Article Free Access Share on Educating the digital artist for the entertainment industry (panel): the collision of academia and business Authors: Charles S. Swartz UCLA Extension UCLA ExtensionView Profile , Edwin E. Catmull Pixar PixarView Profile , Robin King Sheridan College Sheridan CollegeView Profile , Richard Weinberg University of Southern California University of Southern CaliforniaView Profile , Jane Veeder San Francisco State University San Francisco State UniversityView Profile Authors Info & Claims SIGGRAPH '97: Proceedings of the 24th annual conference on Computer graphics and interactive techniquesAugust 1997 Pages 456–458https://doi.org/10.1145/258734.258907Published:03 August 1997Publication History 1citation499DownloadsMetricsTotal Citations1Total Downloads499Last 12 Months21Last 6 weeks1 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteeReaderPDF

An architecture is presented for fast high-quality rendering of complex images. All objects are reduced to common world-space geometric entities called micropolygons, and all of the shading and visibility calculations operate on these micropolygons. Each type of calculation is performed in a coordinate system that is natural for that type of calculation. Micropolygons are created and textured in the local coordinate sysem of the object, with the result that texture filtering is simplified and improved. Visibility is calculated in screen space using stochastic point sampling with a z buffer. There are no clipping or inverse perspective calculations. Geometric and texture locality are exploited to minimize paging and to support models that contain arbitrarily many primitives.

An algorithm is presented that solves the visible surface problem at each pixel independently. This allows motion blur and depth of field blurring to be integrated into the algorithm. It also allows parallel processing. The algorithm works on large numbers of polygons. An analytic Gaussian filter is used. The filter can be elongated or scaled differently for each polygon to adjust for its speed or distance from the focal plane. This is achieved by shrinking or scaling the polygon prior to solving the hidden surface problem so that blurring is correctly presented when objects obscure each other.

Currerntly texture mapping onto projections of 3-D surfaces is time consuming and subject to considerable aliasing errors. Usually the procedure is to perform some inverse mapping from the area of the pixel onto the surface texture. It is difficult to do this correctly. There is an alternate approach where the texture surface is transformed as a 2-D image until it conforms to a projection of a polygon placed arbitrarily in 3-space. The great advantage of this approach is that the 2-D transformation can be decomposed into two simple transforms, one in horizontal and the other in vertical scanline order. horizontal scanline order, Sophisticated light calculation is also time consuming and difficult to calculate correctly on projected polygons. Instead of calculating the lighting based on the position of the polygon, lights, and eye, the lights and eye can be transformed to a corresponding position for a unit square which we can consider to be a canonical polygon. After this canonical polygon is correctly textured and shaded it can be easily conformed to the projection of the 3-D surface.

The generation of computer-shaded pictures requires that a program calculate intensities for each pixel. These intensities are then viewed on a monitor or transferred to film. It is desirable that the intensity values on the monitor or the reflectance values on the film have a linear relationship with the original calculated intensities both to give the user fine control over the appearance of the object and to do anti-aliasing properly. However, after the calculated intensity values are sent to the DAC'c some nonlinear distortions occur both in the CRT and on film. We can compensate for this by adjusting the intensity number just before digital-to-analog conversion. This compensation can be accomplished by a table lookup using the intensity as an index into the table and the corresponding value as the new intensity. Fortunately, most hardware manufacturers have made table lookup features part of their equipment. However, the entries in the table must still be generated. The methods for doing this are not well known in the computer graphics community. This paper presents an algorithm for generating a compensation table.

In recent years we have gained understanding about aliasing in computer generated pictures and about methods for reducing the symptoms of aliasing. The chief symptoms are staircasing along edges and objects that pop on and off in time. The method for reducing these symptoms is to filter the image before sampling at the display resolution. One filter that is easy to understand and that works quite effectively is equivalent to integrating the visible intensities over the area that the pixel covers. There have been several implementations of this method - mostly unpublished - however most algorithms break down when the data for the pixel is complicated. Unfortunately, as the quality of displays and the complexity of pictures increase, the small errors that can occur in a single pixel become quite noticeable. A correct solution for this filter requires a hidden-surface algorithm at each pixel! If the data at the pixel is presented as a depth-ordered list of polygons then the average visible intensity can be found using a polygon clipper in a way similar to that employed by two known hidden-surface algorithms. All of the polygons in a pixel are clipped against some front unclipped edge into two lists of polygons. The algorithm is recursively entered with each new list and halts when the front polygon is clipped on all sides, thereby obscuring the polygons behind. The area weighted colors are then returned as the value to be added to the other pieces in the pixel.

In the last few years several systems have been written for aiding in the conventional two-dimensional animation process. The main purpose of these systems has been to let the computer produce missing drawings based on extreme drawings produced by animators. While there has been some success and a great deal of optimism, the promise of higher output and quality using a computer has not been realized. The transition from simple drawings optimized for use on the computer to the complicated and detailed drawings of quality conventional animation has been much harder than expected. The principle difficulty is that the animators drawings are really two dimensional projections of three dimensional characters as visualized in the animators head, hence information is lost, ie. one leg obscures another. The problem of making a program infer the original object from its projections is akin to extremely difficult artificial intelligence problems. Efforts to overcome this by drawing skeletons or increasing the number of overlays require more manual intervention thereby offsetting the gains of using the computer. One way to analyze an approach is to determine the average time required of an artist or operator at all stages of animation for every frame. A second problem not generally recognized is that a production animation system requires the management of hundreds of thousands of drawings, hence data base management techniques not normally found in experimental animation systems.

No abstract available.