Visualization of multi-dimensional, irregularly sampled data sets is useful in 3-D seismic as well as in many other geophysical problems, for example, to analyze AVO as a function of azimuth in CDP gathers in a 3-D land data set. The problem with most existing seismic visualizers, including SEP's primary multi-dimensional seismic visualizer Biondi and van Trier (1993), is that they operate on regularly sampled data. In the last year, SEP has been developing a generalization of the SEPlib software package, called SEPlib90Biondi et al. (1996), capable of handling irregularly sampled data. Using SEPlib90 utilities, an irregularly sampled data set can be gridded onto a regularized space. We present here four AVS modules: Read Sep90, Infill90, nD to 2D Slicer, and field to plane, built using SEPlib90 data access routines, which can be combined with standard AVS modules to visualize multi-dimensional irregularly sampled data sets. The first module, Read Sep90, reads a SEPlib90 data set in its entirety or allowing a sub-set by specifying windowing parameters. The second module, Infill90, stacks a data set according to a user selected stacking axis and regularizes it by interpolating the data across a user selected dimension. This module also generates information on the reliability of regularized data. The third module, nD to 2D Slicer, takes 2-D orthogonal slices from both the n-dimensional data array and the corresponding reliability information. The fourth module, field to plane, generates from a 2-D slice the corresponding surface in 3-D space, assigning vertex transparency values or modifying saturation values, according to the reliability information. This paper is in two parts. The first part describes the four new modules interface and operation. The second part presents an AVS network built using the new modules and some standard AVS modules, which can be use to visualize multi-dimensional irregularly sampled data.