THE STUDIES OF THE LOST SAMPLE RECONSTRUCTION: TRANSFORM DOMAIN APPROACH

• Main Authors: Ping-Chang Shih, 施並彰
• Other Authors: Prof. Chau-Yun Hsu
• Format: Others
• Language: en_US
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/80019767059001838702

碩士 === 大同大學 === 通訊工程研究所 === 92 === Discrete Fourier Transform (DFT) is a popular tool for digital signal processing. And the other hand, lost sample reconstruction has been proposed by Papoulis-Gerchberg, by using iteration with DFT. Hsu and Chen[12] propose an improved approach from the view of the boundary-matched. But they process the signal into boundary-matched signal in time domain, it need the differential of the original signal. Generally, it is not easy to get the differential data. In this thesis, we will propose new approach that the process the signal in transform domain. The new approach can achieve the same performance as that of Hsu and Chen, with less complexity and flexibility. We do not need the differential of the original signal. It is ease and convenient for processing the recovery of the loss samples.