“Towards a unified approach to 2-D picture manipulation” by Shoup

  • ©Richard G. Shoup




    Towards a unified approach to 2-D picture manipulation



    This talk describes a new architecture for creating and manipulating 2-D raster scanned images. This special-purpose hardware, called Aurora, manipulates images in the form of digital video point streams from a variety of sources (including its own memory) and can drive raster scan devices such as CRT monitors, scanners, etc. It uses a single, simple metaphor for all 2-D point array picture manipulations, thus allowing a consistent implementation scheme for everything from text images to freehand painting to high-resolution printing image generation.

    This new design has grown out of the following collection of thoughts:

    1. In general, a picture is an ongoing process, not a state This is the opposite of the usual view of frame buffers, digital “bit maps”, and dynamics done by sequences of still pictures.

    2. At any instant, however, the current state of the picture process can be represented by a 2-D array of colored points–a point array –which is often raster scanned into a point stream which recirculates through the system.

    3. A “display controller” usually touches every point in the image once per frame time, so why not use that ongoing process to do some operations on the resulting point stream picture?

    4. A great many of the common operations we wish to do in creating and manipulating pictures are simple functions which can be done on a point-by-point basis (cursors, freehand painting and drawing, text composition, masking, merging and overlaying, transparency and translucency, TV special effects, etc.). These operations can be implemented by appropriately combining two or more point streams.

    The talk will describe this architecture in general terms and discuss some simple examples and preliminary results from a prototype system.

ACM Digital Library Publication: