抄録
We will describe a theoretical background on realization of a certain class of stochastic multi-scale processes. Here, in stead of using Haar systems, we will introduce a more general approach in which we will employ multiple tree. This idea is influenced by the recent works on new smoothers based on general wavelets. The main drawback in Haar systems, despite succesful application recently, is the blocky nature. This is due to the rather rough interpolator being used. We can argue that by using longer taps filter, i.e. multiple tree, it is possible to overcome such weakness.