Short-time homomorphic analysis

Homomorphic deconvolution has been successfully applied in a variety of areas. In many cases of interest, including speech and seismic processing, the signals to be analyzed are non-stationary and approximately follow a convolutional model only on a short-time basis. Thus, a window is applied to the data. In this paper a first attempt is made to understand the interaction between short-time windowing and homomorphic deconvolution. A model for short-time homomorphic analysis is proposed, which provides a framework for the interpretation of window effects encountered in speech and seismic data processing.