phase filter banks. The square error of the perfect reconstruction of the filter bank is expressed in quadratic form of filter coefficients and the cost function is minimized by solving linear equation iteratively without nonlinear optimization. With some modifications, this method is extended to the design of paraunitary filter banks. Furthermore, the lattice structure of odd-channel paraunitary filter banks is also derived. Design examples are given to validate the proposed method.

We present a novel approach for single-channel noise reduction of speech signals contaminated by additive noise. In this approach, the system requires speech samples to be uttered in advance by the same speaker as that of the input signal. Speech samples used in this method must have enough phonetic variety to reconstruct the input signal. In the proposed method, which we refer to as referential reconstruction, we have used a small database created from examples of speech, which will be called reference signals. Referential reconstruction uses an example-based approach, in which the objective is to find the candidate speech frame which is the most similar to the clean input frame without noise, although the input frame is contaminated with noise. When candidate frames are found, they become final outputs without any special processing. In order to find the candidate frames, a correlation coefficient is used as a similarity measure. Through automatic speech recognition experiments, the proposed method was shown to be effective, particularly for low-SNR speech signals corrupted with white noise or noise in high-frequency bands. Since the direct implementation of this method requires infeasible computational cost for searching through reference signals, a coarse-to-fine strategy is introduced in this paper.

In this paper, the Lapped Orthogonal Transform (LOT) with unequal length basis function is considered. The proposed unequal length LOT (ULLOT) has both long basis of length 2M and short basis of length M, while the lengths of all bases of the conventional LOT are 2M. A new class of LOT can be constructed with some modifications of Malvar's Fast LOT. Therefore, the fast algorithm for the Discrete Cosine Transform (DCT) will surely facilitate the computation of the ULLOT. Although the computational complexity of the ULLOT is always lower than that of the LOT, there exist some cases where the coding gain of the ULLOT becomes slightly higher than that of the LOT. Its ability to reduce ringing artifacts is an attractive feature as well. The size-limited structure for the finite length signal is investigated and the ULLOTs are tested on image coding application. The simulation results confirm the validity of the proposed ULLOT.